-

A life in common: exploring the causal effect of

living on campus.
Holmes, Joshua Mark
https://iro.uiowa.edu/esploro/outputs/doctoral/A-life-in-common/9983777058302771/filesAndLinks?index=0

Holmes, J. M. (2020). A life in common: exploring the causal effect of living on campus [University of Iowa].
https://doi.org/10.17077/etd.ot8m-gkg2

https://iro.uiowa.edu
Free to read and download
Copyright © 2019 Joshua Mark Holmes
Downloaded on 2025/04/24 10:00:40 -0500

-
 A LIFE IN COMMON:
EXPLORING THE CAUSAL EFFECT OF LIVING ON CAMPUS

by

Joshua Mark Holmes

A thesis submitted in partial fulfillment

of the requirements for the Doctor of Philosophy
degree in Educational Policy and Leadership Studies in the
Graduate College of
The University of Iowa

August 2019

Thesis Supervisor: Professor Nicholas A. Bowman

 Copyright by

JOSHUA MARK HOLMES

2019

All Rights Reserved

For my father, who always supported me in every educational pursuit and
taught me to always be humble and kind.

And to my Mom, Chris, Holly, Izzy, and Chloe

who inspire, guide, and love me every day.

ii
 ACKNOWLEDGEMENTS

My journey to finishing this dissertation was helped along the way by numerous

individuals. I would like to briefly acknowledge their personal or professional support. First, my

residence life families throughout the years that include West Chester University, Michigan State

University, Lake Superior State University, and the University of Delaware. My growth from a

student, to a paraprofessional staff member, to a student affairs professional would not have

happened without their guidance and support. At West Chester University, I would like to thank

the Honors College and their staff Kevin Dean and Donna Carney. Additionally, Drs. Tom Purce

and Idna Corbett whom spent numerous hours encouraging me to pursue student affairs as a

professional and pushed me to travel the country and the world to gain new experiences to

broaden my horizons. The hours of conversations in their offices are something special to me even

decades after graduation.

At Michigan State University, my paths intertwined with some tremendous scholars and

practitioners. Namely, Drs. Kris Renn and Bob Coffey who both furthered my interest in student

affairs and opened my eyes to scholarly research. Additionally, I would like to thank my Masters

cohort, whom provided a rich learning environment, robust debates, and unique perspectives that

pushed me as a scholar and practitioner. Specifically, I would like to thank Krissy Petersen,

Daniel Mathis Spadafore, and Liz Culbertson. My journey next took me to Lake Superior State

University, where Ken Peress, Scott Korb, Sharmay Wood, Carol Schmitigal, and Steph Aho

continued to support me through the good times and the not so good. Also, my student staff there

whom pushed me to continue being better, I will forever be thankful to the Brady Boys and

Angela. Their support and laughter pushed me to keep going and to realize my dream of attaining

a PhD. Finally, the University of Delaware refined me into the professional I am today.

Specifically, Jim Tweedy and Michele Kane, Linda Wooters, Lisa Sorantino, Linda Carey,
iii
Tabitha Groh, Vanessa McClafferty, and Karen DeMonte. Also, I would be remiss if I didn’t

thank some of my amazing colleagues and students I met while working at Delaware including

Trina Nocerino Sokoloski, Danielle Dolan, Paul Hengested, Kelvin Wong, and Tom Robertson.

The conversations I had with these individuals made me a more intentional scholar and

practitioner. I am forever grateful for the time we spent together.

During my time at the University of Iowa, I encountered numerous people whom

supported me throughout my doctoral journey. These include my WineStyles family of Ed and

Janaan, Toni, Marcy, Rodney, Teresa, Brian, Jeff, and Tim. You gave me a place to balance the

demands of academe and allowed me to enjoy some amazing wines. You will forever be a chosen

family to me. Thank you. To my musical friends Wes Habley and Evan Hilsabeck, I am grateful

for the opportunity to engage in a create endeavor that helped me think differently about how

higher education and musical theatre can interact. To my cohort and fellow students at Iowa -

Laila, DaVida, Kari, Alex, Nayoung, Teniell, and KC. I would have not made it through the

program without you. The conversations inside and out of the classrooms helped me think

differently and more critically about our work and making higher education more equitable and

accessible to all. To the faculty at Iowa who developed me as a scholar and a human being,

specifically Drs. Chris Ogren, Sherry Watt, Ernie Pascarella, Lyn Redington, and Cassie

Barnhardt. And an extra special thank you to Dr. Ariel Aloe for pushing me to become a better

quantitative researcher. Additionally, I need to specifically thank the support and guidance of two

amazing faculty – Drs. Jodi Linley and Nick Bowman. There are not enough words to describe

how grateful and lucky I am to have worked with each of you. This dissertation is a reflection of

your mentorship, guidance, and support. Especially without you two, it would not have been

possible.

iv
 Lastly, thank you to my friends and family. To my mom, whose love and support have

made me the human I am today. We have been through a lot, but your tenacity, perseverance, grit,

and dedication to family has always lifted me up to be more and become more. You are forever

with me, as a handprint on my heart. I love you way more. And since this is a published

document, it means it’s official. To my brother and his family, Chris, Holly, Izzy, and Chloe.

Thank you for always being there, through everything. Knowing you’ll always be in my corner

and support me no matter what has allowed me to dream big. To Seth, thank you for pushing me

to think more critically, argue more logically, and reason more thoroughly. Having you by my

side is the best gift I could ever ask for – thank you for loving me through this journey. And last,

to my father. I know completing this degree would have made you proud, because you were

always proud of everything I did. Thank you for teaching me to be kind, thoughtful, generous, and

humble. I know you were with me in spirit the last two years – I hope to continue to make you

proud in all of my future pursuits.

v
 ABSTRACT

This this three-article dissertation sought to explore the potential causal link of students’

collegiate residence with three broad categories of student outcomes. Using data from the

Wabash National Study of Liberal Arts Education, each article employed propensity score

matching in an effort to reduce selection bias associated with a student’s decision to live on

campus. The first manuscript examined academic achievement, retention, four-year graduation,

and satisfaction with the college experience and found that living on campus had no direct effect

on any of these outcomes. The second manuscript explored the effect of living on campus on

students’ overall health, alcohol consumption and binge drinking, smoking behaviors, exercise

frequency, and psychological well-being. Findings suggest that living on campus has a positive

effect on students’ first-year alcohol consumption, frequency of binge drinking, and exercising

behaviors. These findings do not persist beyond the first year. Some conditional effects were

uncovered, with a significant interaction between race and campus residence on some outcomes.

The final study considered the effect living on campus has on student engagement. Living on

campus was found to have a direct effect on positive peer interactions, frequency of interactions

with student affairs staff, and co-curricular involvement. Like the second study, conditional

analyses were conducted and revealed significant interactions mostly among race and campus

residence.

vi
 PUBLIC ABSTRACT

Every fall, millions of students move onto college campuses, often for the first time.

Housing students is not a new phenomenon for colleges and universities; rather, within the

United States this practice traces back to the founding of Harvard University in 1636. However,

the belief that living on campus is positive for students and has a direct effect on their lives has

not been completely decided by research. Prior studies are often limited by analytic samples and

research methodologies in the ability to draw causal associations, meaning that living on campus

is the reason for these differences. This dissertation sought to apply advanced statistical

techniques to estimate the potential causal effect of living on campus on various student

outcomes.

Overall, this three-article dissertation found evidence that living on campus does have a

direct effect on some outcomes. It is beneficial for students’ social connections, co-curricular

involvement, and interaction with student affairs staff. Moving off campus after their first year

led to decreased peer connections and lower amounts of co-curricular involvement for students.

Additionally, living on campus during the first year led students to consume alcohol at higher

rates in addition to more frequently binge drinking. However, these specific behaviors did not

appear to persist past the first year. Finally, collegiate residence had no direct effect on students’

psychological well-being, academic achievement, retention, satisfaction with college, nor did it

increase the probability of graduating within four years.

vii
 TABLE OF CONTENTS

LIST OF TABLES ......................................................................................................................... xi

LIST OF FIGURES ...................................................................................................................... xii

CHAPTER ONE: A BRIEF HISTORY OF COLLEGIATE RESIDENCE ..................................1

History of Campus Residences ..................................................................................................3

Early Origins ....................................................................................................................... 3

Seventeenth and Eighteenth Centuries................................................................................ 5

Nineteenth Century ............................................................................................................. 9

Twentieth Century ............................................................................................................ 13

The Twenty-first Century ................................................................................................. 21

Campus Residence Research ...................................................................................................23

Proposal for Empirical Studies ................................................................................................26

References ................................................................................................................................30

CHAPTER TWO: COLLEGIATE RESIDENCE AND ACADEMIC ACHIEVEMENT,

RETENTION, GRADUATION, AND SATISFACTION ............................................................43

Research on Campus Residence, Academic Achievement, and Retention .............................44

Conceptual Framework ............................................................................................................48

The Current Study ....................................................................................................................49

Methods....................................................................................................................................50

Institutional Sample and Data Collection ......................................................................... 50

Analytic Sample ................................................................................................................ 51

Measures ........................................................................................................................... 53

Analyses ............................................................................................................................ 56

Limitations ........................................................................................................................ 60

Results ......................................................................................................................................61

Descriptive Statistics......................................................................................................... 61
viii
 Collegiate Outcomes ......................................................................................................... 62

Discussion and Recommendations ..........................................................................................63

Appendix: Study Variables ......................................................................................................77

References ................................................................................................................................81

CHAPTER THREE: COLLEGIATE RESIDENCE, HEALTH, AND PSYCHOLOGICAL

WELL-BEING ...............................................................................................................................90

Supporting Literature and Study Purpose ................................................................................90

Method .....................................................................................................................................92

Data Source and Participants ............................................................................................ 92

Measures ........................................................................................................................... 93

Analysis............................................................................................................................. 94

Results ......................................................................................................................................95

Discussion ................................................................................................................................96

References ................................................................................................................................99

CHAPTER FOUR: COLLEGIATE RESIDENCE AND STUDENT ENGAGEMENT ...........103

Relevant Literature.................................................................................................................105

Student Engagement and Student Outcomes .................................................................. 105

Living on Campus and Student Engagement .................................................................. 105

Living on Campus and Peer Interactions ........................................................................ 107

Conceptual Framework ..........................................................................................................111

Current Study .........................................................................................................................112

Methods..................................................................................................................................113

Institutional Sample and Data Collection ....................................................................... 113

Analytic Sample .............................................................................................................. 114

Measures ......................................................................................................................... 115

Limitations ...................................................................................................................... 118

ix
 Analyses .......................................................................................................................... 118

Results ....................................................................................................................................121

Propensity Score Analyses.............................................................................................. 121

Outcomes Analyses......................................................................................................... 122

Discussion ..............................................................................................................................125

Recommendations for Research and Practice ........................................................................128

Appendix: Study Variables ....................................................................................................143

References ..............................................................................................................................147

CHAPTER FIVE: CONCLUDING THOUGHTS .....................................................................158

Research Findings ..................................................................................................................159

Future Directions for Research ..............................................................................................161

Implications for Practice ........................................................................................................165

Conclusion .............................................................................................................................166

References ..............................................................................................................................168

x
 LIST OF TABLES

Table 1. Unadjusted Means and Standard Errors for Variables of Interest Based on Students’
First-year Residential Choice ........................................................................................................ 67

Table 2. Unadjusted Means and Standard Errors for Variables of Interest Based on Students’
Moving Off Campus After the First Year ..................................................................................... 69

Table 3. Significance of and Standardized Mean Differences for Each Propensity Score
Model, Before and After Stratification ......................................................................................... 71

Table 4. Results of Regression Analyses of First-year Residence Predicting College

Academic and Satisfaction Outcomes .......................................................................................... 73

Table 5. Results of Regression Analyses of Collegiate Residence Predicting College

Academic and Satisfaction Outcomes for Moving Off Campus after Students’ First Year ......... 74

Table 6. Unstandardized Regression Coefficients for Relationship of Campus Residence with

Psychological Well-Being and Student Health Outcomes after First and Fourth Year of
College .......................................................................................................................................... 98

Table 7. Unadjusted Means and Standard Errors for Variables of Interest Based on Students’
First-year Residential Choice ...................................................................................................... 131

Table 8. Unadjusted Means and Standard Errors for Variables of Interest Based on Students’
Moving Off Campus After Their First Year ............................................................................... 133

Table 9. Significance of and Standardized Mean Differences for Each Propensity Score
Model, Before and After Stratification ....................................................................................... 135

Table 10. Results of Regression Analyses of First-year Collegiate Residence Predicting

Student Engagement Outcomes .................................................................................................. 137

Table 11. Results of Conditional Effects Analyses Based on Student Demographic

Characteristics and Collegiate Residence ................................................................................... 138

Table 12. Predicted Means and Standard Deviations of Standardized Scaled Student
Engagement Measures Based on Student Demographic Characteristics and Place of
Residence .................................................................................................................................... 139

Table 13. Results of Regression Analyses of Collegiate Residence Predicting Student

Engagement Outcomes for Students Moving Off Campus After Their First Year .................... 140

xi
 LIST OF FIGURES

Figure 1. Propensity Score Distributions for Students Living On Campus (treated) and Off
Campus (untreated) During Their First Year ................................................................................ 75

Figure 2. Propensity Score Distributions for Students Living On Campus (treated) and Off
Campus (untreated) During Their Collegiate Experience ............................................................ 76

Figure 3. Propensity Score Distributions for Students Living On Campus (treated) and Off
Campus (untreated) During Their First Year .............................................................................. 141

Figure 4. Propensity Score Distributions for Students Living On Campus (treated) and Off
Campus (untreated) During Their Collegiate Experience .......................................................... 142

xii
 CHAPTER ONE: A BRIEF HISTORY OF COLLEGIATE RESIDENCE

Depending on the disaggregation method, the number of students living on campus only

accounts for 13 percent of the total population of students at all institutions or 25 students at

four-year institutions (U.S. Department of Housing and Urban Development, n.d). Despite the

fact that this group of students represents a small fraction of the student body, institutional

spending on residence halls and facilities to promote living on campus has skyrocketed (Eaton,

Habinek, Goldstein, Dioun, Godoy, Osley-Thomas, 2016; Saffron, 2013; Seltzer, 2017; Weber,

2016). Institutions are using the new residences as recruitment tools (Saffron, 2013) despite

shortages of beds within many institutions across the United States (Kafka, 2018). Institutional

logics seem to currently support the idea of living on campus, but does living on campus make a

difference in the lives of students? The answer lies in research that shows direct and indirect

benefits that expand beyond traditional academic outcomes. Ryan (2016) writes “we come to

understand that well-organised social life—intensified, where possible, by residence—is not

simply an amenity but a critical component in fully realising the true value of higher education,

to the individual student, and to the larger society” (p. 13). Faculty and administrators have

sought to directly affect the lives of their students, from early times developing pious, moral

leaders (Morison, 1936; Rudolph, 1990; Ryan, 2001, 2016; Thelin, 2004), to contemporary times

of socially conscious citizens (Kerr & Tweedy, 2006). If living on campus has a direct role, it is

up to higher education researchers and practitioners to discover and demonstrate the causal

mechanisms and find ways to provide students with the benefits associated with living on

campus.

“It is to be borne in mind that the provision of residence halls is quite as important and as

essential a part of the work of the University as the provision of libraries, laboratories, and class

1
rooms. The chief purpose of university residence halls is not one of mere housing, but rather of

education and educational influence.” (Columbia Bulletin of Information, 1923, p. 8). President

Butler’s words to the Trustees of Columbia University, during his Annual Report of 1922, are as

true now as they were first printed. Proponents have viewed living on campus as an integral

component of the American approach to higher education since the founding of Harvard in 1636.

Mark Ryan (2001), author of A Collegiate Way of Living: Residential Colleges and a Yale

Education, writes “from that beginning in Massachusetts Bay, American higher education was

concerned not only with the training of minds but also with the molding of character, and the

‘Collegiate Way of Living,’ with its common residence, structured community life, intellectual

exchange, and spiritual purpose and practices, was the path to those complementary goals” (p.

49). The approach of postsecondary education in the United States was about more than just

academic growth; it was founded to develop the whole individual.

A life in common living on campus may act as a causal mechanism to develop students’

community lives, moral and spiritual development, and intellectual growth. Hayes (1932)

attributed the causation to residence halls directly, noting that they “have power greatly to further

the essential purposes of colleges, which include the development of socialized human beings as

well as the promotion of scholarship” (p. 12). Contemporary postsecondary education scholars

interested in the effect of living on campus have found some evidence that living on campus is

related to a variety of outcomes including student satisfaction, retention, growth in terms of

critical thinking, gains in personal development, and openness to diversity (Astin, 1977, 1985,

1993; Chickering, 1974; Pascarella & Terenzini, 1991, 2005; Pike, 2002). However, recent

syntheses of the literature suggest that these effects that were once considered overwhelmingly

2
positive may in fact be more nuanced and mixed in terms of how living on campus affects

students (Mayhew, Rockenbach, Bowman, Seifert, & Wolniak, 2016).

Limitations to the current cannon of literature related to living on campus relate to the

sample sizes, the scope of institutional type (for those few studies using a multi-institutional

dataset), and methodological approaches to analyzing the data. Most published literature uses

samples of students in their first year from single institutions without accounting for self-

selection bias associated with a student’s decision to live on campus. Thus, this dissertation

speaks to these limitations by using a multi-institutional, longitudinal dataset employing a quasi-

experimental approach in an attempt to reduce self-selection bias. The end of the chapter

discusses these ideas more in-depth along with the specific research questions at hand, aiming to

illustrate how I plan to answer the following main question: What is the effect of living on

campus for students?

History of Campus Residences

Early Origins

To understand the philosophy and purpose of housing within early institutions founded in

the United States, one must look into the European roots of higher education as they directly

influence the establishment of Harvard in 1636. Initially, the term university was applied more

broadly to represent a “totality of a group, whether of barbers, carpenters or students” (Haskins,

1957, p. 9). In the thirteenth century, faculty were lecturing to 10,000 students in Bologna,

another 3,000 students were enrolled in Oxford, and the number was approximately 30,000 in

Paris (Cowley, 1934a). With such an influx of students into these medieval cities, housing

became problematic. Students sought many places to live and their success and accommodations

were directly correlated to their financial wealth (Cowley, 1934a). More financially well-off

students were able to secure private housing or board with merchants in the cities, while poorer

3
students received less than adequate housing. As the numbers of students increased, profiteering

from local townspeople over rent became an issue. Students bound together to fight for their

rights and, in turn, also led them to use this newfound agency to make specific demands and

requirements from the masters, those in charge of giving lessons to the students (Haskins).

This collective student body marks the beginnings of the residential college’s history.

Haskins (1957) notes that during this time, particularly in Paris, what was once “originally

merely an endowed hospice or hall of residence, the college early became an established unit of

academic life of many universities…the colleges became normal centres of life and teaching,

absorbing into themselves much of the activity of the university” (p. 18). Communal student

living offered more than just room and board; it was becoming an essential part of the formal

structure of the institution. The early self-governance of the students soon became formalized

and “in the course of two centuries the houses which students had established on their own

initiative had passed entirely from their control into the hands of university authorities” (Cowley,

1934a, p. 706). The informal housing arrangements that students had made in order to partake in

the university had now become an institutionalized facet of the university.

These formalized ideals permeated throughout Europe and soon institutions in England,

France, and Germany modeled these ideals of communal student living. In England and France,

student accommodations in the universities were often private or semiprivate apartments;

whereas, in the German model, Bursen, or large open sleeping rooms akin to monasteries,

became the norm (Cowley, 1934a). How institutions chose to accommodate students is less

relevant; the main point is the idea that institutions had to house students was spreading

throughout Europe. The differences between the countries would become more important as

reformation and revolution occurred.

4
 Revolution and reformation across Europe would directly affect institutional priorities

and missions during the time. As the reformation hit the Germanic areas, the Bursen had come to

be associated with the church and soon fell out of form and were abandoned (Cowley, 1934a).

Luther and his 95 theses would change the way Germanic areas viewed anything that had been

associated with the church – and housing students like monks in large open-halled dormitories

would fall out of favor. In France during the times of revolution, funding for universities shifted.

Funding for universities came from Encyclopedists who “gave all their enthusiasm to scholarship

and the advancement of knowledge” (Cowley, p. 707). The priority was solely on academics and

thus the capacity for institutions to house students fell out of favor. However, the insular nature

of England minimized the impact of revolution and reform, so the ideas of dormitories persisted.

This persistence would become important to the early founding of Harvard as the English settlers

colonized the United States.

Seventeenth and Eighteenth Centuries

There are three reasons living on-campus has been an essential part of United States

higher education since its colonial foundation (Thelin, 2004). The first reason for the adaptation

of residence halls in the early colonies relates to the colonies’ lack of concentrations of

populations. If the settlers in the colonies were to establish institutions of higher learning to bring

forth Christian truth to the new world and develop future leaders and ministers for the colonies,

residential colleges were the only initial way to bring students together and establish these

institutions of higher learning (Rudolph, 1990; Ryan, 2001, 2016; Thelin, 2004). Rudolph argues

that, as cities such as New York and Philadelphia grew, “by then what had been a necessity had

become a tradition” (p. 88). That is, as residential colleges became no longer necessary as cities

developed the capacity to board students, they were still maintained partly for the extracurricular

influence had on the students.

5
 The second reason pertains to the pedigree of the founders of colleges in the United

States. Quite simply, these early founders of institutions replicated what they had experienced.

Of the early institutions founded in the United States, most all of the founders were graduates of

Oxford and Cambridge in England. Crowley (1934a) writes “more than twoscore Cambridge

graduates migrated to the Massachusetts Bay colony during its first three decades, and naturally

enough they brought with them a predilection in favor of the educational structures which they

knew in England” (p. 708). Given the high representation of Cambridge and Oxford graduates in

the building of new institutions in the early institutions, their replication of their educational

experiences makes sense. These gentlemen enjoyed their “collegiate way” of living and desired

for their students to experience the same.

The third reason living on campus has been essential to the mission of higher education

in the United States relates to the “collegiate way” of living, coined by Cotton Mather, an early

Harvard graduate (Morison, 1936). It is “the notion that a curriculum, a library, a faculty, and

students are not enough to make a college. It is an adherence to the residential scheme of things.

It is respectful of quiet rural settings, dependent on dormitories, committed to dining halls,

permeated by paternalism” (Rudolph, 1990, p. 87). The idea of paternalism in the early colleges

served to promote morally developed and pious students would could lead within the United

States.

The founders of Harvard College did not want the students fending for themselves in the

“wicked” and “sinful” city. In the early years of these new institutions, living on campus was the

mechanism by which faculty could remove wickedness and sin from college graduates. In 1671,

Harvard’s governing board stated, “it is well known what advantage to Learning accrues by the

multitude of persons cohabiting for scholasticall communion, whereby to actuate the minds of

6
one another, and other waies to promote the ends of a Colledge-Society.” (Morison, 1936, p. 49).

The Puritan founders of Harvard College recognized importance of the collegiate way with

students in pursuit of this liberal education, under paternal control of the college presidents, and

thus followed the residential collegiate model of the English universities at Cambridge and

Oxford (Rudolph, 1990; Ryan, 2001).

The age of incoming students is also of importance, given that at the time the average age

of an entering student was 14. “Because students were young adolescents, faculty served in loco

parentis and their out-of-class duties focused on developing moral character and regulating

student behavior” (Palmer, Broido, & Campbell, 2008, p. 88). The faculty felt they needed to

serve as parents to these young students. As faculty founded other institutions in the United

States, this notion continued—students would come to study, cohabitate, and worship together

(Rudolph, 1990; Ryan, 2001). Crowley (1934a) offers a more direct opinion, writing “students

had souls to be saved and the early faculties were bent upon saving them…if a youngster

misbehaved they believed with certainty that they were exorcising the devil when they whipped

him” (p. 709).

The purpose of education, in the eyes of the early founders, connects back to this primary

reason for housing students on campus—a moral and developmental one. Faculty were expected

to serve in loco parentis. This all-encompassing role for faculty would differ in the United States

from the structure at Oxford and Cambridge, which saw discipline and teaching as separated.

This lack of separation would later play a role in the argument against providing campus

residence. For students in this time period, however, the reason one attended these early

institutions was to develop character as well as the whole human being. Additionally, Ryan

(2001) argues that Latin “shows another basic notion in the colonial New Englanders’ concept of

7
education…their intention was not only to inculcate protestant orthodoxy, but to perpetuate the

intellectual tradition of Europe, with its grounding in the classics” (p. 38). This classical learning

developed Christian truth within the students and prepared them to become the future leaders of

the colonies within the European tradition (Rudolph, 1990; Ryan, 2001; Thelin, 2011). Living in

the residential colleges allowed the faculty and university presidents to espouse Christian values

and morals to the students controlling and mediating the behaviors of their students, ensuring

adherence to a Christian gentlemanly tradition.

The main role and function of faculty was to ensure safe passage from “boyhood to

manhood” (Rudolph, 1990, p. 88). In this view, the students came to these American institutions

of higher learning to become educated, pious, and moral men. Faculty were responsible for this

moral development of the student. Students came with empty minds to be filled with knowledge

espoused by faculty and tutors (Rudolph). “Moral philosophy emerged as the capstone of the

curriculum, taught by the president to the senior class” (Horowitz, 1987, p. 26). Horowitz notes

the students’ role was “to pay, pray, study, and accept” (p. 26). Living on campus in this

“collegiate way” thus allowed the faculty to exercise control over every facet of their students’

lives, and it became one of paternalism. This paternalistic control manifested as a rigid daily

schedule including morning prayers, classes, meal, formal study hours, and finally evening

prayers. The curriculum demanded compliance to moral and pious behavior, and faculty

punished students not adhering to this strict schedule or these behaviors. The faculty and tutors in

the institution and in the residence halls thus “functioned as spies, policemen, and judges”

(Rudolph, p. 104). However, this paternalistic focus would ebb as institutions of higher learning

entered the 1800s and students began to rebel.

8
Nineteenth Century

The nineteenth century saw tremendous growth in institutional type, purpose, and which

students were able to access higher learning. With the changing of the times, two themes related

to campus housing began to emerge: individuals being critical of campus housing and students

openly rebelling against paternalistic notions otherwise known as in loco parentis. Critiques of

on-campus housing came from a variety of individuals. To the outsider not affiliated with the

university, large “concentrations of young men living together, with so little academic work to

do and so many vices to distract them, led to moral decay and rebellion” (Palmer, Broido, &

Campbell, 2008, p. 88). The system of housing students was coming into question for seemingly

cultivating delinquent and morally corrupt young men. Later in the century, institutional leaders

would echo the public’s sentiment, questioning the utility of housing students on campus. The

President of the University of Michigan would note the residence hall “system is objectionable in

itself. By withdrawing young men from the influences of domestic circles, and separating them

from the community, they are often led to contract evil habits, and are prone to fall into

disorderly conduct. It is a mere remnant of the monkish cloisters of the Middle ages, still retained

in England, but banished from the universities of Germany” (Cowley, 1934a, p. 711). To him,

removing students from their communities and home lives was detrimental to their development

and directly caused negative behaviors to grow; his call would shift the focus and purpose of

higher education within the United States, specifically modeling the German system where the

sole focus should only be on research and knowledge production.

To some administrators, the experience of living on campus was no longer a necessary

component of the curriculum. Part of this challenge arose from faculty returning from German

universities in Berlin, Leipzig, Heidelberg and Göttingen. Having experienced the German

system of higher education, the faculty saw value and now placed their main emphasis on
9
knowledge specialization and research (Nuss, 1996). This change shifted faculty’s attention

away from students towards a substantial amount of time on research. This philosophy left little

time for, and ultimately less interest in, students’ non-curricular endeavors which included their

living environments. As such, the current view of the curriculum, research and knowledge

production differed dramatically from earlier centuries when faculty viewed a large part of their

role as developing pious gentlemen (Rudolph, 1990; Ryan, 2001).

In addition to challenges from the community and institutional leaders and the influence

of the German system of higher education, the nineteenth century also saw increased student

rebellion within the dormitories (Johnson, 2015). Scholars would attribute this rebellion to many

causes, noting that “between the American Revolution and the Civil War, students felt a growing

disenchantment with college authority. These sons of patriots inherited a hunger for liberty but

lacked a national stage and a real crisis to snare their attention” (Ireland, 2012). As students

moved on campus, they felt constrained, not appreciating the role of in loco parentis taken on by

the university.

From the late 1700s throughout the early 1800s, the number of rebellions within

residential halls increased dramatically (Crowley, 1934a; Rudolph, 1990). Allmendinger (1973)

writes “individual acts of violence and terrorism disrupted order in residence halls with what

appears to have been increasing frequency beginning in the 1820s” (p. 76). The rebellions

seemed to start as the result of a revolt against paternalism, interior design, or the food being

provided to the students. At Harvard, the Rebellion of 1818 saw chaos among the students due to

the fact that dining rooms “for each class were connected by large openings that made it easy for

students eventually to throw food, furniture, and handy projectiles at rival classes” (Ireland,

2012). This rebellion started as a common fight between classes, where food and the college

10
crockery were thrown between the freshmen and sophomore classes, facilitated by the interior

design. The food fight would eventually lead to the students criticizing the way the university

was handling the discipline of its students culminating with the sophomore class of 80 students

declaring the student suspensions a form of tyranny in which their response was to resign from

Harvard en masse.

Student rebellion on campus was not limited to Harvard; another example occurred at

Yale in 1828, which was known as the “Bread and Butter Rebellion”. Students became aware of

“cooks concealing the students’ pies and serving them to their friends at midnight suppers almost

every night” (Schiff, 1995, p. 80). These students living on campus protested the contract

between Yale and the students in regard to food, refusing to attend classes until the faculty met

their demands of better food service for which the students were paying. Cowley (1934a) writes

of other rebellion at Dartmouth that “faculty were visited by groups of students who would stand

outside their windows and blow tin horns late into the night” (p. 710). At Princeton, students

burned down Nassau Hall in 1802, and they set fire to one of its outhouses in 1814, nearly

burning more of campus. At Williams, students stayed away from recitations for almost a week

to protest a president not succumbing to their demands. Additionally, at Yale there was the

“Conic Section Rebellion” where students refused to recite mathematics lessons. Finally, George

Bancroft, an American historian during that time period, “lost an eye when as a Harvard tutor he

attempted to quell an incipient uprising” (Crowley, 1934a, p. 710). By the middle of the 19th

century, students were more active in rebellion and for faculty “residence halls became places for

students merely to sleep, to eat and occasionally to study…The opportunity to make them the

core of the educational program has been lost in the disciplinary muddle” (Crowley, 1934a).

11
 Horowitz (1987) adds another layer by arguing that rebellion was a way to seek power on

campus in addition to rebelling against the traditional curriculum and paternalism. One way to

generate power and cultural capital on campus was for students to form fraternal organizations.

In addition to power, these organizations would also create different spaces for students to live

that were outside of the faculty’s control. Horowitz argues that college men created fraternities

for the purpose “to create within the larger college a small group of compatible fellows for

friendship, mutual protection, and good times” (p. 29), which competed against serious study.

The college men in fraternities tended to be the wealthier students on campus, which created an

imbalance in power where “in this bifurcated world, the wealthier students tended to fight

authority, the poorer complied” (p. 30). The rise of the fraternities would shift authority from

faculty and institutions to students themselves. To a degree, faculty promoted this mantel of self-

governance among the students (Ryan, 2016).

As such, the purpose of higher education was moving away from the moral and pious

development of students towards one where students created their own social connections and

social mobility. The formal curriculum and espoused purposes of collegiate study had shifted in

the 1800s. Hevel (2017) notes what was once “a college created for religious end and to train

clergy became, over the course of the nineteenth century, devoted to a more scientific

curriculum” (p. 423). The rise of these social organizations, fraternities and sororities, would

come to fill the void administrators left when disregarding living on campus as part of their

leadership responsibilities. By the mid 1880s, one saw the impact as “fraternity and sorority

houses have become fixtures on most college campuses chiefly because students needed places

to live and eat, and the colleges were unable, if not unwilling, to provide them” (Crowley, 1934a,

p. 712). However, not all faculty at this time were committed to the idea of research and

12
knowledge production at the expense of students, so a committee of faculty and corporation

members of Yale issued a report in 1828 defending the classical curriculum. Within this report,

there was also explicit support for the collegiate way of living, stating the students needed “a

substitute…for parental superintendence” (Conrad & Johnson, 2008, p. 188) and that “the

parental character of college government requires that students should be collected together, as to

constitute one family; that the intercourse between them and their instructors may be frequent

and familiar” (p. 190). There were still individuals that believed in the vital role living on

campus plays in the lives of students and would resurface in the twentieth century.

Twentieth Century

As the United States entered the 20th century, the history of living on campus and

residence halls developed on a variety of fronts: the culminating legal challenges to in loco

parentis, shifts in institutional philosophy regarding residence halls, student support for “college

life”, federal policy affecting students and institutions, increased diversity on campus, as well as

the professionalization of the housing field. Each of these would ultimately affect the purpose of

students living on campus, cementing it as part of postsecondary education.

Throughout the second half of the nineteenth century and the first half of the twentieth

century, the courts heard legal challenges to the concept of in loco parentis that would affect the

role of faculty and administrators. In loco parentis is Latin for “in the place of a parent” and

describes the ways in which both government and non-governmental entities take on the

responsibilities of acting as a parent typically for minors (Lee, 2011). Prior to the 1960s, faculty

and administrators used this legal philosophy as the impetus to develop their students’ moral

character and exercise disciplinary control (Lee), and court systems had ruled in favor of

institutions’ authority to act as parents. Wheaton College had suspended a student for joining a

secret society and the University of Illinois expelled a student for not attending chapel. In both
13
cases, the courts ruled in favor of the institutions, citing their capacity to have disciplinary

control and right to control their students.

The court system explicitly named in loco parentis in their ruling in favor of Berea

College. A local business owner had entered into litigation against the college after the college

had expelled students for entering places of amusement and eating houses that were not directly

under institutional control (Gott v. Berea College, 1913). The Kentucky Supreme Court upheld

lower court’s decision explicitly stating that “College authorities stand in loco parentis

concerning the physical and moral welfare, and mental training of the pupils” (Gott v. Berea

College, p. 379). Further litigation would uphold in loco parentis, for example, at Stetson

University in Florida and Syracuse University in New York, giving great deference to

institutions of higher learning in regard to their students (Lee, 2011). However, this would

change in the 1960s when courts recognized the constitutional rights of students ending in the

foundational case of Dixon v. Alabama. Alabama State College had expelled African American

students for participating in a lunch-time sit in without respect for the students’ due process. The

court case resulted in the judges’ ruling that state institutions (public colleges or universities)

must be given due process, a stark departure from the assumed in loco parentis relationship upon

enrollment (Lee). Further, the 26th amendment contributed to institutional view of who was an

adult. With the voting age lowered to 18, most traditionally-aged students were now considered

adults and no longer in need of faculty and administrative oversights as a temporary parent.

Another facet of the early twentieth century changed the trajectory of residence halls and

students living on campus; that is, administrators now supported residence halls and called for

new facilities to be built. Prior to the turn of the century, the “attacks upon dormitories in eastern

colleges and the indifference toward them of state university administrators in the Middle West

14
and West continued without abatement until the 1890’s” (Cowley, 1934b, p. 758). However, late

in the nineteenth century into the twentieth century, institutions such as the University of

Chicago saw value in housing students, constructing 11 new residence halls (Cowley, 1934b). In

1901, the Princeton graduate college was founded focusing on the educational environment

where students could “mingle freely in common daily association with one another” (Cowley, p.

759) within a community of intellectual interests. These institutions would pave the way for

Princeton President Woodrow Wilson to establish the preceptorial system and then later

President Lowell of Harvard to lay the ground work for a quadrangle plan in 1909 (Cowley).

President Lowell would espouse the value of college life arguing the role in which living on

campus would directly play.

College life would become an ethos to administration, alumni, and current students and

as such all would use it to rally support for the establishment of residence halls. To these groups

of individuals, residence halls were an essential part of institutional life. Going to college was no

longer solely academic; college life now included sports, college colors, and a strong alumni

network (Cowley, 1934b). Cowley writes “The object of the undergraduate department is not to

produce hermits, each imprisoned in the cell of his own intellectual pursuits, but men fitted to

take their places I the community and live in contact with their fellow men” (1934b, p. 760-

761). Students’ social connections were becoming paramount to their education and the leaders

of these institutions knew residence halls and living on campus were just the place for these

connections to occur. The social lives of students had once again become institutional priority

(Cowley, 1934b).

Federal policy has also had an impact in increasing the number of students housed on

campus in the United States through monies for construction projects and acts that promoted

15
attainment for individuals. Early federal policy support of housing construction on campus began

with legislation related to the New Deal. The National Industrial Recovery Act of 1933 created

what was known as the Public Works Administration which, over the course of 10 years, would

spend billions on projects related to a variety of public works including housing (Transcript of

National Industrial Recovery Act, 1933). One such group of public entities that would benefit

from this act were postsecondary institutions that would capitalize on the low interest rates

offered through loans to secure funding and support for new housing projects on campus. Other

federal policy that would have an effect on campus housing includes The Housing and Rent Act

of 1947, later renewed in 1950, along with the Higher Education Facilities Act of 1963.

Broadening access to higher education, the Serviceman’s Readjustment Act of 1944, also

called the G.I. Bill, offered returning veterans a range of financial benefits to pursue higher

education (Rudolph, 1990; Thelin, 2004). These benefits saw a dramatic increase in enrollment

and pushed the institutional capacity to house students. Additionally, the Civil Rights Act of

1964 prompted postsecondary educational institutions in the United States to perform a census

which then highlighted the low number of Black students at historically and predominantly white

institutions (Williamson, 2003). This act required administrators to analyze their college’s

demographics and uncovered enrollment inequities within their institutions. The federal

government, through implementation of the Higher Education Act of 1965, the interpretation of

the Equal Protection Clause, and an affirmative action executive order signed by President

Lyndon Johnson in 1967, recognized these discrepancies and took a stand against discrimination

by forcing institutions to begin to open their doors to students who previously had limited or no

access (Kennedy, 2015). No longer was higher education primarily for White, middle or upper-

class, men (Rudolph, 1990; Thelin, 2004).

16
 Shifts in access and enrollment would also affect the need for housing on campus. From

1900 to 1930, the number of high school graduates dramatically increased every year (Levine,

1986). This rise in the number of high school graduates drove an increase in college enrollment.

For example, land grant institutions saw enrollment increase by 65 percent between 1910 and

1920 (Levine). One force that drove this enrollment increase is that, between 1920 and 1940, “a

coalition of major foundations accelerated their effort to bring both standards and standardization

to American higher education, an initiative started in the early 1890s” (Thelin, 2004, p. 238). As

such, during this time, the purpose of education seemed to become one of “culture of aspiration”

(Levine). That is, the reason students were to seek a postsecondary education was for their social

mobility. He attributes part of this aspirational shift due to a result of a newly diversified

structure of institutional type which allowed more students to access more types of institutions.

Despite the increase in number of institutions, segregation and selective admission

policies explicitly limited access to particular institutions of higher education for women, ethnic

and racial minorities, and particular religiously minority groups of students. Thelin (2004) writes

“the selective-admissions machinery was used to increase the social homogeneity of a campus by

rejecting applicants from religious and ethnic minority groups” (p. 197) and that “this meant that

social exclusion was shifted to the admission office” (p. 196). David Levine (1986), calling to

question these new admissions processes, wrote “WASP educators and their traditional

constituencies clung to their racist views” (p. 158). Further, “in the 1920s and 1930s, American

higher education witnessed the emergence of a national elite liberal arts college…a selective

institution rooted in class and ethnic prejudice, not talent” (Levine, p. 137). It seems many

administrators of these elite institutions rationalized the exclusionary practices because of the

17
newly increased access to other types of institutions. These other institutions, however, often

were young and not as well resourced.

Did the education students received at these other institutions compare to that of the

eastern elite institutions in high demand? For African American students, in particular, separate

was not equal. Rudolph (1990) addresses part of the inequities, noting these “institutions, while

collegiate in name, did not remotely resemble a college in standards or facilities. By 1917 two

institutions alone… approximated what was coming be understood as an American college…but

as late as the 1930’s Negro leaders were deploring the evidence that Negro colleges were

graduating students who were unable to read and write” (Rudolph, p. 488). These institutions did

not have the structures to support the collegiate way (Ryan, 2001, 2016) at the same level as their

peer institutions. Thelin prominently represents these inequities writing that “almost all the so-

called colleges for black students around 1900 in fact offered little in the way of college-level

instruction” (p. 186). He continues to describe access for Black students, noting “enrollment

prospects…remained limited, not only in the segregated states but nationwide…just prior to

World War II a white between the ages of eighteen and twenty was four times more likely than a

black of the same age group to enroll in college” (Thelin, p. 232). Because many students were

excluded from these institutions and relegated to less prestigious institutions, exclusion and

stratification worked to reinforce social stratification.

Shifting demographics also affected the purpose of education and living on campus. The

students enrolling in postsecondary institutions were becoming increasingly diverse with “a

perceived need for increased ‘adult’ supervision to satisfy the parents that the welfare of their

children was being addressed” (Evans & Reason, 2001, p. 360) and as such new administrative

units assumed this responsibility for students’ lives outside of the formal classroom (Palmer,

18
Broido, & Campbell, 2008). These new administrators, in 1937, articulated their philosophical

approach to praxis in the Student Personnel Point of View stating their “emphasis…upon

development of the student as a person rather than upon his [sic] intellectual training alone”

(American Council on Education, 1937, p. 1).

However, the educational focus on developing the whole person would come later in the

century (Palmer, Broido, & Campbell, 2008) as the first step was for student affairs professionals

to provide an “adequate housing program” as well as an “adequate food service” for their

students (American Council on Education, 1937, 1949). Once these basic needs were met,

student affairs practitioners could enact programs and curricula that promote positive outcomes

including critical thinking, personal growth and development, persistence, openness to diversity,

and satisfaction (American Council on Education 1937, 1949; Astin, 1977, 1993; Committee on

the Student in Higher Education, 1968; Pascarella, Bohr, Nora, Zusman, Inman, & Desler, 1993;

Pascarella & Terenzini, 1991; Pike, 2002; THE Project, 1975).

The role of residence life professionals shifted dramatically during the latter part of the

20th century. In 1902, the National Organization of the Deans of Women was founded to ensure

the social welfare and physical wellbeing of students (Cowley, 1934b). This organization started

to give purpose to housing professionals, where they “served primarily as building managers and

student disciplinarians. Most housing staff had no educational preparation to integrate living and

learning experiences or even to work with college students” (Palmer, Broido, & Campbell, 2008,

p. 89). Even if their purpose was not integration of living and learning, this organization helped

professionalize the field, explicitly stating the role for these individuals within the campus

organization.

19
 As housing facilities and student populations grew as a result of the G.I. Bill and the

wave of the baby boomers, so too did the role of living on campus through the further

professionalization of housing administrators. S. Earl Thompson, who was Director of Housing

at the University of Illinois, recognized the challenges of overcrowding on campus and proposed

that his housing colleagues convene to discuss these issues. In 1949, these individuals met and

founded the Association of College and University Housing Officials—International (Moiser,

2012). This organization would prompt housing officials to question purpose and their role on

campus, moving beyond the role of manager to that of an educator. In 1965, Harold Riker (1965)

would publish College Housing as Learning Centers, writing

Living is to be defined as more than a bed and learning as more than a desk; they

are part of a total process, a wholeness of student experience on the campus. To

contribute favorably and consistently to this experience, the living and learning

that go on in student housing have to be stimulated and sustained by planned

programs. (p. V)

Learning in the residence hall had to be planned; it was not something that could be left to

chance or simply managed. Other publications such as DeCoster and Mable’s (1974) Student

Development and Education in College Residence Halls would further the conversation towards

the intentionality needed for planning educational experiences in the residence halls to be able to

directly affect student learning and student development.

As such, scholars and practitioners began seeing residence halls as sites of educational

potential. Schroeder and Mabel (1993) would later write in Realizing the Educational Potential

of Residence Halls:

20
 What distinguishes group living in a campus residence from most other forms of

housing is the involvement of both professional and paraprofessional staff

members in providing intentional, as opposed to random, educational experiences

for students. Students living in residence halls participate in more extracurricular,

social, and cultural events; are more likely to graduate; and exhibit greater

positive gains in psychosocial development, intellectual orientation, and self-

concept than students living at home or commuting. In addition, they demonstrate

significantly greater increases in aesthetic, cultural, and intellectual values; social

and political liberalism; and secularism. (p. 23)

Professional and paraprofessional housing staff were now seen as educators, the ones able to

create these enriching experiences, no longer only a housing manager.

The Twenty-first Century

As institutions enter the twenty-first century, the rich story of housing students on

campus now adds to the narrative an increasingly diverse student body, criticism challenging its

utility given its cost, and an effort to formalize the educational contribution living on campus

adds to a student’s life through residential curricula and living-learning communities.

Marcus (2016) writes “as calls intensify for more diversity at universities and colleges,

some students and researchers say socioeconomic and racial segregation on campuses is instead

on the rise” (¶4). Despite an increasingly diverse student body, students are not interacting.

Tienda (2013) describes the shift from the 1960s to 2000, noting that “[a]fter decades of relative

stability, the U.S. racial landscape changed quickly: in the span of just 40 years, the ‘non-White’

population share more than doubled, rising from less than 17% to 38%” (p. 468). Despite having

had limited access to predominantly White institutions of higher education, Students of Color

often lived in segregated housing or off-campus in housing that was far away from the central
21
parts of campus (Perkins, 1997; Williamson, 2003). These students’ lived experiences differed

dramatically from their majority peers. Where the students live and in what type of setting

influences students’ relationships with their peers (Brandon, Hirt, & Cameron, 2008).

Additionally, as racial diversity has increased, so too have microaggressions and racist incidents

within campus housing (Harper, Davis, Jones, McGowan, Ingram, & Platt, 2011; Harwood,

Huntt, Mendenhall, & Lewis, 2012; Strayhorn & Mullins, 2012). Housing officials must now

work with these diverse students, within federal limitations of which questions can be asked on

housing applications, in an effort to satisfy and challenge their students to grow and develop

(Marcus, 2016).

In addition to a growing diverse student body, the cost of living on campus has also

surged. In the past decade, the expenses for living on campus at four-year public institutions has

gone up 25 percent above inflation (Khrais, 2015). The media has addressed the rising costs of

attending postsecondary institutions, pointing to the increased extravagance associated with

living in on campus housing. Matt Taibbi (2013) notes the “university-tuition system really is

exploitative and unfair, designed primarily to benefit two major actors” (¶ 9): colleges and

universities and those who build “extravagant athletic complexes, hotel-like dormitories and God

knows what other campus embellishments” (¶ 10). La Roche, Flanigan, and Copeland (2010)

describe the situation noting “what were once considered to be luxuries in student housing—

kitchens, private bedrooms, private bathrooms, social spaces and lounges—are now expected”

(p. 46). These lavish new residence halls, to some degree, are ways for institutions to separate

themselves in the college choice process, using fancy buildings as a recruitment technique. To

these critics, the purpose of postsecondary education is training for future careers only, and thus

living on campus plays no role in students’ overall educational experiences. However, faculty

22
and administrators argue that “the tradition of liberal learning has always viewed higher

education as more than training for the marketplace, and the residential principle assumes that it

is more than a training of intellect. A university exists to promote conversation—between

generations, among teachers, and among students.” (Ryan, 2001, p. 23). These new residence

halls are designed to promote social interactions among students (Biemiller, 2017). That is,

living on campus could drive educational gains beyond training for a specific job. With the

increase in what some consider luxury living, it is no wonder people are questioning the benefit

of living on campus.

Additionally, as living on campus has become a mainstay of the campus ethos,

professionals working in residence life have worked to argue for an intentional, curricular

approach towards their practice (Kerr & Tweedy, 2006; Ryan, 2016). This intentional shift from

programmers to educators aims to center student learning and development as a main mission of

residence life and housing. Modeling the sequenced and structured way of the traditional

academic curriculum, a curricular approach to student learning in the residence halls “invokes a

commitment to holistic learning, high-impact practices, and collaborations between student

affairs and academic affairs” (Kerr, Tweedy, Edwards, & Kimmel, 2017, p. 30). This

professional approach to housing and working with students living on campus differs

dramatically from the start of the 1900s and their main scope of safety and security. While safety

and security are still essential functions of housing administrators, professionals are also

asserting the learning potential of living on campus (Keeling, 2004, 2006).

Campus Residence Research

Within the United States, while postsecondary education is facing challenges related to

“economic agendas, shifting demographics, increasingly diverse student populations, public

demand for quality and accountability, and faculty concerns…residence halls have an
23
opportunity to shape the transformation of higher education” (Pascarella, Terenzini, & Blimling,

1994, p. 73). History has suggested that living on campus has been seen as an opportunity to

directly promote broad student outcomes, including psychosocial development, student

engagement, and those related to academics. As such, scholars and student development theorists

have studied living on campus, offering theoretical and empirical support for residence halls and

the act of living on campus (Astin, 1993; Blimling & Whitt, 1999; Evans, Forney, Guido, Patton,

Renn, 2010; Mayhew et al., 2016; Pascarella & Terenzini, 1991, 2005). This section examines

the literature and theory supporting living on campus.

One of the earliest works finding support for living on campus comes from Chickering’s

(1974) Commuting versus Resident Students. He ascertains that resident students’ “presence on

campus, their easy access to pertinent information and to the grapevines that carry it, make it

more possible for them to find educational programs and experiences that suit their interests and

abilities (p. 105). In 1985, Alexander Astin would further Chickering’s argument and suggest

that simply “by virtue of eating, sleeping, and spending their waking hours in the college

campus” (p. 145), students living on campus are more likely to identify as being a college

student and subsequently with their institution. Researchers have used the term “identity

centrality” to describe an enduring tendency to think of oneself consistently through the lens of a

particular identity; in other words, this centrality identity describes a relationship between one’s

self-concept and the ability to answer the question “who am I?” (e.g., Sellers, Rowley, Chavous,

Shelton, & Smith, 1997). Astin’s argument, then, might be more broadly related to and in

support of student identity centrality (Holmes, Bowman, Murphy, & Carver, in press). That is,

for students living on campus, they are more likely to identify themselves as a student because of

their physical location and presence on campus. In fact, both scholars recognized that living on

24
campus may influence college students’ growth and development solely because of the proximity

to the campus itself. That is, living on campus may offer an indirect mechanism by which there

is an effect on a variety of student outcomes.

In the foundational review of college student research, How College Affects Students,

Pascarella and Terenzini (1991) argue for the significance of residence life by concluding that

living on campus was “the single most consistent within-college determinant of the impact of

college” (p. 611). Identifying with an institution and living on campus affords students

opportunities for cognitive, moral, and psycho-social growth and development. Pascarella,

Terenzini, and Blimling (1994) would further note that living on-campus is “qualitatively

different” from living off-campus because “living on campus will maximize opportunities for

social, cultural, and extracurricular involvement, and this increased involvement will account for

residential living’s impact on various indices of student development” (p. 25). These ideas would

spark other researchers to further explore the effect of living on campus.

Pascarella and Terenzini (2005), in the second edition of How College Affects Students,

would refine their earlier argument regarding living on campus, suggesting that the effect is more

likely an indirect effect than a direct one due to increased literature capturing broader student

identities. Pascarella and Terenzini suggest this shift in their findings rests largely in the fact that

a vast majority of the research cited in the first edition favored the experiences of full-time,

traditionally-aged, White male students attending four-year institutions. As such, Pascarella

(2006) would note that conditional effects research as one of the most pressing issues in student

development research after the publication of the second edition of How College Affects

Students. He urges researchers to better understand how different college outcomes affect

minoritized students, differently. In other words, Pascarella pushed researchers to explore the

25
conditional effects of campus experiences. In other words, one could interpret this as a call to

center the stories of those students historically underrepresented in the research.

In the most recent edition of How College Affects Students, Mayhew, Rockenbach,

Bowman, Seifert, and Wolniak (2016) continue the conversation about research focused on

campus residence. The authors note “campus residence promotes retention and possibly

promotes learning, but it predicts decreases in psychological well-being. Evidence is mixed for

cognition, values, diversity attitudes, and academic self-concept” (p. 557). For some of the

outcomes, the researchers question why the relationships are “modest and inconsistent” (p. 545)

relative to the earlier research. The authors suggest several possible reasons: more robust

methodologies reducing positive effects, an increase of technology for communication

diminishing the immersive experience of living on campus, and sparse research in the twenty-

first century focused specifically on living on campus. The new decade of research highlights

that the question regarding the effect of campus residence remains largely unanswered, leaving

opportunity for future scholars and researchers to better understand this collegiate way of living.

Proposal for Empirical Studies

A critical quantitative approach grounds all three studies. In furthering the necessity for

critical quantitative inquiry, Stage and Wells (2014) write that rather than just “focusing on

explanation or fairness, the focus is on equity concerns that can be highlighted through analysis

of large data sets and by examining differences by race, class, and gender” (p. 5). In this regard,

these studies aim to explore the effect of campus residence across a variety of subgroups instead

of just focusing on all students. While critical frameworks were not explicitly cited within each

empirical study as a conceptual framework, they were instrumental in my conception of the

research questions, the specific outcomes analyses I undertook, as well as my interpretations of

the findings.
26
 As a researcher, I aim to disaggregate narratives that cluster all students together and fail

to account for differential experiences of students based on other forms of their identity. As a

first-generation student, I know that my experience on campus and through the educational

system was qualitatively different from my peers who had parents that had enrolled and

completed postsecondary education. Additionally, my work for nearly a decade as a residence

life professional helped me see firsthand how students from majority backgrounds experience

college and the residence halls differently from their peers. I want to help uncover the stories of

those students who have gotten lost in the aggregation of narratives about the student experience

so that we, as educational professionals, can ensure that our work is serving all students

equitably.

These studies also seek to address limitations and concerns in the current body of

literature around campus residence. First, all three empirical studies proposed here contain

several characteristics that permit stronger causal evidence and generalizable conclusions

regarding the effect of living on campus over the current published literature. Relatedly, these

proposed studies do not rely on data from a single institution; rather they are based on data from

a diverse set of institutions. Using data from The Wabash National Study of Liberal Arts

Education (WNS), the dataset includes a diverse set of institutions providing variability in the

types of experiences and living arrangements that students experienced. This diversity improves

the empirical studies’ generalizability. That is, using a large number of diverse institutions better

represents the larger population of four-year postsecondary institutions and the students enrolled

between the late 2000s and early 2010s.

Second, the WNS is longitudinal in design; the first time point includes a host of pretest

measures for a number of measures which allows for better modeling of student growth over the

27
course of four years. The longitudinal design of the study addresses limitations of both cross-

sectional and self-reported gain data that are prevalent in prior studies. Given the ability to use

pretests as covariates, the dataset for this study allows me to control for students entering

abilities and characteristics in analyses that look at either the end of first year or fourth year

outcomes. That is, using these precollege covariates in the analyses allows for the examination of

change during students’ college years. A third, but related point, refers to the instrumentation

captured within the WNS. Students answered measures such as the Collegiate Assessment of

Academic Proficiency (CAAP), the Ryff Scales of Psychological Well-Being (SPWB), the Need

for Cognition Scale (NCS), and Positive Attitude Toward Literacy Scale (PATL) that have been

previously psychometrically evaluated and standardized. The use of these types of valid and

reliable measures as outcome variables offer more sound evidence than single-item measures.

Fourth, the analytic sample for these three proposed empirical studies ranges from

approximately 6,000 to 8,000 students. The large sample sizes provide power for the analyses to

detect small effects. A large portion of the literature published regarding campus residence does

not contain samples of this size. These larger sample sizes also allow me to construct more

complex statistical models. Multilevel models estimating fixed and random effects require more

data to be able to more precisely estimate parameters within the model (Hamilton, 2012; Snijders

& Bosker, 2011). Having access to thousands of individuals, I am able to more confidently

construct these complex models. Additionally, a larger analytic sample allows me to examine

potential moderators within each analysis and see if any meaningful subgroup analyses based on

race, first-generation status, and high-school ability are occurring.

Finally, these studies employ propensity score matching, a technique that is not

traditionally seen in campus residence literature (for perhaps the lone exception, see Schudde,

28
2011). Failure to account for self-selection bias into the treatment, in this case living on campus,

results in parameters and results that are biased. Using propensity score matching, I can

statistically attempt to reduce self-selection bias resulting in analyses that is akin to a randomized

experiment (Guo & Fraser, 2015). While not a true randomized experiment, the results from

these studies using quasi-experimental methodology are better suited for causal arguments than

other research and analytic designs (Angrist & Pischke, 2014; Murnane & Willett, 2011).

The findings from these studies seek to inform three audiences. First, my intent is that

these studies can better inform student affairs practitioners, specifically housing administrators,

about the effect of campus residence on psychological well-being, health related outcomes,

student engagement, and academic outcomes. The findings will help inform narratives about the

effect of living on campus. Second, I aim for these studies to help inform policy. Institutions

across the country are moving to live-on requirements, often using the positive benefits of living

on campus as the logic for these rules. Having studies that can use quasi-experimental

methodology to eliminate bias can better inform these policy makers. Third, I wish to inform

future areas of study for higher education researchers. While the dataset used for the three

empirical studies is longitudinal and includes numerous pre-college measures, it has its

limitations. Thus, I hope this dissertation fills in gaps in the current literature while subsequently

providing future directions for scholars interested in campus residence.

29
 References

Akers, A. (2010). Determination of the optimal number of strata for bias reduction in propensity

score matching (Doctoral dissertation). Retrieved from ProQuest LLC. (ED518529)

Allmendinger, Jr., D. F. (1973). The dangers of ante-bellum student life. Journal of Social

History, 7(1), 75-85.

American Council on Education. (1937). The student personnel point of view. (American

Council on Education Studies, series 1, no. 3). Washington, DC: Author.

American Council on Education, Committee on Student Personnel Work. (1949). The student

personnel point of view. (Rev. ed.: American Council on Education Studies, series 6, no.

13). Washington, DC: Author.

American College Personnel Association. (1996). The student learning imperative: Implications

for student affairs. Retrieved from http://www.myacpa.org/files/acpas-student-learning-

imperativepdf

Angrist, J. D., & Pischke, J. S. (2014). Mastering' metrics: The path from cause to effect.

Princeton, NJ: Princeton University Press.

Astin, A. W. (1977). Four critical years. Effects of college on beliefs, attitudes, and knowledge.

San Francisco, CA: Jossey-Bass Publishers.

Astin, A. (1985). Achieving education excellence: A critical assessment of priorities and

practices in higher education. San Francisco, CA: Jossey-Bass.

Astin, A.W. (1991). Assessment for excellence: The philosophy and practice of assessment and

evaluation in higher education. Washington, DC: American Council on Education/Oryx

Press Series on Higher Education.

Astin, A. W. (1993). What matters in college? Four critical years revisited. San Francisco, CA:

Jossey-Bass Publishers.
30
Austin, P. C. (2011). An introduction to propensity score methods for reducing the effects of

confounding in observational studies. Multivariate Behavioral Research, 46, 399–424.

Biemer, P. P., & Christ, S. L. (2008). Weighting survey data. International handbook of survey

methodology, 317-341.

Blimling, G. S. (1993). The influence of college residence halls on students. Higher education:

Handbook of theory and research, 9, 248-307.

Blimling, G. S. (1999). A meta-analysis of the influence of college residence halls on academic

performance. Journal of College Student Development, 40(5), 551.

Blimling, G. S., & Whitt, E. J. (1999). Good practice in student affairs: Principles to foster

student learning. San-Francisco, CA: Jossey-Bass.

Bowman, N. A., & Felix, V. (2017). It’s who I am: Student identity centrality and college

student success. Journal of Student Affairs Research and Practice, 54, 235-247.

Brandon, A., Hirt, J. B., & Cameron, T. (2008). Where you live influences who you know:

Differences in student interaction based on residence hall design. Journal of College &

University Student Housing, 35(2), 62-79.

Brookhart, M. A., Schneeweiss, S., Rothman, K. J., Glynn, R. J., Avorn, J., & Stürmer, T.

(2006). Variable selection for propensity score models. American Journal of

Epidemiology, 163(12), 1149-1156.

Campbell, C. M., & Cabrera, A. F. (2011). How sound is NSSE?: Investigating the psychometric

properties of NSSE at a public, research-extensive institution. The Review of Higher

Education, 35(1), 77-103.

Chickering, A. W. (1974). Commuting versus resident students. San Francisco, CA: Jossey-

Bass.

31
Cochran, W. G. (1968). The effectiveness of adjustment by subclassification in removing bias in

observational studies. Biometrics, 24, 205-213.

Columbia University Bulletin of Information. (1923). Annual report of the President of

Columbia University. Princeton, NJ: Princeton University.

Committee on the Student in Higher Education. (1968). The student in higher education. New

Haven, CT: Hazen Foundation.

Conrad, C. F., & Johnson, J. (2008). College and university curriculum: Placing learning at the

epicenter of courses, programs and institutions. 2nd Ed. ASHE Reader Series. Boston,

MA: Pearson Custom Publishing.

Cowley, W. H. (1934a). The history of student residential housing. School and Society,

40(1040), 705 – 712.

Cowley, W. H. (1934b). The history of student residential housing II. School and Society,

40(1041), 758 – 764.

DeCoster, D. A., & Mable, P. (1974). Student development and education in college residence

halls. Washington, DC: American College Personnel Association.

Drasgow, F. (1986) Polychoric and polyserial correlations. In S. Kotz and N. Johnson (Eds.), The

Encyclopedia of Statistics, Volume 7, pp. 68-74. Hoboken, New Jersey: Wiley and Sons.

Eaton, C., Habinek, J., Goldstein, A., Dioun, C., Santibáñez Godoy, D. G., & Osley-Thomas, R.

(2016). The financialization of US higher education. Socio-Economic Review, 14(3), 507-

535.

Evans, N. J., Forney, D. S., Guido, F. M., Patton, L. D., & Renn, K. A. (2009). Student

development in college: Theory, research, and practice. Hoboken, NJ: John Wiley &

Sons.

32
Evans, N. J., & Reason, R. D. (2001). Guiding principles: A review and analysis of student

affairs philosophical statements. Journal of College Student Development, 42(4), 359-

377.

Flowers, L. A. (2004). Effects of living on campus on African American students' educational

gains in college. NASPA Journal, 41(2), 277-293.

Gilligan, C. (1982). In a different voice. Boston, MA: Harvard University Press.

Gott v. Berea College, 156 Ky. 376, 161 S.W. 204 (1913).

Groves, R. M., Fowler Jr, F. J., Couper, M. P., Lepkowski, J. M., Singer, E., & Tourangeau, R.

(2009). Survey methodology (2nd ed.). Hoboken, NJ: John Wiley & Sons.

Guo, S., & Fraser, M. (2015). Propensity score analyses: Statistical methods and applications.

2nd ed. Thousand Oaks, CA: SAGE Publications, Inc.

Harwood, S. A., Huntt, M. B., Mendenhall, R., & Lewis, J. A. (2012). Racial microaggressions

in the residence halls: Experiences of students of color at a predominantly white

university. Journal of Diversity In Higher Education, 5(3), 159-173.

Harper, S. R., Davis, R. J., Jones, D. E., McGowan, B. L., Ingram, T. N., & Platt, C. S. (2011).

Race and racism in the experiences of Black male resident assistants at predominantly

White universities. Journal of College Student Development, 52(2), 180-200.

Haskins, C. H. (1957). The rise of universities. Ithaca, NY: Cornell University Press.

Hayes, H. (1932). Planning residence halls for undergraduate students in American colleges and

universities: A handbook for the use of college officers and members of building

committees. New York, NY: Teachers College, Columbia university.

33
Hevel, M. S. (2017). A historiography of college students 30 years after Helen Horowitz’s

campus life. In M. B. Paulsen (Ed.), Higher Education: Handbook of Theory and

Research (pp. 419-484). New York, NY: Springer International Publishing.

Hirano, K., Imbens, G. W., & Ridder, G. (2003). Efficient estimation of average treatment

effects using the estimated propensity score. Econometrica, 71(4), 1161-1189.

Holmes, J. M., Bowman, N. A., Murphy, M., & Carter, E. (in press). Envisioning college

success: The role of student identity centrality. Social Psychology of Education.

Holmes, W. M. (2013). Using propensity scores in quasi-experimental designs. Thousand Oaks,

CA: Sage.

Horowitz, H. L. (1987). Campus life: Undergraduate cultures from the end of the eighteenth

century to the present. New York, NY: Alfred a Knopf Incorporated.

Ireland, C. (2012, April 19). Harvard’s long-ago student risings. The Harvard Gazette. Retrieved

from https://news.harvard.edu/gazette/story/2012/04/harvards-long-ago-student-risings/

Johnson, A. (2015, December 11). Student protests, then and now. The Chronicle of Higher

Education, Retrieved from https://www.chronicle.com/article/Student-Protests-

ThenNow/234542

Johnson, W. G., & Cavins, K. M. (1996). Strategies for enhancing student learning in residence

halls. New Directions for Student Services, 75, 69-82.

Kafka, A. (2018, August 31). This fall’s housing crunch is so urgent one university is asking its

professors for help. The Chronicle of Higher Education. Retrieved from

https://www.chronicle.com/article/This-Fall-s-Housing-Crunch/244431?key=ONA-

J8qTe05O7njbTd0tJ9LJRfHZ1hcjdQZN6NKLzhSto2drMSzOk9P1RAP8OuadNC1IRW

ppeFNSZTZsZmpCZzhQdDczdi1oVHhmZWVwSjVFRXRGTWxocmhoUQ

34
Kanoy, K., & Bruhn, J. (1996). Effects of a first-year living and learning residence hall on

retention and academic performance. Journal of The First-Year Experience & Students in

Transition, 8(1), 7-23.

Keeling, R. P. (Ed.). (2004). Learning reconsidered. Washington, DC: National Association of

Student Personnel Administrators.

Keeling, R. P. (Ed.). (2006). Learning reconsidered 2. Washington, DC: National Association of

Student Personnel Administrators.

Kennedy, R. (2015). For discrimination: Race, affirmative action, and the law. New York, NY:

Vintage Books.

Kerr, K. G., & Tweedy, J. (2006). Beyond seat time and student satisfaction: A curricular

approach to residential education. About Campus, 11(5), 9-15.

Kerr, K. G., Tweedy, J., Edwards, K. E., & Kimmel, D. (2017). Shifting to curricular approaches

to learning beyond the classroom. About Campus, 22(1), 22-31.

Kezar, A. J. (2006). The impact of institutional size on student engagement. NASPA

Journal, 43(1), 87-114.

Khrais, R. (2015, September 15). Where to live when you're a college student. National Public

Radio. Retrieved from https://www.npr.org/sections/ed/2015/09/15/439467590/where-to-

live-when-youre-a-college-student

Kohlberg, L. (1958). The development of modes of moral thinking and choice in the years 10 to

16. Chicago, IL: University of Chicago Press.

Kohlberg, L. (1976). Moral stages and moralization: The cognitive-development

approach. Moral development and behavior: Theory research and social issues, 31-53.

35
Kohlberg, L. (1981). Essays on moral development: The philosophy of moral development (Vol.

1). San Francisco, CA: Harper & Row.

Kuh, G. D. (2009). The national survey of student engagement: Conceptual and empirical

foundations. New directions for institutional research, 2009(141), 5-20.

LaNasa, S. M., Olson, E., & Alleman, N. (2007). The impact of on-campus student growth on

first-year student engagement and success. Research in Higher Education, 48(8), 941-

966.

La Roche, C. R., Flanigan, M. A., & Copeland Jr, P. K. (2010). Student housing: Trends,

preferences and needs. Contemporary Issues in Education Research, 3(10), 45-50.

Lee, P. (2011). The curious life of in loco parentis in American universities. Higher Education

Review, 8, 65-90.

Levine, D. O. (1986). The American college and the culture of aspiration, 1915 – 1940. Ithaca,

NY: Cornell University Press.

Little, R. J. A. (1988). A test of missing completely at random for multivariate data with missing

values. Journal of the American Statistical Association, 83(404), 1198–1202.

López Turley, R. N., & Wodtke, G. (2010). College residence and academic performance: who

benefits from living on campus?. Urban Education, 45(4), 506-532.

Marcus, J. (2016, September 26). Why segregation in college increases after freshman year. The

Atlantic. Retrieved from

https://www.theatlantic.com/education/archive/2016/09/segregated-by-

dormitory/501602/

36
Mayhew, M. J., Bowman, N. A., Rockenbach, A. N., Seifert, T. A., & Wolniak, G. C.

(2016). How college affects students: 21st century evidence that higher education

works (Vol. 3). San Francisco, CA: Jossey-Bass.

Mosier, R. (Ed.). (2012). The history of ACUHO-I. Columbus, OH: Association of College &

University Housing Officers – International.

Morison, S. E. (1936). Harvard college in the seventeenth century I. Cambridge, MA: Harvard

University Press.

Murnane, R. J., & Willett, J. B. (2011). Methods matter: Improving causal inference in

educational and social science research. Oxford, England: Oxford University Press.

Myers, J. A., & Louis, T. A. (2007). Optimal propensity score stratification. Baltimore, MD:

Johns Hopkins University. Retrieved from https://core.ac.uk/download/pdf/61318500.pdf

National Survey of Student Engagement. (2018). Engagement indicators. Retrieved from

http://nsse.indiana.edu/html/engagement_indicators.cfm

Nuss, E. M. (1996). The development of student affairs. In S. R. Komives & D. B. Woodard

(Eds.), Student affairs: A handbook for the profession (3rd ed., pp. 22-42). San Francisco,

CA: Jossey-Bass.

Palmer, C., Broido, E. M., & Campbell, J. (2008). A commentary on "The educational role in

college student housing". Journal of College & University Student Housing, 35(2), 86-99.

Pan, W., & Bai, H. (Eds.). (2015). Propensity score analysis: Fundamentals and developments.

New York, NY: Guilford Publications.

Pascarella, E. T. (1985). College environmental influences on learning and cognitive

development: A critical review and synthesis. In J. Smart (Ed.), Higher education:

Handbook of theory and research (pp. 1-61). New York, NY: Agathon.

37
Pascarella, E. T. (2006). How college affects students: Ten directions for future research. Journal

of college student development, 47(5), 508-520.

Pascarella, E. T. (2007). Methodological re-port for the Wabash National Study of Liberal Arts

Education. Iowa City, IA: Center for Research on Undergraduate Education.

Pascarella, E., Bohr, L., Nora, A., Zusman, B., Inman, P., & Desler, M. (1993). Cognitive

impacts of living on campus versus commuting to college. Journal of College Student

Development, 34, 216-220.

Pascarella, E. T., & Terenzini, P. T. (1991). How college affects students: Findings and insights

from twenty years of research. San Francisco, CA: Jossey-Bass.

Pascarella, E. T., & Terenzini, P. (2005). How college affects students (Volume 2). San

Francisco, CA: Jossey-Bass.

Pascarella, E. T., Terenzini, P. T., & Blimling, G. S. (1994). The impact of residential life on

students. In E. T. Pascarella, P. T. Terenzini, & G. S. Blimling (Eds.). Realizing the

educational potential of residence halls, 22-52.

Patrick, A. R., Schneeweiss, S., Brookhart, M. A., Glynn, R. J., Rothman, K. J., Avorn, J., &

Stürmer, T. (2011). The implications of propensity score variable selection strategies in

pharmacoepidemiology – An empirical illustration. Pharmacoepidemiology and Drug

Safety, 20(6), 551-559.

Perkins, L. M. (1997). The African American female elite: The early history of African

American women in the seven sister colleges, 1880–1960. Harvard Educational Review,

67, 718–756.

38
Pike, G. R. (1999). The effects of residential learning communities and traditional residential

living arrangements on educational gains during the first year of college. Journal of

College Student Development, 40(3), 269.

Pike, G. R. (2002). The differential effects of on-and off-campus living arrangements on

students' openness to diversity. NASPA Journal, 39(4), 283-299.

Radford, A. W., Berkner, L., Wheeless, S. C., & Shepherd, B. (2010). Persistence and

attainment of 2003-04 Beginning Postsecondary Students: After 6 years (NCES 2011-

151). Washington, DC: U.S. Department of Education.

Renn, K. A., & Reason, R. D. (2012). College students in the United States: Characteristics,

experiences, and outcomes. Hoboken, NJ: John Wiley & Sons.

Rest, J. R. (1979). The impact of higher education on moral judgment development. Project

Report 1. Minneapolis, MN: University of Minnesota.

Rest, J. R., Narvaez, D., Thoma, S. J., & Bebeau, M. J. (2000). A neo-Kohlbergian approach to

morality research. Journal of Moral Education, 29(4), 381-395.

Riker, H. C. (1965). College housing as learning centers. Washington, DC: American College

Personnel Association.

Rodger, S. C., & Johnson, A. M. (2005). The impact of residence design on freshman outcomes:

Dormitories versus suite-style residences. The Canadian Journal of Higher

Education, 35(3), 83.

Rosenbaum, P. R., & Rubin, D. B. (1984). Reducing bias in observational studies using

subclassification on the propensity score. Journal of the American statistical Association,

79(387), 516-524.

39
Rubin, D. B. 1974. Estimating causal effects of treatments in randomized and nonrandomized

studies. Journal of Educational Psychology, 66, 688–701.

Rubin, D. B. 2001. Using propensity scores to help design observational studies: Application to

the tobacco litigation. Health Services & Outcomes Research Methodology, 2, 169–188.

Rudolph, F. (1990). The American college and university: A history. Athens, GA: University of

Georgia Press.

Ryan, M. B. (2001). A collegiate way of living: Residential colleges and a Yale education. New

Haven, CT: Yale University Press.

Ryan, M. B. (2016). On the future of the collegiate way. In M. Evans & T. Burt (Eds.), The

collegiate way: University education in a collegiate context (pp. 3-16). Rotterdam, The

Netherlands: Sense Publishers.

Saffron, I. (2013, September 18). Granite countertops, flat-screen TVs, fire pits: The surprising

story of how college dorms got luxe. New Republic. Retrieved from

https://newrepublic.com/article/114744/luxification-college-dorms

Schudde, L. T. (2011). The causal effect of campus residency on college student retention. The

Review of Higher Education, 34(4), 581-610.

Schiff, J. A. (1995, December). The way we ate. Yale Magazine, 80. Retrieved from

http://schwarzman.yale.edu/sites/default/files/files/HistoryCommons.pdf

Schroeder, C. C., Mable, P., & Associates. (1993). Realizing the educational potential of

residence halls. San Francisco, CA: Jossey-Bass.

Schafer, J.L. (1997). Analysis of multivariate incomplete data. London, England: Chapman &

Hall Publishers.

40
Seltzer, R. (2017, July 11). Dorm to cost Bethune-Cookman $306 million. Inside Higher

Education. Retrieved from

https://www.insidehighered.com/quicktakes/2017/07/11/dorm-cost-bethune-cookman-

306-million

Snijders, T. A, & Bosker, R. (2011). Multilevel analysis: An introduction to basic and advanced

multilevel modeling. Thousand Oaks, CA: Sage Publications.

Strayhorn, T. L., & Mullins, T. G. (2012). Investigating black gay male undergraduates'

experiences in campus residence halls. Journal of College and University Student

Housing, 39(1), 140-161.

Taibbi, M. (2013, August 15). The college-loan scandal. Rolling Stone. Retrieved from

https://www.rollingstone.com/politics/news/ripping-off-young-america-the-college-loan-

scandal-20130815

Thelin, J. R. (2004). A history of American higher education. Baltimore, MD: Johns Hopkins

University Press.

THE Project. (1975). A student development model for student affairs in tomorrow's higher

education. Journal of College Student Personnel, 16, 334-341.

Thoemmes, F. J., & Kim, E. S. (2011). A systematic review of propensity score methods in the

social sciences. Multivariate behavioral research, 46(1), 90-118.

Tienda, M. (2013). Diversity ≠ inclusion promoting integration in higher education. Educational

Researcher, 42(9), 467-475.

Tinto, V. (1993). Leaving college: Rethinking the causes and cures of student attrition (2nd ed.).

Chicago, IL: University of Chicago Press.

41
Transcript of National Industrial Recovery Act. (1933). National Archives. Retrieved from

https://www.ourdocuments.gov/doc.php?flash=false&doc=66&page=transcript

U.S. Department of Housing and Urban Development. (n.d.). Barriers to success: Housing

insecurity for U.S. college students. Insights into housing and community development

policy. Retrieved from https://www.huduser.gov/portal/periodicals/insight/insight_2.pdf

Weber, D. (2016, September 13). Fancy dorms aren’t the main reason tuition is skyrocketing.

FiveThirtyEight. Retrieved from https://fivethirtyeight.com/features/fancy-dorms-arent-

the-main-reason-tuition-is-skyrocketing/

Weinberg, A. (2013, September 13). Not just a place to sleep. Inside Higher Ed. Retrieved from

https://www.insidehighered.com/views/2013/09/13/essay-calls-more-ambitious-concept-

residence-life

Westreich, D., Cole, S. R., Funk, M. J., Brookhart, M. A., & Stürmer, T. (2011). The role of the

c-statistic in variable selection for propensity score models. Pharmacoepidemiology and

Drug Safety, 20(3), 317-320.

Williamson, J. A. (2003). Black power on campus: The University of Illinois, 1965-75.

Champaign, IL: University of Illinois Press.

42
 CHAPTER TWO: COLLEGIATE RESIDENCE AND ACADEMIC ACHIEVEMENT,
RETENTION, GRADUATION, AND SATISFACTION

Living on campus is not a new practice within higher education. Historians of higher education

have long marked the role of campus residence in students’ postsecondary experiences (Cowley,

1934a, 1934b; Rashdall, 1936; Rudolph, 1990; Stewart, 1942; Thelin, 2004). However, critics of

higher education question the return of investment students receive from contemporary iterations

of living on campus, pointing to increased housing costs and extravagances associated with

living in on campus housing. La Roche, Flanigan, and Copeland Jr. (2010) describe housing

situations by articulating that “what were once considered to be luxuries in student housing—

kitchens, private bedrooms, private bathrooms, social spaces and lounges—are now expected”

(p. 46). Taibbi (2013) adds that colleges and universities build “extravagant athletic complexes,

hotel-like dormitories and God knows what other campus embellishments” (¶ 10). Saffron

(2013) adds that college residences now have extravagant luxuries such as granite counters,

private bathrooms, flat-screen TVs in lounges, fire pits, and lazy rivers.

The subtext of these arguments and critiques is that these new and elaborate residence

halls with increased room and board expenses are financial burdens to students and offer no

direct value to the educational experience. That is, as universities have shifted to providing

housing with higher end amenities, there seems to not be as much of a focused consideration on

their value within a student’s educational experience and their academic achievement. A flat-

screen TV and granite countertop might attract students in the first place, but do these amenities

keep them enrolled in college and satisfied with the experience? Do these amenities affect

students’ academic achievement, retention, and graduation in any way? More broadly, does

living on campus have a direct effect on students’ achievement, retention, or graduation?

43
 Research on Campus Residence, Academic Achievement, and Retention

A substantial body of work explores the relationship between residence and academic

achievement, retention, and the effect of students' collegiate residence across time. One of the

earliest calls comes from Riker (1965) who published a piece calling for residence halls to be

positioned as centers for learning. A decade later, Chickering (1974) asserted that resident

students’ “presence on campus, their easy access to pertinent information and to the grapevines

that carry it, make it more possible for them to find educational programs and experiences that

suit their interests and abilities” (p. 105). In his research, Chickering found that students living

on campus had a positive effect on students as a result of the fundamental different type of

relationships students form with significant others. Astin (1985) further refined these arguments

noting that by virtue of living on campus and spending their entire lives within a postsecondary

environment, students were more likely to identify as being a college student and also with their

institution. Astin (1991) would later assert that these campus identification play a role in various

student outcomes. Challenging these positive narratives in regard to campus residence, Blimling

(1989) would synthesize empirical literature from the mid-1960s to the mid-1980s and find that

campus residence does not directly affect students’ academic performance. His findings align

with the first volume of How College Affects Students, whose authors asserted that living on

campus is “the single most consistent within-college determinant of the impact of college”

(Pascarella & Terenzini, 1991, p. 611). However, when specifically examining the net effect of

residence on academic achievement and graduation, they note that living on campus positively

influences persistence and graduation but the actual net effect that can be attributed to living on

campus is unclear given a lack of adequate statistical data. Additionally, they wrote that living on

44
campus does not lead to a net benefit in terms of academic achievement, after controlling for

differences in past academic performance.

Additional research from the 1990s further complicates the association between living on

campus and these outcomes. In a volume of New Directions for Student Services, research

findings suggested links between living on campus, academic achievement, and persistence to

graduation (Johnson, 1996). These assertions added support to the argument that residence halls

are less of a place to just sleep and socialize and more a site to promote learning within

institutions of higher learning (Blimling, 1993, 1999; Kanoy, & Bruhn, 1996). First year students

who lived on campus have higher rates of retention, a greater degree of academic progress, and

higher academic performance (Thompson, Samiratedu, & Rafter, 1993). However, in large

college town environments where students who lived off were still within close walking distance

to classes and university resources, the differences in academic did not exist (Delucchi, 1993).

While not measuring academic achievement directly, Pascarella, Bohr, Nora, Zusman, Inman,

and Desler (1993) found that living on campus led students to larger gains in cognitive skills and

critical thinking, with no difference existing in students’ reading comprehension or mathematical

abilities.

Twenty-first century research that pertains to campus residence more broadly focused on

disaggregating this experience by various student identities. For Pike and Kuh (2005), on-

campus residence was important for first-generation students, suggesting that first-generation

students living off campus had lower levels of academic engagement and educational aspiration.

These lower levels could be related to a differential peer network or in response to different time

demands, as students living off campus reported working more hours during the week leaving

them less time to engage with campus activities (Newbold, Mehta, & Forbus, 2011). Focusing

45
solely on Black students, Flowers (2004) found that living on campus in residence halls was

positively related to measures of personal and social development. He suggested that these social

skills and connections were essential for students’ academic achievement. López Turley and

Wodtke (2010) used a broader, more representative sample of college students hoping to better

understand the effect living on campus has on students’ academic achievement, namely students’

GPAs. The authors did not find a significant direct relationship between students’ campus

residence and 1st-year grades despite a sample of 2,011 students from 372 institutions; however,

there was significance when viewing specific subgroups of students. Black students living on

campus had higher GPAs than their Black peers living off campus with family. These research

findings follow conclusions from college impact reviews that shifted the narrative living on

campus from that of having a direct effect to more of an indirect role on student outcomes

(Pascarella & Terenzini, 2005). These indirect effects on academic achievement, retention, and

graduation might be linked to qualitatively different peer networks rather than the specific act of

living on campus.

In the most recent edition of How College Affects Students, the authors note that living on

campus is positively related to retention while possibly having an effect on learning, which

supports some earlier research findings while contradicting others (Mayhew, Rockenbach,

Bowman, Seifert, & Wolniak, 2016). The authors suggest that these differential findings within

the research are related to contemporary methodologies that can better estimate causality, a

substantial limitation of earlier collegiate residence research. Research that uses these more

appropriate methodologies offer more rigorous findings pertaining to students’ collegiate

residence. Oseguera and Rhee (2009) used multi-institutional data and hierarchical generalized

linear models to find that living on campus increased the probability of graduating within six

46
years or still being enrolled within the same institution by 4.5 percent. While these findings add

to literature and suggest positive effects from living on campus, it is worth noting that the

researchers did not differentiate among students were still enrolled after six years at the same

institution and those who actually graduated, possibly skewing their findings. Additionally, the

living on campus variable was based on students’ reported plans upon entering school and may

not have captured whether students truly lived on campus. Finally, the researchers also did not

account for the bias associated with students’ choice to live on campus.

To account for the bias associated with students’ self-selection into the “treatment” of

living on campus, researchers have turned to quasi-experimental methodologies. Using

propensity score modeling, Schudde (2011) found that for students living on campus, the

probability of remaining enrolled for their sophomore year was 3.3 percent higher than compared

to those students living off campus. While her methodology statistically controlled for bias

associated with students’ choice of residence, the study only focused narrowly on second-year

retention and did not explore any conditional effects to interrogate a prevailing assumption that

living on campus is positive and beneficial for everyone. Using instrumental variables, de Araujo

and Murray (2010) found that living on campus had a direct effect on students’ GPAs. The

researchers used two instrumental variables: distance of hometown from campus and a dummy

variable representing whether a student was denied housing due to space limitations. In their

estimation, “having lived on campus during any time in the past caused an increase in semester

GPA and cumulative GPA of up to half a letter grade” (p. 877). To confirm their results, the

authors used multiple instrumental variable techniques, finding consistent patterns across all

models. While promising, their findings are based on a single institutional study and may be a

result of the specific institution rather than the act of living on campus. While these recent

47
studies offer a better estimate of the effect of living on campus, some do not use longitudinal

data, while others are limited in the precollege covariates that are included in the models.

Conceptual Framework

This study was conceptually framed by two theories: (1) college impact models (e.g.,

Pascarella & Terenzini, 2005) and (2) Renn and Arnold’s (2003) adaptation of Bronfenbrenner’s

(1979, 1989, 1993) Process-Person-Context-Time (PPCT) ecological systems theory. First,

existing theories on the impact of college aim to understand broadly the factors by which college

affects students (e.g., Astin, 1991; Tinto, 1993). These models offer researchers broad categories

of covariates for inclusion when estimating the effect of college, suggest using pretests to capture

students’ baseline capabilities, as well as recommending the use of longitudinal datasets, when

possible, to model growth.

Bronfenbrenner’s (1979, 1989, 1993) Process, Person, Context, and Time (PPCT) model

counters these college impact models by situating an individual within a complex series of

systems (macrosystems, exosystems, mesosystems, and microsystems), and it is through

interactions with these complex environments that growth and development occurs.

Bronfenbrenner suggests that the functional relationship between persons and their environments

affect individuals and offers insight as to why, for this study in particular, collegiate residence

could be related to students’ academic achievement, retention, and graduation. Renn and Arnold

(2003) extended Bronfenbrenner’s work to higher education conceptualizing the four types of

systems (micro, meso, exo, and macro) are from a collegiate context. For many students’,

particularly traditionally-aged students whom make up a majority of this analytic sample, their

microsystems are associated within a college or university and classrooms, living environments,

campus facilities, and other daily-life environments. For residential students, their systems are

48
likely to overlap with a consistent group of other students also living on campus. In other words,

one’s living environment creates unique opportunities for person-context interactions that differ

for on and off campus students.

The Current Study

This study aims to address the limitations of previous research through use of quasi-

experimental methodology in an effort to reduce self-selection bias (Guo & Fraser, 2015).

Students’ self-selection into living on campus is a notable problem with studying these

dynamics, since the students who choose to live on campus differ in numerous ways from those

who do not (Newbold, Mehta, & Forbus, 2011; Schudde, 2011). Namely this study uses

information from each of the three categories discussed above to generate a propensity score

balanced on each of the observed covariates. This balance statistically creates groups that are

equal in expectation based on the covariates used allowing for comparison of students who live

on campus to their peers off campus avoiding bias related to these covariates and the treatment of

living on campus (Austin, 2011; Guo & Fraser, 2015; Hong & Raudenbush, 2006; Mitra &

Reiter, 2012; Thoemmes & Kim, 2011). Moreover, conditional effects from a quasi-experimental

approach exploring the effect of student characteristics and collegiate residence have not been

examined. For instance, racial dynamics on campus may be quite salient, and this tension might

impact the experiences and outcomes for those living on campus across multiple students’

identities.

To expand upon the previous literature, the present study uses large, longitudinal, multi-

institutional dataset as well as propensity score modeling to reduce the impact of self-selection

related to students’ choice of collegiate residence. Since it is nearly impossible, and somewhat

unethical, to randomly assign students to live on or off campus, the use of quasi-experimental

49
methodologies attempts to statistically reduce self-selection bias into the “treatment” of living on

campus. The current literature suggests that because living on campus facilitates a comparatively

different experience than living off-campus, students who live on campus might have higher

GPAs during their collegiate experience, have higher levels of satisfaction with their overall

experience, and be more likely to be retained and graduate at their institution.

Namely, the current study seeks to answer the following research questions:

1. What effect does students’ first-year residence have on their GPA, retention, four-

year graduation, and collegiate satisfaction at the end of their first and fourth year?

2. Is the effect of living on campus on these outcomes moderated by students’

identities?

3. Does living on campus all four years (versus moving off campus after the first year)

affect students’ GPA, graduation, or college satisfaction?

Methods

Institutional Sample and Data Collection

This study’s data includes those students who participated in the Wabash National Study

of Liberal Arts Education (WNSLAE) across 46 four-year institutions in the United States. The

WNSLAE is a longitudinal designed study to examine student experiences and outcomes at

institutions which have a stated commitment to liberal arts education. The institutions

represented were selected to cover a wide range of geographic diversity, size, selectivity, tuition

costs, and missions. The WNSLAE study institutions also include religious and single-sex

colleges and universities. With the focus of WNSLAE on institutional commitment to liberal arts

education, the researchers intentionally oversampled liberal arts colleges.

50
 The WNSLAE researchers followed three cohorts of students who entered college

between 2006 and 2008, collecting data from each cohort at three distinct time points. First, the

researchers collected information about students’ demographic characteristics, precollege

abilities, behaviors, experiences, and attitudes. This first wave of data collection occurred either

in the summer prior to their first semester or early in the fall of their first year of college (Time

1). Follow-up data collection occurred at the end of students’ first year of college (Time 2), with

one final data collection occurring at the end of students’ fourth-year of college (Time 3). During

times 2 and 3, students completed the WNSLAE Student Experiences Survey and the National

Survey of Student Engagement, both of which provided information about their college

experiences, attitudes, and behaviors. Participants also completed instruments during all three

waves of data collection that were designed to measure several dimensions of personal and

intellectual development. Data collection at time one is intended to serve as a baseline or pre-test

for those measures at Time 2 and Time 3.

Analytic Sample

Institutions reported data for students’ GPA and retention, but not for their place of

residence, so the sample was limited to those who responded to the campus residence variable.

Thus, the analytic sample for the current study included 8,120 student responses at Time 2 and

4,042 responses at Time 3. The longitudinal nature of the data collection inevitably leads to

missing data. Among the specific study variables, 0% to 14% of responses were missing.

Limiting the propensity score creation to only complete cases within the dataset would result in

losing a substantial number of cases and, as such, the decision was made to impute missing data

using multiple imputation (MI) methods (Mitra & Reiter, 2016). There is debate regarding the

best way of generating a propensity score when using imputed data, but scholars generally

51
suggest that calculating an unbiased treatment effect should include the following steps: (a)

imputing the dataset with the outcome of interest included during the imputation stage; (b) using

each M imputed dataset to estimate the parameters of the propensity score model; (c) using each

M imputed dataset to calculate an individual’s propensity score; and (d) averaging the M

propensity scores for each individual (Leyrat et al., 2019; Mitra & Reiter, 2016). For this study,

MI was used to create 40 valid but different versions of complete data (Collins, Schafer, & Kam,

2001). Variability from multiple imputations tends to produce larger standard errors rather than a

single-imputation approach, which reduces the likelihood of a Type I error. While dependent

variables were included in the imputation equations, these imputed values for outcome variables

were removed before point-estimate analyses (von Hippel, 2007).

This study also utilizes a weighting algorithm developed by the WNS researchers to

make the sample more representative of the incoming first-year cohorts of those participating

institutions in terms of sex, race, and academic ability. Scholars note that students from

underrepresented groups tend to have higher rates of survey nonresponse as well as college

attrition, and survey weighting helps address this issue. Additionally, students in the study from

larger institutions were randomly sampled from institutional records whereas at smaller

institutions all students were invited to participate, so weighting adjusts for potential differences

between the institutional population and study participants as a result of the probability of being

included in the study and unit nonresponse (see Biemer & Christ, 2008; Groves et al., 2009). The

weights for this study were normalized with a mean of 1.0 to prevent changes to the analytic

sample size. Students within this weighted sample attended a liberal arts college (50 percent), a

research university (32 percent), and a regional college or university (18 percent). Regarding

student demographics, 6 percent identified as Hispanic or Latinx, 7 percent identified as Black or

52
African American, and 8 percent of students identified as Asian or Pacific Islander, while 38

percent of students identified their sex as male.

Measures

Dependent variables. College GPA was measured at the end of the first year and the end

of the fourth year on an eight-point scale (1 = C- or lower, to 8 = A). College satisfaction was

computed as the index of two items that were assessed at those same two time points: “How

would you evaluate your entire educational experience at this institution?” (1 = poor, to 4 =

excellent), and “If you could start over again, would you go to the same institution you are now

attending?” (1 = definitely no, to 4 = definitely yes). These satisfaction indices suggested

internally reliability within this sample (Cronbach’s alpha = .72 in the first year and .76 in the

fourth year). College achievement and satisfaction measures were subsequently standardized

with a mean of zero and a standard deviation of one within the analytic sample, allowing for the

results of the outcomes analyses to be interpreted as Cohen ds (i.e., the standardized mean

difference) between the two groups of students (Cohen, Cohen, West, & Aiken, 2003).

Three college retention variables indicated whether students were enrolled at their initial

institution in the Fall term of the second year, third year, and fourth year. Four-year graduation

indicated whether the student had graduated from that institution at the end of their fourth year.

For all retention and graduation variables, a response of 0 equals no and a response of 1 indicates

yes.

Treatment Variables. The primary independent variable for the first two research

questions indicated students’ campus residence during their first year of college. This variable

was recoded from a measure with four levels to a dichotomous variable indicating living on

campus (yes/no) to increase statistical power. Students who reported living in a sorority or

53
fraternity house were excluded from the analyses, as the data are limited in determining if a

student’s particular organization house was on or off campus. The third research question

compared students who reported living on campus at times 2 and 3 to those students who

reported living on at time 2 and living off campus at time 3. Students who always reported living

off campus were not included in the analysis of the third research question. The decision was

made to compare these two groups in order to constitute a better counterfactual group than to

have three groups that included students who did not live on campus at all.

Propensity Score Variables. The current research questions required two distinct

propensity score models. The first estimates students’ propensity to live on campus during the

first year, while the second model examines students’ propensities to move off campus during

their fourth year after having lived on their first year compared to their peers who reported living

on campus at both time 2 and time 3. The first model utilizes only precollege characteristics

while the second model includes additional covariates to capture students’ experiences during the

first year of college. Both models used institutional fixed effects to account for all between-

institution differences (Allison, 2009; Arpino & Mealli, 2011; Li, Zaslavsky, & Landrum, 2013).

To do this, dummy variables were used for each institution except one, which was omitted as a

referent group. The variables included in the first propensity score model were all precollege and

selected based on the expected impact on students’ collegiate residence decision. The second

propensity score model included additional covariates capturing student experiences in addition

to all variables within the first model. As such, both models included constructs that only directly

predicted the outcomes at the end of the first and fourth years and not the treatment.

For the first model predicting students’ probability to live on campus during their first

year, the student demographics and characteristics were precollege and included high school

54
involvement, high school experiences and behaviors, academic motivation, students’ need for

cognition, highest intended degree, and standardized test scores. Controls for students’

background characteristics included their precollege academic ability as measured through high

school GPA, race, sex, and the level of parental education because of their empirical

relationships with college student engagement and outcomes (e.g., Kinzie et al., 2007; Radford,

Berkner, Wheeless, & Shepherd, 2010). In addition to these common predictors, speaking

English as a second language and being an international student were also included, since these

students may choose to live on campus out of convenience or due to a lack of connections at a

new institution.

Pre-college variables that captured social behavior measures including students’ alcohol

behaviors and smoking behaviors were included in both models. Variables representing how

students spent time socializing, on their computers, exercising, using the library, in

extracurricular activities, as well as time studying with friends, talking with teachers,

volunteering, and working for pay were also included. Additional variables capturing students

incoming dispositions and traits included an academic motivation scaled measure (8 items, α =

.69), a need for cognition scaled measure (18 items, α = .89), highest intended degree, a scaled

measure of precollege diversity experiences (4 items, α = .79), perceptions of their overall health

(single item) and psychological wellbeing (Ryff’s Psychological Well-being Scale, 54 items, α =

.89). Finally, students’ reporting of whether the institution was their first choice and their

perception of their high school racial composition were included in the propensity score model.

Further information can be found in the appendix, including scales and question stems for each

measure.

55
 To answer the last research question about the impact of moving off campus after the first

year, a second propensity score was created using each of the aforementioned precollege

variables. Previous literature suggests that social connections may affect student success

outcomes and college satisfaction (Allen & Haniff, 1991; Astin, 1993; Berger & Milem, 1999;

Cohen & Willis, 1985; Pascarella, 1985; Tinto, 1993); to account for these social experiences

during college, additional variables were incorporated within the propensity score model. They

included scaled measures representing quality of non-classroom interactions with faculty (5

items, α = .86), frequency of interactions with faculty (4 items, α = .73), frequency of

interactions with student affairs staff (5 items, α = .85), the degree of positive peer interactions (8

items, α = .88), co-curricular involvement (single item asking the number of hours per week one

participates in co-curricular activities), experiences with diversity outside the classroom (6 items,

α = .70), meaningful discussions with diverse peers (3 items, α = .83), and negative experiences

related to diversity (5 items, α = .82).

Analyses

Propensity Score Creation. The same procedure was followed to generate each

propensity score using a multiply imputed dataset to account for missing data (Allison, 2000;

Crowe, Lipkovich, & Wang; 2010; D’Agostino, 1998; D’Agostino & Rubin, 2000; Leyrat et al.,

2019; Li, 2013; Mitra & Reiter, 2016; Qu & Lipkovich, 2009). A full list of the propensity score

variables and the treatment variable of interest as well as their means and standard errors before

imputation are included in Tables 1 and 2. Additional covariates capturing students’ collegiate

social experiences were added to the second propensity score model in an effort to reduce bias

associated with their first-year experiences. Correlations were assessed for all variables included

within the propensity score model. The correlations among all variables were calculated with

56
results indicating no correlations above .37. Additionally, variance inflation factors were

calculated after each propensity score model. Collinearity diagnostics, as assed by VIF, were

well below the recommended 2.00 threshold (Craney & Surles, 2002) and as such offer further

support for the decision to retain each variable in the propensity model.

A logistic regression model was used to compute a single, linear propensity score (as

opposed to using the predicted probability; see Guo & Fraser, 2015; Pan & Bai, 2015). To

understand the relationship of each individual variable with collegiate residence, each was

entered as a lone predictor of living on campus while including institutional fixed effects within

the logistic regression model. Generally, scholars suggest including variables related to the

outcome within the propensity score even if they are not significantly associated with the

treatment (Brookhart et al., 2006; Patrick et al., 2011; Westreich et al, 2011), so some variables

were used because the literature suggests they are related to college retention and graduation

(Astin, 1991; Berger & Milem, 1999; Credé & Niehorster, 2012; Kuh et al, 2008; Mayhew et al.,

2016; Pascarella & Terenzini, 1995, 2001; Radford, Berkner, Wheeless, & Shepherd, 2010;

Schudde, 2011; Tinto, 1993).

This study employed stratification to reduce bias associated with students’ decisions to

live on or off campus (Austin, 2011; Guo & Fraser, 2015; Hong & Raudenbush, 2006;

Thoemmes & Kim, 2011). To assess if balance was achieved using stratification, the linear

propensity score variable was first divided into five equal strata (Cochran, 1968). Cochran

suggests five strata can eliminate approximately 90 percent of the bias associated with the

observed variables. However, for both propensity score models, balance was not achieved within

each stratum, as the propensity score was significantly predicted by the strata, the treatment, and

the interaction between the two.

57
 Further exploration indicated multiple strata were significant predictors of the propensity

score and as such, the decision was made to use ten strata. Balance was also not initially

achieved, as the interaction of the stratification variable and the treatment remained significant

predictors of the propensity score. However, exploring each the significance of each strata in

predicting the propensity score, it was found that the two conditions only differed within the first

stratum. Thus, the decision was made to trimming extreme scores within this stratum, removing

137 participants. Support for balance was found through a nonsignificant two-way analysis of

variance predicting the linear propensity score from strata and treatment condition (campus

residence) as independent variables. Additionally, there was no significant interaction between

the variables in predicting the propensity score.

Before subclassification by stratification, there were numerous covariates that

significantly predict students’ decisions to live on campus. Within each propensity score model

(predicating first-year collegiate residence and students’ decisions to move off campus after their

first year), visual inspections of common regions of support, assessment of standardized mean

differences before and after stratification, and significant predictors of collegiate residence

suggest the bias associated with the observed covariates was eliminated with the inclusion of the

strata variable. After the inclusion of that stratification variable, none of the covariates

significantly predicted campus residence, and most standardized mean differences were below

.05. Only six predictors in the first model and three predictors in the second model were above

.05, but these were all under the recommended threshold of a .10 standardized mean difference

(Thoemmes & Kim, 2011).

Several approaches were used to demonstrate bias reduction through stratification. These

included visual inspection of the regions of common support for overlap as well as assessing

58
standardized mean differences before and after stratification. Table 3 includes the standardized

mean difference of each observed covariate before and after balancing through subclassification.

Figures 1 and 2 includes a visual depiction of the overlap between treatment and control within

each stratum. Both suggest that the propensity score modeling using subclassification

successfully removed bias regarding collegiate residence associated with the observed covariates

within each propensity score model. Finally, the propensity score stratification variable was

compared to propensity score weights to see the efficiency of bias reduction (Thoemmes & Kim,

2011). The use of weights in place of stratification demonstrated identical results in both

significant predictors and pre and post standardized mean differences, offering further evidence

of the efficacy of the propensity score stratification variable.

Outcomes Analyses. To assess the effect of campus residence on each the outcomes, this

study used ordinary least squares regression models with clustered robust standard errors. Within

each model, the outcome of interest was predicted by the collegiate residence variable, the PSM

strata, and institutional fixed effects entered as independent variables. Each continuous outcome

was standardized so that the coefficients convey the difference between living on campus and off

campus in terms of standard deviation units (Cohen, Cohen, West, & Aiken, 2003).

To explore whether the potential impact of campus residence varies across groups,

additional analyses included interaction terms between the residence variable and each of several

student-level variables (sex, parental education, standardized test scores, and race). Variables for

living on campus, the relevant moderator, and the interaction term were entered simultaneously

into the equation to model these interactions appropriately (Jaccard & Turrisi, 2003). Even

though there were small sample sizes among some racial and ethnic groups, the decision was

59
made to retain five racial categories in an effort to not combine all non-White students as one

monolithic entity.

Limitations

Some limitations should be noted. First, although quasi-experimental designs can yield

results that better estimate causal effects, the strength of bias reduction is limited by the variables

used to generate the propensity score. In other words, the propensity score analyses can only

eliminate the bias associated with observed covariates and are thus only as effective as the

covariates that were included in the model. It is possible that this study does not include other

important variables that might be linked to a students’ propensity to live on campus. Second, the

campus residence variable does not capture the experiences students had while living on or off

campus. Students’ experiences within campus residence might differ across buildings at the same

institution or across institutions. Third, some of the outcomes used in the final analyses are self-

reported. Students reported their grades using scale points that range from “mostly As” to

“mostly Cs or lower.” Some literature suggests that students are biased in self-reporting

(Bowman, 2011) while others suggest self-reported GPAs and therefore the grades are

reasonable approximations of students’ actual GPAs (Kuncel, Crede, & Thomas, 2005). As such,

the self-reported measure used in the outcome analyses may not be representative of student

achievement.

Fourth, the consolidation of the treatment variable into two groups limits the

generalizability of the data. The experience of students living off campus with parents or

caregivers might not be the same as students who live off campus with peers. Additional research

should disaggregate these different environments to see if there is any difference in their impact

on student achievement and satisfaction. Fifth, this study assumes for the final research question

60
that students who responded that they lived on campus during their first and fourth years did not

move off their second or third year. It is possible that students moved off campus during their

sophomore or junior years only to come back on campus during their fourth year. The current

dataset does not offer clarification regarding these two years and as such the assumption is made

that those who answered in the affirmative for both the end of first and fourth years stayed on

campus the entire time. This assumption may not be true, and further research should include

ways of capturing student residence throughout their undergraduate experience.

Results

Descriptive Statistics

Tables 1 and 2 provide summary statistics for the populations of interest based on

students' choice of collegiate residence. Noticeable differences based on campus residence are

apparent within each table. For students’ first year of postsecondary education, non-trivial

differences across conditions exist in terms of students’ background characteristics, including

race, standardized test scores, and parental education. Students who are White, male, have higher

standardized test scores, and higher self-reported parental income are more likely to live on

campus. Students with minoritized backgrounds such as first-generation status, non-native

English speakers, Asian or Pacific Islander, Latinx, and students with lower test scores tend to

live off campus at higher rates. Some high school activities differ based on campus residence

including time spent exercising, high school involvement, socializing, studying with friends,

working for pay, and computer usage.

For residence choice after their first year, time spent in high school activities differs for

students living on and off campus. Within this sample, students who reported exercising,

socializing, studying with friends, and working for pay seem to live off campus at higher rates

61
based on means and standard deviations, while more academically related variables such as high

school involvement, interacting with teachers, using the library, and reading for fun are more

likely to be found in students living on campus. Additionally, it seems that White students and

those with higher parental income move off campus more frequently than their peers.

Collegiate Outcomes

Results of the outcome analyses exploring the direct effect of collegiate residence on

academically related and satisfaction outcomes are presented in Tables 4 and 5. Main effects

before the propensity score adjustment for students’ first-year residence indicate that living on

campus is positively related to and has a small to medium effect on students’ satisfaction at the

end of their fourth year (for effect size guidelines, see Mayhew et al., 2016). Similarly, living on

campus during the first year is positively associated with retention to the fourth year and

graduation within four years before propensity score adjustment. However, after the propensity

score adjustment, none of these outcomes remains significant.

Addressing research question two, four interaction models were assessed across first-

generation status, race and ethnicity, sex, and high school ability. Each student identity

characteristic was interacted with the living on campus variable and did not resulted in any

significant interactions. This finding seems to suggest that living on campus in the first year, on

the whole, has no effect on all students or differential effects across students. In other words,

after adjusting for bias associated with decisions to live on campus, there is no direct effect on

student achievement, retention, graduation within four years, or college satisfaction.

Results exploring the effect of students moving off campus after their first year offers a

similar pattern. Students’ who moved off campus were significantly less satisfied with their

collegiate experience at the end of their fourth year before the propensity score adjustment. After

62
the adjustment, however, there is no statistical difference between the two groups of students. In

fact, none of the outcomes is significant after the propensity score adjustment, suggesting that

students who choose to move off campus after their first year or stay on campus during their time

in college do not differ in their grades, graduation rates, or satisfaction with their collegiate

experience. Additional interaction models assessed the interaction between residence during

college and student characteristics previously used and again found no significant interactions.

Discussion and Recommendations

This study built upon the existing literature and theoretical perspectives as it relates to

students’ choice of residence and their subsequent academic achievement, retention, graduation,

and satisfaction with their collegiate experience. The current study’s strength comes through the

use of a multi-institutional sample, accounting for self-selection through propensity score

analyses, and exploring differential effects across several groups of students. After propensity

score adjustment, it appears that living on campus does not have a direct effect on students’

retention, graduation, grades, or satisfaction with the collegiate experience. Further, it seems

these nonsignificant findings hold across students’ race, first-generation status, sex, and ability.

These results are surprising given the amount of previous literature implying that living

on campus is a beneficial experience for students, particularly in terms of retention. In the

present study, before conditioning the data on precollege behaviors and characteristics, living on

campus was positively associated with collegiate satisfaction; however, after conditioning the

data through propensity score subclassification, these relationships no longer held. This

conditioning of the data might be the reason for the contrasting findings. The bias associated

with social variables was eliminated by the propensity score stratification which might explain

why these findings seem to be consistent with other studies that find no significant effect on

63
retention or graduation with the inclusion of social engagement variables within the statistical

models (Mayhew et al., 2016). In other words, with social engagement indicators included in

both propensity score models, living on campus was found to have no direct effect on the

academic achievement, retention, probability of graduation, or satisfaction with the collegiate

experience which indicates the importance of accounting for social covariates when estimating

these outcomes (Berger & Milem, 1999; D’Agostino, 1998; Kuh et al, 2008; Milem & Berger,

1997; Tinto, 1993).

There were two other research questions for this study that did not yield any significant

findings the warrant brief discussion. First, sex, race, high school ability, and first-generation

status did not have any significant interactions with living on campus. Given literature that

contends there are hostile campus environments for students with minoritized identities (Bates &

Bourke, 2016; Kulick, Wernick, Woodford, & Renn, 2017; Rankin et al., 2010; Strayhorn &

Mullins, 2012), the lack of significance in this study indicates that, regardless of identity, living

on campus does not have a direct effect. It is possible that students with marginalized identities

experience hostile climates regardless of their residence on or off campus. Second, the research

question comparing students who move off campus after their first year to those students who

stay on was not significant. In other words, this study suggests that moving off campus does not

negatively affect students’ academic achievement or probability of graduation in four years. This

finding may be interesting to higher education administrators considering live on requirements

for sophomore students, or beyond. Often these policies are implemented using rhetoric claiming

that those students on campus achieve higher GPAs and graduation rates.

While these nonsignificant results are noteworthy, there are some considerations to note

and room for future research. First, while this sample included over 40 institutions in the sample,

64
the institutions did have a committed statement to liberal arts, which may positively affect all

students, not just those living on campus. Institutional mission and resources may be better

utilized in helping all students succeed, not just those living on campus. Second, this study was

limited by a lack of data that captures institutional requirements of living on campus. If possible,

future studies should include in the sampling frame both types of institution to better compare

the direct effect of living on campus. Additionally, the research should capture data to

statistically model whether living on campus is a choice or a requirement. With this information,

statistical models could partial out variance in the student outcomes that might be attributed to

institutional policy. Third, while the current study has a robust set of covariates, there could be

additional student engagement and social connection information not included in the model.

Researchers should continue to examine social connections and engagement among college

students to ensure this model was not missing any information in the causal model. Researchers

might understand what specifically leads to the formation of social connections. Fourth, these

findings further the need to consider the impact of living on campus separately by group. While

the current study found no conditional effects, it could be due to limitations in power and small

samples within each subgroup. Future research should seek to increase the number of minoritized

students within their analytic samples.

In short, this study provides an important contribution by supporting strong causal claims

and more generalizable evidence to the literature on living on campus and college student

engagement. By using a quasi-experimental methodology, specifically propensity score

subclassification, the current study reduced the self-selection bias that may be associated with a

student’s choice to live on campus based on a number of observed variables. The study sought to

further the campus residence literature by better causally represent the effect of living on campus

65
and after achieving balance, and it found that living on campus during the first year and

subsequently moving off campus had no direct effect on student retention, grades, graduation, or

their feelings associated with their collegiate experience.

66
Table 1. Unadjusted Means and Standard Errors for Variables of Interest Based on Students’
First-year Residential Choice

Variable Observations On Campus Off Campus

Outcome Variables
College Grades, end of first year 7,678 6.06 (0.02) 5.94 (0.08)
College Grades, end of fourth year 7,732 6.43 (0.02) 6.16 (0.10)
College Satisfaction, end of first year 8,023 3.33 (0.01) 3.23 (0.03)
College Satisfaction, end of fourth year 7,732 3.46 (0.01) 3.24 (0.05)
Retention to Fall Semester, second year 7,911 0.92 (0.00) 0.91 (0.01)
Retention to Fall Semester, third year 7,911 0.84 (0.00) 0.80 (0.02)
Retention to Fall Semester, fourth year 7,911 0.83 (0.00) 0.70 (0.02)
Four-year Graduation 7,911 0.70 (0.01) 0.46 (0.02)

Demographic Characteristics
Sex (1 = male) 8,045 0.38 (0.01) 0.32 (0.02)
Black/African-American (1=yes) 7,732 0.09 (0.00) 0.11 (0.01)
Asian/Pacific Islander (1=yes) 7,732 0.05 (0.00) 0.11 (0.01)
Latinx (1=yes) 7,732 0.04 (0.00) 0.10 (0.01)
White (1=yes) 7,732 0.78 (0.00) 0.64 (0.02)
Other race/ethnicity (1=yes) 7,732 0.05 (0.00) 0.04 (0.01)
First generation (1 = no) 7,516 0.28 (0.01) 0.50 (0.02)
High school test scores (1 = above median) 7,473 0.58 (0.01) 0.28 (0.02)
Self-reported high school GPA 7,857 4.57 (0.01) 4.33 (0.03)
Self-reported parental income 7,198 5.07 (0.03) 4.39 (0.11)
English is native language (1 = yes) 7,285 0.93 (0.00) 0.83 (0.02)
Self-reported disability (1 = yes) 7,897 0.11 (0.00) 0.11 (0.01)

Time spent in high school activities

Exercising 7,818 3.84 (0.01) 3.47 (0.05)
High school involvement 7,805 3.84 (0.01) 3.31 (0.06)
Socializing 7,821 4.41 (0.01) 4.33 (0.04)
Studying with friends 7,809 2.80 (0.01) 2.67 (0.04)
Interacting with teachers 7,800 3.39 (0.01) 3.19 (0.05)
Volunteering 7,806 3.14 (0.01) 2.94 (0.05)
Working for pay 7,807 3.19 (0.02) 3.43 (0.06)
Playing on computer 7,807 4.35 (0.01) 4.22 (0.05)
Using computer for homework 7,807 4.28 (0.01) 4.17 (0.04)
Using the library 7,796 2.79 (0.01) 2.89 (0.05)
Reading for fun 7,802 3.16 (0.01) 3.07 (0.05)
Smoking (1 = did not smoke) 7,974 1.06 (0.00) 1.09 (0.01)
Binge drinking (0 = did not binge drink) 7,960 0.56 (0.01) 0.60 (0.05)
Drinking alcohol (0 = did not consume 7,963 0.48 (0.01) 0.56 (0.05)
alcohol)
Psychosocial and psychological measures
Goal aspirations 7,775 4.43 (0.01) 4.30 (0.05)
Psychological well-being 7,842 4.52 (0.01) 4.49 (0.03)
Need for cognition 7,959 3.47 (0.01) 3.40 (0.03)
Overall health 7,977 4.27 (0.01) 4.17 (0.03)

67
 Table 1—continued

Diversity experiences 7,988 3.97 (0.01) 3.92 (0.03)

Academic motivation 7,974 3.61 (0.01) 3.65 (0.02)
Institutional Information
High school racial composition 7,895 2.06 (0.01) 2.39 (0.05)
Institutional choice 7,857 3.46 (0.01) 3.32 (0.04)
a
Note full variable descriptions and values can be found in the Appendix.

68
Table 2. Unadjusted Means and Standard Errors for Variables of Interest Based on Students’
Moving Off Campus After the First Year

Variable Observations On Campus Off Campus

Outcome Variables
College GPA in the fourth year 3,607 6.49 (0.03) 6.39 (0.03)
Collegiate Satisfaction, end of fourth year 3,737 3.57 (0.01) 3.51 (0.01)
Four-year Graduation 3,660 0.93 (0.01) 0.88 (0.01)

Demographic Characteristics
Sex (1 = male) 3,744 0.35 (0.01) 0.38 (0.01)
Black/African-American (1=yes) 3,636 0.06 (0.01) 0.05 (0.01)
Asian/Pacific Islander (1=yes) 3,636 0.05 (0.00) 0.04 (0.01)
Latinx (1=yes) 3,636 0.05 (0.01) 0.04 (0.00)
White (1=yes) 3,636 0.78 (0.01) 0.82 (0.01)
Other race/ethnicity (1=yes) 3,636 0.06 (0.01) 0.04 (0.00)
First generation (1 = no) 3,531 0.24 (0.01) 0.23 (0.01)
High school test scores (1 = above median) 3,478 0.73 (0.01) 0.65 (0.01)
Self-reported high school GPA 3,645 4.71 (0.01) 4.68 (0.01)
Self-reported parental income 3,388 5.05 (0.06) 5.37 (0.06)
English is native language (1 = yes) 3,649 0.92 (0.01) 0.94 (0.01)
Self-reported disability (1 = yes) 3,671 0.12 (0.01) 0.10 (0.01)

Time spent in high school activities

Exercising 3,632 3.78 (0.03) 3.99 (0.03)
High school involvement 3,624 4.00 (0.03) 3.87 (0.03)
Socializing 3,633 4.31 (0.02) 4.47 (0.02)
Studying with friends 3,627 2.81 (0.02) 2.86 (0.02)
Interacting with teachers 3,619 3.47 (0.02) 3.39 (0.02)
Volunteering 3,624 3.28 (0.03) 3.22 (0.02)
Working for pay 3,625 2.90 (0.03) 3.23 (0.03)
Playing on computer 3,627 4.36 (0.02) 4.35 (0.02)
Using computer for homework 3,625 4.32 (0.02) 4.27 (0.02)
Using the library 3,623 2.86 (0.03) 2.76 (0.02)
Reading for fun 3,625 3.30 (0.03) 3.11 (0.03)
Smoking (1 = did not smoke) 3,718 1.04 (0.01) 1.04 (0.00)
Binge drinking (0 = did not binge drink) 3,713 0.32 (0.02) 0.62 (0.03)
Drinking alcohol (0 = did not consume 3,610 0.29 (0.02) 0.53 (0.02)
alcohol)
Psychosocial and psychological measures
Goal aspirations 3,610 4.56 (0.03) 4.48 (0.03)
Psychological well-being 3,667 4.55 (0.01) 4.57 (0.01)
Need for cognition 3,714 3.59 (0.01) 3.45 (0.01)
Overall health 3,722 4.30 (0.01) 4.33 (0.01)
Diversity experiences 3,724 4.07 (0.01) 3.96 (0.02)
Academic motivation 3,717 3.69 (0.01) 3.57 (0.01)
Institutional Information
High School racial composition 3,654 2.05 (0.03) 1.98 (0.02)

69
 Table 2—continued

Institutional choice 3,644 3.45 (0.02) 3.51 (0.02)

Collegiate Experiences
End of first-year grades 3,625 6.33 (0.03) 6.34 (0.03)
Entire experience again, end of first year 3,730 3.53 (0.01) 3.44 (0.01)
Choose same college, end of first year 3,731 3.46 (0.02) 3.40 (0.02)
Interactions with faculty 3,689 0.17 (0.02) 0.01 (0.02)
Peer interactions 3,679 0.16 (0.02) 0.14 (0.01)
Co-curricular involvement 3,732 0.15 (0.02) 0.17 (0.02)
Campus diversity experiences and events 3,701 0.12 (0.01) -0.03 (0.01)
Positive diversity experiences 3,671 0.09 (0.02) -0.02 (0.02)
Negative diversity experiences 3,672 -0.10 (0.02) -0.09 (0.02)
Interactions with student affairs staff 3,671 0.02 (0.02) 0.01 (0.02)

a
Note full variable descriptions and values can be found in the Appendix.

70
Table 3. Significance of and Standardized Mean Differences for Each Propensity Score Model,
Before and After Stratification

Standardized Mean Differences

First Propensity Second Propensity
Score Model Score Model
Before After Before After
Predictor Variablea
Balancing Balancing Balancing Balancing
Demographic Characteristics
Sex (1 = male) -0.06 -0.02 -0.08 0.01
Black/African-American -0.09 0.04 -0.07 -0.03
Asian/Pacific Islander -0.24*** -0.09 0.06 0.05
Latinx -0.20*** -0.01 0.01 0.04
White 0.35*** 0.06 0.04* -0.02
Other race/ethnicity -0.07 -0.05 -0.08** -0.03
First generation (1 = no) -0.28*** 0.01 -0.03 -0.02
High school test scores (1 = above median) 0.40*** -0.03 -0.08** -0.01
Self-reported high school GPA 0.20 0.07 0.22 0.08
Self-reported parental income 0.26*** 0.04 0.13* 0.08
English is native language (1 = yes) 0.38*** 0.08 -0.01 -0.01
Self-reported disability (1 = yes) 0.09 0.03 0.03* 0.05
Time spent in high school activities
Exercising 0.21*** 0.04 0.15*** 0.03
High school involvement 0.34*** 0.05 -0.10** 0.01
Socializing 0.19*** 0.1 0.15*** 0.02
Studying with friends 0.06 -0.02 0.05*** 0.01
Interacting with teachers 0.03 -0.03 -0.06* 0.01
Volunteering -0.05 -0.05 -0.01 0.01
Working for pay -0.05 0.05 0.14*** -0.05
Playing on computer 0.03 0.00 0.01 0.04
Using computer for homework 0.08* -0.01 -0.09* 0.01
Using the library -0.03 0.02 -0.06 -0.02
Reading for fun 0.16** 0.06 -0.11*** 0.02
Smoking -0.01 0.00 0.01 0.01
Binge drinking 0.00 -0.01 0.16*** -0.04
Drinking alcohol -0.04 0.03 0.18*** -0.06
Psychosocial and psychological measures
Goal aspirations -0.02* -0.03 -0.02 -0.02
Psychological well-being 0.09* 0.02 0.00 -0.02
Need for cognition 0.13*** -0.03 -0.20*** 0.01
Overall health 0.09 0.01 0.04 0.01

71
 Table 3—continued

Diversity experiences 0.11* -0.01 -0.16*** -0.03

Academic motivation -0.06 0.01 -0.15*** 0.01
Additional Information
High school racial composition -0.31*** 0.01 -0.04 0.02
Institutional choice 0.15** 0.03 0.04 -0.02
Collegiate Experiencesb
End of first-year grades -0.02 -0.02
Entire experience again, end of first year -0.12** -0.03
Choose same college, end of first year -0.05 -0.02
Interactions with faculty -0.21*** -0.01
Peer interactions -0.04 -0.01
Cocurricular involvement -0.03 -0.01
Diversity experiences on campus -0.19*** 0.04
Positive diversity experiences -0.14*** 0.02
Negative diversity experiences 0.01 0.01
Interactions with student affairs staff -0.03 0.00

a
Significant predictors of the treatment variable, living on campus, are noted as follows *p < .05
**p < .01 ***p < .001

b
Institutional fixed effects were incorporated into the propensity score model by including
dummy codes for each institution while leaving one institution out as the referent group.

72
Table 4. Results of Regression Analyses of First-year Residence Predicting College Academic
and Satisfaction Outcomes
_____________________________________________________________________________________

No PSM adjustment PSM adjustment

Outcome variable B SE Delta-p B SE

College GPA in the first year -0.02 0.09 -0.27 0.14
College GPA in the fourth year 0.11 0.09 -0.04 0.12
College Satisfaction, end of first year 0.11 0.13 -0.02 0.08
College Satisfaction, end of fourth year 0.25** 0.09 0.17 0.16
Retention to fall of second year 0.06 0.24 0.11 0.19
Retention to fall of third year 0.25 0.14 0.12 0.19
Retention to fall of fourth year 0.51** 0.19 0.04 0.11 0.21
Graduated within four years 0.93*** 0.17 0.05 0.14 0.24

Note. Institutional fixed effects were included in all analyses. College GPA and college satisfaction were
examined with ordinal least squares multiple regression analyses; these outcomes were standardized with
a mean of zero and a standard deviation of one. Retention and graduation were examined with logistic
regression analyses to predict these dichotomous outcomes. Based on Cruce’s (2009) recommendation,
the only delta-p values reported were for significant binary outcomes. *p < .05 **p < .01 ***p < .001
_____________________________________________________________________________________

73
Table 5. Results of Regression Analyses of Collegiate Residence Predicting College Academic
and Satisfaction Outcomes for Moving Off Campus after Students’ First Year

_____________________________________________________________________________________

No PSM
adjustment PSM adjustment

Outcome variable B SE B SE
College GPA in the fourth year 0.08 0.29 0.04 0.06
College satisfaction, end of fourth year 0.12* 0.06 0.08 0.08
Graduated within four years -0.48 0.25 -0.14 0.23

Note. Students who reported living on campus at both Time 2 and 3 were coded 1, while students who
reported moving off campus after Time 2 were coded as 0. Institutional fixed effects were included in all
analyses. The continuous outcomes representing academic achievement and satisfaction were
standardized with a mean of zero and a standard deviation of one. Each outcome was analyzed using
ordinal least squares multiple regression analyses with robust standard errors.
*p < .05 **p < .01 ***p < .001
_____________________________________________________________________________________

74
Figure 1. Propensity Score Distributions for Students Living On Campus (treated) and Off
Campus (untreated) During Their First Year
Proportion of Sample

.2 .4 .6 .8 1
Propensity Score

Untreated Treated: On support

Treated: Off support

75
Figure 2. Propensity Score Distributions for Students Living On Campus (treated) and Off
Campus (untreated) During Their Collegiate Experience

.2 .4 .6 .8 1
Propensity Score

Untreated Treated

76
 Appendix: Study Variables

Demographic Characteristics
Sex (1 = male) Student’s institution provided data from their school file for sex and race / ethnicity
Black/African-American (1=yes)
Asian/Pacific Islander (1=yes)
Latinx (1=yes) Dummy variables for race / ethnicity entered separately into the model
White (1=yes)
Other race/ethnicity (1=yes)
Recoding of variable asking what is the highest level of education each of your parents
First generation (1 = no) or guardians completed? First generation was coded as students who selected did not
finish high school or high school graduate / GED.
Variable converted SAT scores using the COMPASS conversion, so all scores were on a
High school test scores (1 = above median)
common metric

Which of the following best describes your overall grade range in high school? 1 = A- to
Self-reported high school GPA
A+ to 5 = Below D-

What is the best estimate of your parents’ totally annual income and your annual
Self-reported parental income
income? 1 = less than $14,999 to 9 = $300,000 or more

English is native language (1 = yes) Is English your native language?

Self-reported disability (1 = yes) Aggregate of question, mark all of the following diagnosed disabilities that apply to you

Time spent in high school activities

Exercising Question asked, during the last year in high school, how often did you engage in each of
High school involvement the following activities? Scale was 1 = very often to 5 = never
Socializing
Studying with friends
Interacting with teachers

77
 Volunteering
Working for pay
Playing on computer
Using computer for homework
Using the library
Reading for fun
How many cigarettes do you smoke a day? 1 = I don’t smoke cigarettes to 5 = 2 or more
Smoking
packs a day

In a typical week of your last year of high school, how often did you consume 5 or more
Binge drinking
drinks in one sitting? 1 = 0 times to 5 = 5 or more times

In a typical week of your last year of high school, how often did you consume alcoholic
Drinking alcohol
beverages? 1 = 0 times per week to 9 = more than 7 times per week
Psychosocial and psychological measures
What is the highest academic degree you intend to earn in your lifetime? 1 = vocational /
Goal aspirations
technical certificate or diploma to 6 = Doctorate degree

Overall Health Overall, how would you rate your health? 1 = excellent to 5 = very poor

Scaled measure developed by Carol Ryff (1989). Items included were from all six
subscales including autonomy, environmental mastery, personal growth, positive
Psychological well-being, 54 items, α = .89
relations with others, purpose in life, and self-acceptance. Scales ranged from 1 =
strongly disagree to 5 = strongly agree

Degree to which one enjoys engaging in effortful cognitive activities. Sum of eighteen
items on the Need for Cognition short form (Cacioppo, Petty, & Kao, 1984). Scaled from
Need for cognition, 18 items, α = .89
1 = Extremely characteristic to 5 = Extremely uncharacteristic

Items include I enjoy having discussions with people whose ideas and values are
different from my own; The real value of a college education lies in being introduced to
Precollege Diversity experiences, 4 items, α = .79 different values; Contact with individuals whose backgrounds (e.g. race, national origin,
sexual orientation) are different from my own is an essential part of my college
education; and I enjoy talking with people who have values different from mine because

78
 it helps me better understand my values. Scale is 1 = Strongly Agree to 5 = Strongly
Disagree.

Scale (1 = Strongly Agree to 5 Strongly Disagree) includes items such as: I am willing to
work hard in a course to learn the material even if it won’t lead to a higher grade; When I
do well on a test, it is usually because I am well-prepared not because the test is easy; In
Academic motivation, 8 items, α = .70 high school, I frequently did more reading in a class than was required simply because it
interested me; In high school, I frequently talked to my teachers outside of class about
ideas presented during class; Getting the best grades I can is very important to me; I
enjoy the challenge of learning complicated new material

Additional Information
How would you describe the racial composition of the last high school you attended? 1 =
High school racial composition
almost all white students to 5 = almost all students of color

Institutional choice Was this college your… 1 = first choice to 3 = third choice

Collegiate Experiences
What have most of your grades been up to now at this institution? 1 = C- or lower to 8 =
End of first-year grades
A
How would you evaluate your entire educational experience at this institution? 1 = poor
Entire experience again, end of first year
to 4 = excellent
If you could start over again, would you go to the same institution you are now
Choose same college, end of first year
attending? 1 = definitely no to 4 = definitely yes

About how many hours in a typical week do you spend doing the following: Participating
Co-curricular involvement in co-curricular activities (organizations, campus publications, student government,
fraternity or sorority, intercollegiate or intramural sports, etc.)

Questions related to the quality of non-classroom interactions with faculty including the
extent students agreed that non-classroom interactions had a positive influence on
Interactions with faculty, outside of class, 5 items, α = .86
personal growth, values, and attitudes. Response options were 1 = strongly agree to 5 =
strongly disagree.

Scale representing how often students discussed grades, assignments, career plans, ideas
Frequency of interactions with faculty, 4 items, α = .73 from readings outside of the classroom, or worked on activities other than coursework
with faculty. Response options were 1 = very often to 5 = never.

79
 Scale representing students’ relationships with other students, personally satisfying
relationships, the degree other students have had a positive influence on intellectual
growth and interest in ideas, quality of relationships with other students, ability to meet
Peer interactions, 8 items, α = .88
and make friends with other students, perceptions of other students willing to listen and
help with a personal problem, and the degree to which other students’ values align with
the respondent. Response options were 1 = strongly agree to 5 = strongly disagree.

Scale representing how often students attended debates or lectures on a current political
or social issue during the academic year, had serious discussions with staff whose
political, social, or religious opinions were different from own, degree to which the
institution emphasizes contact among students from different economic, social, and racial
Campus diversity events and experiences, 6 items, α = .70 or ethnic backgrounds, how often the student has had serious conversations with students
who are very different from them in terms of religious beliefs, political opinions, or
personal values, and how often student participated in a racial or cultural awareness
workshop during the current academic year. Response options were 1 = very often to 5 =
never

Scale representing how often student had discussions regarding inter-group relations with
diverse students, had meaningful and honest discussions about issues related to social
Positive diversity experiences, 3 items, α = .82 justice with diverse students, and how often they shared personal feelings and problems
with diverse students while attending this college. Response options were 1 = very often
to 5 = never

Scale representing how often students had guarded or cautious interactions with diverse
students, how often they felt silenced by prejudice and discrimination from sharing
personal experiences with diverse others, how often they had hurtful or unresolved
Negative diversity experiences, 5 items, α = .82 interactions with diverse students, had somewhat hostile interactions with diverse
students, and how often they felt insulted or threatened based on race, national origin,
values or religion with diverse students while attending this college. Response options
were 1 = very often to 5 = never

Scale representing how often students discussed personal problems or concerns, worked
on out-of-class activities (e.g. committees, orientation, student life activities), talked
Interactions with student affairs staff, 5 items, α = .85 about career plans, discussed ideas from readings or classes, or discussed grades or
assignments with student affairs professionals. Response options were 1 = very often to 5
= never

80
 References

Allen, W. R. (1991). Introduction. In W. R, Allen, E. G. Epps, & N. Z. Haniff (Eds.), College in

Black and White: African American students in predominantly White and in historically

Black public universities (pp. 1-14). Albany, NY: State University of New York Press

Allison, P. D. (2000). Multiple imputation for missing data: A cautionary tale. Sociological

methods & research, 28(3), 301-309.

Allison, P. D. (2009). Fixed effects regression models. Thousand Oaks, CA: Sage.

Arpino, B., & Mealli, F. (2011). The specification of the propensity score in multilevel

observational studies. Computational Statistics & Data Analysis, 55(4), 1770-1780.

Astin, A. (1985). Achieving education excellence: A critical assessment of priorities and

practices in higher education. San Francisco, CA: Jossey-Bass.

Astin, A.W. (1991). Assessment for excellence: The philosophy and practice of assessment and

evaluation in higher education. Washington, DC: American Council on Education/Oryx

Press Series on Higher Education.

Astin, A. W. (1993). What matters in college? Four critical years revisited. San Francisco, CA:

Jossey-Bass Publishers.

Austin, P. C. (2011). An introduction to propensity score methods for reducing the effects of

confounding in observational studies. Multivariate Behavioral Research, 46, 399–424.

Bates, A. K., & Bourke, P. Q. (2016). 2016 Your First College Year Survey. Los Angeles, CA:

Higher Education Research Institute.

Berger, J. B., & Milem, J. F. (1999). The role of student involvement and perceptions of

integration in a causal model of student persistence. Research in Higher

Education, 40(6), 641-664.

81
Biemer, P. P., & Christ, S. L. (2008). Weighting survey data. International handbook of survey

methodology, 317-341.

Blimling, G. S. (1989). A Meta-Analysis of the Influence of College Residence Halls on

Academic Performance. Journal of College Student Development, 30(4), 298-308.

Blimling, G. S. (1993). The influence of college residence halls on students. Higher education:

Handbook of Theory and Research, 9, 248-307.

Blimling, G. S., & Whitt, E. J. (1999). Good practice in student affairs: Principles to foster

student learning. San-Francisco, CA: Jossey-Bass.

Bronfenbrenner, U. (1977). Toward an experimental ecology of human development. American

Psychologist, 32, 513-531.

Bronfenbrenner, U. (1979). The ecology of human development: Experiments by nature and

design. Cambridge, MA: Harvard University Press.

Bronfenbrenner, U. (1989). Ecological systems theory. In R. Vasta (Ed.), Six theories of

development (pp. 187-249). Greenwich, CT: JAI Press.

Bronfenbrenner, U. (1993). The ecology of cognitive development: Research models and

fugitive findings. In R. H. Wozniak & K. W. Fischer (Eds.), Development in context:

Acting and thinking in specific environments (pp, 3-44). Hillsdale, NJ: Erlbaum

Brookhart, M. A., Schneeweiss, S., Rothman, K. J., Glynn, R. J., Avorn, J., & Stürmer, T.

(2006). Variable selection for propensity score models. American Journal of

Epidemiology, 163(12), 1149-1156.

Bowman, N. A. (2011). Validity of college self-reported gains at diverse

institutions. Educational Researcher, 40(1), 22-24.

82
Chickering, A. W. (1974). Commuting versus resident students. San Francisco, CA: Jossey-

Bass.

Cochran, W. G. (1968). The effectiveness of adjustment by subclassification in removing bias in

observational studies. Biometrics, 295-313.

Cohen, J., Cohen, P., West, S. G., & Aiken, L. S. (2003). Applied multiple regression/correlation

analysis for the behavioral sciences. Mahwah, NJ: Lawrence Erlbaum Associates.

Cohen, S., & Wills, T. (1985). Stress, social support, and the buffering hypothesis. Psychological

Bulletin, 98, 310-357

Collins, L. M., Schafer, J. L., & Kam, C.-M. (2001). A comparison of inclusive and restrictive

strategies in modern missing data procedures. Psychological Methods, 6(4), 330-351.

http://dx.doi.org/10.1037/1082-989X.6.4.330

Cowley, W. H. (1934a). The history of student residential housing. School and Society,

40(1040), 705 – 712.

Cowley, W. H. (1934b). The history of student residential housing II. School and Society,

40(1041), 758 – 764.

Craney, T. A., & Surles, J. G. (2002). Model-dependent variance inflation factor cutoff values.

Quality Engineering, 14(3), 391-403.

Credé, M., & Niehorster, S. (2012). Adjustment to college as measured by the student adaptation

to college questionnaire: A quantitative review of its structure and relationships with

correlates and consequences. Educational Psychology Review, 24(1), 133-165.

Crowe, B. J., Lipkovich, I. A., & Wang, O. (2010). Comparison of several imputation methods

for missing baseline data in propensity scores analysis of binary

outcome. Pharmaceutical statistics, 9(4), 269-279.

83
d'Agostino, R. B. (1998). Propensity score methods for bias reduction in the comparison of a

treatment to a non‐randomized control group. Statistics in Medicine, 17(19), 2265-2281.

D'Agostino Jr, R. B., & Rubin, D. B. (2000). Estimating and using propensity scores with

partially missing data. Journal of the American Statistical Association, 95(451), 749-759.

de Araujo, P., & Murray, J. (2010). Channels for improved performance from living on campus.

American Journal of Business Education, 3(12), 57.

Delucchi, M. (1993). Academic performance in a college town. Education, 114(1), 96 – 100.

Flowers, L. A. (2004). Effects of living on campus on African American students' educational

gains in college. NASPA Journal, 41(2), 277-293.

Groves, R. M., Fowler Jr, F. J., Couper, M. P., Lepkowski, J. M., Singer, E., & Tourangeau, R.

(2009). Survey methodology (2nd ed.). Hoboken, NJ: John Wiley & Sons.

Guo, S., & Fraser, M. (2015). Propensity score analyses: Statistical methods and applications.

2nd ed. Thousand Oaks, CA: SAGE Publications, Inc.

Hong, G., & Raudenbush, S. W. (2006). Evaluating kindergarten retention policy: A case study

of causal inference for multilevel observational data. Journal of the American Statistical

Association, 101(475), 901-910.

Jaccard, J., & Turrisi, R. (2003). Interaction effects in multiple regression. Thousand Oaks, CA:

Sage Publications.

Johnson, W. G. (1996). Strategies for enhancing student learning in residence halls. New

Directions for Student Services, 1996(75), 69-82.

Kanoy, K., & Bruhn, J. (1996). Effects of a first-year living and learning residence hall on

retention and academic performance. Journal of The First-Year Experience & Students in

Transition, 8(1), 7-23.

84
Kinzie, J., Thomas, A. D., Palmer, M. M., Umbach, P. D., & Kuh, G. D. (2007). Women

students at coeducational and women's colleges: How do their experiences

compare?. Journal of College Student Development, 48(2), 145-165.

Kuh, G. D., Cruce, T. M., Shoup, R., Kinzie, J., & Gonyea, R. M. (2008). Unmasking the effects

of student engagement on first-year college grades and persistence. The Journal of

Higher Education, 79(5), 540-563.

Kulick, A., Wernick, L. J., Woodford, M. R., & Renn, K. (2017). Heterosexism, depression, and

campus engagement among LGBTQ college students: Intersectional differences and

opportunities for healing. Journal of Homosexuality, 64(8), 1125-1141.

Kuncel, N. R., Credé, M., & Thomas, L. L. (2005). The validity of self-reported grade point

averages, class ranks, and test scores: A meta-analysis and review of the

literature. Review of Educational Research, 75(1), 63-82.

La Roche, C. R., Flanigan, M. A., & Copeland Jr, P. K. (2010). Student housing: Trends,

preferences and needs. Contemporary Issues in Education Research, 3(10), 45-50.

Leyrat, C., Seaman, S. R., White, I. R., Douglas, I., Smeeth, L., Kim, J., & Williamson, E. J.

(2019). Propensity score analysis with partially observed covariates: How should

multiple imputation be used?. Statistical Methods in Medical Research, 28(1), 3-19.

Li, M. (2013). Using the propensity score method to estimate causal effects: A review and

practical guide. Organizational Research Methods, 16(2), 188-226.

Li, F., Zaslavsky, A. M., & Landrum, M. B. (2013). Propensity score weighting with multilevel

data. Statistics in Medicine, 32(19), 3373-3387.

López Turley, R. N., & Wodtke, G. (2010). College residence and academic performance: who

benefits from living on campus?. Urban Education, 45(4), 506-532.

85
Mayhew, M. J., Bowman, N. A., Rockenbach, A. N., Seifert, T. A., & Wolniak, G. C.

(2016). How college affects students: 21st century evidence that higher education

works (Vol. 3). San Francisco, CA: Jossey-Bass.

Milem, J. F., & Berger, J. B. (1997). A modified model of college student persistence: Exploring

the relationship between Astin's theory of involvement and Tinto's theory of student

departure. Journal of College Student Development, 38(4), 387.

Mitra, R., & Reiter, J. P. (2016). A comparison of two methods of estimating propensity scores

after multiple imputation. Statistical Methods in Medical Research, 25(1), 188-204.

Newbold, J.J., Mehta, S.S., & Forbus, P. (2011). Commuter students: Involvement and

identification with an institution of higher education. Academy of Educational Leadership

Journal, 15(2), 141-153

Oseguera, L., & Rhee, B. S. (2009). The influence of institutional retention climates on student

persistence to degree completion: A multilevel approach. Research in Higher

Education, 50(6), 546-569.

Pan, W., & Bai, H. (Eds.). (2015). Propensity score analysis: Fundamentals and developments.

New York, NY: Guilford Publications.

Pascarella, E., Bohr, L., Nora, A., Zusman, B., Inman, P., & Desler, M. (1993). Cognitive

impacts of living on campus versus commuting to college. Journal of College Student

Development, 34, 216-220.

Pascarella, E. T. (1985). College environmental influences on learning and cognitive

development: A critical review and synthesis. In J. Smart (Ed.), Higher education:

Handbook of theory and research (pp. 1-61). New York, NY: Agathon.

86
Pascarella, E. T., & Terenzini, P. T. (1991). How college affects students: Findings and insights

from twenty years of research. San Francisco, CA: Jossey-Bass.

Pascarella, E. T., & Terenzini, P. (2005). How college affects students (Volume 2). San

Francisco, CA: Jossey-Bass.

Pascarella, E., Bohr, L., Nora, A., Zusman, B., Inman, P., & Desler, M. (1993). Cognitive

impacts of living on campus versus commuting to college. Journal of College Student

Development, 34, 216-220.

Pike, G. R., & Kuh, G. D. (2005). First-and second-generation college students: A comparison of

their engagement and intellectual development. The Journal of Higher Education, 76(3),

276-300.

Qu, Y., & Lipkovich, I. (2009). Propensity score estimation with missing values using a multiple

imputation missingness pattern (MIMP) approach. Statistics in Medicine, 28(9), 1402-

1414.

Radford, A. W., Berkner, L., Wheeless, S., & Shepherd, B. (2010). Persistence and attainment of

2003–04 beginning postsecondary students: After six years. Washington, DC: National

Center for Education Statistics.

Rankin, S., Weber, G., Blumenfeld, W., & Frazer, M. S. (2010). 2010 state of higher education

for lesbian, gay, bisexual, & transgender people. Charlotte, NC: Campus Pride.

Rashdall, H. (1936). The universities of Europe in the middle ages (vol. 3): A new edition in

three volumes. English universities-student life. Cambridge, England: Oxford University

Press.

Renn, K. A., & Arnold, K. D. (2003). Reconceptualizing research on college student peer

culture. The Journal of Higher Education, 74(3), 261-291.

87
Riker, H. C. (1965). College housing as learning centers. Washington, DC: American College

Personnel Association.

Rudolph, F. (1990). The American college and university: A history. Athens, GA: University of

Georgia Press.

Saffron, I. (2013, September 18). Granite countertops, flat-screen TVs, fire pits: The surprising

story of how college dorms got luxe. New Republic. Retrieved from

https://newrepublic.com/article/114744/luxification-college-dorms

Schudde, L. T. (2011). The causal effect of campus residency on college student retention. The

Review of Higher Education, 34(4), 581-610.

Stewart, H. Q. (1942). Some social aspects of residence halls for college women. Professional &

Technical Press.

Strayhorn, T. L., & Mullins, T. G. (2012). Investigating black gay male undergraduates'

experiences in campus residence halls. Journal of College and University Student

Housing, 39(1), 140-161.

Taibbi, M. (2013, August 15). The college-loan scandal. Rolling Stone. Retrieved from

https://www.rollingstone.com/politics/news/ripping-off-young-america-the-college-loan-

scandal-20130815

Thelin, J. R. (2004). A history of American higher education. Baltimore, MD: Johns Hopkins

University Press.

Thoemmes, F. J., & Kim, E. S. (2011). A systematic review of propensity score methods in the

social sciences. Multivariate behavioral research, 46(1), 90-118.

Thompson, J., Samiratedu, V., & Rafter, J. (1993). The effects of on-campus residence on first-

time college students. NASPA Journal, 31(1), 41-47.

88
Tinto, V. (1993). Building community. Liberal Education, 79(4), 16-21.

Von Hippel, P. T. (2007). 4. Regression with missing Ys: An improved strategy for analyzing

multiply imputed data. Sociological Methodology, 37(1), 83-117.

Westreich, D., Cole, S. R., Funk, M. J., Brookhart, M. A., & Stürmer, T. (2011). The role of the

c‐statistic in variable selection for propensity score models. Pharmacoepidemiology and

Drug Safety, 20(3), 317-320.

89
 CHAPTER THREE: COLLEGIATE RESIDENCE, HEALTH, AND PSYCHOLOGICAL
WELL-BEING

Transitioning into college is a major life event. For many new students, it involves leaving home

and moving onto a college campus, often their first time away from parents or caregivers. While

there is substantial evidence citing the effect of peer culture on a host of outcomes (e.g., Mayhew

et al., 2016; Pascarella & Terenzini, 2005), ecological models of student development offer

insight as to how interactions among the most immediate of students’ environments create forces

of campus peer cultures that affect students (Renn & Arnold, 2003). In other words, students

living on campus have different environments than their off-campus peers that directly affect

their experiences. This new collegiate environment also offers students numerous experiences

that can promote, or hinder, their psychological well-being (PWB), health, and health-related

behaviors. Given that college might be the first time that students take full responsibility for their

health and well-being, understanding the direct effects particular environments have on students

and the peer groups they foster is necessary, especially since the patterns established during these

years could form the foundation for their future health-risk behaviors including alcohol and

tobacco-use (Timberlake et al., 2007; Wetter et al, 2004).

Supporting Literature and Study Purpose

Scholarship focused on health-related outcomes among college students explores alcohol

consumption patterns, sexual health behaviors, physical health, and changes in eating patterns.

Drinking patterns among college students in the United States are important due to their negative

consequences that range from sleep interruptions to physical assault or even death (e.g.,

Wechsler & Nelson, 2008). Often alcohol consumption is directly linked to campus peer

cultures, especially those created within campus residence halls or Greek-lettered houses. In

fact, students living in suite style rooms reported higher negative alcohol behaviors compared to

90
other residential environments (i.e. frequency of drinking, binge drinking, and heavy episodic

drinking) (Cross, Zimmerman, & O’Grady, 2009). Often, student staff are utilized in an attempt

to affect these peer alcohol norms (Perkins, 2002). While some literature indirectly suggests

residence halls related to increased alcohol consumption (Borsari, Murphy, & Barnett, 2007),

others point to the negative effect of living off campus (Benz et al., 2017). Other studies have

explored relationships between campus residence, diet, and exercise finding decreased

consumption of fruits and vegetables as well as daily activities exacerbated by living off campus

(Small et al., 2012) in addition to differential eating patterns (Brevard & Rickets, 1996). While

the literature on health-related outcomes suggests where a student lives matters, these studies are

limited by small sample sizes and a failure to account for a students’ self-selection into campus

residence halls.

Additionally, there are advocates who argue for a broader conception of health beyond

just physical health. According to the Center for Disease Control and Prevention (2018), this

conception of well-being includes both physical and mental health. Among college students,

well-being is sometimes studied using Ryff’s (1989) scaled measure of psychological well-being.

To her, psychological well-being is based on the premise that "being well" encompasses a range

of characteristics and perceptions beyond just happiness. Using these theoretical foundations,

Ryff proposed a model of psychological wellbeing that includes six distinct dimensions:

autonomous functioning and decision making, mastery of one's environment, seeking

opportunities for personal growth, maintaining positive relations with others, having a sense of

purpose in life, and accepting and thinking positively about oneself. From this perspective, PWB

encompasses the use of certain skills and perspectives that are helpful for overcoming life’s

challenges as well as effectively navigating one's own life (Ryff, Keyes, & Hughes, 2003;

91
Smider, Essex, & Ryff, 1996). Subsequent research has indicated that psychological well-being

contributes to a range of essential outcomes in college and adult life, including increased social

support, greater life satisfaction, and improved physical health (Bowman & Kitayama, 2009;

Ryff, 2008). Within the literature, living on campus has been linked with negative effects on

students PWB (see Mayhew et al., 2016), however a vast majority of the research is based upon

students first year in college. As such, more research is needed to examine the longitudinal

effects of residence on PWB.

Largely, prior literature suggests that living on campus seems to have a negative effect on

these particular outcomes. It might be that many traditionally-aged students are on their own for

the first time and might not make productive decisions in terms of their health and well-being.

To extend further literature, this study conceptualizes the interaction between students and their

collegiate residence as the causal mechanism that affects alcohol consumption, health, and PWB

(Renn & Arnold, 2003) and seeks to minimize potential bias associated with students’ residential

choice, this study utilizes propensity score modeling to reduce self-selection bias into the

“treatment” of living on campus (e.g. Guo & Fraser, 2015). Variables that are included in this

model are based on theoretical or empirical assertions of their relation to the impact college has

on students (e.g., Pascarella & Terenzini, 1991, 2005). This model includes variables related to

students’ precollege traits and dispositions, student demographic information, as well as

information representing precollege health, alcohol, and PWB outcomes.

Method
Data Source and Participants

This study used participants of the Wabash National Study, which was designed to

explore the relationships among a variety of college experiences and liberal arts outcomes. The

study encompasses institutions that included religiously affiliated, single-sex, and minority-

92
serving schools. Institutional characteristics spanned a variety of selectivity, tuition costs, and

geographic diversity. Students were surveyed at three time points: at the beginning of their first

year(Time 1), the end of their first year (Time 2), and the end of students’ fourth year. The

analytic sample for the current study was restricted to the 39 four-year institutions offering

campus housing to students. For analyses predicting fourth-year psychological well-being and

health outcomes, the sample was also limited to the students who completed all three survey

waves. Thus, the analytic sample contained 4,814 students at time 2 and 2,630 students at time 3.

Of these students, 62% were female, 6% were Asian American/Pacific Islander, 9% were

Black/African American, 5% were Latino/Hispanic, and 2% were from another race/ethnicity.

Measures

Dependent Variables. Psychological well-being was indicated using the Ryff Scales of

Psychological Well-Being (PWB) (α = .89). This 54-item instrument includes six dimensions of

psychological well-being: self-acceptance, personal growth, purpose in life, positive relations

with others, environmental mastery, and autonomy. The overall PWB scale score was used to

mitigate potential concerns with construct validity (Springer & Hauser, 2006). Health outcomes

and behaviors were measured through single items regarding tobacco use (1 = I don’t smoke to 5

= 2 or more packs per day), alcohol consumption (1 = 0 times per week to 9 = more than 7 times

per week), frequency of binge drinking in a typical week (1 = 0 times to 5 = 5 or more times),

frequency of aerobic exercise (1 = I don’t exercise regularly to 5 = more than 6 hours per week),

and overall health quality (5 = Excellent to 1 = Very poor).

Independent Variable. The primary independent variable indicated students first year of

residence (0 = off-campus, 1 = on-campus). The original item included a response option for

fraternities and sororities; these students were removed from the analytic sample due to

ambiguity about whether this housing was on or off campus.

93
 Propensity Score. The current sample contained students nested within institutions, so

institutional fixed effects were used to account for all between-institution differences. This

approach uses dummy variables for all but one institution, omitted as the referent group (see

Allison, 2009). In the propensity score model, all variables included were precollege

characteristics selected based on their expected impact on participation in the treatment and/or

outcome variables (see Mayhew et al., 2016; Pascarella & Terenzini, 2005). These variables

include students’ academic motivation (8 items, α = .70), need for cognition (18 items, α = .89),

time spent in high school activities (10 individual items), highest intended degree, standardized

test scores, high school GPA, race, sex, and level of parental education . Additionally, speaking

English as a second language and being an international student were also included, since these

may be associated with students’ decisions regarding campus residence. Finally, pretest variables

of the outcomes of interest included students’ perception of their overall health before college,

students’ high school alcohol and smoking habits, and students’ precollege psychological well-

being (α = .89).

Analysis

The first step was to create a score representing one’s propensity to live on campus.

Logistic regression was used to compute a propensity score (see Guo & Fraser, 2015; Pan & Bai,

2015). To create this score, each precollege variable was entered as the lone student-level

predictor of campus residence while also including institutional fixed effects. This study used

stratification, which is one of several propensity score approaches used to account for self-

selection bias (e.g., see Austin, 2011; Guo & Fraser, 2015; Hong & Raudenbush, 2006). The

propensity score was divided into 10 equal strata; within each stratum, cases were trimmed so

that each contained students in both conditions that were comparable. Support for balance was

assessed through a two-way analysis of variance predicting the linear propensity score with strata
94
and treatment condition (campus residence) as independent variables, which yielded no

significant main effect of treatment condition nor any interaction effect (Thoemmes & Kim,

2010). Additional support for balance included visual inspection of the region of common

support across strata as well as assessing standardized mean differences before and after

stratification. After stratification, each predictor variables had a standardized mean difference of

less than .05 except for one, which was less than .10.

To assess the effect of campus residence on the study outcomes, ordinary least squares

regression models with clustered robust standard errors predicted each outcome, with campus

residence, the PSM strata, and institutional fixed effects as predictors. The outcomes were treated

as continuous and standardized so that the coefficients would convey the difference between

living on campus and off campus in terms of standard deviation units. To explore whether the

potential impact of campus residence varies across groups, additional analyses included

interaction terms for the residence variable and several student-level variables (race, sex, parental

education, and high school test scores). These analyses examined each interaction separately to

avoid issues with multicollinearity.

Results

Table 1 contains results for the full analytic sample before and after balancing with

propensity score stratification. Analyses without a propensity score adjustment indicate that

living on campus predicts greater binge drinking in both years, as well as greater alcoholic

beverage consumption, psychological well-being, overall health, and time spent exercising in the

first year. After the propensity score adjustment, only exercise, alcoholic consumption, and rates

of binge drinking remained significant. It appears living on campus has a small to medium effect

on students’ frequency of exercise and alcohol use, with standardized mean differences of .18 to

.26 (Mayhew et al., 2016).

95
 Moderation analyses between campus membership and four student characteristics (sex,

race/ethnicity, parental education, and standardized test scores) found some significant

interactions. The effect of campus residence on PWB was significantly more positive for female

students than for male (B = -.25, SE = .12, p = .049). Additionally, living on campus is more

positive for higher ability students than their counterparts (B = .45, SE = .20, p = .035).

Exploring race, there were three significant interactions pertaining to living on campus during

the first year meaning that the effect of living on campus is larger for these students: smoking

among Black students at the end of the first year (B = .063, SE = .02, p = .001), exercise among

Asian American and Pacific Islander students at the end of their fourth year (B = .694, SE = .113,

p = .001), and alcohol consumption among Latinx students at the end of their first year (B = .644,

SE = .12, p = .0001).

Discussion

This study adds to the existing campus residence literature through utilization of a multi-

institutional sample, accounting for self-selection through quasi-experimental methodology, and

exploring differential effects across several groups of students. The findings suggest campus

residence has the strongest effect on students’ alcohol behaviors and exercise habits at the end of

their first year. For students living on campus, new peer associations, along with a lack of direct

parental or guardian supervision, could explain the increase in reported alcohol consumption and

binge-drinking behaviors during this first year. Additionally, these students might be gravitating

towards peer norms and cultures that affect decisions pertaining to exercise and alcohol.

However, the effect of living on campus on these specific outcomes did not persist

beyond the first year. It might be that after their first year, students develop a sense of self and

are less reliant on the campus peer culture, thus any direct effect that campus residence had is no

longer relevant. It might be that the novelty of new peer groups or the new campus environment
96
may wear off. Student affairs professionals, especially those working with first-year students,

might find ways to promote social norming marketing campaigns to affect students’ perceptions

of alcohol behaviors on campus to better represent the campus norm (e.g., DeJong et al., 2006;

Turner, Perkins, & Bauerle, 2008). In brief, these findings suggest that student affairs

practitioners need to better understand peer cultures in residence halls and why they lead students

to drink more frequently and engage in higher rates of binge drinking.

This study further suggests that there may be differential effects as a function of

demographics; that is, the effect of living on campus is larger for some groups than others. The

effect of living on campus on PWB was more positive for female students than for males which

is contrary to prior literature that found uniformly negative associations with PWB (Mayhew et

al., 2016). Understanding why this finding is the case might lead housing professionals to create

some intentional programming related to psychological well-being and potentially target male

students. Looking at the effect of campus residence within racial categories illustrates differential

impacts on smoking for Black students, exercise for Asian American and Pacific Islander

students, and drinking for Latinx students. Further research with a larger analytic sample is

needed to verify these findings and ensure they were not a product of random chance.

97
Table 6. Unstandardized Regression Coefficients for Relationship of Campus Residence with
Psychological Well-Being and Student Health Outcomes after First and Fourth Year of College
First Year Fourth Year
Before PS After PS Before PS After PS
Balance Balance Balance Balance
Outcome B SE B SE B SE B SE
Psychological 0.18* 0.07 0.07 0.10 0.09 0.12 0.05 0.12
well-being
Overall health -0.15* 0.06 -0.11 0.10 0.01 0.08 0.02 0.11
Smoking -0.01 0.97 -0.05 0.08 -0.24 0.22 -0.08 0.14
Binge-drinking 0.22*** 0.07 0.22* 0.10 0.34*** 0.08 0.17 0.09
Alcohol 0.23** 0.07 0.18* 0.07 0.31 0.16 0.21 0.12
consumption
Exercise 0.29*** 0.08 0.26*** 0.07 0.25 0.13 0.09 0.10

Note. Each year includes coefficients before and after propensity score adjustment. Institutional fixed
effects were included in all analyses. All outcomes were examined with ordinal least squares multiple
regression analyses; these outcomes were standardized with a mean of zero and a standard deviation
of one. *p<0.05 **p<0.01 ***p<0.001

98
 References

Allison, P. D. (2009). Fixed effects regression models. Thousand Oaks, CA: Sage.

Austin, P. C. (2011). An introduction to propensity score methods for reducing the effects of

confounding in observational studies. Multivariate Behavioral Research, 46, 399–424.

Benz, M. B., DiBello, A. M., Balestrieri, S. G., Miller, M. B., Merrill, J. E., Lowery, A. D., ... &

Carey, K. B. (2017). Off-campus residence as a risk factor for heavy drinking among

college students. Substance Use & Misuse, 52(9), 1236-1241.

Brevard, P. B., & Ricketts, C. D. (1996). Residence of college students affects dietary intake,

physical activity, and serum lipid levels. Journal of the American Dietetic

Association, 96(1), 35-38.

Brookhart, M. A., Schneeweiss, S., Rothman, K. J., Glynn, R. J., Avorn, J., & Stürmer, T.

(2006). Variable selection for propensity score models. American Journal of

Epidemiology, 163(12), 1149-1156.

Borsari, B., Murphy, J. G., & Barnett, N. P. (2007). Predictors of alcohol use during the first year

of college: Implications for prevention. Addictive Behaviors, 32(10), 2062-2086.

Bowman, N. A. (2010). The development of psychological well-being among first-year college

students. Journal of College Student Development, 51(2), 180-200.

Centers for Disease Control. (2018). Well-being concepts. Retrieved from

https://www.cdc.gov/hrqol/wellbeing.htm

Cook, T. D., & Steiner, P. M. (2010). Case matching and the reduction of selection bias in quasi-

experiments: the relative importance of pretest measures of outcome, of unreliable

measurement, and of mode of data analysis. Psychological Methods, 15(1), 56-68.

99
Cross, J. E., Zimmerman, D., & O'Grady, M. A. (2009). Residence hall room type and alcohol

use among college students living on campus. Environment and Behavior, 41(4), 583-

603.

DeJong, W., Schneider, S. K., Towvim, L. G., Murphy, M. J., Doerr, E. E., Simonsen, N. R., &

Scribner, R. A. (2006). A multisite randomized trial of social norms marketing campaigns

to reduce college student drinking. Journal of Studies on Alcohol, 67(6), 868-879.

Guo, S., & Fraser, M. (2015). Propensity score analyses: Statistical methods and applications.

2nd ed. Thousand Oaks, CA: SAGE Publications, Inc.

Hong, G., & Raudenbush, S. W. (2006). Evaluating kindergarten retention policy: A case study

of causal inference for multilevel observational data. Journal of the American Statistical

Association, 101(475), 901-910.

Mayhew, M. J., Bowman, N. A., Rockenbach, A. N., Seifert, T. A., & Wolniak, G. C.

(2016). How college affects students: 21st century evidence that higher education

works (Vol. 3). San Francisco, CA: Jossey-Bass.

Pan, W., & Bai, H. (Eds.). (2015). Propensity score analysis: Fundamentals and developments.

New York, NY: Guilford Publications.

Pascarella, E. T., & Terenzini, P. T. (1991). How college affects students: Findings and insights

from twenty years of research. San Francisco, CA: Jossey-Bass.

Pascarella, E. T., & Terenzini, P. (2005). How college affects students (Volume 2). San

Francisco, CA: Jossey-Bass.

Perkins, H. W. (2002a). Social norms and the prevention of alcohol misuse in collegiate

contexts. Journal of Studies on Alcohol, (14), 164-172.

100
Perkins, H. W. (2002b). Surveying the damage: a review of research on consequences of alcohol

misuse in college populations. Journal of Studies on Alcohol, (14), 91-100.

Renn, K. A., & Arnold, K. D. (2003). Reconceptualizing research on college student peer

culture. The Journal of Higher Education, 74(3), 261-291.

Ryff, C. D. (1989). Happiness is everything, or is it? Explorations on the meaning of

psychological well-being. Journal of Personality and Social Psychology, 57(6), 1069-

1081.

Ryff, C. D., Keyes, C. L., & Hughes, D. L. (2003). Status inequalities, perceived discrimination,

and eudemonic well-being: Do the challenges of minority life hone purpose and

growth?. Journal of health and Social Behavior, 275-291.

Small, M., Bailey-Davis, L., Morgan, N., & Maggs, J. (2013). Changes in eating and physical

activity behaviors across seven semesters of college: living on or off campus

matters. Health Education & Behavior, 40(4), 435-441.

Smider, N. A., Essex, M. J., & Ryff, C. D. (1996). Adaptation to community relocation: The

interactive influence of psychological resources and contextual factors. Psychology and

Aging, 11(2), 362-372.

Springer, K. W., & Hauser, R. M. (2006). An assessment of the construct validity of Ryff’s

scales of psychological well-being: Method, mode, and measurement effects. Social

Science Research, 35(4), 1080-1102.

Thoemmes, F. J., & Kim, E. S. (2011). A systematic review of propensity score methods in the

social sciences. Multivariate behavioral research, 46(1), 90-118.

101
Timberlake, D. S., Hopfer, C. J., Rhee, S. H., Friedman, N. P., Haberstick, B. C., Lessem, J. M.,

& Hewitt, J. K. (2007). College attendance and its effect on drinking behaviors in a

longitudinal study of adolescents. Alcoholism: Clinical and Experimental

Research, 31(6), 1020-1030.

Turner, J., Perkins, H. W., & Bauerle, J. (2008). Declining negative consequences related to

alcohol misuse among students exposed to a social norms marketing intervention on a

college campus. Journal of American College Health, 57(1), 85-94.

Wechsler, H., & Nelson, T. F. (2008). What we have learned from the Harvard School of Public

Health College Alcohol Study: Focusing attention on college student alcohol

consumption and the environmental conditions that promote it. Journal of Studies on

Alcohol and Drugs, 69(4), 481-490.

Westreich, D., Cole, S. R., Funk, M. J., Brookhart, M. A., & Stürmer, T. (2011). The role of the

c‐statistic in variable selection for propensity score models. Pharmacoepidemiology and

Drug Safety, 20(3), 317-320.

Wetter, D. W., Kenford, S. L., Welsch, S. K., Smith, S. S., Fouladi, R. T., Fiore, M. C., & Baker,

T. B. (2004). Prevalence and predictors of transitions in smoking behavior among college

students. Health Psychology, 23(2), 168-177.

102
 CHAPTER FOUR: COLLEGIATE RESIDENCE AND STUDENT ENGAGEMENT

Student engagement matters in contemporary residence halls. Housing administrators and

campus architects posit that residence halls affect student engagement through their proximity to

the center of campus as well as the room type in which students live (Biemiller, 2017). George

Kuh (2009a) argues that this focus on student engagement is not new, rather “when the history of

American higher education is rewritten years from now, one of the storylines of the first

decade…will be the emergence of student engagement as an organizing construct” (p. 5). He

continues to suggest that engagement is paramount for students because it “helps to develop

habits of the mind and heart that enlarge them for continuous learning and personal

development” (Kuh, p. 5). Tracing the history of on campus housing illustrates why engagement,

as an organizing construct, is not new.

Higher education administrators, rather, operate from an implicit assumption that living

on campus is positively related to student engagement. Among the earliest institutions within the

United States, historians document that one of easiest ways to engage students was to house them

on campus (Rudolph, 1990; Thelin, 2011). As higher education historian Frederick Rudolph

termed it, students and faculty lived and learned together in “the collegiate way,” characterized

by frequent contact and close community. This sense of community was said to be beneficial to

the student experience and promoted by living on campus in residential colleges. Professors and

higher education administrators suggested that this collegiate way afforded residents a common

sense of purpose towards a pursuit of scholarship and that by forming residential colleges the

institution becomes “small enough to enable its members to experience university life on a

smaller and more human scale – a scale that is both manageable and intimate” (Ryan, 2016, p.

XI). This engagement within the institution was thought to be important for students’ social

103
connections and academic achievement. Thus, this collegiate way of living permeated and

became part of an institutional ethos regrading postsecondary education.

As faculty priorities shifted from the out of class experiences of students towards an

interested in knowledge specialization and research, the field of student affairs began to emerge

in order to fill this subsequent out-of-classroom gap and to promote student engagement

(American College Personnel Association, 1996; MacKinnon & Associates, 2004). Modern

theorists have suggested that professionals should attend to where and how students spend their

time. Pace (1984) wrote that focus be payed to the quality of students’ effort, while Astin (1984)

added that students’ involvement is a function of both quantity and quality of this effort. The two

perspectives imply that how and where students spend their time matters. As such, higher

education theorists and administrators alike have assumed that students living on campus are

able to invest their time and effort differently than their peers off campus because of residence

and proximity to campus alone.

However, these assumptions have rarely been examined in the literature and warrant

further exploration. For many higher education administrators across the United States, student

engagement has become a top priority due to the apparent link to student success (Kuh, 2003,

2009b; Pascarella, Seifert, & Blaich, 2010). Engagement is a term “used to represent constructs

such as quality of effort and involvement in productive learning activities” (Kuh, 2009b, p. 6)

however others might argue these measures really reflect quantity of effort not quality. While

there is a substantial body of literature suggesting specific educationally purposeful activities are

positively associated with a variety of student outcomes, there is a dearth of literature that has

specifically studied the direct role that living on campus plays in student engagement.

104
 Relevant Literature

Student Engagement and Student Outcomes

Research on student engagement has demonstrated positive links to a variety of student

outcomes. Often the research links particular types of engagement with specific outcomes that

include learning and achievement, psychosocial growth and development including interpersonal

development. Student engagement on campus contributes to sense of belonging, satisfaction with

collegiate experiences, and thriving (Berger & Milem, 1999; Braxton et al., 2004; Cuevas, 2015;

Kuh et al., 2006; Pascarella & Terenzini, 2005; Strayhorn, 2012). Other literature examines links

between educationally purposeful activities and grades and persistence, with positive findings for

both (Kuh, Cruce, Shoup, Kinzie, & Gonyea, 2008; McCormick, Kinzie, & Gonyea, 2013;

Reason, 2009). Questioning quantity over quality, there is research that illustrates a seemingly

linear relationship between engagement and academic achievement (Huang & Chang, 2004),

while others have found there can be overinvolvement leading to negative effects on GPAs

(Emerick, 2005). Finally, higher levels of student engagement are positively associated with a

sense of belonging on campus (Cheng, 2004; Elkins et al., 2011). In short, the literature suggests

that student engagement still matters.

Living on Campus and Student Engagement

One way living on campus is theorized to positively affect student engagement is through

proximity to campus resources. Chickering (1974) asserted that resident students’ “presence on

campus, their easy access to pertinent information and to the grapevines that carry it, make it

more possible for them to find educational programs and experiences that suit their interests and

abilities” (p. 105). Astin (1985) added that simply spending their entire lives on campus, students

in residence halls are more likely to identify as being a college student and subsequently with

105
their institution. Each recognized that living on campus is influential on college students’

engagement with their peers, academic environments, learning, and connection to the institution.

These theories are based on students in the past and may not apply to contemporary students

living in the times of social media and technology that allows for immediate connections and

social integration to the university in a different capacity.

Often literature that is focused on student engagement reports its relationship with

campus residence as a result of inclusion of that information as a covariate in researchers’

statistical models instead of it being the researcher’s primary question of interest. Synthesizing

contemporary literature, the most recent edition of How College Affects Students supports the

notion that most often involvement is used as a predictor variable to explain various outcomes,

rather than the outcome itself. One particular instance suggests that social involvement seemed to

explained the link between campus residence and retention (Mayhew, Rockenbach, Bowman,

Seifert, & Wolniak, 2016).

One of the only studies to specifically look at campus residence as it directly relates to

student engagement comes from Graham and colleagues. Using data from the National Survey of

Student Engagement, Graham, Hurtado, and Gonyea (2018) examined how living on campus is

associated with a variety of student engagement outcomes including collaborative learning,

discussions with diverse others, student-faculty interactions, quality of interactions on campus,

supportive campus environments, and co-curricular gains measured by a scale representing

working effectively with others, solving complex real-world problems, understanding people of

different backgrounds, being an active and informed citizen, and developing or refining a

personal code ethics. For all seven outcome variables, students who lived on campus reported

higher levels of engagement when compared to their off-campus peers. Additionally, coding the

106
living variable as living on campus, within walking distance, and further than walking distance

allowed the researchers to explore the effect of proximity to campus. The further a student

reported living from campus, the more negative of an effect on their engagement. While the

study finds power through the utilization of a large institutional sample, it is limited in by a

cross-sectional first-year sample that is unable to account for any potential longitudinal

engagement gains within college. Additionally, all of the data are self-reported and only included

a few student precollege characteristics in the final model.

Living on Campus and Peer Interactions

In addition to proximity to campus environments, living in residence halls is thought to

affect students’ peer interactions and formations. Astin (1993) defined peer groups as “any group

of individuals in which the members identify, affiliate with, and seek acceptance and approval

from each other” (p. 401). For students living on campus, their peer groups are made of students

also living on campus, whereas this type of peer group may be smaller or nonexistent for off-

campus students (Astin, 1985; Chickering, 1974). Residence life professionals have long

believed an important benefit to living on campus is the educational programming and peer

connections unique to residence hall living (Blimling, Whitt, & Associates, 1999). Terenzini and

Pascarella (1984) wrote:

Residential units were presumed to be significant environments through both their

sociological structures and the normative influences exerted by their occupants.

Structurally, residence units might be expected to influence the nature of students’

collegiate experiences both through their physical configurations and consequent

influence on the nature and extent of students’ interactions with one another, and through

the sorts of rules that govern student behaviors, as well as the academic social

107
 experiences afforded students through the nature of the social and academic

programming conducted within the residence unit. (p. 114).

This earlier research focuses on the direct effect of peer groups upon their peers’ growth,

development, and success (Astin, 1993; Milem, 1998). The assumption was that residence halls

“maximize opportunities for social, cultural, and extracurricular involvement, and this increased

involvement will account for residential living’s impact on various indices of student

development” (Pascarella, Terenzini, & Blimling, 1994, p. 25). These classic theories highlight

the importance of social integration and sense of belonging for students and its role in students’

decisions to stay in or leave postsecondary education (e.g., Museus, 2014; Tinto, 1993). In other

words, higher education professionals have long believed peer groups can have a strong

influences on the student experience, especially with whom and in how students interact as well

as how campus is perceived (Antonio, 2001).

Contemporary literature exploring peer interactions for students living on campus has

often focused on the physical layout of on campus residence halls, specifically as it pertains to

student engagement. As residence hall spaces have grown from traditional dormitory style living

to offer enhanced residence hall spaces, amenities, and other features, these changing styles of

living spaces offer another complicating factor in peer interactions and student engagement. La

Roche, Flanigan, and Copeland Jr. (2010) put it this way: “What were once considered to be

luxuries in student housing—kitchens, private bedrooms, private bathrooms, social spaces and

lounges—are now expected” (p. 46). Different types of residence halls, such as suite-style versus

traditional dormitories, can have an impact on student interactions. Recently constructed or

renovated residence halls now include amenities such as workout facilities, full-service dining

options, coffee shops, private bathrooms, and walk-in closets (Kavehkar, 2013; Lederman, 2009)

108
as well as technological updates to limit outsider access and increase safety (O’Neil, 2014). As

such, many institutions have moved away from the traditional residence hall with long hallways

and community bathrooms and replaced them with suite-style or apartments (Palmer, Broido, &

Campbell, 2008).

However, despite the increase of student apartments and suites, those students living in

traditional halls are found to have more frequent interactions with other residents than their

counterparts in suite-style halls (Brandon, Hirt, & Cameron, 2008). Additionally, others found

that students living in traditional halls have more positive and mature interpersonal relationships

than the students living in super-suite and adjoined suite halls (Owen, 2010). In terms of sense of

belonging, satisfaction with college, and a student’s intent to persist, research suggests, there is

no difference between traditional residence halls and suite-style apartments (Bronkema &

Bowman, 2017). Additional research corroborates the notion that students living in traditional

residence hall spaces have a greater sense of community compared to their peers in suite-style

rooms (Devlin, Donovan, Nicolov, Nold, & Zandan, 2008). However, these findings suggest that

it less about the physical structure of the halls and more about which students are within the

environment. For spaces with only first year students, those students had better outcomes

compared to peers in mixed-year halls and for students living in upper class spaces, there was a

decreased sense of belonging and satisfaction with the institution. In other words, the peer

cultures affect student outcomes.

For student engagement specifically, students in suite-style buildings exhibited higher

levels of activity than their peers in traditional dorm-style residences (Rodger & Johnson, 2005),

with the researchers operationalizing activity level through a checklist of activities on campus

that fall into the following groups: co-curricular experiences, interactions with faculty, and

109
engagement with campus programming. It might be that these students sought opportunity

outside of their living spaces to connect and engage with the broader community whereas

students in traditional spaces already had a fair amount of social connection as a function of their

community. One might expect the same dynamic for those students who live in apartments near,

but not on campus. That is, given the proximity to campus, these near-campus apartment

dwellers might seek social connections outside of their living environments.

Complicating the narrative of positive effects of living on campus in regard to peer group

interactions is an acknowledgement of uneven benefits for students of differing identities. While

residence halls have the potential to foster positive interactions with students from diverse

backgrounds, they can alternatively encourage groupthink and incubate spaces for hostile

discriminatory practices to persist (Blimling, 1993; Harper et al., 2011; Pascarella & Terenzini,

2005; Strayhorn & Mullins, 2012). For example, when examining the experiences of Black, gay,

male undergraduate students, residence hall policies and programming perpetuated heterosexism,

homophobia, and isolation (Strayhorn & Mullins). The physical environment for Black students

living on campus is often times challenging. For these students, the spaces on campus often lack

Black cultural representation and little to no places that offer escapes from racial micro-

aggressions (Hotchkins & Dancy, 2017). In a similar study to Hotchkins and Dancy, students

identifying as African American, Asian American, Latino, and Native American experienced

over 70 distinct racial microaggressions while living in residence halls (Harwood, Huntt,

Mendenhall, & Lewis, 2012). For those living on campus, one study found that white women

seemed to have the highest sense of belonging compared to white men, and men and women of

color (Garvey et al., 2018). These findings highlight the varied benefits of living on campus by

110
race. While living on campus increases the frequency of social interactions among peers, not all

are positive or lead to greater levels of engagement.

Conceptual Framework

The current study relies on three complimentary conceptual frameworks to guide the

study: (1) Kuh’s (2009b) work on student engagement; (2) Chickering and Gamson’s (1987)

seven principles for good practice in undergraduate education; and (3) Renn and Arnold’s (2003)

adaptation of Bronfenbrenner’s (1979, 1989, 1993) Process-Person-Context-Time (PPCT)

ecological systems theory. Contemporary work regarding student engagement traces its roots

back to earlier work on engagement theory (Astin, 1991; Pace, 1980, 1984; Tyler, 1932), which

posits that how a student spends their time can be directly related to their educational success. As

such, the current study operationalizes student engagement based on Kuh (2009b) and the

National Study of Student Engagement (NSSE, 2018). Additionally, Chickering and Gamson’s

(1987) add that the importance of contact among faculty and students as well as reciprocity and

cooperation between students in helping improve undergraduate education. As such outcomes of

interest can be represented by collaborative learning, time spent with faculty, interactions with

staff, the quality of peer interactions, engaging with diverse others, and co-curricular

involvement. While these frames suggest experiences that are beneficial to students, they do not

necessarily account for behavior change or the causal mechanism by which change occurs.

Ecological theories can offer insight as to why these particular experiences are important and

how living on campus can affect student engagement.

Bronfenbrenner’s (1979, 1989, 1993) Process, Person, Context, and Time (PPCT) model

counters these college impact models by situating an individual within a complex series of

systems (macrosystems, exosystems, mesosystems, and microsystems), and it is through

111
interactions with these complex environments that growth and development occurs. Renn and

Arnold (2003) extended Bronfenbrenner’s work to higher education. Their conceptualization of

the four types of systems (micro, meso, exo, and macro) are defined from a collegiate context.

For many students, particularly traditionally-aged students whom make up a majority of the

analytic sample, microsystems are associated within a college or university and classrooms,

living environments, campus facilities, and other daily-life environments. For residential

students, their systems are likely to overlap with a consistent group of other students also living

on campus. In other words, one’s living environment creates unique opportunities for person-

context interactions that differ for on and off campus students because residence halls create

additional opportunities for microsystems to overlap.

Current Study

Addressing shortcomings of past residential literature that specifically looks at student

engagement as an outcome, this study utilizes propensity score modeling to reduce bias

associated with students’ self-selection into campus residence in hopes of better estimating the

causal impact of living on campus (Austin, 2011; Guo & Fraser, 2015; Hong & Raudenbush,

2006; Mitra & Reiter, 2012; Thoemmes & Kim, 2011). Additionally, this study leverages the

benefits of a large, longitudinal, multi-institutional dataset that captured a host of precollege

constructs and pretests on a number of outcome variables. Ultimately, this study seeks to address

what role collegiate residence plays in shaping various forms of student engagement.

Additionally, this study seeks to explore the conditional effect of living on campus by first-

generation status, race/ethnicity, socioeconomic status, and prior academic achievement.

Specifically, then, the study seeks to answer three main research questions:

112
 1) Does living on campus affect students’ levels of engagement at the end of their first

and fourth years?

2) Are the effects of student engagement and place of residence moderated by students’

minoritized identities?

3) Does moving off campus after a students’ first year affect student engagement

outcomes at the end of students’ fourth year?

Methods

Institutional Sample and Data Collection

The data for the current study come from students who participated in the Wabash

National Study of Liberal Arts Education (WNSLAE). The WNS is a multi-institutional,

longitudinal study that included 46 four-year institutions across the United States; it was

designed to examine student experiences and outcomes within institutions stating a commitment

to liberal arts education. Institutions represent a wide range of geographic diversity, size,

selectivity, tuition costs, and missions as well as religious and single-sex colleges and

universities.

The respondents included three cohorts of students who entered college between 2006

and 2008. Within each cohort, the WNS researchers collected data at three distinct time points.

The first wave of data, occurring either in the summer prior to their first semester or early in the

fall of their first year of college (Time 1), gathered information about students’ demographic

characteristics, precollege abilities, behaviors, experiences, and attitudes. The second data

collection happened at the end of students’ first year of college (Time 2), with one final data

collection at the end of students’ fourth-year of college (Time 3). During both time 2 and 3,

students completed the WNSLAE Student Experiences Survey and the National Survey of

113
Student Engagement, both of which provided information about their college experiences,

attitudes, and behaviors.

Analytic Sample

For the first two research questions, the analytic sample includes 7,855 students who

responded to questions at Time 2 and 5,747 students answering the measure at Time 3. The

analytic sample addressing the third research question includes 3,752 students who responded at

Times 1, 2, and 3. The first two research questions use collegiate residence in the first year to

predict student outcomes at the end of their first year and fourth year. The third research question

uses a comparison group of students who lived on campus throughout their experience to those

who moved off campus after their first year on their fourth-year outcomes. Due to the

longitudinal nature of the data as well as the inclusion of a number of control variables in the

propensity score models, up to 14% of information was missing across several variables and

timepoints. Multiple imputation using 40 datasets was employed to account for the missing data

(e.g., Little & Rubin, 2002). Larger number of imputed datasets create more variability from

these imputations in order to produce larger standard errors and reduce the likelihood of a Type I

error. While dependent variables were included in the imputation equations, the imputed values

for the outcome variables were removed before conducting any primary analyses (von Hippel,

2007).

This study also utilizes weighting to make the sample more representative of the

incoming first-year cohorts of those participating institutions. Specifically, the variables used to

create the weights were sex, race, and academic ability. Weights were used to help account for

higher rates of survey nonresponse as well as college attrition among underrepresented students

(Biemer & Christ, 2008; Groves et al., 2009). The weights also adjusted for potential differences

114
between the institutional population and study participants as a result of the probability of being

included in the study. To prevent changes to the analytic sample size, weights for this study were

normalized with a mean of 1.0. Within the weighted sample, 6 percent of students identified as

Hispanic or Latinx, 7 percent identified as Black or African American, and 8 percent of students

identified as Asian or Pacific Islander, while 38 percent of students identified their sex as male.

Students attended a liberal arts college (50 percent), a research university (32 percent), or a

regional university (18 percent).

Measures

Dependent Variables. For this study, student engagement was operationalized using

scales to capture facets of Chickering and Gamson’s (1987) good practices as well as items that

capture the NSSE’s engagement indicators (National Survey of Student Engagement, 2018).

These scales at times 2 and 3 were quality of interactions with faculty outside of the classroom (5

items, α = .85 and .87, respectively), frequency of interactions with faculty (4 items, α = .71 and

.75), degree of positive peer interactions (8 items, α = .87 and .85), cooperative learning (4 items,

α = .71 and .70), diversity experiences (6 items, α = .72 and .75), meaningful discussions with

diverse peers (3 items, α = .83 and .84), negative diversity interactions with peers (5 items, α =

.84 and .86), and frequency of interactions with student affairs staff (5 items, α = .85 and .88).

Additionally, a single item measured co-curricular involvement via the number of hours in these

activities.

Treatment Outcome Variables. The primary independent variable indicated students’

collegiate residence. The variable was recoded from a variable with several categories to a

dichotomous variable indicating living on campus or off campus. Students who reported living in

a fraternity or sorority house at Time 2 were dropped from analyses due to the inability to

115
determine if the house was located on or off campus (n = 265). The third question of interest was

answered by creating a variable in which students who responded they lived on campus at both

Time 2 and Time 3 were compared to those students who lived on campus at Time 2 but then

reported moving off campus at Time 3. Students who reported living off campus at Time 2 who

reported lived on campus at Time 3 were excluded from analysis, as the research question was

focused on the effect moving off campus after having lived on campus during the first year.

Propensity Score Covariates. The research questions of this study require two distinct

propensity score models; the first estimates students’ propensity to live on campus during the

first year, and the second model examines students’ propensities to move off campus during their

fourth year after having lived on their first year. The first model incorporates student

demographic information and precollege characteristics that could affect students’ decisions to

live on or off campus, while the second model adds additional covariates that capture students’

collegiate experiences during their first year of college, as these social experiences might affect

decisions to stay on campus or move off. Within each model, institutional fixed effects were

used to account for the between-institution differences due to the nested nature of the analytic

sample, with students within institutions (Allison, 2009; Arpino & Mealli, 2011; Li, Zaslavsky,

& Landrum, 2013).

Propensity score covariates were used based off of college impact theories and include

student demographics, precollege characteristics, and institutional fixed effects (Astin, 1991;

Credé & Niehorster, 2012; Kuh et al, 2008; Mayhew, et al., 2016; Pascarella and Terenzini,

1995, 2001; Radford, Berkner, Wheeless, & Shepherd, 2010; Schudde, 2011). For the first model

predicting students’ propensity to live on or off campus for the first year, covariates included

precollege student demographics and characteristics, high school academic and co-curricular

116
involvement, experiences in high school with peers and teachers, social behaviors (including

smoking and drinking), academic motivation, need for cognition, highest intended degree,

standardized test scores, academic motivation, and precollege experiences with diversity (full

variable descriptions are available in the appendix). Controls for students’ background

characteristics consist of precollege academic ability (measured through high school GPA), race,

sex (male or female), and the level of parental education (first-generation or not) and were

specifically included as each of these have been theorized to predict college student engagement

and outcomes (e.g., Astin, 1993; Mayhew et al., 2016; Pascarella & Terenzini, 2005).

College impact theories mainly guided the covariate selection for the second propensity

score model. This second model aimed to reduce bias associated with students’ decisions to

move off campus after the first year. In addition to these covariates, the model included scales

that were related to students’ first-year student engagement. These time 2 measures act as

pretests for the corresponding Time 3 outcomes. Specifically, the measures were quality

interactions with faculty outside of the classroom, frequency of interactions with faculty, degree

of positive peer interactions, cooperative learning, diversity experiences, meaningful discussions

with diverse peers, negative diversity interactions with peers, avoidance of negative diversity

experiences, and frequency of interactions with student affairs staff. It is recommended that

pretests be included in the propensity score model (Cook & Steiner, 2010). Additional literature

also offered theoretical justification or empirical support for the inclusion of these covariates into

the propensity score model (Astin, 1993; Berger & Milem, 1999; Garza & Fullerton, 2018;

Graham, Hurtado, & Gonyea, 2018; Kuh, 2009; Krause & Coates, 2008; Simpson & Burnett,

2017; Walsh & Kurpius, 2016). For both propensity scores, the decision was made to keep all

117
theoretically warranted variables used regardless of pre-stratification significance or standardized

mean difference (Brookhart et al., 2006; Patrick et al., 2011; Westreich et al, 2011).

Limitations

Despite the advances this study offers to the literature, there are some limitations that

need to be noted. First, although quasi-experimental methodologies can offer results that are

better estimates of causal effects, that ability corresponds to the observed covariates included in

the model and it is possible that the current model does not include other important variables that

could be linked to a students’ propensity to live on campus. Second, the campus residence

variable only indicates where a student reports living and not the experience had in that

environment. Institutional resources and professional staffing vary by institution, affecting the

student experience across the institutions. Third, the variables used in the outcome analyses are

self-reported. Some literature suggests that students may not be able to accurately report their

experiences and, as such, the results might be biased (Porter, 2011). Fourth, the data is limited in

answering the last research question in that there is no way of knowing when students reporting

moving off campus after the first year. The assumption was made that those who answered in the

affirmative for both the end of first and fourth years stayed on campus the entire time and did not

move off campus.

Analyses

Descriptive Statistics. Tables 7 and 8 offer summary statistics based on collegiate

residence during the first year (Table 7) as well as moving off campus after the first year (Table

8). These tables illustrate differences between those students living on and off campus their first

year. In most cases, those living on campus have higher means on the standardized scaled

student engagement indicators. For example, students living on campus have a higher mean

118
(0.05) compared to those off campus (-0.52) in terms of co-curricular involvement at the end of

the first year. Similar patterns emerge at both the end of the first and fourth years for non-

classroom interactions with faculty, positive peer interactions, and diversity experiences on

campus. In terms of students’ background characteristics, those who are more likely to live on

campus are White students, students with higher standardized test scores, and higher self-

reported parental income. Students with minoritized backgrounds including those who are first-

generation, non-native English speakers, Asian or Pacific Islander, Latinx, or with lower-ability

tend to live off campus in their first year at higher rates.

In terms of collegiate residence after students’ first year, it appears that those who spent

more time in social activities such as exercising, socializing, studying with friends, and working

for pay tended to live off campus at higher rates. In terms of academic covariates, interacting

with faculty inside and out of the classroom differed the most among those who moved off

campus versus those staying on campus. Additionally, comparing the two groups, it seems that

White students and those with higher parental income move off campus more frequently than

their peers.

Propensity Score Creation. The removal of bias from propensity score modeling allows

the research to mimic random allocation within experiments under ideal circumstances. As such,

this study uses information from precollege experiences, behaviors, and demographics to

generate a propensity score that is balanced on each of the observed covariates. If balancing is

achieved, the quasi-experimental methodology statistically creates groups that are equal in

expectation based on the observed covariates to compare students who live on campus to their

peers off campus on a variety of student engagement measures (Austin, 2011; Guo & Fraser,

119
2015; Hong & Raudenbush, 2006; Mitra & Reiter, 2012; Thoemmes & Kim, 2011). The same

steps were followed for each model to generate the propensity score for collegiate residence.

Overall, first a logistic regression model was fit to compute a single, linear propensity

score, followed by bias reduction assessment, ending with determining if balance had been

achieved as a result of stratification based on the propensity score. Prior to the logistic

regression, each predictor to be included in the propensity score model was correlated with all

the other variables to explore if any variables were collinear. Within the correlation matrix, there

were no correlations above 0.37. Additionally, collinearity was further examined by exploring

variance inflation factors after each propensity score model. These collinearity diagnostics, as

assed by VIF, were well below the recommended 2.00 threshold and as such offer further

support for the decision to retain each variable in the propensity model (Craney & Surles, 2002).

To create the propensity score, a logistic regression model was fit to determine a single,

linear propensity score for each participant (rather than using one’s predicted probability; see

Guo & Fraser, 2015; Pan & Bai, 2015). Each variable was entered as the lone student-level

predictor of living on campus while including institutional fixed effects within the model.

Scholars suggest that including all variables, not only the significant predictors related to the

treatment or outcome, does not affect the propensity score (Brookhart et al., 2006; Patrick et al.,

2011; Westreich et al, 2011) and as such all theoretically driven variables that might predict

living on campus or student engagement were included in the final propensity score model.

Outcome Analyses. To assess the effect of campus residence on student engagement

outcomes, this study used an ordinary least squares regression model with clustered robust

standard errors to predict each outcome separately. Within each model, the outcome of interest

was predicted by the collegiate residence variable, the PSM strata variable accounting for bias

120
associated with residential choice, and the institutional fixed effects variables with each entered

as independent variables. Each student engagement outcome was treated as continuous and

standardized within the analytic sample so that each beta coefficient would convey the difference

between collegiate residence on and off campus in terms of standard deviation units.

Results

Propensity Score Analyses

This study employed stratification to determine whether the bias associated with

covariates predicting students’ collegiate residential choice has was successfully controlled for

after propensity score subclassification. Subclassification, also known as stratification, is one

method used in propensity score modeling to eliminate bias associated with observed covariates

(Austin, 2011; Guo & Fraser, 2015; Hong & Raudenbush, 2006; Thoemmes & Kim, 2011).

Cochran (1968) suggests five equal strata can remove a majority of bias; however, within this

study, the interaction of strata and treatment was still significant indicating imbalance. Ten strata

were then assessed (Akers, 2010) and after trimming extreme scores within the last stratum,

balance was achieved (as determined by nonsignificant main effects from a two-way analysis of

variance predicting the linear propensity score with strata and campus residence as independent

variables as well as a nonsignificant interaction term between collegiate residence and strata).

Support was also assessed through visual inspection of regions of common support for

overlap and by assessing standardized mean differences before and after stratification. Figures 1

and 2 illustrate the propensity scores for students living on and off campus. Table 9 provides the

standardized mean differences of each variable before and after balancing. Prior to stratification,

18 covariates in the first propensity score model and 24 in the second model were significant

predictors of collegiate residence. After stratification, none of the variables remained significant

121
and nearly all standardized mean differences between the treatment and the control were less

than 0.05 with none were greater than .10.

Outcomes Analyses

Table 10 provides results for the effect of living on and off campus during the first year

on student engagement outcomes throughout college. Prior to balancing, living on campus was

significantly related to increases in non-classroom faculty interactions at the end of students’

fourth year, positive peer interactions at all time points, diversity experiences at the end of

students’ first year, frequency of interactions with student affairs staff at each time, and co-

curricular involvement at the end students’ first and fourth years. After the inclusion of the

propensity score strata variable, the socially related variables remained significant. Students

living on campus reported greater positive peer interactions at Time 2 and Time 3 (B = 0.54, SE

= 0.11, p = 0.001; B = 0.26, SE = 0.12, p = 0.05), frequency of interactions with student affairs

staff (Time 2 B = 0.34, SE = 0.14, p = 0.05; Time 3 B = 0.24, SE = 0.11, p = 0.05), and higher

rates of co-curricular involvement (Time 2 B = 0.32, SE = 0.08, p = 0.001; Time 3 B = 0.16, SE

= 0.08, p = 0.05). Additionally, in the presence of propensity score adjustment, students living

on campus engaged in more cooperative learning during their first year (B = 0.32, SE = 0.12, p =

0.05). These standardized coefficients represent medium to large effects (Mayhew et al., 2016),

suggesting that students who live on campus their first year receive notable benefits in socially

related outcomes.

This study also explored the conditional effect of campus residence. Table 11 provides

results for analyses that included interaction terms between collegiate residence and several

student-level variables including first generation status, ability (measured through standardized

test scores), sex, and race. To test the interactions, the variable for collegiate residence, the

moderator of interest, and the interaction term all were entered simultaneously into the equation
122
for each model (Jaccard & Turrisi, 2003). Despite small subgroup sample sizes within the racial

categories, the decision was made to retain the five racial categories (Black, Latinx, Asian

American / Pacific Islander, other races, and White) in an effort to not assume that all non-White

students experience campus residences as a monolithic entity. Within each subgroup analysis,

separate models examined each interaction to avoid issues with multicollinearity.

The results of the conditional effects are mixed. Table 11 highlights the significant

interactions indicating for some subgroups of students the effect of living on campus is larger

than their comparative group. Nonsignificant results were not reported and as such, the findings

suggest that there is no conditional effect based on student ability or first-generation status, since

the interactions of those characteristics with the campus residence variable yielded no significant

results for any of the student engagement outcomes at either time point. To better understand the

magnitude of the significant interactions, Table 12 provides the predicted means from the

regression analyses based on sex or race and campus residence. For the interaction between

students’ sex and campus residence, only the outcome of cooperative learning at time 3 (B = .26,

SE = .11, p = .05) was significant. The effect of living on campus during a student’s first year is

significantly more positive for males than females. There are the most interactions when viewing

the combination of students’ race and status of living on campus upon a host of student

engagement outcomes. Significant differences were found in the effect of campus residence for

Black versus White students predicting co-curricular involvement during at the end of the first

year (B = -0.37, SE = 0.09, p = 0.05), interacting with peers (B = -0.28, SE = 0.12, p = 0.05),

negative diversity interactions with peers including instances of hostility and negative

environments at the end of the first (B = .37, SE = 0.16, p = 0.05) and fourth year (B = 0.63, SE

= 0.20, p = 0.01). The predicted margins from Table 12 suggest that living on campus seems to

123
be positively related to co-curricular involvement among White students, whereas it appears to

have no effect for Black students. The same pattern holds these students when viewing negative

diversity interactions with peers, the interaction of living on campus and race leads to

significantly more amounts for Black students compared to their White peers. In this case, the

positive number reflects greater frequency of guarded and cautious interactions with peers as

well as feeling silenced by prejudice and discrimination, among other experiences.

In terms of other racial identities, there were fewer significant interactions. Specifically,

meaningful discussions with their diverse peers had a significant interaction between campus

residence and Asian American and Pacific Islander students (B = 0.42, SE = 0.16 p = 0.05) as

well as Latinx students (B = 0.34, SE = 0.11 p = 0.01). For both groups, the effect of living on

campus was more positive in terms of meaningful interactions with peers for these groups than

their White peers at the end of the first year. However, the interaction between Latinx and White

students with living on campus saw a significant result in terms of non-classroom student faculty

interactions (B = -0.51, SE = 0.11, p = 0.001). Table 12 indicates that the effect of living on

campus is significantly worse for Latinx students in terms of non-classroom interactions with

faculty during their first year. In other words, it appears that campus residence only matters for

Latinx students and it is opposite of what one might expect.

Finally, to answer the question pertaining to students who move off campus after their

first year, Table 13 illustrates those results. Prior to subclassification, the results suggest moving

off campus only has a direct effect on students’ diversity experiences; however, after balancing

on the propensity score, this result is no longer significant. That said, after accounting for the

bias associated with the observed covariates, students staying on campus reported lower levels of

positive peer interactions (B = -0.18, SE = 0.08, p = 0.05) as well as co-curricular involvement

124
(B = -0.23, SE = 0.11, p = 0.05). When interpreting the coefficients, it appears that staying on

campus seems to have a small effect in regard to these student engagement outcomes (Mayhew

et al., 2016).

Discussion

Students living on campus are more proximal to student affairs staff and peer networks.

This findings from this study suggest that the proximity of living on campus has a direct effect

on some of these student engagement outcomes. Namely, as a result of living on campus their

first year, students develop greater positive peer relationships that are more rewarding,

personally satisfying, and had a positive influence on their lives. These benefits persist through

the end of their fourth year, with students who reported living on campus their first year still

reporting higher amounts of positive peer interactions at the end of their fourth year.

Additionally, living on campus during this time connects students to student affairs staff and

increased co-curricular involvement, but it has no impact on their connection to faculty or the

frequency in which they interact with faculty.

These findings somewhat support earlier theories that suggest living on campus provides

greater opportunities and access to social networks and peer groups (Astin, 1985; Chickering,

1974; Terenzini & Pascarella, 1984). While living on campus has a direct effect on peer group

relationships and the frequency in which students interact with student affairs staff, it seems the

proximity has no relations to faculty. This nonsignificant finding might be a function of the

course enrollment patterns of first year students, with a greater likelihood of larger lecture style

classes and introductory courses not related to students’ majors. Thus, for students in their first

year, there is no additional impetus to connect with faculty that first year.

125
 The juxtaposition of nonsignificant findings related to students’ engaging with diversity

events as well as significant findings pertaining to meaningful discussions with diverse peers is

surprising given institutions’ focus on promoting connections across difference and providing

opportunities on campus for students through events such as lectures. Many, however, have

argued that this emphasis is more about rhetoric as opposed to actually affecting reality. The

juxtaposition of these nonsignificant findings with significant peer interactions scales is

noteworthy. Taken together, these findings might imply that increases in co-curricular

involvement might only occur within a homogenous peer (in)group and that students are not

interacting across difference when comparing collegiate residences. In short, the main effects

find that living on campus increases some connections to peers, co-curricular involvement, and

to student affairs staff, but not student engagement outcomes associated with diversity.

However, as contemporary research illustrates, the effect of living on campus during a

student’s first year is not entirely equitable. The results illustrate that when compared to with

their White peers, Black students’ who lived on campus had social connections that seemed to be

less rewarding. These findings are not surprising given potentially overtly and implicitly racist

campus environments for students of color compared to their White peers (Harper, 2012;

Hotchkins & Dancy, 2017; Strayhorn & Mullins, 2012). For Black students, these toxic

environments might be the reason co-curricular involvement is significantly less than their White

peers during the first year and why these students report higher rates of diversity experiences

which include attending events geared towards social issues, or serious conversations with staff,

or awareness of institutional policies promoting increased contact among all students. For Asian

American and Pacific Islander and Latinx students, the effect of living on campus was more

positive than the interaction for White students. These students report higher rates of meaningful

126
interactions and conversations with diverse peers. One explanation might be that campus

environments are more hostile for Black students than for other minoritized students. Further

research with large samples of students or further in-depth qualitative research is needed to better

understand the effect of residential campus environments and students’ race as they affect

student engagement outcomes.

Aside from the socially related outcomes, living on campus during the first year directly

affected students’ levels of co-curricular involvement that persists throughout students’ four

years. After propensity score adjustment, living on campus has a small to medium, positive

effect on co-curricular involvement. Again, this finding illustrates the importance of students’

proximity to student clubs and organizations. Living on campus provides students the ability to

participate in club and student organization meetings that might occur during nights and

weekends that could be more problematic for students living off campus.

Moving off campus after students’ first year had a small effect on a couple of student

engagement outcomes. Prior to the propensity score adjustment, diversity experiences were the

only significant outcome, with those students moving off campus reporting fewer experiences.

However, after the propensity score adjustment, diversity experiences were no longer significant

but positive peer interactions and co-curricular involvement became significant. Students who

moved off campus after their first year had more positive peer interactions and co-curricular

involvement. In regard to positive peer interactions, the questions making up the scale pertain to

connections made with peers at the institution, and it might be these students moved off campus

with their close peers and best friends. Students moving into apartments off campus are more

likely to live with their close friends, which might be the reason moving off campus has a

positive effect on these interactions. Additionally, co-curricular involvement was higher for

127
students who moved off campus than for students who lived on campus in their fourth year. It

might be these off-campus students use involvement as a means to stay connected to the

institution.

Recommendations for Research and Practice

While the findings from this study advance the understanding of the effect of living on

campus, future researchers should continue to work in understanding the role of residence halls

in students’ lives. First, scholars need to better understand the interaction of race and on campus

living as it relates to social connections. Does living on campus only promote homogenous peer

group interactions, and if so, how do these interactions compare to students off campus? It might

be beneficial for only Latinx, Asian American, or Pacific Islander students, while not beneficial

for others. It could be, too, that these results might also be driven in large part by the

overwhelming White student population within the sample. Additionally, are students really

interacting with diversity through intentional programming and conversation with staff and

peers? Again, the results of this study suggest they might not be, which is disappointing given

institutional sample’s commitment to liberal arts education. Higher education administrators

might better assess their programming and connections with students. Does requiring resident

assistant staff to engage in diversity programming or campus-wide programming such as a one

book, one community or diversity speaker really affect students or the campus environment?

Does living on campus really allow students to tap into and engage with these programmatic

entities, or do contemporary students just want to stay connected to like-minded individuals?

Administrators and researchers alike should work to better understand the phenomena within

higher education and specifically on their campus, respectively. As such there should be further

inquiry with a larger and more diverse student population to better explore campus residence and

128
peer groups. Researchers should continue to represent wide varieties of institutions in their

studies and should continue to explore how on campus experiences differ by race, especially at

predominantly White institutions.

Second, while the WNS dataset is robust and offers researchers the ability to explore

college impact longitudinally, a more recent set of data would add more recent understandings of

how living on campus might affect students. Moving on campus ten years ago was a different

transition than it is today. Ten years ago, social media was not nearly as prevalent. Students did

not have Snapchat or Twitter to stay connected and engaged with their precollege friends or with

the collegiate community. Using data from contemporary students might offer a different pattern

of connections on campus. Students may not leave their precollege communities behind in an

attempt to integrate within campus (Tinto, 1993); rather they may actively engage those deep

connections with peers outside the institution. Researchers can work to better understand how

students use technology and with whom they stay connected throughout their collegiate

experiences. It might be students rely on those other friends rather than establishing new

connections within the institution. For higher education administrators, engaging students past

their screens might be difficult, especially if it is challenging content such as that related to

diversity. Hostile campus cultures might force students to find belonging and connection

elsewhere. For Black students who live on campus at predominantly White institutions,

technology might be a way to mitigate hostile campus climates and cope with (daily)

microaggressions.

Third, this study employed a quasi-experimental methodology to better estimate the

causal effect of living on campus. After adjustment, some of the outcomes remained significant,

while others became nonsignificant. Quasi-experimental methodology allows higher education

129
researchers to better estimate the effect of particular experiences and outcomes, provided there is

a robust set of observed covariates. When designing future studies, scholars should take time to

reflect on what data should be collected to better understand the phenomena of study. Finally,

future research into living on campus should explore what occurs when students live in residence

halls. That is, are there common programmatic experiences or staffing structures that best

promote these social connections?

In short, this study provides an important contribution by supporting strong causal claims

and more generalizable evidence to the literature on living on campus and college student

engagement. By using propensity score weighting to reduce the self-selection bias associated

with living on campus, this study attempted to better causally estimate the effect of living on

campus. After achieving balance through propensity score weighting, it was shown that living on

campus has a direct effect on positive peer interactions, cooperative learning in the first year,

frequency of interactions with student affairs staff, and co-curricular involvement.

130
Table 7. Unadjusted Means and Standard Errors for Variables of Interest Based on Students’
First-year Residential Choice

Variablea Observations On Campus Off Campus

Outcome Variables
Non-classroom faculty interactions in first year 7,937 0.01 (0.01) -0.05 (0.03)
Non-classroom faculty interactions in fourth year 4,099 0.03 (0.01) -0.11 (0.06)
Frequency of faculty interactions, end of first year 8,113 0.01 (0.01) -0.01 (0.03)
Frequency of faculty interactions, end of fourth year 4,078 -0.01 (0.01) -0.05 (0.05)
Positive peer interactions, end of first year 7,914 0.04 (0.01) -0.33 (0.03)
Positive peer interactions, end of fourth year 4,094 0.06 (0.01) -0.33 (0.05)
Cooperative learning, end of first year 7,901 0.01 (0.01) -0.04 (0.03)
Cooperative learning, end of fourth year 4,090 0.02 (0.01) 0.05 (0.05)
Diversity experiences, end of first year 7,965 0.02 (0.01) -0.11 (0.02)
Diversity experiences, end of fourth year 4,071 -0.01 (0.01) -0.10 (0.04)
Meaningful discussions with diverse peers, first year 7,879 0.01 (0.01) -0.02 (0.03)
Meaningful discussions with diverse peers, fourth year 4,086 -0.03 (0.01) -0.01 (0.06)
Negative diversity interactions with peers, first year 7,883 -0.00 (0.01) 0.03 (0.04)
Negative diversity interactions with peers, fourth year 4,088 -0.06 (0.01) 0.01 (0.06)
Frequency of interactions with student affairs staff, 7,890 0.01 (0.01) -0.11 (0.03)
first year
Frequency of interactions with student affairs staff, 4,089 0.01 (0.01) -0.05 (0.03)
fourth year
Co-curricular involvement, end of first year 8,088 0.05 (0.01) -0.52 (0.03)
Co-curricular involvement, end of fourth year 4,069 0.01 (0.02) -0.53 (0.05)

Demographic Characteristics
Sex (1 = male) 8,045 0.38 (0.01) 0.32 (0.02)
Black/African-American (1=yes) 7,732 0.09 (0.00) 0.11 (0.01)
Asian/Pacific Islander (1=yes) 7,732 0.05 (0.00) 0.11 (0.01)
Latinx (1=yes) 7,732 0.04 (0.00) 0.10 (0.01)
White (1=yes) 7,732 0.78 (0.00) 0.64 (0.02)
Other race/ethnicity (1=yes) 7,732 0.05 (0.00) 0.04 (0.01)
First generation (1 = no) 7,516 0.28 (0.01) 0.50 (0.02)
Standardized test score (1 = above median) 7,473 0.58 (0.01) 0.28 (0.02)
Self-reported high school GPA 7,857 4.57 (0.01) 4.33 (0.03)
Self-reported parental income 7,198 5.07 (0.03) 4.39 (0.11)
English is native language (1 = yes) 7,285 0.93 (0.00) 0.83 (0.02)
Self-reported disability (1 = yes) 7,897 0.11 (0.00) 0.11 (0.01)

Time spent in high school activities

Exercising 7,818 3.84 (0.01) 3.47 (0.05)
High school involvement 7,805 3.84 (0.01) 3.31 (0.06)
Socializing 7,821 4.41 (0.01) 4.33 (0.04)
Studying with friends 7,809 2.80 (0.01) 2.67 (0.04)
Interacting with teachers 7,800 3.39 (0.01) 3.19 (0.05)
Volunteering 7,806 3.14 (0.01) 2.94 (0.05)
Working for pay 7,807 3.19 (0.02) 3.43 (0.06)
Playing on computer 7,807 4.35 (0.01) 4.22 (0.05)
Using computer for homework 7,807 4.28 (0.01) 4.17 (0.04)

131
 Table 7—continued

Using the library 7,796 2.79 (0.01) 2.89 (0.05)

Reading for fun 7,802 3.16 (0.01) 3.07 (0.05)
Smoking 7,974 1.06 (0.00) 1.09 (0.01)
Binge drinking 7,960 0.56 (0.01) 0.60 (0.05)
Drinking alcohol 7,963 0.48 (0.01) 0.56 (0.05)
Psychosocial and psychological measures
Goal aspirations 7,775 4.43 (0.01) 4.30 (0.05)
Psychological well-being 7,842 4.52 (0.01) 4.49 (0.03)
Need for cognition 7,959 3.47 (0.01) 3.40 (0.03)
Overall health 7,977 4.27 (0.01) 4.17 (0.03)
Diversity experiences 7,988 3.97 (0.01) 3.92 (0.03)
Academic motivation 7,974 3.61 (0.01) 3.65 (0.02)
Institutional Information
High school racial composition 7,895 2.06 (0.01) 2.39 (0.05)
Institutional choice 7,857 3.46 (0.01) 3.32 (0.04)

a
Note full variable descriptions and values can be found in the Appendix.

132
Table 8. Unadjusted Means and Standard Errors for Variables of Interest Based on Students’
Moving Off Campus After Their First Year

Variablea Observations On Campus Off Campus

Fourth-Year Outcome Variables

Non-classroom faculty interactions 3,730 0.12 (0.02) -0.05 (0.02)
Frequency of faculty interactions 3,743 0.05 (0.02) -0.07 (0.02)
Positive peer interactions 3,724 0.06 (0.02) 0.06 (0.02)
Cooperative learning 3,720 -0.02 (0.02) 0.06 (0.02)
Diversity experiences 3,737 0.08 (0.01) -0.10 (0.01)
Meaningful discussions with diverse peers 3,717 0.07 (0.02) -0.11 (0.02)
Negative diversity interactions with peers 3,718 -0.04 (0.02) -0.09 (0.02)
Frequency of interactions with student affairs staff 3,720 0.07 (0.02) -0.07 (0.02)
Co-curricular involvement 3,733 0.04 (0.02) -0.06 (0.02)

Demographic Characteristics
Sex (1 = male) 3,744 0.35 (0.01) 0.38 (0.01)
Black/African-American (1=yes) 3,636 0.06 (0.01) 0.05 (0.01)
Asian/Pacific Islander (1=yes) 3,636 0.05 (0.00) 0.04 (0.01)
Latinx (1=yes) 3,636 0.05 (0.01) 0.04 (0.00)
White (1=yes) 3,636 0.78 (0.01) 0.82 (0.01)
Other race/ethnicity (1=yes) 3,636 0.06 (0.01) 0.04 (0.00)
First generation (1 = no) 3,531 0.24 (0.01) 0.23 (0.01)
High school test scores (1 = above median) 3,478 0.73 (0.01) 0.65 (0.01)
Self-reported high school GPA 3,645 4.71 (0.01) 4.68 (0.01)
Self-reported parental income 3,388 5.05 (0.06) 5.37 (0.06)
English is native language (1 = yes) 3,649 0.92 (0.01) 0.94 (0.01)
Self-reported disability (1 = yes) 3,671 0.12 (0.01) 0.10 (0.01)

Time spent in high school activities

Exercising 3,632 3.78 (0.03) 3.99 (0.03)
High school involvement 3,624 4.00 (0.03) 3.87 (0.03)
Socializing 3,633 4.31 (0.02) 4.47 (0.02)
Studying with friends 3,627 2.81 (0.02) 2.86 (0.02)
Interacting with teachers 3,619 3.47 (0.02) 3.39 (0.02)
Volunteering 3,624 3.28 (0.03) 3.22 (0.02)
Working for pay 3,625 2.90 (0.03) 3.23 (0.03)
Playing on computer 3,627 4.36 (0.02) 4.35 (0.02)
Using computer for homework 3,625 4.32 (0.02) 4.27 (0.02)
Using the library 3,623 2.86 (0.03) 2.76 (0.02)
Reading for fun 3,625 3.30 (0.03) 3.11 (0.03)
Smoking 3,718 1.04 (0.01) 1.04 (0.00)
Binge drinking 3,713 0.32 (0.02) 0.62 (0.03)
Drinking alcohol 3,610 0.29 (0.02) 0.53 (0.02)
Psychosocial and psychological measures
Goal aspirations 3,610 4.56 (0.03) 4.48 (0.03)
Psychological well-being 3,667 4.55 (0.01) 4.57 (0.01)
Need for cognition 3,714 3.59 (0.01) 3.45 (0.01)
Overall health 3,722 4.30 (0.01) 4.33 (0.01)

133
 Table 8—continued

Diversity experiences 3,724 4.07 (0.01) 3.96 (0.02)

Academic motivation 3,717 3.69 (0.01) 3.57 (0.01)
Institutional Information
High school racial composition 3,654 2.05 (0.03) 1.98 (0.02)
Institutional choice 3,644 3.45 (0.02) 3.51 (0.02)
First-year Collegiate Experiences
Grades 3,625 6.33 (0.03) 6.34 (0.03)
Entire experience again 3,730 3.53 (0.01) 3.44 (0.01)
Choose same college 3,731 3.46 (0.02) 3.40 (0.02)
Non-classroom faculty interactions 3,689 0.17 (0.02) 0.01 (0.02)
Frequency of faculty interactions 3,683 0.22 (0.02) 0.02 (0.02)
Positive peer interactions 3,679 0.16 (0.02) 0.14 (0.01)
Cooperative learning 3,732 0.15 (0.02) 0.17 (0.02)
Campus diversity experiences and events 3,682 0.09 (0.01) 0.00 (0.01)
Meaningful discussions with diverse peers 3,701 0.12 (0.01) -0.03 (0.01)
Negative diversity interactions with peers 3,671 0.09 (0.02) -0.02 (0.02)
Frequency of interactions with student affairs 3,671 0.02 (0.02) 0.01 (0.02)
staff
a
Note full variable descriptions and values can be found in the Appendix.

134
Table 9. Significance of and Standardized Mean Differences for Each Propensity Score Model,
Before and After Stratification

Standardized Mean Differences

First Propensity Second Propensity
Score Model Score Model
Before After Before After
Predictor Variablea
Balancing Balancing Balancing Balancing
Demographic Characteristics
Sex (1 = male) -0.06 -0.02 -0.08 0.01
Black/African-American -0.09 0.04 -0.07 -0.03
Asian/Pacific Islander -0.24*** -0.09 0.06 0.05
Latinx -0.20*** -0.01 0.01 0.04
White 0.35*** 0.06 0.04* -0.02
Other race/ethnicity -0.07 -0.05 -0.08** -0.03
First generation (1 = no) -0.28*** 0.01 -0.03 -0.02
High School Test Scores (1 = above median) 0.40*** -0.03 -0.08** -0.01
Self-reported High School GPA 0.20 0.07 0.22 0.08
Self-reported parental income 0.26*** 0.04 0.13* 0.08
English is native language (1 = yes) 0.38*** 0.08 -0.01 -0.01
Self-reported disability (1 = yes) 0.09 0.03 0.03* 0.05
Time spent in high school activities
Exercising 0.21*** 0.04 0.15*** 0.03
High school involvement 0.34*** 0.05 -0.10** 0.01
Socializing 0.19*** 0.1 0.15*** 0.02
Studying with friends 0.06 -0.02 0.05*** 0.01
Interacting with teachers 0.03 -0.03 -0.06* 0.01
Volunteering -0.05 -0.05 -0.01 0.01
Working for pay -0.05 0.05 0.14*** -0.05
Playing on computer 0.03 0.00 0.01 0.04
Using computer for homework 0.08* -0.01 -0.09* 0.01
Using the library -0.03 0.02 -0.06 -0.02
Reading for fun 0.16** 0.06 -0.11*** 0.02
Smoking -0.01 0.00 0.01 0.01
Binge drinking 0.00 -0.01 0.16*** -0.04
Drinking alcohol -0.04 0.03 0.18*** -0.06
Psychosocial and psychological measures
Goal aspirations -0.02* -0.03 -0.02 -0.02
Psychological well-being 0.09* 0.02 0.00 -0.02
Need for cognition 0.13*** -0.03 -0.20*** 0.01
Overall health 0.09 0.01 0.04 0.01

135
 Table 9—continued

Diversity experiences 0.11* -0.01 -0.16*** -0.03

Academic motivation -0.06 0.01 -0.15*** 0.01
Additional Information
High School Racial Composition -0.31*** 0.01 -0.04 0.02
Institutional Choice 0.15** 0.03 0.04 -0.02
Collegiate Experiencesb
End of first-year grades -0.02 -0.02
Entire experience again, end of first year -0.12** -0.03
Choose same college, end of first year -0.05 -0.02
Interactions with faculty -0.21*** -0.01
Peer Interactions -0.04 -0.01
Cocurricular involvement -0.03 -0.01
Diversity experiences on campus -0.19*** 0.04
Positive diversity experiences -0.14*** 0.02
Negative diversity experiences 0.01 0.01
Interactions with student affairs staff -0.03 0.00

a
Significant predictors of the treatment variable, living on campus, are noted as follows *p < .05
**p < .01 ***p < .001

b
Institutional fixed effects were incorporated into the propensity score model by including
dummy codes for each institution while leaving one institution out as the referent group.

136
Table 10. Results of Regression Analyses of First-year Collegiate Residence Predicting Student
Engagement Outcomes

____________________________________________________________________________________

No PSM adjustment PSM adjustment

Outcome variable B SE B SE
Non-classroom faculty interactions in first year 0.07 0.05 -0.07 0.08
Non-classroom faculty interactions in fourth year 0.17* 0.07 0.01 0.13
Frequency of faculty interactions, end of first year -0.03 0.06 -0.01 0.12
Frequency of faculty interactions, end of fourth year 0.09 0.08 -0.04 0.10
Positive peer interactions, end of first year 0.50*** 0.07 0.54*** 0.11
Positive peer interactions, end of fourth year 0.44*** 0.08 0.26* 0.12
Cooperative learning, end of first year 0.09 0.08 0.32* 0.12
Cooperative learning, end of fourth year -0.02 0.08 0.01 0.14
Diversity experiences, end of first year 0.16* 0.06 0.12 0.12
Diversity experiences, end of fourth year 0.10 0.06 0.09 0.06
Meaningful discussions with diverse peers, first year 0.03 0.05 0.12 0.14
Meaningful discussions with diverse peers, fourth year 0.01 0.08 0.07 0.10
Negative diversity interactions with peers, first year 0.01 0.12 -0.01 0.21
Negative diversity interactions with peers, fourth year -0.01 0.09 -0.03 0.13
Frequency of interactions with student affairs staff, 0.13* 0.06 0.34* 0.14
first year
Frequency of interactions with student affairs staff, 0.18* 0.07 0.24* 0.11
fourth year
Co-curricular involvement, end of first year 0.53*** 0.10 0.32*** 0.08
Co-curricular involvement, end of first year 0.44*** 0.07 0.16* 0.08

Note. Institutional fixed effects were included in all analyses. Student engagement outcomes were
standardized with a mean of zero and a standard deviation of one. Each outcome was analyzed using
ordinal least squares multiple regression analyses with robust standard errors. *p < .05 **p < .01 ***p <
.001
_____________________________________________________________________________________

137
Table 11. Results of Conditional Effects Analyses Based on Student Demographic
Characteristics and Collegiate Residence

______________________________________________________________________________

Outcome variable B SE
Cooperative learning
Male, Time 3 0.26* 0.11
Non-classroom Student Faculty Contact
Latinx, Time 2 -0.51*** 0.13
Co-curricular Involvement
Black, Time 2 -0.37* 0.20
Interacting with Peers
Black, Time 2 -0.28* 0.12
Meaningful Peer Interactions with Diverse Others
Asian American and Pacific Islander, Time 2 0.42* 0.16
Latinx, Time 2 0.34** 0.11
Negative Diversity Interactions with Peers
Black, Time 2 0.37* 0.16
Black, Time 3 0.63** 0.20

Note. The results contain interactions between the outcome of interest along with the student
demographic characteristic. Institutional fixed effects were presented in all analyses along with
the moderator, campus residence variable, and the interaction term. Student engagement
outcomes were examined with ordinal least squares multiple regression analyses and were
standardized with a mean of zero and a standard deviation of one. Given the substantial number
of nonsignificant findings, only interactions with significant findings are presented in the table
shown here.
*p < .05 **p < .01 ***p < .001
______________________________________________________________________________

138
Table 12. Predicted Means and Standard Deviations of Standardized Scaled Student Engagement
Measures Based on Student Demographic Characteristics and Place of Residence

Living Off Campus Living On Campus

Cooperative learning
Sex, Time 3
Female 0.12 (0.05) 0.33 (0.12)
Male 0.07 (0.05) 0.21 (0.13)
Non-classroom Student Faculty Contact
Race, Time 2
Latinx 0.64 (0.12) 0.12 (0.09)
White 0.00 (0.11) -0.01 (0.05)
Co-curricular Involvement
Race, Time 2
Black -0.32 (0.10) -0.33 (0.03)
White -0.27 (0.11) 0.13 (0.05)
Interacting with Peers
Race, Time 2
Black 0.06 (0.09) 0.07 (0.20)
White -0.21 (0.07) -0.10 (0.04)
Meaningful Peer Interactions with Diverse Others
Race, Time 2
Asian American and Pacific Islander 0.001 (0.17) 0.31 (0.07)
White -0.04 (0.08) -0.11 (0.04)
Race, Time 2
Latinx -0.17 (0.17) 0.36 (0.07)
White -0.04 (0.08) -0.11 (0.04)
Negative Diversity Interactions with Peers
Race, Time 2
Black 0.06 (0.13) 0.26 (0.06)
White 0.14 (0.11) -0.02 (0.04)
Race, Time 3
Black 0.22 (0.24) 0.81 (0.36)
White 0.09 (0.02) 0.13 (0.15)

139
Table 13. Results of Regression Analyses of Collegiate Residence Predicting Student
Engagement Outcomes for Students Moving Off Campus After Their First Year

_____________________________________________________________________________________

No PSM adjustment PSM adjustment

Outcome variable B SE B SE
Non-classroom faculty interactions 0.03 0.07 0.08 0.07
Frequency of faculty interactions 0.04 0.06 0.11 0.09
Positive peer interactions 0.06 0.08 -0.18* 0.08
Cooperative learning -0.02 0.07 0.01 0.06
Diversity experiences 0.16*** 0.03 0.05 0.04
Meaningful discussions with diverse peers -0.01 0.05 -0.05 0.10
Negative diversity interactions with peers -0.03 0.06 0.06 0.12
Frequency of interactions with student affairs staff 0.10 0.07 -0.14 0.10
Co-curricular involvement 0.05 0.08 -0.23* 0.11

Note. Students who reported living on campus at both Time 2 and 3 were coded 1, while students who
reported moving off campus after Time 2 were coded as 0. Institutional fixed effects were included in all
analyses. Student engagement outcomes were standardized with a mean of zero and a standard deviation
of one. Each outcome was analyzed using ordinal least squares multiple regression analyses with robust
standard errors. *p < .05 **p < .01 ***p < .001
_____________________________________________________________________________________

140
Figure 3. Propensity Score Distributions for Students Living On Campus (treated) and Off
Campus (untreated) During Their First Year
Proportion of Sample

.2 .4 .6 .8 1
Propensity Score

Untreated Treated: On support

Treated: Off support

141
Figure 4. Propensity Score Distributions for Students Living On Campus (treated) and Off
Campus (untreated) During Their Collegiate Experience

.2 .4 .6 .8 1
Propensity Score

Untreated Treated

142
 Appendix: Study Variables

Demographic Characteristics
Sex (1 = male) Student’s institution provided data from their school file for sex and race / ethnicity
Black/African-American (1=yes)
Asian/Pacific Islander (1=yes)
Latinx (1=yes) Dummy variables for race / ethnicity entered separately into the model
White (1=yes)
Other race/ethnicity (1=yes)
Recoding of variable asking what is the highest level of education each of your parents
First generation (1 = no) or guardians completed? First generation was coded as students who selected did not
finish high school or high school graduate / GED.
Variable converted SAT scores using the COMPASS conversion, so all scores were on a
High school test scores (1 = above median)
common metric

Which of the following best describes your overall grade range in high school? 1 = A- to
Self-reported high school GPA
A+ to 5 = Below D-

What is the best estimate of your parents’ totally annual income and your annual
Self-reported parental income
income? 1 = less than $14,999 to 9 = $300,000 or more

English is native language (1 = yes) Is English your native language?

Self-reported disability (1 = yes) Aggregate of question, mark all of the following diagnosed disabilities that apply to you

Time spent in high school activities

Exercising
High school involvement
Question asked, during the last year in high school, how often did you engage in each of
Socializing
the following activities? Scale was 1 = very often to 5 = never
Studying with friends
Interacting with teachers

143
 Volunteering
Working for pay
Playing on computer
Using computer for homework
Using the library
Reading for fun
How many cigarettes do you smoke a day? 1 = I don’t smoke cigarettes to 5 = 2 or more
Smoking
packs a day

In a typical week of your last year of high school, how often did you consume 5 or more
Binge drinking
drinks in one sitting? 1 = 0 times to 5 = 5 or more times

In a typical week of your last year of high school, how often did you consume alcoholic
Drinking alcohol
beverages? 1 = 0 times per week to 9 = more than 7 times per week
Psychosocial and psychological measures
What is the highest academic degree you intend to earn in your lifetime? 1 = vocational /
Goal aspirations
technical certificate or diploma to 6 = Doctorate degree

Overall Health Overall, how would you rate your health? 1 = excellent to 5 = very poor

Scaled measure developed by Carol Ryff (1989). Items included were from all six
subscales including autonomy, environmental mastery, personal growth, positive
Psychological well-being, 54 items, α = .89
relations with others, purpose in life, and self-acceptance. Scales ranged from 1 =
strongly disagree to 5 = strongly agree

Degree to which one enjoys engaging in effortful cognitive activities. Sum of eighteen
items on the Need for Cognition short form (Cacioppo, Petty, & Kao, 1984). Scaled from
Need for cognition, 18 items, α = .89
1 = Extremely characteristic to 5 = Extremely uncharacteristic

Items include I enjoy having discussions with people whose ideas and values are
different from my own; The real value of a college education lies in being introduced to
Precollege Diversity experiences, 4 items, α = .79 different values; Contact with individuals whose backgrounds (e.g. race, national origin,
sexual orientation) are different from my own is an essential part of my college
education; and I enjoy talking with people who have values different from mine because

144
 it helps me better understand my values. Scale is 1 = Strongly Agree to 5 = Strongly
Disagree.

Scale (1 = Strongly Agree to 5 Strongly Disagree) includes items such as: I am willing to
work hard in a course to learn the material even if it won’t lead to a higher grade; When I
do well on a test, it is usually because I am well-prepared not because the test is easy; In
Academic motivation, 8 items, α = .70 high school, I frequently did more reading in a class than was required simply because it
interested me; In high school, I frequently talked to my teachers outside of class about
ideas presented during class; Getting the best grades I can is very important to me; I
enjoy the challenge of learning complicated new material

Additional Information
How would you describe the racial composition of the last high school you attended? 1 =
High school racial composition
almost all white students to 5 = almost all students of color

Institutional choice Was this college your… 1 = first choice to 3 = third choice

Collegiate Experiences
What have most of your grades been up to now at this institution? 1 = C- or lower to 8 =
End of first-year grades
A
How would you evaluate your entire educational experience at this institution? 1 = poor
Entire experience again, end of first year
to 4 = excellent
If you could start over again, would you go to the same institution you are now
Choose same college, end of first year
attending? 1 = definitely no to 4 = definitely yes

About how many hours in a typical week do you spend doing the following: Participating
Co-curricular involvement in co-curricular activities (organizations, campus publications, student government,
fraternity or sorority, intercollegiate or intramural sports, etc.)

Questions related to the quality of non-classroom interactions with faculty including the
extent students agreed that non-classroom interactions had a positive influence on
Interactions with faculty, outside of class, 5 items, α = .86
personal growth, values, and attitudes. Response options were 1 = strongly agree to 5 =
strongly disagree.

Scale representing how often students discussed grades, assignments, career plans, ideas
Frequency of interactions with faculty, 4 items, α = .73 from readings outside of the classroom, or worked on activities other than coursework
with faculty. Response options were 1 = very often to 5 = never.

145
 Scale representing students’ relationships with other students, personally satisfying
relationships, the degree other students have had a positive influence on intellectual
growth and interest in ideas, quality of relationships with other students, ability to meet
Peer interactions, 8 items, α = .88
and make friends with other students, perceptions of other students willing to listen and
help with a personal problem, and the degree to which other students’ values align with
the respondent. Response options were 1 = strongly agree to 5 = strongly disagree.

Scale representing how often students attended debates or lectures on a current political
or social issue during the academic year, had serious discussions with staff whose
political, social, or religious opinions were different from own, degree to which the
institution emphasizes contact among students from different economic, social, and racial
Campus diversity events and experiences, 6 items, α = .70 or ethnic backgrounds, how often the student has had serious conversations with students
who are very different from them in terms of religious beliefs, political opinions, or
personal values, and how often student participated in a racial or cultural awareness
workshop during the current academic year. Response options were 1 = very often to 5 =
never

Scale representing how often student had discussions regarding inter-group relations with
diverse students, had meaningful and honest discussions about issues related to social
Positive diversity experiences, 3 items, α = .82 justice with diverse students, and how often they shared personal feelings and problems
with diverse students while attending this college. Response options were 1 = very often
to 5 = never

Scale representing how often students had guarded or cautious interactions with diverse
students, how often they felt silenced by prejudice and discrimination from sharing
personal experiences with diverse others, how often they had hurtful or unresolved
Negative diversity experiences, 5 items, α = .82 interactions with diverse students, had somewhat hostile interactions with diverse
students, and how often they felt insulted or threatened based on race, national origin,
values or religion with diverse students while attending this college. Response options
were 1 = very often to 5 = never

Scale representing how often students discussed personal problems or concerns, worked
on out-of-class activities (e.g. committees, orientation, student life activities), talked
Interactions with student affairs staff, 5 items, α = .85 about career plans, discussed ideas from readings or classes, or discussed grades or
assignments with student affairs professionals. Response options were 1 = very often to 5
= never

146
 References

Akers, A. (2010). Determination of the optimal number of strata for bias reduction in propensity

score matching (Doctoral dissertation). Retrieved from ProQuest LLC. (ED518529)

Allison, P. D. (2009). Fixed effects regression models. Thousand Oaks, CA: Sage.

Antonio, A. L. (2001). The role of interracial interaction in the development of leadership skills

and cultural knowledge and understanding. Research in

Higher Education, 42, 593–617.

American College Personnel Association. (1996). The student learning imperative: Implications

for student affairs. Retrieved from http://www.myacpa.org/files/acpas-student-learning-

imperativepdf

Arpino, B., & Mealli, F. (2011). The specification of the propensity score in multilevel

observational studies. Computational Statistics & Data Analysis, 55(4), 1770-1780.

Astin, A. (1985). Achieving education excellence: A critical assessment of priorities and

practices in higher education. San Francisco, CA: Jossey-Bass.

Astin, A.W. (1991). Assessment for excellence: The philosophy and practice of assessment and

evaluation in higher education. Washington, DC: American Council on Education/Oryx

Press Series on Higher Education.

Astin, A. W. (1993). What matters in college? Four critical years revisited. San Francisco, CA:

Jossey-Bass Publishers.

Austin, P. C. (2011). An introduction to propensity score methods for reducing the effects of

confounding in observational studies. Multivariate Behavioral Research, 46, 399–424.

Berger, J. B., & Milem, J. F. (1999). The role of student involvement and perceptions of

integration in a causal model of student persistence. Research in Higher

Education, 40(6), 641-664.

147
Biemer, P. P., & Christ, S. L. (2008). Weighting survey data. International handbook of survey

methodology, 317-341.

Biemiller, L. (2017, August 4). What’s new in freshman housing? Buildings that help students

make friends. The Chronicle of Higher Education. Retrieved from

http://www.chronicle.com/article/What-s-New-in-Freshman/240864

Blimling, G. S. (1993). The influence of college residence halls on students. Higher education:

Handbook of theory and research, 9, 248-307.

Blimling, G. S., & Whitt, E. J. (1999). Good practice in student affairs: Principles to foster

student learning. San-Francisco, CA: Jossey-Bass.

Brandon, A., Hirt, J. B., & Cameron, T. (2008). Where you live influences who you know:

Differences in student interaction based on residence hall design. Journal of College &

University Student Housing, 35(2), 62-79.

Braxton, J. M., Hirschy, A. S., & McClendon, S. A. (2004). Understanding and reducing college

student departure. ASHE-ERIC Higher Education Report, Vol. 30, No. 3. Washington,

DC: School of Education and Human Development, The George Washington University.

Bronfenbrenner, U. (1979). The ecology of human development: Experiments by nature and

design. Cambridge, MA: Harvard University Press.

Bronfenbrenner, U. (1989). Ecological systems theory. In R. Vasta (Ed.), Six theories of

development (pp. 187-249). Greenwich, CT: JAI Press.

Bronfenbrenner, U. (1993). The ecology of cognitive development: Research models and

fugitive findings. In R. H. Wozniak & K. W. Fischer (Eds.), Development in context:

Acting and thinking in specific environments (pp, 3-44). Hillsdale, NJ: Erlbaum.

148
Bronkema, R., & Bowman, N. A. (2017). A residential paradox? Residence hall attributes and

college student outcomes. Journal of College Student Development, 58, 624-630.

Brookhart, M. A., Schneeweiss, S., Rothman, K. J., Glynn, R. J., Avorn, J., & Stürmer, T.

(2006). Variable selection for propensity score models. American Journal of

Epidemiology, 163(12), 1149-1156.

Cheng, D. X. (2004). Students' sense of campus community: What it means, and what to do

about it. NASPA Journal, 41(2), 216-234.

Chickering, A. W. (1974). Commuting versus resident students. San Francisco, CA: Jossey-

Bass.

Chickering, A. W., & Gamson, Z. F. (1987). Seven principles for good practice in undergraduate

education. AAHE Bulletin, 3(7).

Cochran, W. G. (1968). The effectiveness of adjustment by subclassification in removing bias in

observational studies. Biometrics, 24, 205-213.

Cook, T. D., & Steiner, P. M. (2010). Case matching and the reduction of selection bias in quasi-

experiments: the relative importance of pretest measures of outcome, of unreliable

measurement, and of mode of data analysis. Psychological Methods, 15(1), 56-68.

Craney, T. A., & Surles, J. G. (2002). Model-dependent variance inflation factor cutoff

values. Quality Engineering, 14(3), 391-403.

Credé, M., & Niehorster, S. (2012). Adjustment to college as measured by the student adaptation

to college questionnaire: A quantitative review of its structure and relationships with

correlates and consequences. Educational Psychology Review, 24(1), 133-165.

Cuevas, A. E. P. (2015). Thriving in college: Predictors of honors student academic,

psychological, and social well-being. Azusa, CA: Azusa Pacific University.

149
Devlin, A. S., Donovan, S., Nicolov, A., Nold, O., & Zandan, G. (2008). Residence hall

architecture and sense of community: Everything old is new again. Environment and

Behavior, 40(4), 487-521.

Emerick, S. M. (2005). What types of out-of-classroom involvement are related to students' self-

efficacy and grade point: An investigation of the constructs of quantity, quality and

connectedness of involvement. (Doctoral dissertation). Retrieved April 23, 2019, from

Dissertations & Theses: A&I. (Publication No. AAT 3175325).

Garvey, J. C., Guyotte, K. W., Latopolski, K. S., Sanders, L. A., & Flint, M. A. (2018).

Belongingness in Residence Halls: Examining Spaces and Contexts for First-Year

Students Across Race and Gender. Journal of The First-Year Experience & Students in

Transition, 30(2), 9-25.

Garza, A. N., & Fullerton, A. S. (2018). Staying Close or Going Away: How Distance to College

Impacts the Educational Attainment and Academic Performance of First-generation

College Students. Sociological Perspectives, 61(1), 164-185.

Graham, P. A., Hurtado, S. S., & Gonyea, R. M. (2018). The benefits of living on campus: Do

residence halls provide distinctive environments of engagement?. Journal of Student

Affairs Research and Practice, 55(3), 255-269.

Groves, R. M., Fowler Jr, F. J., Couper, M. P., Lepkowski, J. M., Singer, E., & Tourangeau, R.

(2009). Survey methodology (2nd ed.). Hoboken, NJ: John Wiley & Sons.

Guo, S., & Fraser, M. (2015). Propensity score analyses: Statistical methods and applications.

2nd ed. Thousand Oaks, CA: SAGE Publications, Inc.

150
Harwood, S. A., Huntt, M. B., Mendenhall, R., & Lewis, J. A. (2012). Racial microaggressions

in the residence halls: Experiences of students of color at a predominantly white

university. Journal of Diversity In Higher Education, 5(3), 159-173.

Harper, S. R. (2012). Race without racism: How higher education researchers minimize racist

institutional norms. The Review of Higher Education, 36(1), 9-29.

Harper, S. R., Davis, R. J., Jones, D. E., McGowan, B. L., Ingram, T. N., & Platt, C. S. (2011).

Race and racism in the experiences of Black male resident assistants at predominantly

White universities. Journal of College Student Development, 52(2), 180-200.

Hong, G., & Raudenbush, S. W. (2006). Evaluating kindergarten retention policy: A case study

of causal inference for multilevel observational data. Journal of the American Statistical

Association, 101(475), 901-910.

Hotchkins, B. K., & Dancy, T. (2017). A House Is Not A Home: Black Students' Responses to

Racism in University Residential Halls. Journal of College & University Student

Housing, 43(3), 42 - 53.

Huang, Y. R., & Chang, S. M. (2004). Academic and cocurricular involvement: Their

relationship and the best combinations for student growth. Journal of College Student

Development, 45(4), 391-406.

Kavehkar, K. (2013, June 28). The top 10 coolest dorms in the country. The Huffington Post.

http://www.huffingtonpost.com/2013/06/25/coolest-college-dorms_n_3497147.html

Krause, K. L., & Coates, H. (2008). Students’ engagement in first‐year university. Assessment &

Evaluation in Higher Education, 33(5), 493-505.

151
Kuh, G. D. (2003). What we're learning about student engagement from NSSE:

Benchmarks for effective educational practices. Change: The Magazine of Higher

Learning, 35(2), 24-32.

Kuh, G. D. 2009a. The national survey of student engagement: Conceptual and empirical

foundations. New Directions for Institutional Research, 141: 5–20.

Kuh, G. D. 2009b. What student affairs professionals need to know about student

engagement. Journal of College Student Development, 50: 683–706.

Kuh, G. D., Cruce, T. M., Shoup, R., Kinzie, J., & Gonyea, R. M. (2008). Unmasking the effects

of student engagement on first-year college grades and persistence. The Journal of

Higher Education, 79(5), 540-563.

Kuh, G. D., Kinzie, J. L., Buckley, J. A., Bridges, B. K., & Hayek, J. C. (2006). What matters to

student success: A review of the literature (Vol. 8). Washington, DC: National

Postsecondary Education Cooperative.

La Roche, C. R., Flanigan, M. A., & Copeland Jr, P. K. (2010). Student housing: Trends,

preferences and needs. Contemporary Issues in Education Research, 3(10), 45-51.

Lederman, D. (2009, September 2). New heights for dorm luxury at BU. Inside Higher Ed.

Retrieved from https://www.insidehighered.com/quicktakes/2009/09/02/new-heights-

dorm-luxury-bu

Li, F., Zaslavsky, A. M., & Landrum, M. B. (2013). Propensity score weighting with multilevel

data. Statistics in Medicine, 32(19), 3373-3387.

MacKinnon, F. J. D., & Associates (Eds.). (2004). Rentz’s student affairs practice in higher

education. Springfield, IL: Charles C. Thomas.

152
Mayhew, M. J., Bowman, N. A., Rockenbach, A. N., Seifert, T. A., & Wolniak, G. C.

(2016). How college affects students: 21st century evidence that higher education

works (Vol. 3). San Francisco, CA: Jossey-Bass.

McCormick, A. C., Kinzie, J., & Gonyea, R. M. (2013). Student engagement: Bridging research

and practice to improve the quality of undergraduate education. In M. B. Paulsen (Ed.),

Higher education: Handbook of theory and research (pp. 47-92). Netherlands: Springer.

Milem, J. F. (1998). Attitude change in college students: Examining the effect of college peer

groups and faculty normative groups. The Journal of Higher Education, 69(2), 117-140.

Mitra, R., & Reiter, J. P. (2016). A comparison of two methods of estimating propensity scores

after multiple imputation. Statistical Methods in Medical Research, 25(1), 188-204.

Museus, S. D. (2014). The culturally engaging campus environments (CECE) model: A new

theory of success among racially diverse college student populations. In Higher

education: Handbook of theory and research (pp. 189-227). Springer, Dordrecht.

National Survey of Student Engagement. (2018). Engagement indicators. Retrieved from

http://nsse.indiana.edu/html/engagement_indicators.cfm

O’Neil, M. (2014, MAY 19). Door by door, colleges install systems for online control of

building access. The Chronicle of Higher Education. Retrieved from

http://chronicle.com/article/Door-by-Door-Colleges-Install/146647

Owens, J. T. (2010). The impact of university housing construction type on psychosocial

development of first-year students. ProQuest Dissertations and Theses. Retrieved from

http://ezproxy.lib.indiana.edu/login?url=http://search.proquest.com/docview/850523784?

accountid=11620

153
Pace, C. R. (1980). Measuring the quality of student effort. Current Issues in Higher Education,

2(1), 10-16.

Pace, C. R. (1984). Measuring the quality of college student experiences. An account of the

development and use of the college student experience questionnaire. (Report No. 141).

Los Angeles, CA: Higher Education Research Institute.

Palmer, C., Broido, E. M., & Campbell, J. (2008). A commentary on "The educational role in

college student housing". Journal of College & University Student Housing, 35(2), 86-99.

Pan, W., & Bai, H. (Eds.). (2015). Propensity score analysis: Fundamentals and developments.

New York, NY: Guilford Publications.

Pascarella, E. T., Seifert, T. A., & Blaich, C. (2010). How effective are the NSSE benchmarks in

predicting important educational outcomes?. Change: The Magazine of Higher

Learning, 42(1), 16-22.

Pascarella, E. T., & Terenzini, P. T. (1991). How college affects students: Findings and insights

from twenty years of research. San Francisco, CA: Jossey-Bass.

Pascarella, E. T., & Terenzini, P. (2005). How college affects students (Volume 2). San

Francisco, CA: Jossey-Bass.

Pascarella, E. T., Terenzini, P. T., & Blimling, G. S. (1994). The impact of residential life on

students. In E. T. Pascarella, P. T. Terenzini, & G. S. Blimling (Eds.). Realizing the

educational potential of residence halls, 22-52.

Patrick, A. R., Schneeweiss, S., Brookhart, M. A., Glynn, R. J., Rothman, K. J., Avorn, J., &

Stürmer, T. (2011). The implications of propensity score variable selection strategies in

pharmacoepidemiology – An empirical illustration. Pharmacoepidemiology and Drug

Safety, 20(6), 551-559.

154
Porter, S. R. (2011). Do college student surveys have any validity?. The Review of Higher

Education, 35(1), 45-76.

Radford, A. W., Berkner, L., Wheeless, S. C., & Shepherd, B. (2010). Persistence and

attainment of 2003-04 Beginning Postsecondary Students: After 6 years (NCES 2011-

151). Washington, DC: U.S. Department of Education.

Reason, R. D. (2009). An examination of persistence research through the lens of a

comprehensive conceptual framework. Journal of College Student Development, 50(6),

659-682.

Renn, K. A., & Arnold, K. D. (2003). Reconceptualizing research on college student peer

culture. The Journal of Higher Education, 74(3), 261-291.

Rodger, S. C., & Johnson, A. M. (2005). The impact of residence design on freshman outcomes:

Dormitories versus suite-style residences. The Canadian Journal of Higher

Education, 35(3), 83-99.

Rudolph, F. (1990). The American college and university: A history. Athens, GA: University of

Georgia Press.

Ryan, M. B. (2016). On the future of the collegiate way. In M. Evans & T. Burt (Eds.), The

collegiate way: University education in a collegiate context (pp. 3-16). Rotterdam, The

Netherlands: Sense Publishers.

Schudde, L. T. (2011). The causal effect of campus residency on college student retention. The

Review of Higher Education, 34(4), 581-610.

155
Simpson, D. B., & Burnett, D. (2017). Commuters Versus Residents: The Effects of Living

Arrangement and Student Engagement on Academic Performance. Journal of College

Student Retention: Research, Theory & Practice.

https://doi.org/10.1177/1521025117707516

Strayhorn, T. L. (2012). College students' sense of belonging: A key to educational success for

all students. Abingdon, United Kingdom: Routledge.

Strayhorn, T. L., & Mullins, T. G. (2012). Investigating black gay male undergraduates'

experiences in campus residence halls. Journal of College and University Student

Housing, 39(1), 140-161.

Thelin, J. R. (2004). A history of American higher education. Baltimore, MD: Johns Hopkins

University Press.

Thoemmes, F. J., & Kim, E. S. (2011). A systematic review of propensity score methods in the

social sciences. Multivariate behavioral research, 46(1), 90-118.

Terenzini, P. T., & Pascarella, E. T. (1984). Freshman attrition and the residential context. The

Review of Higher Education, 7(2), 111-124.

Tinto, V. (1993). Building community. Liberal Education, 79(4), 16-21.

Tyler, R. W. (1932). Measuring the results of college instruction. Educational Research Bulletin,

253-260.

Walsh, K. J., & Robinson Kurpius, S. E. (2016). Parental, residential, and self-belief factors

influencing academic persistence decisions of college freshmen. Journal of College

Student Retention: Research, Theory & Practice, 18(1), 49-67.

156
Westreich, D., Cole, S. R., Funk, M. J., Brookhart, M. A., & Stürmer, T. (2011). The role of the

c-statistic in variable selection for propensity score models. Pharmacoepidemiology and

Drug Safety, 20(3), 317-320.

157
 CHAPTER FIVE: CONCLUDING THOUGHTS

The purpose of this three-article dissertation was to explore the effect of living on campus on

students’ engagement, health, psychological well-being, academic achievement, retention, four-

year graduation, and satisfaction with the collegiate experience. As I reflect on the findings from

this dissertation, I am reminded of the beliefs I developed having worked in Residence Life and

Housing for over a decade. My idea about the value of an on-campus experience reminds me of

words written by Terenzini and Pascarella (1984) almost 40 years ago that seem to guide many

residence life departments, at least those I have worked in. They argue for the direct, positive

benefit of living on campus, namely that

residence units might be expected to influence the nature of students’ collegiate

experiences both through their physical configurations and consequent influence on the

nature and extent of students’ interactions with one another, and through the sorts of rules

that govern student behaviors, as well as the academic social experiences afforded

students through the nature of the social and academic programming conducted within

the residence unit. (p. 114)

They positioned campus residential environments as “significant” environments that affect

college students and their work influenced higher educational administrators and student affairs

professionals, like me, for years to come. I believed that the collegiate experience of living on

campus was the beneficial experience. For countless student and parent orientation sessions, I

espoused the value of residence halls in developing community, interacting with diverse others,

connecting to campus, enjoying the postsecondary experience more, achieving higher grades, as

well as being retained to a greater degree. I worked with colleagues to design intentional

programs help students grow and develop. Our work was guided by (mostly) correlational

research that often did not account for self-selection into campus living environments. As such,
158
my dissertation used quasi-experimental methodology to explore what effect (if any) living on

campus has for students.

Research Findings

Throughout the twentieth century, research and theoretical justifications considered

specifically how residence halls affect students (e.g., Astin, 1977; Chickering, 1974; Pascarella,

1984; Riker, 1965). While these early theories and findings continued to shape higher education,

they were somewhat problematic in that they were often based on students with privileged

identities (i.e., male, White) and did not represent the interaction of other student identities with

the experience of living on campus. Contemporary research that examines the relationship of

living on campus with a variety of outcomes finds some significant results and some

nonsignificant results depending on the variables used within the model (Mayhew et al., 2016).

Additionally, this research is often limited to correlational findings, based on limited sample

sizes or single institution studies, or lacks the use of quasi-experimental methodologies to

statistically reduce bias associated with students’ self-selection into living on campus. As such,

this dissertation sought to use propensity score stratification to estimate the causal effect of living

on campus on student outcomes.

To address this question, I separated the dissertation into three empirical studies

representing outcomes capturing student engagement, health and psychological well-being, and

academic success. Chapter Two found that living on campus does not have a direct effect on

students’ academic achievement, retention, graduation, or their satisfaction with the collegiate

experience. Additionally, there were not significant interactions between living on campus and

students’ identity characteristics in regard to these outcomes. I also wanted to know if staying on

campus until the fourth year was beneficial in terms of these outcomes. Again, there were no

significant relationships between where a student lived and these particular outcomes. These
159
findings problematize institutions that use these arguments in support of campus policy requiring

students to live on campus. Promoting living on campus to parents and to students might not

include a touted benefit of higher grades, increased retention, graduation, or college satisfaction.

Chapter Three examined the effect living on campus might have on students’ overall

health, psychological well-being, exercise habits, and alcohol-related behaviors, including rates

of consumption and binge drinking. This study found that students who lived on campus their

first year reported higher rates of exercising compared to their off-campus peers. In this case,

proximity to resources is beneficial in that these campus residents can more easily access

exercising spaces. Additionally, this student found no significant direct effect of collegiate

residence on students’ overall health.

However, living on campus had a direct effect on students’ alcohol behaviors. Within this

chapter, I found that students living on campus in their first year reported binge drinking more

frequently and higher rates of alcohol consumption. However, the effect of first year residence

became nonexistent after the first year for these outcomes, suggesting that campus peer cultures

might be to blame. For Latinx students, living on campus was particularly detrimental in terms of

alcohol consumption. Living on campus provides students with a different environment in which

to engage and participate within the university. This residential environment might include peer

cultures that promote drinking as part of the normative collegiate experience.

In Chapter Four, I explored the role living on campus played in students’ engagement

within postsecondary educations. After propensity score adjustment, living on campus increased

students’ positive peer interactions, frequency of interaction with student affairs staff, and co-

curricular involvement. Additionally, the effect of first-year campus residence persists through

the end of the fourth year, with these same outcomes still positive and significant. The findings

160
also indicate that there is no conditional effect based on student ability or first-generation status,

since the combination of those characteristics with the campus residence variable yielded no

significant interactions for any of the student engagement outcomes at either time point. For race,

the study revealed there are some interactions between racial identities and living on campus.

Future Directions for Research

The findings from these three empirical studies led me to consider, broadly, directions for

future research. First, future research should better continue to ensure broader representation in

research and question assumptions about specific research methodologies. Quantitative research

on the impact of higher education is often based on epistemological, theoretical, and quantitative

models that center students who are White, middle class, male, and often of traditional age,

failing to promote the engagement and development of a substantial percentage of the

undergraduate population in the United States (Alcantar, 2014; Bensimon, 2007; Chavez, Ke, &

Herrera, 2012; Teranishi, 2007). With increasing numbers of diverse students enrolling in

postsecondary education, it is imperative that researchers of higher education reframe their work

to continue including these marginalized students and work to center their narratives within

research and findings.

Second, future research could explore the specific type of recreational facility, either

within the hall or on campus, to better understand proximity affects usage and ultimately how

much time students spend exercising and ultimately their health and psychological well-being.

Further understanding of where students exercise, in relation to their collegiate residence, might

determine if peer groups and social connections are the driving force behind recreational

participation. Research should also include health outcomes in analyses as reporting greater rates

of exercise might be a proxy for social connections and belonging, rather than exercising for

health.
161
 Notably missing from the present analyses were results by students’ gender identity and

sexuality. Studies have consistently shown how LGBTQ college students experience

discrimination interpersonally and systemically (Dugan, Kusel, & Simounet, 2012; Rankin et al.,

2010) and that LGBTQ+ students felt more unsafe after their first year on campus compared to

their heterosexual and cisgender peers (Bates & Bourke, 2016). LGBTQ+ students at research

universities also report low levels of institutional support, often feeling as if their institutions did

not respect their identities (Tyler Clementi Center, 2017). Hostile climates for LGBTQ+ students

are related to various negative outcomes, specifically consequences for student learning,

persistence, and mental health and wellness (Kulick, Wernick, Woodford, & Renn, 2017; Rankin

et al., 2010). While research explores campus environments for LGBTQ+ students, there have

been few studies examining the specific role living on campus plays in the lives of these students

when compared directly to their off-campus peers. Within the WNS, this information was not

asked of students and as such I was unable to utilize it within my analyses. Further large-scale,

longitudinal studies are employed to capture data to represent students based on their genders

and sexualities to continue to understand how LGBTQ+ collegians microsystems interact to

affect various outcomes.

Diversity on campus is more broad today than it has ever been, with students’ visible and

invisible identities more openly discussed and represented. Another facet of the critical

quantitative paradigm calls on researchers to “not seek merely to verify models; it seeks new

models and ways of measuring” (Stage & Wells, 2014, p. 5). This shift calls on researchers to

reflect on their ways of undertaking data analyses to see if the methods are equitable or if there

are better ways of capturing the student experience. The last part of this section reflects on the

use of propensity score models and critical quantitative frameworks.

162
 In 2007, New Directions for Institutional Research offered a special issue regarding the

use of quantitative data to answer critical questions. Stage, the editor of the issue, suggested that

critical quantitative inquiry has two major tasks. First, she called on researchers to use large-

scale datasets to reveal inequities and the ways these are systematically perpetuated by

institutions. Second, she noted that the task of the researcher was to “question the models,

measures, and analytic practices of quantitative research in order to offer competing models,

measures, and analytic practices that better describe the experiences of those who have not been

adequately represented” (p. 10). Both perspectives informed these empirical studies, namely

through the use of a longitudinal, multi-institutional dataset and to explore the conditional effects

of living on campus by first generation status, standardized test score ability, sex, and race.

However, I am left questioning how the propensity scores were created and if the conditional

analyses are truly the most equitable way to capture the student experience.

All three empirical studies employed a quasi-experimental methodology in an effort to

make better causal inference regarding living on campus. This statistical method emulates the

“gold standard” randomized controlled trial (RCT). In a perfect RCT, participants are randomly

assigned to either the control or treatment group and these groups are equal in expectation. That

is, the random allocation to the two groups eliminates any outside bias and any identified effect

is attributable to the treatment. Propensity score modeling seeks to statistically replicate this

random allocation and as such to reduce bias that is associated with self-selection into the

“treatment” or in this case, living on campus.

In creating propensity scores, theoretically justified variables are included into a

statistical model, often a logistic regression, in effort to make the treatment and control groups

equal in expectation based on variables that are theoretically linked or thought to be related to

163
participation into the treatment. Within each of my studies, I then transformed teach person’s

predicted probability of treatment into a single, linear propensity score (see Guo & Fraser, 2015;

Pan & Bai, 2015). Scholars note that the variables that should be included in this model are ones

that are theoretically related to the treatment and the outcome(s). As such, I included variables in

my model that represented marginalized identities within the WNS data. One particular variable

was students’ racial identities because of its noted relation to the outcomes of this dissertation

(e.g. Mayhew et al., 2016).

However, as I reflect on my modeling choices and the data used in this dissertation, I am

reminded by the large number of White students represented. While these students are mostly

reflective of the institutional sample’s population percentages, I wondered about the impact on

propensity score adjustment. To explore this idea, I re-created the propensity score model from

the dissertation but restricted the sample to only Black students. Afterwards, I correlated the

Black-only propensity score with score originally calculated. If the only difference between the

models is whether race is used as a balancing variable, I would expect students with high

propensities to continue having those high scores. However, preliminary findings suggest that

Black students’ propensity scores created using the full sample with white students had only a

modest correlation with that score created only using Black students (r = .28). This correlation

might suggest that the estimated propensity for Black students’ participation in the treatment

changes depending on which students are included within the model. Critical quantitative

frameworks call researchers to reflect on their models and assumptions, and these findings could

suggest that how propensity scores are generated are largely a function of majority students. If

that is the case, when estimating the effects of college impact for students at the margins,

different models might be needed. Future research could compare the outcomes from models

164
using full and restricted propensity scores to see if consistent patterns hold. Additionally,

conditional analyses, like those conducted in this dissertation, could be compared to results from

these conditioned datasets. If the patterns are similar across each estimation method, it might be

that propensity score generation among full analytic samples are equitable. However, as

researchers, we should not assume this is the case. Regardless, these thoughts remind me that is

important to think about the research question and the analytic technique beforehand to ensure

that it promotes equity and justice for all students. As researchers, we should take care to not

reproduce systematic inequities in our findings.

Implications for Practice

The three empirical studies within this dissertation offer implications for practice. For

housing professionals, discussions around alcohol may need to change. Higher education

administrators, specifically those working with residence life, might work to better understand

peer culture in their buildings, how it gets reinforced, and ways to positively norm different

behaviors among their students. A further implication comes from the literature that examines

how students on campus are housed. While living in an all first-year hall seems to be positive for

those students in terms of GPA, college satisfaction, and possibly intent to persist (Bronkema &

& Bowman, 2017), might there be a negative impact on social behaviors? Could mixing first-

year students in with upper division students positively affect alcohol behaviors, namely a

decrease in first year students’ alcohol consumption and rates of binge drinking? Further

research on this topic might clarify these questions.

In regard to student engagement, the findings of increased interaction with student affairs

staff that persists throughout all four years is promising. Many housing officials aim to promote

campus connections through residence hall programming and student staff. It might be that

expecting resident assistants to be experts in referral to student affairs offices on campus might
165
sway students’ perceptions of student affairs professionals. That is, through connection with a

student staff member on campus, students in their fourth year might be more aware of or more

inclined to use campus resources. Living on campus also seems to promote positive peer

interactions. These interactions might also promote positive health-related behaviors, such as the

finding in chapter three that students living on campus reported exercising more frequently than

their off-campus peers. When I think about the overall trend of significance among the three

studies, the findings indicate that living on campus only directly affects outcomes that are related

to students’ social behaviors. In other words, why does living on campus have no direct effect on

students’ academic achievement, psychological well-being, or any of the other outcomes I

examined? It could be that these patterns would not persist if the study were replicated with

students in 2019 who have grown up with different forms of technology, different life

experiences, and an increased political divide hinder these social connections?

In terms of academic achievement, retention, four-year graduation, and collegiate

satisfaction, the lack of significant findings implies that programmatic efforts surrounding

academics do not work. Additionally, residence hall environments designed to engage students

academically that include community and private study spaces seem to have no effect.

Practitioners might consider if that is a function of the space or of the programmatic efforts being

offered by student affairs divisions.

Conclusion

Nearly 100 years ago, the case was being reargued for the necessity of college residence

halls. In the Columbia Bulletin, the case was made for campus residence, noting that it is “quite

as important and as essential a part of the work of the University as the provision of libraries,

laboratories, and class rooms. The chief purpose of university residence halls is not one of mere

166
housing, but rather of education and educational influence.” (Columbia Bulletin of Information,

1923, p. 8). I do not believe that the perspective regarding the educational influence of residence

halls will wane in the near future. However, as this dissertation finds, this educational facet of

residence hall living is more social in nature rather than academic, perhaps contrary and to the

chagrin of what campus administrators would hope. Or, it may be that living on campus and

these social connections is the essential piece in the promoting an educated society, as noted with

the founding of Harvard some 400 years ago.

167
 References

Alcantar, C. M. (2014). Civic engagement measures for Latina/o college students. New

Directions for Institutional Research, 2013(158), 23-35.

Astin, A. W. (1977). Four critical years. Effects of college on beliefs, attitudes, and knowledge.

San Francisco, CA: Jossey-Bass Publishers.

Bates, A. K., & Bourke, P. Q. (2016). 2016 Your First College Year Survey. Los Angeles, CA:

Higher Education Research Institute.

Bensimon, E. M. (2007). The underestimated significance of practitioner knowledge in the

scholarship on student success. The Review of Higher Education, 30(4), 441-469.

Bronkema, R., & Bowman, N. A. (2017). A residential paradox? Residence hall attributes and

college student outcomes. Journal of College Student Development, 58, 624-630.

Chávez, A. F., Ke, F., & Herrera, F. A. (2012). Clan, sage, and sky: Indigenous, Hispano, and

Mestizo narratives of learning in a New Mexico context. American Educational Research

Journal, 49(4), 775-806.

Chickering, A. W. (1974). Commuting versus resident students. San Francisco, CA: Jossey-

Bass.

Dugan, J. P., Kusel, M. L., & Simounet, D. M. (2012). Transgender college students: An

exploratory study of perceptions, engagement, and educational outcomes. Journal of

College Student Development, 53(5), 719-736.

Kulick, A., Wernick, L. J., Woodford, M. R., & Renn, K. (2017). Heterosexism, depression, and

campus engagement among LGBTQ college students: Intersectional differences and

opportunities for healing. Journal of Homosexuality, 64(8), 1125-1141.

168
Garvey, J. C., Rankin, S., Beemyn, G., & Windmeyer, S. (2017). Improving the Campus Climate

for LGBTQ Students Using the Campus Pride Index. New Directions for Student

Services, 2017(159), 61-70.

Guo, S., & Fraser, M. (2015). Propensity score analyses: Statistical methods and applications.

2nd ed. Thousand Oaks, CA: SAGE Publications, Inc.

Mayhew, M. J., Bowman, N. A., Rockenbach, A. N., Seifert, T. A., & Wolniak, G. C.

(2016). How college affects students: 21st century evidence that higher education

works (Vol. 3). San Francisco, CA: Jossey-Bass.

Pan, W., & Bai, H. (Eds.). (2015). Propensity score analysis: Fundamentals and developments.

New York, NY: Guilford Publications.

Pascarella, E. T. (1984). Reassessing the effects of living on-campus versus commuting to

college: A causal modeling approach. The Review of Higher Education, 7(3), 247-260.

Rankin, S., Weber, G., Blumenfeld, W., & Frazer, M. S. (2010). 2010 state of higher education

for lesbian, gay, bisexual, & transgender people. Charlotte, NC: Campus Pride.

Renn, K. A. (2017, April 10). LGBTQ student on campus: Issues and opportunities for higher

education leaders. Higher Education Today. Retrieved from

https://www.higheredtoday.org/2017/04/10/lgbtq-students-higher-education/

Riker, H. C. (1965). College housing as learning centers. Washington, DC: American College

Personnel Association.

Stage, F. K. (2007). Answering critical questions using quantitative data. New Directions for

Institutional Research, 2007(133), 5-16.

Stage, F. K., & Wells, R. S. (2014). Critical quantitative inquiry in context. New Directions for

Institutional Research, 2013(158), 1-7.

169
Teranishi, R. T. (2007). Race, ethnicity, and higher education policy: The use of critical

quantitative research. New Directions for Institutional Research, 2007(133), 37-49.

Terenzini, P. T., & Pascarella, E. T. (1984). Freshman attrition and the residential context. The

Review of Higher Education, 7(2), 111-124.

Tyler Clementi Center. (2017, October 26). Emerging research on queer-spectrum and trans-

spectrum students in higher education. New Brunswick, NJ: Author

170