This article was downloaded by: [141.214.17.

222]
On: 07 March 2015, At: 16:43
Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered
office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Communication Education
Publication details, including instructions for authors and
subscription information:
http://www.tandfonline.com/loi/rced20

Toward A Model of Sources of Influence

in Online Education: Cognitive Learning
and the Effects of Web 2.0
Caleb T. Carr , Paul Zube , Eric Dickens , Carolyn A. Hayter &
Justin A. Barterian
Published online: 24 Sep 2012.

To cite this article: Caleb T. Carr , Paul Zube , Eric Dickens , Carolyn A. Hayter & Justin
A. Barterian (2013) Toward A Model of Sources of Influence in Online Education: Cognitive
Learning and the Effects of Web 2.0, Communication Education, 62:1, 61-85, DOI:
10.1080/03634523.2012.724535

To link to this article: http://dx.doi.org/10.1080/03634523.2012.724535

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the
“Content”) contained in the publications on our platform. However, Taylor & Francis,
our agents, and our licensors make no representations or warranties whatsoever as to
the accuracy, completeness, or suitability for any purpose of the Content. Any opinions
and views expressed in this publication are the opinions and views of the authors,
and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content
should not be relied upon and should be independently verified with primary sources
of information. Taylor and Francis shall not be liable for any losses, actions, claims,
proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or
arising out of the use of the Content.

This article may be used for research, teaching, and private study purposes. Any
substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,
systematic supply, or distribution in any form to anyone is expressly forbidden. Terms &
Conditions of access and use can be found at http://www.tandfonline.com/page/terms-
and-conditions
 Communication Education
Vol. 62, No. 1, January 2013, pp. 6185

Toward A Model of Sources of

Influence in Online Education:
Cognitive Learning and the Effects
of Web 2.0
Caleb T. Carr, Paul Zube, Eric Dickens, Carolyn A. Hayter
& Justin A. Barterian
Downloaded by [] at 16:43 07 March 2015

To explore the integration of education processes into social media, we tested an initial
model of student learning via interactive web tools and theorized three sources of
influence: interpersonal, intrapersonal, and masspersonal. Three-hundred thirty-seven
students observed an online lecture and then completed a series of scales. Structural
equation modeling supported several individual hypotheses and partially supported the
overall model. Findings indicated that instructor credibility has a significant positive
effect on content area knowledge, whereas social identification with online colearners
has a negative effect on learning outcomes. Findings are discussed with respect to both
theoretical and practical implications of the integration of interactive media as a
classroom resource.

Keywords: Online Education; Social Media; Cognitive Learning; Social Influence

Educators are increasingly turning to online tools to supplement and supplant

traditional methods of delivering educational content. Scholars have explored the
ways shifting educational processes online affects students and their education

Caleb T. Carr (Ph.D., Michigan State University) is an Assistant Professor at Illinois State University. Paul Zube
(M.A., State University of New York*Albany) is a Visiting Assistant Professor at Ferris State University. Eric
Dickens (M.A., Michigan State University) is a Visiting Assistant Professor at Monmouth College. Carolyn
A. Hayter (M.A., Michigan State University) and Justin A. Barterian (M.A., Michigan State University) are
Ph.D. students in the Department of Counseling, Educational Psychology, & Special Education at Michigan
State University. An earlier version of this study was presented at the 2012 annual meeting of the National
Communication Association in Orlando, FL. The authors are grateful to two anonymous reviewers and the
editor for their assistance refining this manuscript. We also thank Glenn Hansen for his statistical analysis
assistance as well as John Banas and his graduate teaching assistants for aiding in data collection. Caleb T. Carr
can be contacted at ctcarr@ilstu.edu

ISSN 0363-4523 (print)/ISSN 1479-5795 (online) # 2013 National Communication Association

http://dx.doi.org/10.1080/03634523.2012.724535
 62 C. T. Carr et al.
(e.g., Bach, Haynes, & Lewis-Smith, 2007), particularly the use of online technologies
under the control of the educational institution, like course management systems
(CMSs; e.g., Blackboard, WebCT, D2L), instructor-created websites, and publishers’
online supplemental materials. These technologies represent interesting research
contexts but are constrained as they reflect limited-access tools modeled on tradi-
tional classroom techniques and resources, minimizing potential exposure to broad
outside ideas and influences.
A variety of increasingly social tools are used to present course content (e.g.,
lectures, course notes, and practice quizzes) and provide opportunities for student
interaction, active engagement, and the development of a learning dialogue. How-
ever, as technological affordances (the actions and purposes the device serves for its
users) and pedagogical theory converge to promote active online learning, it is likely
the processes involved in students’ assimilation of knowledge are concomitantly
affected. Beyond changing modality, the platforms used to achieve educational goals
are becoming more public and user-generated. As online tools become increasingly
user-friendly and publically accessible, content is commonly delivered outside a
Downloaded by [] at 16:43 07 March 2015

formal classroom context. Faculty, student groups, and scholars can post content on
sites such as YouTube from which any user can draw knowledge (Young, 2008). The
pace with which educators are adapting and integrating new technologies into and
beyond classrooms elicits the question of how these technologies are affecting the
learning process.
We based this study on the need to better understand communication and
education in interactive online settings. Given the complex interactions, multilevel
processes, and evolving technologies involved in education, a unifying theory of
online learning was difficult to articulate. However, guided by the assumption that
many learning processes are fundamental regardless of medium (Clark, 1994), our
model drew heavily from earlier educational models, replicating previous research
to extend findings to a novel educational context. Moreover, the model considered
several factors unique to emergent media*including the influence of anonymous
third parties*on both learning and its antecedents. We developed and tested this
model not to explain the online educational process holistically, but rather as a
starting point from which future scholarship may draw to conceptualize and explain
the process of online education as it shifts from the brick-and-mortar classroom to
the click-and-mortar interactive web classroom.

Shifting Education to Web 2.0

Given the rapid growth of both access to and affordance of the Internet, educators
have seen a rush to provide educational experiences online. The early World Wide
Web*Web 1.0*comprised static web pages created by a single user for public
dissemination of content. Early adopters in the field of education limited online
content to syllabi and study guides, but as available storage and bandwidth for
content increased, so too did educator’s breadth of use. Currently, many educators
use online CMSs to make lecture notes and course grades available outside of
 Testing a Model of Online Learning 63

traditional classrooms (Dutton, Cheong, & Park, 2004; Harrington, Gordon, &
Schibik, 2004), and sometimes supplement in-class interactions with online
discussion boards. Initial studies of mediated education concluded that media’s
increasing penetration into educational contexts would not change the process of
educating nor the educational experience for students (e.g., Clark, 1994). As such,
most Web 1.0 research focused solely on how to adapt extant educational practices
for online delivery.
Yet, to quote Harasim (2000, p. 41), ‘‘Shift happens.’’ Advances in Internet access
and online interaction have given rise to the more social context of Web 2.0 (O’Reilly,
2005), opening new possibilities for changes in the way online content is delivered by
instructors and interpreted by students. Although precise definitions and opera-
tionalization of Web 2.0 are difficult to identify (Burke, 2009), Web 2.0 tools are
typically conceptualized as those exploiting the social, interactive, and participatory
nature of Web 2.0 tools (Walther et al., 2011). These social web tools allow users,
rather than programmers alone, to create the content of the website through
interaction. A current exemplar of the social web is Facebook*a service that, if
Downloaded by [] at 16:43 07 March 2015

unused, becomes relatively unvalued. As an increasing number of people add content

to this social network site (in the form of personal profiles, multimedia uploads, and
interactions in public forums), the site becomes more valuable to both current and
prospective users. As articulated in Anderson’s (2008) long tail approach, which
argues many online tools may get the majority of their use and data from a minority
of users, even these gains may not require all users to participate equally, and enough
use from a limited number of users may be sufficient to substantiate the site and
engage nonparticipatory users. The online encyclopedia Wikipedia is frequently
viewed by many users, but entries are created and edited by only a small percentage
of users (Kittur & Kraut, 2008). However, any user can contribute to Wikipedia by
editing or commenting on content, reflecting a paradigmatic shift in the general use
of the Web and in its application and influence within education (Harasim, 2000).
Web 2.0 has made online collaboration*including education*more dynamic,
interactive, and accessible to a broader public than mere CMSs. An increasing
number of academic programs are conducted entirely online (Allen & Seaman, 2011),
and it is now possible to complete a degree without ever collocating with fellow
students (Harasim, 2000). The participatory web has also enabled large-scale
educational interaction with more diverse colearners than those typically found on
college campuses. For example, Stanford professors conducted a graduate course in
computer science in Fall 2011 with over 58,000 globally distributed students using
publically available videos and automated assignment collection, evaluation, and
feedback (Leckart, 2012). Given the opportunities for asynchronous, anonymous,
and large-scale interaction, Web 2.0 can alter classroom communication, even if
instructors rely on traditional teaching and content delivery methods.
Instructors in both traditional and online classrooms are using publically accessible
online tools to supplement and enhance classroom content and discussions, often
turning to services like YouTube for recorded lectures or humorous examples of class
concepts. A survey of faculty at a mid-sized university (Burke, Snyder, & Rager, 2009)
 64 C. T. Carr et al.
found the majority of instructors perceived YouTube as an effective teaching resource
and concomitantly integrated YouTube videos into their course design. However, as
institutions and instructors use social media for content delivery, they are faced with
questions regarding the effect of Web 2.0 on educational processes. Specifically, social
media tools may change the way individuals communicate with others (Asterhan &
Eisenmann, 2011) and perceive themselves (Walther et al., 2011). It is therefore
important to consider factors that may influence the messages and their effects in
participatory educational media. Thus, in this study we focused on cognitive learning
as an outcome of interest to many educators and administrators, and then developed
an initial model to explore interpersonal, intrapersonal, and masspersonal effects
on cognitive learning via Web 2.0. These concepts are defined and described in the
following sections.

Cognitive Learning
Education researchers have identified several types of learning, including affective,
Downloaded by [] at 16:43 07 March 2015

behavioral, and cognitive learning (e.g., Bloom, 1956; Gagné, 1972; Krathwohl, 2002).
Affective learning focuses on students’ perceptions toward teacher communication
and course content (Gagné, 1972; Pogue & Ahyun, 2006), and behavioral learning
focuses on behavioral modification such as adolescents’ social skills or operant
conditioning in animals (Shuell, 1986). Though both affective and behavioral learn-
ing outcomes are important in educational practices, the present research focuses
on cognitive learning. Bloom (1956) defined cognitive learning as the ‘‘recall or
recognition of knowledge and the development of intellectual abilities and skills’’
(p. 7). Practically, cognitive learning has been conceptualized as the ability of an
individual to retain and understand information (Edwards, Edwards, Shaver, &
Oaks, 2009). Given the prominence of knowledge retention and standardized test-
ing as learning outcomes (Dumont, 1996; Oner, 1995), recall is typically used as an
operationalization of the construct of cognitive learning, and it was the focus of
the present work. Decades of research have focused on cognitive learning outcomes,
with recent lines of research and academic journals focusing on cognitive learning
as it migrates to the Internet (e.g., Rovai, Wighting, Baker, & Grooms, 2009; Yang,
Richardson, French, & Lehman, 2011). Yet, research into learning processes within
the emergent social web requires a reexamination of cognitive learning, as traditional
sources of influence (i.e., teachers and students themselves) may alter communica-
tive and learning processes by altering interactions with instructors and students
experiences and providing additional sources of influence such as anonymous
others.

Toward a Model of Learning in Social Media

Because many stakeholders view cognitive learning as an essential outcome,
maximizing cognitive learning goals remains at the core of multidisciplinary research,
including education, communication, and psychology. Most cognitive learning
research has focused on traditional face-to-face instructional contexts, but many of
 Testing a Model of Online Learning 65

those findings have yet to be confirmed in online, social environments. The purpose
of this study is to retest some of those previously established relationships and to
examine a conceptual model of learning in the Web 2.0 environment. We proffer
cognitive learning via Web 2.0 is broadly influenced by three sources of influence:
interpersonal, intrapersonal, and masspersonal communication. Given the different
communicative processes involved in these three sources, it may be inappropriate for
a single theory to unify and articulate these influences. Consequently, in the following
sections we address these sources, briefly establish their proposed influence on
students’ cognitive learning (guided by the construct of instructor credibility, SIDE
model, and the masspersonal perspective, respectively), and derive hypotheses.

Interpersonal influence of instructors online. Instructors have always been recognized

as interpersonally influential in the education of students and their achievement of
academic goals such as cognitive learning. One of the most important variables affecting
the studentteacher relationship, and therefore student’s learning, is instructor
credibility (Finn et al., 2009; Myers, 2001). Schrodt et al. (2009) noted that instructor
Downloaded by [] at 16:43 07 March 2015

credibility reflects ‘‘students’ attitudes toward the instructor as a source of commu-

nication’’ (p. 351). Students who believe a teacher is a source of competent, trustworthy,
and caring communication demonstrate increased cognitive learning (McCroskey,
Valencic, & Richmond, 2004; Teven, 2001; Tibbles, Richmond, McCroskey, & Weber,
2008). Although web-hosted education may not require the physical presence of an
instructor, perceptions of the virtual instructor still significantly influence class
engagement and learning (Bach et al., 2007). Even when represented with avatars*
digital representations of individuals*in public virtual spaces, instructors must
manage their communication to enhance their perceived credibility (Sherblom,
Withers, & Leonard, 2009). The importance of perceiving the primary source of course
content as believable and trustworthy, even in online social spaces, guided the first
hypothesis:
H1: In the Web 2.0 learning context, increased perceived instructor credibility leads
students to increased cognitive learning.

Interpersonal influence online. An individual’s idiosyncratic and internalized view of

a course and its content, or educational affect, can also have a significant effect on
cognitive learning (Allen, Witt, & Wheeless, 2006) and has become a variable of
significant consideration in instructional communication research. Kearney (1994)
referred to educational affect as ‘‘an increasing internalization of positive attitudes
toward the content of subject matter’’ (p. 81). Affective learning has been identified
as an antecedent to cognitive learning and positively correlated with student’s
motivation to learn (Rodriguez, Plax, & Kearney, 1996). Previous research has
demonstrated the positive relationship between instructor credibility and affective
learning (e.g., Finn et al., 2009; Martin, Chesebro, & Mottet, 1997; Mottet, Parker-
Raley, Beebe, & Cunningham, 2007; Pogue & Ahyun, 2006). For example, students
self-reported greater affective learning and motivation when exposed to written
scenarios of interactions with highly credible teachers, even when the teachers used
 66 C. T. Carr et al.
few immediacy cues (Pogue & Ahyun, 2006). As instructors and students are
expected to interact similarly online and offline, the positive influence of perceived
teacher credibility on a student’s educational affect is also expected, guiding the
second hypothesis:
H2: In the Web 2.0 learning context, increased perceived instructor credibility leads
students to increased educational affect.
In addition to being affected by instructor credibility, intrapersonal perceptions
of class content also influence cognitive learning effects. For example, Edwards and
colleagues (Edwards, Edwards, Qing, & Qahl, 2007; Edwards et al., 2009) detected
a positive relationship between student’s affective learning expectations (as mani-
pulated by online word-of-mouth statements) and state motivation to learn.
Particularly online, where persistent text and public self-presentation facilitate a
feedback loop reinforcing one’s self-perception (Gonzales & Hancock, 2009; Walther
et al., 2011), we predicted a similar effect of educational affect on learning outcomes:
H3: In the Web 2.0 learning context, increased educational affect leads students to
Downloaded by [] at 16:43 07 March 2015

increased cognitive learning.

Masspersonal influence online. Whereas instructors’ outside influence and students’
own self-perceptions are established antecedents to cognitive learning offline, the
influence of others may be radically different online as compared to offline. Unlike
traditional mass media and early Web development in which a sender unidirection-
ally transmitted a message to all receivers with only limited feedback (Berlo, 1960),
new interactive web media enable users to rapidly engage and comment on media
messages (Rafaeli, 1988). Social media allow an individual to post a message, but then
allow others within the network to comment on the message, and in doing so
coconstruct the content ultimately available to and influencing those observing
the site. This one-to-many communication from an individual to the individual’s
social network that enables feedback from the social network defines masspersonal
communication (O’Sullivan, 2005), as messages are interpersonally communicative
in context yet distributed through mass communication channels. Studies have
demonstrated the different effects of masspersonal interactions compared to mass
or interpersonal effects. For example, Walther et al. (2011) found that individuals
processed self-statements differently when messages were presented either in private
(in the form of a Word document) or publically in a social forum (in the form of a
blog). Findings indicated that messages posted masspersonally had a larger influence
on the subject’s self-view and perceived identity than when messages were posted
privately via Word document. In an educational context, masspersonal communica-
tion via social media may lead to the dual influence of others, from their portrayed
social identity online and the comments they present.

Others’ social identities. Offline, the influence of individuals other than teachers in
classroom settings has been empirically supported, usually referring to classmates.
For example, network analyses of traditional classrooms have demonstrated that
gains in classroom performance by semester’s end are higher in classrooms with
 Testing a Model of Online Learning 67

better-performing students than in classrooms with lower-performing students

(Baldwin, Bedell, & Johnson, 1997; Heck, Price, & Thomas, 2004). These results
suggest that aggregated classmates can influence an individual student’s cognitive
learning simply through exposure, so that better classmates can pull up even lower-
performing colearners. Discussion and interaction with classmates allow students to
engage with and apply course materials, enabling learning and the perception of the
course as valuable (Applebee, Langer, Nystrand, & Gamoran, 2003), and this is true
even when discussions with classmates occur online (Russo & Koesten, 2005; Swan,
2002).
However, in Web 2.0 learning environments where individuals can anonymously
join, comment on, and engage in learning processes even if not formal members of
the class or even academically capable of the course content, third-party ‘‘others’’ may
be both anonymous and visually unidentifiable. One effective means of theorizing
the effects of these anonymous others is through the social identity model of
deindividuation effects (SIDE; Reicher, Spears, & Postmes, 1995). SIDE theorists have
Downloaded by [] at 16:43 07 March 2015

argued that when individuals interact with visual anonymity (Lea, Spears, & deGroot,
2001; Lee, 2006) as is often the case online, they identify and relate based on their
social characteristics and traits rather than using idiosyncratic similarities with
particular individuals (Postmes, Spears, Sakhel, & deGroot, 2001; Reicher et al., 1995).
Depersonalized individuals relate based on ingroup/outgroup dynamics (Douglas &
McGarty, 2001), which in turn leads to greater valuation of the social ingroup’s
attitudes and norms, in turn leading to potentially greater social influence by group
members. In many Web 2.0 services (e.g., YouTube) users post pseudonymous
messages often indicating social self-categorization or personality traits without
providing personally identifying or individuating information, similar to instant
message and personal email user names (Bechar-Israeli, 1995; Heisler & Crabill, 2006).
In the absence of personal cues to distinguish others online, students are forced to rely
on others’ social identities to guide perceptions and affect influence.
Social identification positively influences an individual’s perception of an instructor
and engagement in traditional classroom settings (Edwards et al., 2007, 2009). First,
greater identification with colearners’ social identity leads to a student’s perception of
classroom community and control, which is likely to increase perceptions of instructor
credibility Thus, we posited the following hypothesis:
H4: Increased social identification with online commenters of an online lecture leads
students to increased perceived instructor credibility.

Second, social identification with colearners has been shown to increase perceptions
of the classroom as a community (Mazer, Murphy, & Simonds, 2007). As students
identify with peers, they feel more engaged in the learning process and consequently
demonstrate greater learning affect. Therefore, students feeling more a part of the
classroom ingroup are expected to feel greater engagement in the online course.
H5: Increased social identification with online commenters of an online lecture leads
students to increased educational affect.
 68 C. T. Carr et al.
Social identification with online others is also expected to facilitate cognitive
learning. In a comparison of staff versus peer tutoring programs, Moust and Schmidt
(1994) detected differences in students’ social identification with tutors, finding that
peer tutors were able to relate more readily to students’ experiences and educational
challenges. Differences in course and content engagement were attributed, in part,
to student’s identification with peer tutors’ experiences and interpretation of course
content. As students demonstrate greater social identification in social learning
environments (i.e., perceive themselves as part of the classroom ingroup), they will
likely be guided by normative behavior to engage in class content, increasing learning
outcomes (De Simone, Oka, & Tischer, 1999; Paulus, Horvitz, & Shi, 2006). Because
students perceive themselves as part of the Web 2.0 classroom, they may be more
involved and therefore demonstrate more cognitive learning in an online session.
H6: Increased social identification with online commenters of an online lecture leads
students to increased cognitive learning.

Others’ online comments. Beyond mere presence, these anonymous or pseudonymous

Downloaded by [] at 16:43 07 March 2015

others online may influence online learners via the comments they contribute to the
course discussion. The ability to leave feedback in the form of comments that are
copresent with the original media content is unique to social media tools. The nature
of comments posted online is as varied as their foci and includes user-generated
statements regarding a target individual (Walther, Van Der Heide, Westerman, &
Tong, 2008), system-generated amalgamations of user feedback such as rating systems
(Lampe, 2006), and metareviews of reviewers (Resnick, Kuwabara, Zeckhauser, &
Friedman, 2000). In this study, we focused on comments that address the quality of
and engagement with media content, such as comments posted to a YouTube video.
A cursory interaction with YouTube.com quickly illustrates that comments are
common in occurrence and broad in quality and topicality within the web tool.
The positivity of an online comment may be expected to influence a student’s
attitudes toward course content. Although user-generated comments on message
content can reflect interactive exchanges (i.e., discussion between commenters), often
comments reflect reactive exchanges whereby individuals respond to the central
message of the online content or video clip. Walther and colleagues (Walther, Carr,
et al., 2010; Walther, DeAndrea, Kim, & Anthony, 2010) have suggested that these
reactive exchanges shape the way individuals exposed to the central message and
resultant comments interpret and perceive the content. Indeed, the positivity of
comments in social media can influence others’ attitudes (Lee & Youn, 2009; Walther,
Carr, et al., 2010). Consequently, we predict the valence of a comment about an
online lesson can influence students’ perceptions of that lesson, so that a positive
comment (e.g., ‘‘This is a great lecture’’) may result in students’ feeling more positive
toward the content, while a negative comment (e.g., ‘‘This is a terrible lecture’’) may
result in students feeling less positive toward the content.
H7: The positivity of a comment addressing the content of an online lecture is
positively related to the educational affect of students observing the comment.
 Testing a Model of Online Learning 69

Others’ comments can further influence a user’s perceptions, such as those of a

course instructor. Anonymous RateMyProfessor.com teacher evaluations influence
students’ perceptions of a teacher’s abilities and credibility (Edwards et al., 2007),
so that positive reviews result in more positive perceptions of the course instructor and
negative reviews induce more negative perceptions of the course instructor, compared
to when students are not exposed to online ratings. Even statements not directly
addressing a specific perception or target can tacitly influence attitudes. The halo effect
(Asch, 1946) posits that when a target is believed to possess desirable characteristics,
additional positive attributions are overlaid onto the target. Research findings indicate
that statements about the class, its content, and even colearners can color an
individual’s view of the instructor (Myers, 2004). Consequently, instructors’ comment
about course content may influence students’ perceptions of instructor credibility.
H8: The positivity of a comment addressing the content of an online lecture is
positively related to comment observers’ perceptions of the instructor’s
credibility.
Downloaded by [] at 16:43 07 March 2015

Finally, comment valence in social media may influence individuals’ social

identification with the commenter. Negative comments online are often the result
of outgroup members’ seeking to incite dissention or divesture of group goals
(Herring, Job-Sluder, Scheckler, & Barab, 2002). Consequently, it may be expected
that positive comments regarding course material are attributed to ingroup members
(i.e., enrolled students) while negative comments are discounted as being authored by
outgroup members. Recalling Edwards et al.’s (2007) finding that online professor
ratings influence students’ perceptions of course instructors, one likely explanation
for the significant influence of ratings is the perceived social identification with
raters*presumed fellow students. The expectation of a positive relationship between
the positivity of comments and social identification with a commenter guided our
final hypothesis.
H9: The positivity of an online comment positively influences perceptions of social
identification with a commenter.

Structural model. Although testable independently, it is in the aggregate that these

hypotheses constitute an initial model of learning as it occurs in social media, as the
antecedent influences presented occur together in situ. As such, the validity of each
hypothesis is important in addition to validating the complete model (Figure 1).
Consequently, an experiment was designed to test the overall proposed model in
addition to individual hypotheses.

Method
Participants
Three-hundred thirty-seven participants were drawn from several sections of a
single undergraduate communication survey course at a large Midwestern univer-
sity. Participants took part in this research as part of regular classroom activity,
 70 C. T. Carr et al.

Figure 1. Hypothesized model of online learning.

although participation occurred outside of the classroom as a supplement to a course

lecture. Participants’ age ranged from 18 to 92 (M19.67, SD 4.27), 49% of
respondents were female, and respondents indicated a 3.12 (SD.51) average cumu-
lative grade point average. Only 10% of participants were declared communication
Downloaded by [] at 16:43 07 March 2015

majors, as measured by a binary questionnaire item. Six participants were ultimately

excluded from analysis as outliers (see Data Analysis), resulting in a final sample of
331 participants.

Stimulus Materials
Video lecture. A video lecture was created with the aid of a department faculty
member familiar with the course content. Participants confirmed that they had had
no previous contact with the video instructor by negative responses to the item,
‘‘Have you ever met or had a class with the instructor in the video?’’ The video
instructor, a casually dressed, middle-aged white male seated in a faculty office in
front of a shelf of textbooks, presented a simulated 8-minute lecture addressing
elements of nonverbal communication modeled to be delivered as a typical course
lecture. This video lecture content was held constant across conditions and uploaded
to YouTube, a popular Web 2.0 site. The experimental manipulations consisted of
simulated user comments (see next section) that appeared under the video. These
pages were then copied to a local server, and the underlying source code was modified
so that although the hyperlinks were visible, they did not actually navigate the user
to other sites. The video was active and began playing automatically. In all other
respects, the stimuli appeared to be actual YouTube content, although it is not certain
that all subjects believed they were actually interacting with YouTube.com.

User comments. Video comments were adopted from stimuli used by Edwards et al.
(2009), in their test of the effect of instructor-related comments designed to produce
low and high expectations of learning in a course. In positively valenced conditions,
the lecture video was followed by two comments: ‘‘You’ll learn a lot about nonverbal
communication from this video. It gives great tips. Watching it has helped me make
sense of a lot of the readings,’’ and, ‘‘It was easy to learn so much from this lecture.
 Testing a Model of Online Learning 71

I still remember everything that was covered. You can imagine how easy it was to pass
the next exam.’’ In negatively valenced conditions, the lecture video was followed by
two comments: ‘‘You’ll learn nothing about nonverbal communication from this
video. It gives worthless tips. Watching it hasn’t helped me make sense of a lot of
the readings,’’ and, ‘‘It wasn’t easy to learn much from this lecture. I can’t remember
a single thing that was covered. You can imagine how hard it was to pass the next
exam.’’

Commenters. In order to manipulate social identification with commenters on the

YouTube video lecture, each comment was posted by one of four fictitious users,
whose deindividuated user names were manipulated to influence social identifica-
tion. Comments were posted by pseudonymous users. Pseudonyms were modified
from earlier research (Carr, Vitak, & McLaughlin, in press) to activate varied levels
of social identification. Usernames ‘‘Sooner15,’’ and ‘‘ITakeCOMM1113,’’ reflected
the university and course affiliation (i.e., ingroup members), while ‘‘CLB006,’’ and,
‘‘MI_Buttons’’ were used as less socially identifiable (i.e., outgroup) user names.
Downloaded by [] at 16:43 07 March 2015

Across conditions, the two messages were posted by pseudonymous users of like
social identification: Commenters were either both ingroup members or both
outgroup members.

Procedures
Course instructors informed all enrolled students via email and in-class announce-
ments that, in anticipation of the difficulty of upcoming class content, a video had
been made available that students could use to preview course content before reading
or discussing it in class. Moreover, students were informed they would take a brief
questionnaire and practice exam following the video lecture to prepare them for an
upcoming exam. An email including the announcement and link to the study was
sent to participants at least one week before course content was covered in class.
Upon clicking the link, participants were directed to a website that randomly
redirected participants’ browsers to one of four conditions. In each condition, parti-
cipants first watched a video lecture on nonverbal communication. The video was
uniform across conditions, although the commenter and valence of the comments
posted below the video differed based on the experimental treatment condition. After
watching the video, participants were directed to a survey instrument which included
measures and demographic information. After completing the survey measures,
participants completed a 10-item multiple choice quiz developed by course staff to
evaluate their retention of course content presented in the video lecture. After com-
pleting the quiz, all participants were automatically redirected to a single site which
provided quiz answers and collected identifying information to award class credit.

Measures
Several items were included in the survey instrument to assess the variables of interest
in this research. Social identification was assessed using a previously validated 5-item
 72 C. T. Carr et al.
social identification scale (Wang, 2007; Wang, Walther, & Hancock, 2009). Using
7-point Likert-type items with endpoints of 1 (strongly disagree) and 7 (strongly
agree), participants indicated their agreement with statements including, ‘‘The com-
menters belong to a similar social group as me,’’ and, ‘‘The commenters and I are
part of the same social group.’’ The mean of item responses was used to assess social
identification with comment posters, with higher means indicating greater social
identification. The scale demonstrated strong reliability, Cronbach’s a.82.
Affect was measured using a 4-item subset of McCroskey’s (1994) Affective
Learning Scale focusing on affect toward content. Items asked respondents to respond
to their perceptions of class content using 7-point semantic differential items with
anchor points including, Bad/Good, Valuable/Worthless, Unfair/Fair, and Positive/
Negative. The mean of item responses was used to assess affect toward content
presented in the video, with higher means indicating greater educational affect. The
scale demonstrated strong reliability, a.88.
Recalling the importance of standardized testing in education, often operationa-
lized as knowledge retention (Dumont, 1996; Oner, 1995), and following previous
Downloaded by [] at 16:43 07 March 2015

research (Edwards et al., 2009; Frymier & Houser, 1999), cognitive learning outcomes
were measured using a 10-item multiple choice quiz. Specifically, items assessed the
knowledge domain of Bloom’s (1956) taxonomy, asking participants to recall specific
terminology and statistics presented in the video lecture using multiple-choice and
truefalse items for which there was only one right answer. The sum of correct
responses was used as an interval-level indicator of knowledge recall that could range
from zero to ten, with higher scores indicating greater cognitive learning. Participants
averaged 7.89 (SD 1.93) correct responses. A postlecture test allowed between-
subjects tests for differences in the dependent measure based on exposure to different
experimental stimuli while controlling for prior knowledge through random assign-
ment to conditions (Keppel & Wickens, 2004), and allowing attribution of observed
differences to experimental treatments.
The 10-item cognitive learning measure demonstrated moderate but acceptable
reliability (Kuder-Richardson 20 .63) according to guidelines from Subkoviak
(1988). However, the KR-20 should be evaluated cautiously, as it is sensitive to
several factors such as the grouping of questions or the difficulty of the test (Kuder
& Richardson, 1937). Consequently, additional standards were utilized to assess the
reliability and validity of the dependent measure. First the KR-20 value, though
lower than traditional reported reliability measures, is commensurate with previous
research utilizing novel post-test measures of recall such as subject-material tests
(e.g., Chamorro-Premuzic & Furnham, 2003; De Grez, Valcke, & Roozen, 2009;
Ellis, 2000). Second, Saupe (1961) noted that most common classroom tests, such
as the one used in this experiment, have a mean score of around 70 percent of
questions answered correctly. The present results map well onto this heuristic, with
79% of questions answered correctly (SD19%), suggesting an externally valid
cognitive learning measure. Finally, the results of the postlecture test were compared
against the final course grades in course sections from which participants were
 Testing a Model of Online Learning 73

drawn, revealing no significant differences between semester-long performance

(M 78.7, SD11.65) and the present study’s test scores t(336) .47, p .68.
Given these standards and considerations of the reliability and validity of the
measure, analysis progressed using the 10-item postlecture test as the dependent
measure.
Finally, two manipulation checks were used to ensure participants had read and
recalled comments following the video lecture. First, a bivariate yes/no response item
asked participants directly if they had read the comments posted to the video.
Second, participants were presented with four comments and asked to identify one
of the two comments posted below the video lecture they had watched. Eighty-
eight percent of participants responded affirmatively they had read the comments
posted to the video, and 80% of participants who read the comments (73% of all
participants) correctly identified the comment included in their stimuli. These figures
suggest the manipulation was successful, and that most participants attended to the
comments. To maximize external validity and to account for small, subtle effects,
all participants’ responses were analyzed, regardless of whether they explicitly noted
Downloaded by [] at 16:43 07 March 2015

reading the comments.

Data Analysis
To test hypothesized relationships, we used LISREL 8.80 to conduct structural
equation modeling (SEM) with maximum likelihood estimates. Before conducting
the analysis, data were examined to assess the validity of statistical assumptions
for SEM. Measurement error variance was addressed by first examining data for
outliers. Six participants were identified as providing outlying data assessed by the
Mahalanobis’ D2, and their responses were excluded from analysis, leaving n331.
Next, a chi-square test (328.866, p B.001) revealed multivariate skewed and kurtosis
data, and therefore nonnormality. As maximum likelihood estimation assumes
normal distribution, data were normalized (du Toit, du Toit, Mels, & Cheng, 2007)
to meet statistical assumptions.
Finally, the hypothesized relationships were estimated using a combination of
a latent composite and a hybrid model (cf. Stephenson & Holbert, 2003). As cogni-
tive learning was modeled as an endogenous (i.e., influenced by other variables in
the model) latent variable, its error term was specified by first fixing the path from
the latent construct to its observed variable (i.e., exam score) to 1.0, and then error
variance of the observed index was fixed to [(1  reliability)variance of 1.38] to
‘‘reflect the proportion of variance in the index attributable to measurement error’’
(Stephenson & Holbert, 2003, p. 335). Instructor credibility, educational affect, and
social identification with poster were modeled with the hybrid approach, specifying
relationships between the scale items and their respective latent concepts (Table 1).
Combining these latent composite and hybrid models allowed evaluation of the
hypotheses as well as the proposed model, while accounting for measurement error
and variance of the constructs.
 74 C. T. Carr et al.
Table 1 Nonstandardized Coefficients of Latent Variable Indicators in Structural
Equation Model
Instructor Educational Social Identification Cognitive
Indicator item Credibility Affect with Commenter Learning

Credibility 1 .49
Credibility 2 .51
Credibility 3 .58
Credibility 4 .21
Credibility 5 .41
Affect 1 1.07
Affect 2 .77
Affect 3 .78
Affect 4 1.07
Social Identification 1 .65
Social Identification 2 1.06
Social Identification 3 .98
Social Identification 4 .47
Social Identification 5 .97
Social Identification 6 1.03
Downloaded by [] at 16:43 07 March 2015

Cognitive Learning 1.51

Results
The initial structural model demonstrated good model fit, x2(112, n 331)222.33,
p B.001, NNFI .96, CFI 0.97, RMSEA .055(90% CI: .044.065). Although the pro-
bability associated with the chi-squared statistic was lower than desired, the statistic
is problematic to interpret with large samples (Stephenson & Holbert, 2003) as were
present in this study. The non-normed fit index (NNFI), comparative fit index (CFI),
and root mean square error of approximation (RMSEA) are more robust statistics
less influenced by sample size, and demonstrated good fit (Byrne, 2007) of the over-
all model. Figure 2 illustrates the model, including standardized effects, and Table 2
provides correlations among key study variables.
We used SEM to test individual hypotheses in addition to the omnibus model.
Instructor credibility significantly predicted cognitive learning, so that participants
who perceived the teacher in the video as more credible scored higher on the post-
test, b .32, p B.001, supporting H1. Instructor credibility significantly predicted
educational affect, so that participants who perceived the teacher in the video as more
credible also perceived the course content as more valuable and engaging, b .33,
p B.001, supporting H2. Educational affect did not significantly predict cognitive
learning, so that there was no significant effect of a students’ perceived value on
post-test performance, b .02, ns, and thus H3 was rejected. Social identification
with commenters did not significantly predict perceptions of instructor credibility,
so that there was no significant effect of a student perceiving the commenter was of
a similar social group on the perceived trustworthiness and competency of the
presenter, b .01, ns, not supporting H4. Social identification with commenters
increased educational affect, so that participants who perceived themselves as closer
to commenters’ ingroup felt the content addressed in the lecture was more valuable,
 Testing a Model of Online Learning 75

Figure 2. Hypothesized structural equation model results. x2 (112, n 331) 222.33,

p B.001, NNFI .96, CFI 0.97, RMSEA .055(90% CI: .044.065). *p B.05, **p B.01,
***p B.001. All parameter estimates are standardized.

b .23, p B.001, supporting H5. Social identification with commenters signifi-

Downloaded by [] at 16:43 07 March 2015

cantly predicted cognitive learning, but in the opposite direction as predicted, so that
as participants identified more with a commenter’s social group, they performed
more poorly on the post-test, b  .25, p B.01, not supporting H6. Comment
positivity did not significantly influence educational affect, so that comments
favorable toward course material did not lead to more favorable attitudes regarding
course material, b  .04, ns, and thus H7 was rejected. Comment positivity
significantly influenced perceived instructor credibility, so that participants exposed
to positive comments assessed the lecturer as more trustworthy and competent,
b .13, p B.05, supporting H8. Finally, comments favorable toward course material
led to greater social identification with the commenter and comments negative
toward course material led to reduced social identification with the commenter,
b .15, p B.05, supporting H9.
The model additionally infers several indirect effects not tested in the SEM analysis.
As direct effects of variables may be affected in the presence of a mediating variable
(Baron & Kenny, 1986), Sobel (1982) tests assessed whether the mediator affected
the influence of an antecedent to the dependent variable. Although bootstrapping is
often recommended as a technique for indirect effects, Sobel tests allowed testing for
indirect effects based on the b-values and standard errors provided by the SEM, rather
than conducting separate regression equations. Of the five possible indirect effects,

Table 2 Correlation Matrix of Key Study Variables

1 2 3 4

1. Instructor Credibility 
2. Educational Affect .33* 
3. Social Identification with Poster .01 .23* 
4. Cognitive Learning .49* .10 .37* 

*p B.001.
 76 C. T. Carr et al.
three mediators were identified with a Sobel test. The effect of social identification
on cognitive learning was mediated by participants’ perceptions of instructor credibi-
lity (z  2.55, p B.05); the effect of comment positivity on educational affect was
mediated by participants’ perceptions of instructor credibility (z 2.06, p B.05); and
the effect of comment positivity on educational affect was mediated by participants’
perceptions of social identification with the commenter (z2.11, p B.05). Educa-
tional affect did not mediate the relationship between social identification with
commenter and cognitive learning (z .22, p .83); and social identification with the
commenter did mediate the effect of comment positivity on instructor credibility
(z  .18, p B.05).

Discussion
Taken together, these results depict an interesting and somewhat surprising model
for online education, reinforcing some previous findings in educational research while
revealing novel processes and sources of influence in interactive media. Specifically, the
Downloaded by [] at 16:43 07 March 2015

model reinforces the importance of instructors in driving scholarship yet innovatively

indicates that the sense of active learning among a community of peers*so venerated
in previous research*may not exert similar effects online. Perhaps most surprising
is the lack of self-influence: Both instructors and peers can influence students’
educational affect in online learning communities, but there appears to be no main
effect of an individual’s educational affect on learning outcomes. Consequently, while
educators may still seek to utilize instructors and others in online interactive learning
environs to engage learners, individual engagement may simply serve as intrinsic,
idiosyncratic benefit without the direct externality of increased learning performance.
However, this model still affords a novel, albeit tenuous, perspective regarding the
educational process and sources of communicative influence via interactive web
tools.
Findings of the present study support earlier research demonstrating a moderate
to high relationship between instructor credibility and student learning outcomes;
Schrodt et al. (2009) demonstrated a correlation of b.69 between the two variables.
The present results indicate a smaller, yet still significant, effect of instructor credibility
on cognitive learning, b .32. Of the three sources of communicative influence
posited to affect cognitive learning, instructor credibility was the only significant
positive antecedent, reinforcing the need for instructors to present themselves to be
commensurate with knowledgeable, trustworthy sources of information (Schrodt
et al., 2009), even when mediated. In the classroom, educators are able to synchro-
nously receive and react to student feedback regarding credibility issues, which may
range from explicit comments validating or challenging the instructor to subtle
cues expressed via body language or facial expressions. However, when presenting
themselves in static, stable, asynchronous media (YouTube video, discussion forum
posts, etc.), instructors are prohibited from nimble assessment of and reaction to
student perceptions. These findings indicate that instructors need to carefully self-
present themselves to maximize their perceived credibility, which may necessitate
 Testing a Model of Online Learning 77

multiple recordings of lectures, selective self-presentation and disclosures in messages,

and even self-presentation in alternate relational contexts (e.g., carryover effects of
credibility perceptions from RateMyProfessor.com to course content). Such careful,
idealized self-presentations may diminish a student’s ability to identify with the course
instructor, but the substantive effect of perceived credibility on learning outcomes
may necessitate such a trade-off.
Counter to a priori hypothesizing, there was no main effect of educational affect on
learning outcomes. Studies of offline learning have supported the strong influence of
an individual’s perception of course content as valuable on the individual’s learning
the material (e.g., McCroskey, Richmond, & Bennett, 2006; Paulus et al., 2006);
however, this effect was not detected when learning occurred via social media,
suggesting new or at least altered engagement processes in online learning as
compared to offline learning. Interestingly, although instructor credibility and social
identification with commenters both positively influenced educational affect, it
remains unclear how gains in class engagement do not translate into greater cognitive
learning effects. Online learning may not facilitate active learning as in traditional
Downloaded by [] at 16:43 07 March 2015

classrooms, as individuals may reread textual content or review audiovisual materials

in online formats, an explanatory effect echoing previous findings of asynchronous
computer-mediated education (Zhao, Lei, Yan, Lai, & Tan, 2005). Alternately, this
finding may suggest a need to revise notions of edutainment (Okan, 2003) to earlier
conceptualizations that perceived educational affect pessimistically and empirically
demonstrated less engaging instructors (and therefore less engaged students) led
students to greater cognitive learning outcomes (Peck & Veldman, 1973).
Surprisingly, there was a significant effect of social identification with commenter
on cognitive learning, but the effect was the inverse of the effect predicted in H5,
so that as participants identified more with commenters’ social group they were able
to recall less material from the online lesson in the post-test. A potential explanation
for this finding can also be drawn from SIDE (Reicher et al., 1995). SIDE predicts
that the social group with whom an individual identifies should guide behavioral
and attitudinal norms. Stimuli material used commenter pseudonyms meant to
activate course and institutional identities; however, post hoc analysis did not reveal
significant differences in social identification with what should have been in-group
members (e.g., ITakeCOMM1113 and Sooner15; M 4.05, SD .98) and with what
should have been outgroup members (e.g., MI_Buttons and CLB006; M 4.08,
SD .92), t(335) .266, ns. Consequently, differences in participants’ perceptions
of commenters’ social identities may have been guided by superordinate or sub-
ordinate group categorization, resulting in unanticipated effects. For example,
participants may have recognized commenters as members of subordinate social
groups (e.g., course section similar or different from theirs) or superordinate social
groups (e.g., trollers seeking to simply poison an online discussion; cf., Herring et al.,
2002). If participants categorized commenters by a superordinate social category,
the inverse effect identified may have caused participants to seek diversity rather than
community in the online educational situation. Particularly online, where people
expect to interact with depersonalized, diverse populations (Merryfield, 2001),
 78 C. T. Carr et al.
participants may not have sought the traditional class of homogeneous peers, but
rather learn better when they perceive they are part of a larger community of
heterogeneous individuals. This explanation would be consistent with earlier work
(Carr, 2010; Tanis & Postmes, 2003) which has noted the micro-, macro-, and meso-
level social identification with others problematic in SIDE-based research, such as
the research questions surrounding the effect of social identification.
Finally, the indirect effects obtained in the post hoc analysis are notable. Comment
valence had a positive influence on instructor credibility and social identification
(i.e., favorable comments increased instructor perceptions and the belief that the
commenter was of a similar social group), yet had no direct effect on educational
effect. Although previous research has identified the interaction of comment positi-
vity and social identification with commenter on resultant perceptions (Walther,
DeAndrea, et al., 2010), this interaction did not manifest in the present research with
regard to perceptions of course content. However, the effect of comment positivity did
directly influence perceived instructor credibility and social identification, suggesting
others’ comments are not taken at face value to influence perceptions of content,
Downloaded by [] at 16:43 07 March 2015

and intrapersonal effects may be divorced from interpersonal and intergroup effects.
As such, it seems that comments to online lectures may serve as interpersonal, rather
than intrapersonal, sources of attitudinal influence.
Moreover, that the instructor credibility and social identification mediate the
relationship between positivity of comment and educational affect contributes to
understanding the process of engaging learners in online social contexts. Consistent
with previous research, positively valenced comments regarding course content more
positively influenced both instructor credibility (Edwards et al., 2009) and social
identification (Herring et al., 2002), but without any direct effect on educational
affect. Yet in both cases, students may experience cognitive dissonance (Festinger,
1957) when exposed to positive comments about the course or instructor while
possessing negative perceptions of the course, and vice versa. Therefore, the mediating
effects of instructor credibility and social identification on educational affect may be
explained as individuals’ intrapersonal attitudes toward course content are aligned
with others’ statements.

Future Research
The educational process is complex, and we would be naı̈ve to believe this model
comprehensively encapsulates the totality of online learning, regardless of its explana-
tory power. The model is admittedly limited, focused primarily on the psychosocial
effects of various sources of influence frequently occurring online. Yet, this limited
focus provides a baseline on which future research agenda may build and extend
to include additional influences, including those intrinsic, extrinsic, and systemic. For
example, future research may more narrowly study the intrapersonal effects of
individual students, such as technological self-efficacy and previous social media
use (LaRose, Mastro, & Eastin, 2001), on learning outcomes via social media. As an
additional example, this research focused on the communicative effects of messages
 Testing a Model of Online Learning 79

and their senders on student receivers; however, additional studies into usability and
accessibility could consider design implications of the content delivery system on
student experience. It may be that YouTube, an interface and experience commonly
associated with entertainment yet used here to deliver educational stimuli, may not
be the ideal interactive tool for educators and students. Though this work presents
an initial model of interactive web-based learning, it is intended to be neither
comprehensive nor conclusive, and as such we encourage future research to push
beyond the foundational model to explore additional dimensions and outcomes
of online learning.
As an initial framework, this study contains shortcomings which future research
may seek to overcome to further validate, extend, or even refute its findings. Perhaps
most important for future research to address is its single-exposure design. Only
utilizing one video lecture may not have allowed individuals to fully engage in the
online learning experience as would occur in online classes or on-ground classes
heavily supplemented with social media. Consequently, future research with long-
itudinal educational exposures may afford greater environmental validity as well as
Downloaded by [] at 16:43 07 March 2015

increase student engagement, as the nonsignificant effect of educational affect in the

present research may be due to the lack of participants’ involvement in the novel and
single-exposure online lesson.
Likewise, the experimental design used a single post-test measure to assess
influences on cognitive learning, functionally measuring short-term recall of lecture
material. Though such a post-test measure is consistent with common classroom
practices and maps well onto Bloom’s (1956) taxonomic focus of knowledge, future
research would be wise to test long-term learning effects of the proposed model.
Online classes conducted in tandem with on-site classes could assess the influence of
these factors in situ and over the course of an entire semester, testing both cumula-
tive knowledge gains and students’ retention of specific knowledge over a long-term
period. Longer-term studies may help overcome the present study’s weakness of
relying on measurement immediately following exposure to the content being tested,
and help assess the influence of immediacy effects. Longitudinal and iterative test-
ing may further allow assessment of cognitive learning to increase their reliability.
While the reliability and validity of our measure are commensurate with typical
in-classroom tests (Saupe, 1961), future studies may find more reliable measures,
particularly to aid subsequent empirical tests and extensions of our model.
Additionally, this research utilized participants from multiple sections of a single
on-ground course to collect data. There may be a learning effect for those students
who regularly engage in online learning as they become more familiar with the
requisite technology and study techniques for online, self-directed learning. As
individuals spend less time engaging with a novel online tool, they may spend more
time engaging in the learning content, a concern that calls for replication of this study
in a totally online course to consider the effects of students more familiar with online
learning. A related concern is the need to understand effects at the meso-educational
level, considering how learners across various classes may differ in their engagement
and the appropriateness of delivering course content online. For example, in more
 80 C. T. Carr et al.
experiential coursework such as theater and architecture, online communication
with the instructor and class may actually problematize the education process given
problems encoding and relaying messages. Our model may be scaled up to an entire
online curriculum to consider meso-level influences of not only specific instructors,
but also of various courses, departments, or applications of online tools to engage
learners, and the effects on learning outcomes.
Finally, our research and model focused on cognitive learning as an educational
outcome. While recall is a significant dependent variable in much educational
research, Prensky (2010) argued other conceptualizations of learning outcomes, such
as linking and applying concepts (rather than rote memorization) and technological
self-efficacy, may be more critical outcome measures, particularly given the rise of
information technology alongside Web 2.0. Indeed, Metzger, Flanagin, and Zwarun
(2003) found that while college students rely heavily on Internet search engines to
find answers to questions, they often do not critically evaluate search results for
credibility and accuracy. Students’ recall of content may not matter if they are unable
to use technology and assess the veracity and quality of information obtained, and
Downloaded by [] at 16:43 07 March 2015

future research may seek to validate this model with other outcomes equally relevant
in modern education practices.

Conclusion
In this research, we proposed and tested an initial model of online education via
social media. As more educators and institutions seek to design and host online
educational programs (Koehler & Mishra, 2005; Koehler, Mishra, Hershey, & Peruski,
2004), this model affords researchers an initial framework to understand how the
learning process may occur when education exists in an environment beyond the
sterile and controllable walls of brick-and-mortar classrooms. This model presents
scholars in the fields of communication and education an initial conceptualization
of critical factors influencing student learning via social media, specifically the
influences of instructors and online others. As educators increasingly turn to these
emergent web tools to enhance, supplement, and deliver educational content, this
research provides an initial model from which future work may draw.

References
Allen, I. E., & Seaman, J. (2011). Going the distance: Online education in the United States, 2011.
Newburyport, MA: Sloan Consortium.
Allen, M., Witt, P. L., & Wheeless, L. R. (2006). The role of teacher immediacy as a motivational
factor in student learning: Using meta-analysis to test a causal model. Communication
Education, 55, 2131. doi: 10.1080/03634520500343368
Anderson, C. (2008). The long tail: Why the future of business is selling less of more. New York, NY:
Hyperion Books.
Applebee, A. N., Langer, J. A., Nystrand, M., & Gamoran, A. (2003). Discussion-based approaches
to developing understanding: Classroom instruction and student performance in middle and
high school English. American Educational Research Journal, 40, 685730. doi: 10.3102/
00028312040003685
 Testing a Model of Online Learning 81

Asch, S. E. (1946). Forming impressions of personality. The Journal of Abnormal and Social
Psychology, 41, 258290. doi: 10.1037/h0055756
Asterhan, C. S. C., & Eisenmann, T. (2011). Introducing synchronous e-discussion tools in co-
located classrooms: A study on the experiences of ‘‘active’’ and ‘‘silent’’ secondary school
students. Computers in Human Behavior, 27, 21692177. doi: 10.1016/j.chb.2011.06.011
Bach, S., Haynes, P., & Lewis-Smith, J. (2007). Online learning and teaching in higher education.
Buckingham, U.K.: Open University Press.
Baldwin, T. T., Bedell, M. D., & Johnson, J. L. (1997). The social fabric of a team-based MBA
program: Network effects on student satisfaction and performance. Academy of Management
Journal, 40, 13691397. doi: 10.2307/257037
Baron, R. M., & Kenny, D. A. (1986). The moderatormediator variable distinction in social
psychological research: Conceptual, strategic, and statistical considerations. Journal of
Personality and Social Psychology, 51, 11731182. doi: 10.1037/0022-3514.51.6.1173
Bechar-Israeli, H. (1995). From BBonehead  To BcLoNehEAd: Nicknames, play, and identity
on Internet Relay Chat. Journal of Computer-Mediated Communication, 1(2). doi: 10.1111/j.
1083-6101.1995.tb00325.x
Berlo, D. K. (1960). The process of communication. New York, NY: Rinehart & Winston.
Bloom, B. S. (1956). A taxonomy of educational objectives: Handbook 1: The cognitive domain.
New York, NY: Longmans Green.
Downloaded by [] at 16:43 07 March 2015

Burke, M. (2009). The semantic web and the digital library. Perspectives, 61, 316322. doi: 10.1108/
00012530910959844
Burke, S. C., Snyder, S., & Rager, R. C. (2009). An assessment of faculty usage of YouTube as a
teaching resource. The Internet Journal of Allied Health Sciences and Practice, 7(1), 18.
Byrne, B. (2007). Structural equation modeling with AMOS: Basic concepts, applications, and
programming. New York, NY: Taylor & Francis.
Carr, C. T. (2010, June). The diametrics and modality of SIDE: A review and extension. Paper pre-
sented at the annual meeting of the International Communication Association, Singapore.
Carr, C. T., Vitak, J., & McLaughlin, C. (in press). Strength of social cues in online impression for-
mation: Expanding SIDE research. Communication Research. doi: 10.1177/0093650211430687
Chamorro-Premuzic, T., & Furnham, A. (2003). Personality predicts academic performance:
Evidence from two longitudinal university samples. Journal of Research in Personality, 37,
319338. doi: 10.1016/S0092-6566(02)00578-0
Clark, R. E. (1994). Media will never influence learning. Educational Technology Research and
Development, 42, 2129. doi: 10.1007/BF02299088
De Grez, L., Valcke, M., & Roozen, I. (2009). The impact of an innovative instructional intervention
on the acquisition of oral presentation skills in higher education. Computers & Education, 53,
112120. doi: 10.1016/j.compedu.2009.01.005
De Simone, C., Oka, E. R., & Tischer, S. (1999). Making connections efficiently: a comparison of
two approaches used by college students to construct networks. Contemporary Educational
Psychology, 24, 5569. doi: 10.1006/ceps.1998.0983
Douglas, K. M., & McGarty, C. (2001). Identifiably and self-presentation: Computer-mediated
communication and intergroup interaction. British Journal of Social Psychology, 40, 399416.
doi: 10.1348/014466601164894
du Toit, S., du Toit, M., Mels, G., & Cheng, Y. (2007). LISREL for Windows: PRELIS user’s guide.
Lincolnwood, IL: Scientific Software International.
Dumont, R. (1996). Teaching and learning in cyberspace. IEEE Transactions on Professional
Communications, 39, 192204. doi: 10.1109/47.544575
Dutton, W. H., Cheong, P. H., & Park, N. (2004). The social shaping of a virtual learning
environment: The case of a university-wide course management system. Electronic Journal
of e-Learning, 2(1), 6980.
Edwards, A., Edwards, C., Shaver, C., & Oaks, M. (2009). Computer-mediated word-of-mouth
communication on RateMyProfessors.com: Expectancy effects on student cognitive and
 82 C. T. Carr et al.
behavioral learning. Journal of Computer-Mediated Communication, 14, 368392. doi: 10.
1111/j.1083-6101.2009.01445.x
Edwards, C., Edwards, A., Qing, Q., & Qahl, S. (2007). The influence of computer-mediated word-
of-mouth communication on student perceptions of instructors and attitudes toward
learning course content. Communication Education, 55, 255277. doi: 10.1080/0363452070
1236866
Ellis, K. (2000). Perceived teacher confirmation: The development and validation of an instrument
and two studies of the relationship to cognitive and affective learning. Human Communica-
tion Research, 26, 264291. doi: 10.1111/j.1468-2958.2000.tb00758.x
Festinger, L. (1957). A theory of cognitive dissonance. Stanford, CA: Stanford University Press.
Finn, A. N., Schrodt, P., Witt, P. L., Elledge, N., Jernberg, K. A., & Larson, L. M. (2009). A meta-
analytical review of teacher credibility and its association with teacher behaviors and student
outcomes. Communication Education, 58, 516537. doi: 10.1080/03634520903131154
Frymier, A. B., & Houser, M. L. (1999). The revised learning indicators scale. Communication
Studies, 50, 112. doi: 10.1080/10510979909388466
Gagné, R. M. (1972). Domains of learning. Interchange, 3(1), 18. doi: 10.1007/BF02145939
Gonzales, A. I., & Hancock, J. T. (2009). Identity shift in computer-mediated environments. Media
Psychology, 11, 167185. doi: 10.1080/15213260802023433
Harasim, L. (2000). Shift happens: Online education as a new paradigm in learning. The Internet
Downloaded by [] at 16:43 07 March 2015

and Higher Education, 3, 4161. doi: 10.1016/S1096-7516(00)00032-4

Harrington, C. F., Gordon, S. A., & Schibik, T. J. (2004). Course management system utilization and
implications for practice: A national survey of department chairpersons. Online Journal of
Distance Learning Administration, 7(4).
Heck, R. H., Price, C. L., & Thomas, S. L. (2004). Tracks as emergent structures: A network analysis
of student differentiation in a high school. American Journal of Education, 110, 321353.
doi: 10.1086/422789
Heisler, J. M., & Crabill, S. L. (2006). Who are ‘‘stinkybug’’ and ‘‘Packerfan4’’? Email pseudonyms
and participants’ perceptions of demography, productivity, and personality. Journal of
Computer-Mediated Communication, 12, 114135. doi: 10.1111/j.1083-6101.2006.00317.x
Herring, S., Job-Sluder, K., Scheckler, R., & Barab, S. (2002). Searching for safety online: Managing
‘‘trolling’’ in a feminist forum. The Information Society, 18, 371384. doi: 10.1080/
01972240290108186
Kearney, P. (1994). Affective learning. In R. B. Rubin, P. Palmgreen, & H. E. Sypher (Eds.), Com-
munication research measures: A sourcebook (pp. 8185). New York, NY: Guilford Press.
Keppel, G., & Wickens, T. D. (2004). Design and analysis: A researcher’s (4th ed.). Upper Saddle
River, NJ: Prentice Hall.
Kittur, A., & Kraut, R. E. (2008, November). Harnessing the wisdom of crowds in Wikipedia: Quality
through coordination. Paper presented at the ACM conference on Computer-Supported
Cooperative Work, San Diego, CA.
Koehler, M. J., & Mishra, P. (2005). What happens when teachers design educational technology?
The development of technological pedagogical content knowledge. Journal of Educational
Computing Research, 32, 131152. doi: 10.2190/0EW7-01WB-BKHL-QDYV
Koehler, M. J., Mishra, P., Hershey, K., & Peruski, L. (2004). With a little help from your students:
A new model for faculty development and online course design. Journal of Technology and
Teacher Education, 12, 2556.
Krathwohl, D. R. (2002). A revision of Bloom’s taxonomy: An overview. Theory Into Practice, 41,
212218. doi: 10.1207/s15430421tip4104_2
Kuder, G. F., & Richardson, M. W. (1937). The theory of the estimation of test reliability.
Psychometrika, 2, 151160. doi: 10.1007/BF02288391
Lampe, C. A. (2006). Ratings use in an online discussion system: The Slashdot case (Doctoral
dissertation, University of Michigan). Retrieved from http://hdl.handle.net/2027.42/39369
 Testing a Model of Online Learning 83

LaRose, R., Mastro, D., & Eastin, M. S. (2001). Understanding Internet usage: A social-cognitive
approach to uses and gratifications. Social Science Computer Review, 19, 395413. doi: 10.
1177/089443930101900401
Lea, M., Spears, R., & deGroot, D. (2001). Knowing me, knowing you: Effects of visual anonymity
on self-categorization, stereotyping and attraction in computer-mediated groups. Personality
and Social Psychology Bulletin, 27, 526537. doi: 10.1177/0146167201275002
Leckart, S. (2012, March 20). The Stanford education experiment could change higher learning
forever. Wired, 20.
Lee, E.-J. (2006). When and how does depersonalization increase conformity to group norms in
computer-mediated communication?. Communication Research, 33, 423447. doi: 10.1177/
0093650206293248
Lee, M., & Youn, S. (2009). Electronic word of mouth (eWOM): How eWOM platforms influence
consumer product judgement. International Journal of Advertising, 28, 473499. doi: 10.2501/
S0265048709200709
Martin, M. M., Chesebro, J. L., & Mottet, T. P. (1997). Students’ perceptions of instructors’ socio-
communicative style and the influence on instructor credibility and situational motivation.
Communication Research Reports, 14, 431440. doi: 10.1080/08824099709388686
Mazer, J. P., Murphy, R. E., & Simonds, C. J. (2007). I’ll see you on ‘‘Facebook’’: The effects of
computer-mediated teacher self-disclosure on student motivation, affective learning, and
Downloaded by [] at 16:43 07 March 2015

classroom climate. Communication Education, 56, 117. doi: 10.1080/03634520601009710

McCroskey, J. C. (1994). Assessment of affect toward communication and affect toward instruction
in communication. In S. Morreale & M. Brooks (Eds.), 1994 SCA summer conference pro-
ceedings and prepared remarks: Assessing college student competence in speech communication.
Annandale, VA: Speech Communication Association.
McCroskey, J. C., Richmond, V. P., & Bennett, V. E. (2006). The relationships of student end-of-
class motivation with teacher communication behaviors and instructional outcomes.
Communication Education, 55, 403414. doi: 10.1080/03634520600702562
McCroskey, J. C., Valencic, K. M., & Richmond, V. P. (2004). Toward a general model of
instructional communication. Communication Quarterly, 52, 197210. doi: 10.1080/
01463370409370192
Merryfield, M. M. (2001). The paradoxes of teaching a multicultural education course online.
Journal of Teacher Education, 52, 283299. doi: 10.1177/0022487101052004003
Metzger, M. J., Flanagin, A. J., & Zwarun, L. (2003). College student Web use, perceptions of
information credibility, and verification behavior. Computers & Education, 41, 271290.
doi: 10.1016/S03601315(03)000496
Mottet, T. P., Parker-Raley, J., Beebe, S. A., & Cunningham, C. (2007). Instructors who resist
‘‘college lite’’: The neutralizing effect of instructor immediacy on students’ course-workload
violations and perceptions of instructor credibility and affective learning. Communication
Education, 56, 145167. doi: 10.1080/03634520601164259
Moust, J. H. C., & Schmidt, H. G. (1994). Facilitating small-group learning: A comparison of student
and staff tutors’ behavior. Instructional Science, 22, 287301. doi: 10.1007/BF00891782
Myers, S. A. (2001). Perceived instructor credibility and verbal aggressiveness in the college class-
room. Communication Research Reports, 18, 354364. doi: 10.1080/08824090109384816
Myers, S. A. (2004). The relationship between perceived instructor credibility and college student
in-class and out-of-class communication. Communication Reports, 17, 129137. doi: 10.1080/
08934210409389382
O’Reilly, T. (2005). What is Web 2.0: Design patterns and business models for the next generation
of software. Retrieved from http://oreillynet.com/pub/a/oreilly/tim/news/2005/09/30/what-
is-web-20.html
O’Sullivan, P. B. (2005, May). Masspersonal communication: Rethinking the mass interpersonal divide.
Paper presented at the the annual meeting of the International Communication Association,
New York, NY.
 84 C. T. Carr et al.
Okan, Z. (2003). Edutainment: Is learning at risk?. British Journal of Educational Technology, 34,
255264. doi: 10.1111/1467-8535.00325
Oner, S. (1995). Learning outcomes in higher education. In J. Burke (Ed.), Outcomes, learning,
and the curriculum: Implications for NVQs, GNVQs, and other qualifications (pp. 273284).
London, UK: Falmer Press.
Paulus, T. M., Horvitz, B., & Shi, M. (2006). ‘Isn’t It Just Like Our Situation?’ Engagement and
learning in an online story-based environment. Educational Technology Research and Develop-
ment, 54, 355385. doi: 10.1007/s11423-006-9604-2
Peck, R. F., & Veldman, D. J. (1973). Effects of teacher characteristics on cognitive and affective gains of
pupils. Austin, TX: Research and Development for Teacher Education, University of Texas.
Pogue, L. L., & Ahyun, K. (2006). The effect of teacher nonverbal immediacy and credibility on
student motivation and affective learning. Communication Education, 55, 331344. doi: 10.
1080/03634520600748623
Postmes, T., Spears, R., Sakhel, K., & deGroot, D. (2001). Social influence in computer-mediated
communication: The effects of anonymity on social behavior. Personality and Social
Psychology Bulletin, 27, 12431254. doi: 10.1177/01461672012710001
Prensky, M. (2010). Teaching digital natives: Partnering for real learning. Thousand Oaks, CA:
Corwin.
Rafaeli, S. (1988). Interactivity: From new media to communication. In R. P. Hawkins, J. M.
Downloaded by [] at 16:43 07 March 2015

Weimann, & S. Pingree (Eds.), Advancing communication science: Merging mass and inter-
personal processes (pp. 110134). Newbury Park, CA: Sage.
Reicher, S. D., Spears, R., & Postmes, T. (1995). A social identity model of deindividuation
phenomena. European Review of Social Psychology, 6, 161198. doi: 10.1080/14792779443
000049
Resnick, P., Kuwabara, K., Zeckhauser, R., & Friedman, E. (2000). Reputation systems.
Communications of the ACM, 43(12), 4548. doi: 10.1145/355112.355122
Rodriguez, J. I., Plax, T. G., & Kearney, P. (1996). Clarifying the relationship between teacher
nonverbal immediacy and student cognitive learning as the central causal mediator. Com-
munication Education, 45, 293305. doi: 10.1080/03634529609379059
Rovai, A. P., Wighting, M. J., Baker, J. D., & Grooms, L. D. (2009). Development of an instrument
to measure perceived cognitive, affective, and psychomotor learning in traditional and
virtual classroom higher education settings. The Internet and Higher Education, 12(1), 713.
doi: 10.1016/j.iheduc.2008.10.002
Russo, T. C., & Koesten, J. (2005). Prestige, centrality, and learning: A social network analysis of
an online class. Communication Education, 54, 254261. doi:10.1080/03634520500356394
Saupe, J. L. (1961). Some useful estimates of the Kuder-Richardson formula number 20 reliability
coefficient. Educational and Psychological Measurement, 21, 6371. doi: 10.1177/00131644
6102100105
Schrodt, P., Witt, P. L., Turman, P. D., Myers, S. A., Barton, M. H., & Jernberg, K. A. (2009). Instructor
credibility as a mediator of instructors’ prosocial communication behaviors and students’
learning outcomes. Communication Education, 58, 350371. doi: 10.1080/03634520902926851
Sherblom, J. C., Withers, L. A., & Leonard, L. G. (2009). Communication challenges and oppor-
tunities for educators using Second Life. In C. Wankel & J. Kingsley (Eds.), Higher Education
in Virtual Worlds: Teaching and Learning in Second Life (pp. 2946). Bingley, U.K.: Emerald
Group.
Shuell, T. J. (1986). Cognitive conceptions of learning. Review of Educational Research, 56, 411436.
doi: 10.3102/00346543056004411
Sobel, M. E. (1982). Asymptotic confidence intervals for indirect effects in structural equation
models. Sociological Methodology, 13, 290312. doi: 10.2307/270723
Stephenson, M. T., & Holbert, R. L. (2003). A Monte Carlo simulation of observable versus latent
variable structural equation modeling techniques. Communication Research, 30, 332354.
doi: 10.1177/0093650203030003004
 Testing a Model of Online Learning 85

Subkoviak, M. J. (1988). A practitioner’s guide to computation and interpretation of reliability

indices for mastery tests. Journal of Educational Measurement, 25, 4755. doi: 10.1111/j.1745-
3984.1988.tb00290.x
Swan, K. (2002). Building learning communities in online courses: The importance of interaction.
Education, Communication & Information, 3, 2349. doi: 10.1080/1463631022000005016
Tanis, M., & Postmes, T. (2003). Social cues and impression formation in CMC. Journal of
Communication, 53, 676693. doi: 10.1093/joc/53.4.676
Teven, J. J. (2001). The relationships among teacher characteristics and perceived caring. Com-
munication Education, 50, 159169. doi: 10.1080/03634520109379241
Tibbles, D., Richmond, V. P., McCroskey, J. C., & Weber, K. (2008). Organizational orientations in
an instructional setting. Communication Education, 57, 389407. doi: 10.1080/0363452080
1930095
Walther, J. B., Carr, C. T., Choi, S., DeAndrea, D., Kim, J., Tong, S., & Van Der Heide, B. (2010).
Interaction of interpersonal, peer, and media influence sources online: A research agenda for
technology convergence. In Z. Papacharissi (Ed.), The networked self (pp. 1738). New York,
NY: Routledge.
Walther, J. B., DeAndrea, D., Kim, J., & Anthony, J. C. (2010). The influence of online comments on
perceptions of antimarijuana public service announcements on YouTube. Human Commu-
nication Research, 36, 469492. doi: 10.1111/j.1468-2958.2010.01384.x
Downloaded by [] at 16:43 07 March 2015

Walther, J. B., Liang, Y. J., DeAndrea, D. C., Tong, S. T., Carr, C. T., Spottswood, E. L., & Amichai-
Hamburger, Y. (2011). The effect of feedback on identity shift in computer-mediated
communication. Media Psychology, 14, 126. doi: 10.1080/15213269.2010.547832
Walther, J. B., Van Der Heide, B., Westerman, D., & Tong, S. T. (2008). The role of friends’ behavior
on evaluations of individuals’ Facebook profiles: Are we known by the company we keep?.
Human Communication Research, 34, 2849. doi: 10.1111/j.1468-2958.2007.00312.x
Wang, Z. (2007, November). The assessment of interpersonal attraction and group identification
in virtual groups. Paper presented at the annual meeting of the National Communication
Association, Chicago.
Wang, Z., Walther, J. B., & Hancock, J. T. (2009). Social identification and interpersonal com-
munication in computer-mediated communication: What you do versus who you are in
virtual groups. Human Communication Research, 35, 5985. doi: 10.1111/j.1468-2958.2008.
01338.x
Yang, D., Richardson, J. C., French, B. F., & Lehman, J. D. (2011). The development of a content
analysis model for assessing students’ cognitive learning in asynchronous online discus-
sions. Educational Technology Research and Development, 59, 4370. doi: 10.1007/s11423-
010-9166-1
Young, J. R. (2008, January 9). Thanks to YouTube, professors are finding new audiences.
The Chronicle of Higher Education. Retrieved from http://chronicle.com/article/Thanks-to-
YouTube-Professo/381/
Zhao, Y., Lei, J., Yan, B., Lai, C., & Tan, S. (2005). What makes the difference? A practical analysis
of research on the effectiveness of distance education. The Teachers College Record, 107,
18361884. doi: 10.1111/j.1467-9620.2005.00544.x