Hindawi Publishing Corporation

ISRN Education
Volume 2013, Article ID 958530, 29 pages
http://dx.doi.org/10.1155/2013/958530

Review Article
Curriculum-Based Measurement: A Brief History of
Nearly Everything from the 1970s to the Present

Gerald Tindal
College of Education, University of Oregon, Eugene, OR 97403, USA

Correspondence should be addressed to Gerald Tindal; geraldt@uoregon.edu

Received 8 October 2012; Accepted 13 November 2012

Academic Editors: N. Dumais, F. Jimenez, and L. McCall

Copyright © 2013 Gerald Tindal. This is an open access article distributed under the Creative Commons Attribution License, which
permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

This paper provides a description of 30 years of research conducted on curriculum-based measurement. In this time span, several
subject matter areas have been studied—reading, writing, mathematics, and secondary content (subject) areas—in developing
technically adequate measures of student performance and progress. This research has been conducted by scores of scholars across
the United States using a variety of methodologies with widely differing populations. Nevertheless, little of this research has moved
from a “measurement paradigm” to one focused on “training in data use and decision making paradigm.” The paper concludes with
a program of research that is needed over the next 30 years.

1. Introduction in secondary schools) summarize findings from the research

and then address different aspects of CBM research. For oral
In 2003, Bill Bryson wrote A Short History of Nearly Every- reading fluency, the focus and reflective summary is on the
thing and detailed the history of the cosmos, big bang, analytical models used to determine the growth within a
earth, and the composition and development of life and the year and factors influencing reliability of measurement. For
population of the planet. The book was massive in detail
early literacy, the focus and reflective summary is on the rich
and comprehensive in scope. In this paper, I take a similar
nomological net depicting the interrelations among measures
approach and provide a complete guide to curriculum-
(variables) used to document change in skills as students
based measurement, its beginnings, a review of research
conducted over the past 30 years, and a glimpse into issues left learn to read. The writing section addresses administration
unattended. Since its beginning in the late 1970s, a number of measures and metrics for scaling behavior in a manner
of researchers have come together and provided new insights that is sensitive to change for various populations (elementary
into understanding how teachers can monitor learning and versus secondary students). In mathematics, the central
how learning can inform teaching. issue is domain sampling for developing general outcome
I have divided the paper into sections. Section 1 provides measures. Finally, secondary content CBM considers access
a historical coverage of CBM: the beginnings at the Uni- and target skills as students move from learning basic skills
versity of Minnesota from 1977 through 1984. The next five to application of them in learning subject specific content.
Sections 2–6 then address research in various subject areas: Each section ends with a reflective interpretation on this
oral reading fluency, early literacy, writing, mathematics, issue of measurement methodology that is slightly different
and secondary content areas. Finally, Section 7 presents a than that proposed by Fuchs [3]. In her research analysis,
program of research needed to guide future development, studies were depicted in terms of stages. In the first stage, the
and Section 8 provides a theoretical perspective using the focus was on the technical features of static scores (traditional
logic of validation from Messick [1] and Kane [2]. An reliability and criterion validity studies). Then, in a second
important structural note is that the five subject area sections phase, the technical features of growth were considered in
(reading, early literacy, writing, mathematics, and content documenting the relation between change and proficiency
2 ISRN Education

in an academic domain. Finally, the last stage (three) was 2.1. Institute for Research on Learning Disabilities with Curric-
reached when the focus of research was on the relevance ulum-Based Measurement. Curriculum-based measurement
of the data for making instructional decisions (including (CBM) was originally coined in the mid-1970s by Deno and
teacher acceptance in using these data). Mirkin [5] in which they articulated two systems for teachers
While Fuchs’ heuristic is certainly useful and viable, the to use in systematically monitoring students’ progress over
manner in which research on CBM is analyzed in this paper time: (a) mastery monitoring (mm) and (b) general outcome
is quite different. Rather than thinking of the research in a measurement (GOM). To this day, the two forms are con-
sequential manner (with the earlier phases considered more sidered by the National Center on Response to Intervention
basic and the last phase more applied), the depiction in this (RTI) (http://www.rti4success.org/).
MM is the traditional model in which teachers provide
paper is more iterative and targeted at various and more
instruction in a specific skill unit and then test students’
encompassing critical dimensions of a measurement system
mastery. As a result, MM has two unique features from a
within a decision making framework, with some dimensions measurement perspective: (a) the domain for sampling items
more critical in CBM, given its design for classroom use by is limited to those taught, and (b) the graphic display reflects
teachers in monitoring teaching and learning. Finally, in the a stair step function with time on the 𝑥-axis and mastery of a
last two sections, CBM is depicted in the context of response unit (the dependent variable) on the 𝑦-axis. The steeper the
to intervention (RTI) for evaluating instructional programs steps, the faster students are mastering units.
and within a theoretical context using an argumentative In contrast, GOM is based on comparable alternate forms
approach to the process of validation. from long-range goal (LRG) material (skills to be taught over
The big idea in this document is to summarize the an extended period of time). Because CBM was developed
research and findings from the past 30 years to rearrange within special education, the LRG is often drawn from the
the focus of curriculum-based measurement for the next 30 Individual Education Program (IEP). The two unique features
years. In short, the past 30 years of research on curriculum- of this model are that (a) the domain for sampling items spans
based measurement privileged measurement (as noted in the the material being taught over an extended time with items
summary of findings in the five subject area sections) over that both preview and review skills for any given student and
an argument-based validation scheme on training and data- (b) the forms are considered comparable to each other so
based decision making (as noted in the last two sections). the graphic display can include slope and variability as the
primary dependent variable.
2. CBM: The Beginnings at the These early features are important to consider as the
eventual development of CBM moved on to a more norm-
University of Minnesota referenced perspective for comparing students to each other
The early research on curriculum-based measurement at the in order to identify those at risk of failing-to-learn. Because
Institute for Research on Learning Disabilities at the University of the preferred psychometric properties of alternate forms
of Minnesota established two major lines of reports focusing over time leading to slope and variability as outcomes as well
as the difficulty in defining and defending “mastery”, the field
on (a) technical adequacy and (b) teacher decision making.
has generally focused on GOM methodologies.
Following is a brief summary of this research, most of which
The immediate aftermath of the original publication by
still stands today as relevant and of value in considering
Deno and Mirkin [5] was a five-year program of research
continued research that needs to be done in both explicating conducted at the Institute for Research on Learning Dis-
and extending the findings. abilities (IRLD) at the University of Minnesota from 1979
Although not specifically based on the Standards for Edu- through 1983. The focus of the institute was to (a) investigate
cational and Psychological Testing [4], the original research current practices in identification of learning disabilities,
addressed the reliability and validity components as defined and (b) document the technical characteristics of CBM for
at that time. In addition to this focus on technical adequacy, monitoring student progress. In total, 23 monographs and 144
several other characteristics were demanded of the measure- technical reports were published through the IRLD with six
ment system so teachers could use the measures on a regular monographs and 62 research reports written on CBM. The
basis. Deno and Mirkin [5] emphasized specific features of monographs established the critical dimensions of systematic
these measures to allow alternate forms so that a time series formative classroom measurement system [11, 12].
data display could be used to determine if programs were At that time, few alternative options were present for
working or needing to be adjusted. For example, the measures teachers to use in formatively evaluating their instructional
had to be easy to create, quick to administer, usable by all, programs. Basal series had mastery tests with little technical
from the curriculum, and with sufficient technical adequacy. adequacy (reliability and validity) in decision making and
In the end, they laid the groundwork for an experimental actually misinformed teachers about skills being learned and
view of teaching in which any intervention was considered practiced by students [13–17]. Informal reading inventories
a hypothesis that needed vindication” [6, p. 32]. This last and readability-based estimates of passage difficulty were not
component was deemed to be a critical component of for- (and are still not) reliable [18, 19].
mative evaluation [7, 8] and development of Individualized CBMs and time series data on student progress are effec-
Education Programs (IEPs) [9] and assessment of short- and tive setting events for influencing teachers’ decision making
long-term goals [10]. and practices as well managing Individualized Education
ISRN Education 3

Programs (IEPs) [7, 8, 10, 20–30]. Some of this research was and decision making. In this early review, the models were
done because of the unique (statistical) issues present in time compared on several dimensions: relationship to assessment
series data [31–34] the effects of which are summarized by and decision making, underlying premises about the relation-
Marston et al. [35]. ship of assessment data to instruction (primarily addressing
Training and structure of instruction is an important the prescriptive nature of recommendations to teach), type
component of effective data utilization for systematic use in of student response, focus of material for monitoring student
school settings [30, 36, 37]. progress (the “shelf life” of the materials), test format (degree
The practice of implementing CBM should be planned in of standardization, length, and test construction skills), and
its adoption in the schools and may successfully be used by technical adequacy (reliability and validity).
both teachers and students to improve performance [27, 29,
With the exception of curriculum-based measurement
38–40].
(CBM), these models have had a limited presence in the
CBMs can be effectively used in screening students at
risk of failure to learn basic skills [41, 42]. Programs and research literature in special education. CBM, however, has
decisions may be appropriately evaluated at a systems level in been extensively researched over the past 30 years, in both
a manner that comprehensive ties different decisions together general and special education. It has become a term used
in a manner [35, 42–44]. widely in the special education literature and has become
Domain sampling (breadth and representation) in the the backbone of response-to-intervention (RTI), serving as
curriculum is an important issue to consider in developing the preferred form of progress monitoring used to evaluate
progress monitoring probes, influencing sensitivity to change instructional programs. Indeed, its staying power has been a
and rate of progress [45–48]. function of (testimony to) its compatibility with traditional
CBMs are reliable and are both related to other critical measurement development and the applications of standards
measures of basic academic skills and can serve as adequate for educational and psychological testing [4]. Indeed, one of
measures of progress in learning basic skills [13, 20–22, 49– the core focuses on CBM was the need for technical adequacy,
56]. something which was not embedded into the fiber of the
In the end, the research is summarized in three mono- other systems. Indeed, the proliferation of investigations on
graphs focusing on establishing and monitoring progress on CBM has been most recently acknowledged in a Festschrift,
IEP goals [57], systemic use of student classroom perfor- A Measure of Success: How Curriculum-Based Measurement
mance and progress in making multiple decisions [28, 29, has Influenced Education and Learning [67].
57, 58], and a review of the research over the five years of And with this core attention to traditional measurement
funding establishing curriculum-based evaluation in reading, development, a number of tractable issues can be pursued
spelling, and written expression [59]. that would not be possible with non-GOM systems. For
The early research also spawned three books in the area example, it is difficult to address slope with mastery monitor-
of systematic classroom assessment, each addressing unique ing or ratios of correct to incorrect. In contrast, with a solid
aspects of this behavioral measurement approach. The books measurement system based on general outcomes, further
synthesized the research from the institute and researchers research can be conducted on slope, which is addressed
responsible for the initial research as follows: (a) Shinn [60], next in the context of oral reading fluency. Similarly, major
(b) Shinn [61], and (c) Tindal and Marston [62]. The former developmental trajectories in skill development are difficult
two books summarized the research completed during the to document with measurement systems across brief periods
funding of the IRLD and its immediate aftermath, while the of time that is typical of non-GOM systems. Yet, with the
latter book extended CBM into a larger, test driven model for research on early literacy, it is possible to measure the
measurement development. development of early skills (in letter names and sounds)
and their transformation into decoding and fluent reading.
2.2. Other Models: Non-General Outcome Measurement. With non-GOM systems, research on administration and
From the beginning, the term begged the question, and sev- metrics for scoring and reporting outcomes is restricted, an
eral alternative terms were offered. In a special issue of Excep- issue addressed in the area of written expression based on
tional Children, the term CBM was subjected to a debate. a GOM approach. Ironically, with restricted domains from
For Gickling and Thompson [63], the term curriculum-based a non-GOM approach, it is nearly impossible to research
assessment was used to denote administration of mastery the effects of domains; in coverage of mathematics, this
monitoring tasks and then using the ratio of known (cor- area is addressed. Finally, with secondary school measures
rect items) to unknown (incorrect items) to diagnostically (of access and target content skill), a non-GOM approach
determine what to teach. At about the same time, Howell can be used but reflects the traditional manner in which
[64] developed curriculum-based evaluation in which two measurement is conducted in the classroom. In fact, a GOM
types of assessments were administered: survey level and approach in these settings is present with only a tentative
specific level. In addition to the models just noted, these foothold accomplished in the past 20 years. These are the
authors added criterion-referenced curriculum-based assess- topics for the next several sections after which two major
ment [65] to describe a mastery-monitoring model with issues are addressed: what do we know and need to know in
sequentially arranged objectives. Eventually, various models implementing a measurement system for decision making in
were compared [66], and since those early days, most of the classrooms and how can a validity argument be deployed in
research has been on various dimensions of measurement the development of the curriculum-based measurement.
4 ISRN Education

3. Oral Reading Fluency CBM when using slope to prompt change, more consistent
progress monitoring occurs across conditions of goal
Oral reading fluency (ORF) has been extensively studied over [76].
the past three decades. Since the original study [50]. In a con-
ference paper published by Tindal and Nese [68], the history As Tindal and Nese [68] note, the sampling plans for
of research is summarized on slope of improvement for ORF. students have been very inconsistent, primarily representing
In the following section, critical text from that publication is convenience samples for researchers, considerably different
summarized. In addition, other topics of generalizability and in size (from several dozen to thousands), and with different
standard error of measurement are addressed. demographic characteristics documented. The results have
been reported by grade level and, if special education was
3.1. Within Year Average Slope (Growth). The earliest study noted, it was dummy coded. The measures used in these
on average growth in oral reading fluency was by Fuchs studies have been equally varied as the students measured
et al. [69]. They documented that weekly growth in words and described as generic, standard, or grade appropriate.
correct per minute was 1.5 and 2.0 in first and second grades, Likewise, frequency of measurement has been inconsistent
1.0 and 1.5 in third grade, 0.85 and 1.1 in fourth grade, from weekly to triannually, with seasonal effects widely
0.5 and 0.8 in fifth grade, and 0.3 and 0.63 in sixth grade. interpreted. In general, they report the following.
In 2001, Deno et al. [70] documented growth from nearly
(1) Growth is not linear within a year but likely quadratic.
3,000 students tested from four regions of the country. They
Even in the early research by Deno et al. [70], a large
reported 2.0 words correct per minute growth per week
percentage of students showed nonlinear growth.
until students achieved 30 WRCM; thereafter, students in
the general education population improved at least 1.0 word (2) Growth also is not linear across years but reflects
correct per minute/week. much greater improvement in the early grades that
In the past six years, seven studies have been published tapers off in the later elementary years.
on slope of progress in oral reading fluency [71–77]. This (3) Students receiving special education services enter at
recent rise in publications has led to various findings and a lower level (intercept) and generally grow at a slower
conclusions. rate.
More growth occurs from the beginning of the year to
the middle of the year than from the middle to the end of (4) Within year (weekly) growth in oral reading, fluency
the year, with overall growth being moderate [74] and more ranges from .50 words correct per minute per week to
specifically from fall to spring than winter to spring [71]. 2.0 words correct per minute per week.
When quadratic models cannot be computed (because of An important issue behind this research is the growing
the limited number of measurement occasions), a piecewise sophistication of the estimates of growth. In the original
model of growth can be conducted and also fits better than research, growth is simply calculated in a linear manner either
a linear model, with all slopes being positive but negatively as raw score gains or using an ordinary least squares (OLS)
accelerated and decreasing as grade level increases [73]. regression estimate. Furthermore, this growth is artificially
And, as confirmed by Nese et al. [77], slopes are nonlinear converted to a weekly average by taking the values and
for students in grades 3–5 taking benchmark easyCBM dividing by the number of weeks within the span (e.g., year or
measures, again with more growth occurring in the fall than season). In only two studies were students actually measured
in the winter, at least for students in grades three and four, and on a weekly basis (see the two studies by Jenkins et al. [75]
more growth for students in earlier grades; students of color, and Jenkins and Terjeson [76]).
with a disability (SWD), and eligible for free and reduced In the more recent research, hierarchical linear modeling
price lunch perform considerably lower at the beginning of (HLM) is being used [78] for making growth estimates
the year with slope only significant for SWD. conditioned on (nested in) various levels [79]. Typically, time
Most of this research involves measurement on only two is considered level one, student characteristics are considered
or three occasions, either in the fall and winter or in the level two, and teacher or school is considered level three.
fall, winter, and spring. Rarely is instruction considered or This kind of estimation provides a more precise estimate
progress monitoring being used. In other research on slope as the confounding effects of various levels are controlled.
that is more oriented to teacher use in the classroom, two For example, the study by Nese et al. [77] specifically model
explanations have been invoked for different findings on time in both a linear and nonlinear fashion. This finding of
slope. nonlinear growth (within years or over grades) has profound
implications for establishing aimlines (or long-range goals) or
(1) Steeper slopes may occur because of the lower perfor-
interpret results. If indeed students grow more from the fall
mance obtained when only one (versus four) baseline
to the winter than they do from the winter to the spring, then
measure is used in calculating slope. Furthermore,
teachers would be wise to use this information as mid-year
more accurate estimates occur when measurement is
checks on their decision making. More progress should be
either pre-post or every three weeks [75].
expected of students in this first four to five months, without
(2) Instructional changes are made more often as a func- which teachers would be making false positive decisions
tion of weekly goal ambitiousness (1 to 1.5 word per of growth and not changing instruction because they are
week) but interact with frequency of measurement; assuming linear growth. What appears to be an effective
ISRN Education 5

program actually may not be effective, particularly if the of error reliability coefficients and the SEM given a specified
teacher is using a normative reference of growth (e.g., how number of probes when making relative and absolute deci-
other students are doing). sions. Using 20 passages from DIBELS, they found that “the
largest amount of variation in scores, 81%, was attributable
3.2. Generalizability Theory and Standard Error of Measure- to the person facet. Item, or probe, accounted for 10% of the
ment. Separate from research on slope and expected growth variance, and 9% of the variation was located in unaccounted
for students within the school year, another line of research sources of error” (p. 331) with the index of dependability
has developed in the post-IRLD era. In the original IRLD ranging from .81 (SEM of 18 WCPM) with one passage to .97
research, the focus was on traditional analyses of reliability (SEM of 6 WCPM) with nine passages. In a slight variation
(test-retest, inter-judge, and alternate form). In the past of analytical research on reliability of oral reading fluency
decade, however, advancement in the research continues to measures, Christ and Ardoin [86] compared four passage
focus on reliability but with an emaphasis on the conditions construction procedures: “(1) random selection, (2) selection
under which oral reading fluency is measured. Specifically, based on readability results, (3) selection based on mean
two areas of research (with ORF measures) are G-theory and levels of performance from field testing, and (4) use of ED
the influence of various data characteristics on the standard procedures. . .(an estimate of inconsistencies in performance
error of measurement (SEM). within and across both students and passages)” (p. 59). They
reported generalizability coefficients consistent with previous
3.2.1. G-Theory Studies of ORF. In generalizability theory (G- research (.91–.97 for 1–3 passages), but the D study found
theory), reliability is analyzed in terms of the effects of various considerable differences in the passage compositions in favor
measurement facets on estimates of performance [80]. Rather of the ED procedures in estimating the level of student
than considering a score on a performance assessment to performance.
be comprised of observed score and error, in G-theory, For Ardoin and Christ [87], passage stability was inves-
the error is further parsed into constituent components or tigated (FAIP-R, AIMSweb, and DIBELS) using four depen-
facets. For example, performance can be a function of the dent variables: intercept, weekly growth, weekly SE(b), and
task (or passage in the context of oral reading fluency), the SEE. Statistically significant differences were found “with the
occasion (conditions), or the rater (administrator). These FAIP-R passage set resulting in the least amount of error
three facets are most typically studied in G-theory; however, and the DIBELS passage set resulting in the greatest amount
the consistent finding is that raters are rarely influential of error” (p. 274). In addition, the FAIP-R passages had the
whereas tasks are nearly always influential [81]. Generally, lowest intercepts. Finally, measurement error increased with
this research involves two phases in which coefficients from fluency rates. Francis et al. [88] also studied form effects
these facets are estimated in a G-study, and then a D-study (using DIBELS passages) noting the problems with readabil-
is conducted to ascertain the number of such facets to make ity formulae and the lack of actual student performance data
a reliable decision (e.g., how many tasks, occasions, or raters to ensure comparability of passages over time. They reported
are needed?). that the passages were not substitutable (nor comparable)
While traditional research continued on aspects of ORF in the first assessment and that passages were not equally
measurement as a function of passage difficulty, [82], other likely to yield students’ median fluency. Finally, they found
perhaps more sophisticated research has employed G-theory a significant group by wave interaction: “growth trajectories
to understand the effects of facets on the reliability of per- are significantly influenced by the placement of the more
formance estimates [82–85]. For example, Hintze et al. [84] difficult stories (p. 333). Finally, and most recently, Christ
used G-theory to examine person, grade, method, occasion, [89] reported that the number of data points, quality of data,
and various interactions. In the first study, participants and and method used to estimate growth each influenced the
developmental changes explained most of the variance with reliability, validity, and precision of estimated growth” (p. 15)
a generalizability coefficient of .90 with two sets of materials with 14 weeks of intervention required for reliable and valid
and .82 with one set of materials in making intraindividual estimates of slope.
decisions. For interindividual decisions, the generalizability
coefficients were also high (.98) using only three reading 3.3. Summary and Reflection on Analytical Models for Under-
passages. In the second study, participants and grade level standing CBM. The most significant problem from this last
again explained most of the variance with very little influence study and all other (previous) analyses on reliability is that
from CBM progress monitoring procedures. Generalizability nothing about instruction is ever presented or considered.
coefficients were .80. Another study by Hintze and Christ Rather, a limited set of variables for CBM is analyzed: quality
[83] used generalizability theory to study passages (both of the data set, schedule and administration, and trend line
uncontrolled (randomly sampled) and controlled (purpo- estimation. No questions are asked such as the following, all
sively sampled)) from the reading curriculum. They found an of which would influence the reliability of score estimation.
interaction of probe by grade with controlled passages having What is the fidelity of instruction implementation? How
smaller SE(b). well does the intervention match the progress monitoring
of oral reading fluency? What grade level of measure is
3.2.2. G-Theory and Standard Error of Measurement on ORF. used to progress monitor (e.g., it is hardly ideographic if
Poncy et al. [85] documented variability of CBM scores due all students are monitored on grade level passages)? How
to student skill, passage difficulty, and unaccounted sources many data points per intervention phase are present? What
6 ISRN Education

curriculum is used during intervention and how well does that the research to follow was really spurred by three events.
this curriculum represent the progress monitoring measures? First, the National Reading Panel [93] emphasized alphabetics
How often is an intervention changed? How much time (phoneme awareness and phonics), fluency, and comprehen-
are students engaged in what kinds of activities during sion (vocabulary and text), the big five ideas behind effective
an intervention? How many students are in instructional reading, teaching, and learning. Another reason for the spur
groups? How are students identified for Tier II (e.g., what is in development of such measurement research was the No
their initial level of performance on benchmark measures)? Child Left Behind Act [94], with its emphasis on reading and
What seasonal growth is likely present (i.e., in which season mathematics in grades three to eight (and a high school year).
data points are collected, given that several studies have been With this form of high stakes testing, educators realized that
published or presented at conferences indicating that growth reaching proficiency in grade three required attention to the
on oral reading fluency is nonlinear)? What cohort effect foundation skills in earlier grades. Finally, Reading First was
might be present but is ignored (by data collection year funded immediately after NCLB, again providing attention
or nested within geography)? How many days are students to measures of early literacy. Most of the original research
present during the year? How stable is the sample of students behind CBM did not include measurement of these early
across simulation draws (when used)? How many moves reading skills; the landmark review provided by Adams [95]
across schools are present by individual students? How many was still six years out from the end of IRLD funding. As
students are with various (and specific) disabilities (or not Fuchs et al. [96] noted, “in reading, most CBM research has
with any disability)? What are the IEP goals for students with focused on the passage reading fluency task which becomes
disabilities? If students are with a disability, what is the level appropriate for most students sometime during the second
and type of instruction implementation using response to semester of first grade. Additional research is needed to
intervention (RTI) and how do special and general education examine the tenability of reading tasks that address an earlier
teachers plan instruction and interact in its delivery? phase of reading” (p. 7).
The post-IRLD research on the technical characteristics
of ORF has made great advancements and moved the field 4.1. Evidence on Skill Relations. When students are just begin-
much further along using sophisticated analytical techniques ning to read, students need to master both the graphemic and
that has uncovered increasingly nuanced issues. Unfortu- phonemic components of reading. Furthermore as reading
nately, this research has been almost entirely concentrated develops, not only are letter names and sounds important, as
on ORF and not applied to other CBMs. Furthermore, the well as their concatenation, but they form digraphs, rimes,
research has not yet been contextualized into a nomological and syllables that are the building blocks of mono- and poly-
net of practice in which important instructional and decision syllabic words. Not only are letter names, letter sounds, and
making issues have been presented. For example, all of the phoneme segmentation critical skills in developing readers,
previous questions relate equally well to reliability. Finally, in these sublexical features are interrelated, “especially letter
using a Messick frame of validity as argument with a claim sound fluency, may act as the mechanism that connects letter
supported by warrants and evidence, it appears that the major names, letter sounds, and phonemic segmentation to word
claims are instructionally free and with little attention to reading and spelling” (p. 321).
generalizability issues that underlie external and construct However, it is fluency not accuracy that serves as the
validity [90]. essential metric. “Letter name fluency and letter sound
fluency, but not phoneme segmentation fluency, uniquely
4. Early Literacy CBM predicted word reading and were stronger predictors than
their accuracy counterparts. . . Fluent recognition of letter-
A number of skills have been considered in the measurement sound associations may provide the mechanism that supports
of early reading, including alphabet knowledge of the names phonological recoding, blending, and accurate word identi-
and sounds associated with printed letters, phonological fication” [97, pag 321]. These skills are viewed as sublexical
awareness (e.g., ability to distinguish or segment words, but importantly predictive in laying the groundwork for later
syllables, or phonemes), rapid automatic naming (RAN) of word and passage reading fluency [98].
letters or digits, rapid automatic naming (RAN) of objects or This fluency in the building blocks then becomes impor-
colors, writing names or letters in isolation, and phonological tant in reading words and passages. Word recognition
memory (remembering spoken information). “Six variables includes two processes: (a) the ability to decode written words
representing early literacy skills or precursor literacy skills and (b) the ability to decode words instantly and automati-
[have] had medium to large predictive relationships with later cally, in addition to psychological and ecological components
measures of literacy development. These six variables not only [99]. Speed in word reading again is a factor in assessment
[have] correlated with later literacy as shown by data drawn of reading and has long been of interest to researchers [100]
from multiple studies with large numbers of children but and over the years consistently documented its relation with
also maintained their predictive power even when the role comprehension [101]. Generally, authors invoke the logic of
of other variables, such as IQ or socioeconomic status (SES) Laberge and Samuels [102] in that automaticity is important
were accounted for” [91, p. vii]. from an information processing view.
Even though there had been a call to arms on the need Letter sound knowledge also has been studied using
to emphasize reading in our public schools with publication nonsense word fluency (NWF). In the late 1990s, the research
of the Report of the Commission on Reading [92], it appears on curriculum-based measurement expanded to include
ISRN Education 7

measures of early reading literacy. Probably the most ubiqui- results on the DIBELS benchmark assessment scores indi-
tous measure to appear is the Dynamic Indicators of Early Lit- cated that the treatment-intensive/strategic students scored
eracy Skills (DIBELS) [103]. DIBELS has been used to explore significantly lower on the PSF and NWF subtests in the winter
the importance of phonological awareness and phonics as (pretest), when compared to the other two benchmark groups
predictors of reading difficulty [104, 105]. with or without treatment” [109, p. 23] and, in a study of
In a study by Fuchs et al. [96], higher coefficients were the influence of early phonological awareness and alphabet
found both predictively and concurrently for word identi- knowledge, Yesil-Dagli Ummuhan [110] studied 2481 ELL
fication fluency (WIF) over nonsense word fluency (NWF) students and found, on average, that the ELL students showed
with the Woodcock Reading Mastery Test-Revised. More a 38 words per minute growth in English ORF throughout the
importantly, they found that “word identification fluency first grade year and were able to read 53 words per minute
accounts for more unique variance than nonsense word at the end of the first grade. Finally and more generally,
fluency. . . and word identification fluency slope dominated phonemic awareness (PA) instruction has been found to be
nonsense word fluency slope in 3 of 4 comparisons” (pp. 16- large and statistically significant for direct measures (𝑑 =
17). For Ritchey [98], however, both measures were moder- 0.86), as well as more indirect measures of reading (.53)
ately correlated with each other and with word identifica- and spelling (.59). Importantly, PA instruction was effective
tion and word attack subtests from the Woodcock Reading for students from various backgrounds, including from low
Mastery Test. Importantly, a ROC analysis showed both socioeconomic backgrounds, at-risk and disabled readers,
measures to have relatively high positive predictive power preschoolers, kindergartners, and first graders, as well as
(in identifying students at risk). NWF was not found to be normally developing readers. Instruction was more effective
particularly useful in distinguishing students’ emerging skill when accompanied with letters, with only a few PA skills, in
in learning to blend over time, as the manner in which student small groups and over 5–18 hours [111].
respond may change. Likewise, beta weights for DIBELS
subtests in kindergarten were moderately correlated with 4.2. Growth Trajectories in Learning to Read. Sensitivity to
other published measures of reading, and only 32% to 52% instruction and reflection of progress is related to the larger
of the variance was explained for literacy constructs at the issue of growth trajectories possible (and necessary) in the
end of first grade [106, p. 349]. Furthermore, letter naming, early years of elementary school. For example, Stage et al.
nonsense word, and phoneme segmentation fluency were as [112] investigated letter sounds and its “relative predictive
highly correlated with motivation, persistence, and attitude as relationship of kindergarten letter-naming and letter-sound
they were with reading submeasures. Nevertheless, Clemens fluency to first-grade growth in ORF” (p. 227). Using four
et al. [107] administered word identification fluency (WIF), measurement occasions over first grade, they found both
letter naming fluency (LNF), phoneme segmentation fluency letter names and sounds to predict growth in oral reading
(PSF), and nonsense word fluency (NWF) as screening fluency (even above that of the first ORF occasion to predict
measures for 138 first grade students in the fall and another set later growth in ORF). They also reported that, with eight
of reading measures (TOWRE) at the end of first grade. Using or fewer letter names, 81% of the students were correctly
a ROC analysis, they reported that the measure with the classified as at risk.
greatest classification accuracy was the WIF as a significant Fluent readers in first grade have also been found to be
predictor for each of the outcome variables. Only 3-4 stu- fluent in second grade [97]. This finding is consistent with
dents per classroom were falsely classified. AUC values were other researchers who have reported that letter naming speed
shown “ranging from .862 to .909 across outcome measures, appears to be greatest in second grade (accounting for 11% of
followed by LNF (range = .821–.849), NWF (range = .793– the variance to just more than 2% in fifth grade on a reading
.843), and PSF (range = .640–.728)” (p. 238). Only modest comprehension test [99]. In fact, phonemic awareness in
improvements were made in combining the measures. kindergarten has been found to be an important predictor
Even though moderate correlations have been reported of reading and spelling in fourth and fifth grades [113]. In
between various DIBELS measures and published tests, the a study using hierarchical linear models of growth with
use of these measures for monitoring progress is really what is three levels (time, student, and classroom), Linklater et al.
critical. And some sensitivity to progress (given instruction) [114] documented growth on initial sound fluency, phoneme
has been reported. Hagan-Burke et al. [108] documented segmentation fluency, combined phoneme segmentation, and
the sensitivity of growth using the DIBELS measures of nonsense word fluency. “ISF in the beginning of kindergarten
letter naming fluency, nonsense word fluency, phoneme significantly predicted and accounted for variability on end-
segmentation fluency, and word use fluency in relation to the of-kindergarten measures of nonsense words, word iden-
Test of Word Reading Efficiency. They reported significant tification, and reading comprehension” (p. 389). However,
correlations among the measures (with the DIBELS measures considerable variation in growth was also apparent, as a
loading on a single factor) and significant influence of NWF function of time, gender, and initial language status.
in predicting the TOWRE (accounting for over 50% of the In an analysis of growth in reading from kindergarten
variance). In another study, both phoneme segmentation through third grade, Speece et al. [115] used several measures
fluency (PSF) and nonsense word fluency (NWF) subtests of oral language, syntax, listening comprehension, phono-
of the DIBELS have been documented to be sensitive to logical awareness, decoding, emergent literacy, and spelling
intensive instruction for 47 kindergarten students across in addition to several background measures (race, gender,
four classes in an urban K-5 Title I school [109]. “The SES as measured by free/reduced lunch, family literacy, and
8 ISRN Education

primary language spoken by the child). They found that “the letter sounds per minute as benchmark goals for nonsense
prediction of third grade performance and growth varied by words (with less than 30 indicating the need for intensive
type of reading skill assessed. . .only phonological awareness, instructional support). In the end, oral reading fluency may
emergent reading (TERA-2), and family literacy retained be more sensitive to monitoring progress than nonsense word
their significance as predictors of the intercept parameter in fluency [96, 120]. Nevertheless, more research is needed to
the conditional model. . .It appears that the unique linguistic document the handshake between the relations among the
roots of word-level reading at third grade are limited to the skills and the trajectory supporting normal development.
influence of phonological awareness skill” (p. 328). These
findings on the importance of fluency in the basic skills of 5. Writing CBM
reading (sublexical features) help explain a finding over 20
years ago: word decoding skills in first grade accounts for 44% Written expression curriculum-based measures (administra-
of first grade and 12% of fourth grade reading comprehension tion and scoring) were investigated in the original research
variance [116]. at the IRLD for their reliability and criterion relatedness
to other published measures [52]. The initial findings indi-
4.3. Summary and Reflection on Skill Relations in CBM. cated that correlations were moderate to high with a three-
Measurement of early literacy skills appears to be consistent minute writing time to an unfinished story starter (a prompt
with the traditional criteria of curriculum-based measures. providing the student with a scenario to which they were
These skills are straightforward in operationalizing, can directed to write a narrative) and words written and words
generate alternate forms for frequent administration, can be spelled correctly. Later, correct word sequences and correct
administered by others with a modicum of training, and letter sequences were added as metrics for scoring student
reflect important general outcomes with sufficient technical CBM outcomes [56]. In short order, a number of other
adequacy. Unlike oral reading fluency, however, the skills researchers expanded nearly all aspects of measurement:
reflect a complex constellation with a relatively brief shelf administration time, prompts, scoring procedures, and crite-
life. For example, letter naming, one of the earliest skills rion measures. This initial research also was conducted only
to develop in kindergarten, is probably not sensitive to with students in elementary schools and eventually expanded
progress across years (beyond first grade) for most students. to secondary students [121–123], who further expanded the
Likewise, in first grade, association of sounds with letters outcome indicators to both number and percentage of correct
likely has a similar short period for which it is sensitive. word sequences and correctly spelled words. These later
Other measures like initial sound fluency as well as phoneme researchers reported stronger correlations (and more obvious
segmentation, nonsense word fluency, or elision tasks are group differences) when percentages were used rather than
likely to be short lived even though they represent slightly straight counts.
more advanced skills than letter naming or sounding. And,
as Vloedgraven and Verhoeven [117] note, three problems 5.1. Administration and Scoring Systems for Elementary School
still exist with measurement of phonological awareness: Students. The initial research on written expression [51,
(a) the theoretical explanation among various measures as 52] was conducted with 82 students from five elementary
related to a larger coherent construct, (b) the inaccuracy in schools in Minneapolis and St Paul school districts in grades
measurement (and its developmental shelf life), and (c) the three to six. Like all the initial validation studies at the
difficulty in documenting growth. They reported difficulty IRLD in reading and spelling, the focus was on various
parameters that showed the easiest task to be rhyming, then parameters of a CBM, including amount of time (one to five
phoneme segmentation, followed by phoneme blending, and minutes), directions for administration and scoring, as well
finally phoneme segmentation as the most difficult task. They as collection of criterion-related evidence. A number of story
also noted differences in performance on the tasks as a starters and topic sentences were used, and students’ writing
function of grade. samples were scored for average 𝑡-unit length (essentially
Furthermore, the relation of these skills in development independent clauses), use of mature (infrequent) words, large
is quite complex. As Mark Twain wrote, “what is needed words, words spelled correctly, and total words written. The
is that each letter of the alphabet shall have a perfectly latter four measures correlated the highest with the published
definite sound, and that this sound shall never be changed or test (Test of Written Language). In a later study, the emphasis
modified without the addition of an accent, or other visible was on growth with significant linear trends reported for
sound. . .But the English alphabet is pure insanity. It can words spelled correctly, number of correct letter sequences,
hardly spell any word in the language with any degree of and total words written [35]. Finally, the reliability of various
certainty” [118, pp. 168-169]. Therefore, it is quite uncertain outcome indicators were established by Marston and Deno
the degree to which the progression of skills advances and the [54], including alternate form, test-retest, inter-judge, and
thresholds that are needed for eventual transformation into internal consistency. The only research to expand the mea-
actual reading (decoding) or fluency (reading accurately with sures from the original outcome indicators was conducted
speed). Even though proficiency levels have been identified, by Videen et al. [56] who established correct word sequences
the data supporting their application is not inviolate. For as a “valid” indicator of writing proficiency in grades three
example, L. Fuchs and D. Fuchs [119] suggested that students through six. Later research validated these original findings
need to complete 35 letter sounds per minute by the end for total words written, words spelled correctly, and correct
of kindergarten, and Good et al. [104] recommended 50 word sequences with elementary students [124]; ironically,
ISRN Education 9

the countable indices correlated more highly with ratings of length of all continuous strings of correctly spelled and
story and organization than with conventions. sequenced words, proportion of the total words written that
Nearly a decade later, the focus of research on written are legible, and the proportion of words written that are
expression continued to address various metrics for scoring correctly spelled. They reported that the percent of legible
performance [125]. In this study, total words written, correct words, correct word sequences, and mean length of correct
word sequences, words spelled correctly, correct punctuation word sequences were the strongest predictors of holistic
marks, and correct capitalizations, complete sentences, and ratings, although growth over six months was limited to total
words in complete sentences were investigated by correlating number of words written, number of legible words written,
the measures with teacher ratings (ideas, organization, voice, and number of correctly spelled words.
word choice, sentence fluency, and conventions). Although In a later refinement of this work, correct minus incorrect
inter-rater reliability of all measures was adequate (above word sequences were added as an outcome indicator with
90% on most measures and 80% on two of them), test-retest appropriate criterion-related evidence [128, 129]. In the for-
reliability was quite modest except for total words written mer study, a number of different scoring metrics were used
and words spelled correctly. At best, moderate correlations (total words, words correct, correct word sequences, number
have been reported among these CBM measures and a of characters, and number sentences written). Correlations
standardized test with the relation between teacher ratings (with published subtests in both reading and math and
and the standardized test much higher. English GPA as well as a rating of quality) were moderate (and
In a study similar to the original validation research, a higher with reading). “In sum, the results of the correlational
number of scoring systems was investigated: words written, analyses revealed that four measures—characters per word,
words spelled correctly, correct word sequences, and correct sentences, CWS, and MLCWS—had a fairly consistent and
minus incorrect word sequences along with different stimulus reliable pattern of relations with other measures of writing
tasks. The focus was correlating these measures (and tasks) proficiency. Of these four, only sentences and CWS showed
with a standardized test of writing and a state test, as well as divergent validity with correlations higher for the writing
language arts GPA. “Students completed two passage-copying measures than for the reading and mathematics measures”
tasks and responded to two picture, two narrative, and two (p. 20). Group differences were documented that reflected
expository prompts” [126, p. 554]. Moderately high correla- a sensible pattern with students with learning disabilities
tions were found among all the measures. In grade 3, none significantly below basic, regular, and enriched students. In
of the measures showed change from fall to spring, while the the latter study [129], a slightly younger group of students was
5th grade students (and to a lesser extent 7th grade students) studied (in grades six through eight), an expanded admin-
showed noticeable growth on several of the measures. istration time was used (three and five minutes), and story
Very little research has been done on sensitivity to versus descriptive writing was considered. Similar dependent
progress with writing CBMs. In one of the few such studies, variables were used as the previous study in 1999 (total words,
students were administered traditional story starters and then words correct and incorrect, correct and incorrect word
provided an intervention focusing on the writing process sequences, number of characters per word, number words per
(brainstorming and writing complete sentences) [127]. As sentence, as well as correct minus incorrect word sequences)
in most of the research on written expression, a number of along with teacher ratings (purpose, tone, and voice; main
scores were analyzed from the writing samples before and idea, details, and organization; structure mechanics and leg-
after the intervention: total words written, total punctua- ibility) and a district writing test used as criterion measures.
tion marks, correct punctuation marks, words in complete They reported the highest reliability and criterion validity for
sentences, correct word sequences, and simple sentences. correct word sequences and correct minus incorrect word
Ironically, although the number of total words written was sequences with few differences in administration time or type
the only measure to show a difference from pre-to-post of writing.
intervention, it did not correlate with a standardized test (in This change in scoring systems across grade levels
addition to simple sentences). indicates that “curriculum-based measures need to change
as students become older and more skilled” [130, p. 151].
5.2. Administration and Scoring Systems for Secondary School With students in grades four, eight, and ten, the alter-
Students. This research on written expression expanded in nate form reliability was investigated with different sample
both the populations studied and the outcome indicators duration or scoring procedures (total words, correct word
used to reflect proficiency with research by Parker and sequences, and correct minus incorrect word sequences),
Tindal. For example, percentage of words spelled correctly and criterion-related evidence was collected with a state
and percentage of correct word sequences were found to be test. They reported that “alternate-form reliability correlation
suitable indicators for students in late elementary and middle coefficients decreased with grade level” (p. 159). Although
school grades [121]. In another study with only middle school all three scoring systems correlated equally well with the
students [122], a number of production and production- state test (with a slight advantage for CWS-ICWS) and the
independent indicators were studied: total number of words three administration times (three, five, and ten minutes), they
written regardless of spelling or handwriting legibility, num- also found “a general pattern of decreasing coefficients with
ber of letter groupings recognizable as real English words, shorter sample lengths for older students” (p. 163).
number of correctly spelled words, number of adjacent and In another study using a state test as a criterion measure,
correctly spelled word pairs that make sense together, average Espin et al. [131] varied the following variables to investigate
10 ISRN Education

the impact on reliability and validity: time (3 to 10 minutes), sequences, number of correct capitalizations, and number of
scoring procedures, and criterion reference (state writing punctuation marks (both total and correct). Using the Test of
test). In addition, they investigated impact on primary Written Language (TOWL) as a criterion, they reported high
language of the students. “Although [reliability] correlations reliability for the CBMs (among 10 scorers) and moderate
were similar across scoring procedure, differences were seen correlations. Only percentage of correct word sequences,
for time-frame” (p. 181) and “the scoring procedure with the number of correct capitalizations, and number of correct
strongest reliability and validity coefficients was (CIWS)” (p. punctuation marks were significant.
182).
In a study using holistic judgments and a state test 5.3. Summary and Reflection on Administration and Scor-
as criterion measures, a number of unique scoring proce- ing Systems in Writing CBM. McMaster and Espin [136]
dures were used with high school students, including incor- reviewed this body of work including the initial IRLD
rect word sequences (ICWSs), correct punctuation marks research and the decade in which the research was initially
(CPMs), adjectives (ADJs), and adverbs (ADVs), all scored expanded by Tindal and colleagues, as well as Espin and
for each 10-minute sample [132]. They reported that “both colleagues. Researchers focused on reliability, validity, and
CPM and ICWS were moderately to strongly related to sensitivity to growth with the initial research. In their
the holistic scores. . .[but only] ICWS was moderately and research with elementary students, they addressed students
inversely correlated (𝑟 = −.51) with the norm-referenced at different skill levels, screening decisions, scoring proce-
scores” (p. 366), though considerable misidentification of dures, and beginning writers. In their coverage of secondary
learning disabilities was present using various percentile rank students, the issues were the integration of reading and
cut offs (15th to 30th PR). The number of adjectives or adverbs writing, attention to students requiring remedial and special
was not sufficiently reliable or correlated with either holistic education, a focus on screening and progress monitoring,
judgments or the state test score. consideration of scoring procedures, type task and sample
This line of research expanded to include expository duration, and, finally, predictions of performance on school-
essays and a much longer period of time to write (35 minutes based indicators. They concluded by noting the need to
versus the traditional three to five minutes). In this last study address reliability in progress monitoring and attention
with a relatively small sample size (only 22 middle school to generalizability and consequential validity of measures.
students), “criterion variables in the study were the number “Finally, if measures are used by teachers to monitor progress
and make instructional decisions, it is necessary to demon-
of functional essay elements and quality ratings of the essays”
strate that student performance improves as a result” (p. 82).
[133, p. 210]. They reported moderately strong correlations
For the past 30 years, the research on written expression
among correct and correct-incorrect word sequences and has addressed a number of test administration variables.
both functional elements as well as quality ratings. Interest- Clearly, the writing process needs to address the stimulus
ingly, the correlations decreased only somewhat when only prompt for soliciting student writing. In addition, the amount
the first 50 words of the essays were scored (particularly of time to write is a critical variable. Finally, the system used to
with the quality rating and for the students with learning generate a score needs to be carefully considered. Obviously,
disabilities). Finally, growth over time was documented from all three of these variables need to be studied in terms of
pre- to posttest. reliability and validity, particularly criterion-related evidence
To use criterion measures that teachers would find more to be both predictive of performance on other measures
credible than standardized tests or state tests, Fewster and but perhaps more importantly to be credible for teachers
Macmillan [134] studied student performance from late (and therefore be predictive of their own judgments obtained
elementary school through high school and correlated it with through grades or holistic ratings). Unlike all other areas in
teacher grades. Students were given three (annual) adminis- CBM, these test administration variables have been studied,
trations of both a reading measure (oral reading fluency) and often with similar results.
a three-minute story starter from which to write a composi- In summary, by the late 2000s, a sufficient number of
tion (scored by counting the total number of words written studies had been conducted in written expression CBM to
(TWW) and the number of words spelled correctly (WSC)). indicate that (d) scoring written products needed to be
different for middle school students (from the previous use
In correlating these measures with teacher grades (in English
of words written and words written correct with elemen-
and social studies), they reported that “in nearly every case,
tary students, (c) percentages of words written correctly
WRC [words read correctly] correlated more highly with (WWC) and correct word sequences (CWS), as well as
course grades than WSC (p. 152)” with most in the moderate correct minus incorrect word sequences (CIWS) were more
range and accounting for 15% to 21% of the variance. sensitive metrics (in terms of criterion related evidence), (b)
Finally, in the most recent study with 447 eighth grade amount of writing time influenced reliability but not validity,
students, Amato and Watkins [135] studied the predictive particularly for older students, and (a) genre for writing was
validity of traditional CBMs in writing. A number of metrics not an influential variable (on either reliability or validity).
were used: total words written, words spelled correctly (both Furthermore, the various measures of written expression
count and percent, correct word sequences (both count appeared to be correlated with published writing tests, state
and percent), correct-incorrect word sequences, number of tests, and teacher grades.
ISRN Education 11

6. Mathematics CBM However, given that even computation skills require

even more fundamental skills without the use of symbols,
As of 2008, most of the research on CBM has been in reading, considerable research also has been conducted on these skills.
not mathematics. In fact, “an exhaustive literature search on For example, Clarke et al. [148] included four types of tasks in
the development, uses, and implementation of CBM yielded their research: (a) oral counting, (b) number identification,
578 reports, dissertations, chapters, and articles (Foegan et al., (c) quantity discrimination, and (d) missing numbers. In
[137]). A breakdown by subject area showed that only 32 were addition to analyzing growth, two published standardized
conducted in the area of mathematics. Additionally, of the 32 tests were used as criterion measures. Although they did not
studies exploring CBM in the area of mathematics, 28 were find linear growth, “correlations between the experimental
conducted at the early mathematics and elementary levels, measures in the fall and criterion measures in the spring were
whereas only were 4 implemented at the middle school level” in the moderate range” (p. 10). Likewise, Chard et al. [149]
[138, p. 236]. used five tasks: counting 1–20, from 3 and 6 and by 2 s, 5 s,
The issue of domain sampling is critical in mathematics and 10 s, identifying numbers (1–20), writing numbers (1–
given the somewhat qualitative shifts in content represented 100), discriminating quantity (1–20), and identifying missing
in this subject area over the grades. That is, students in numbers (1–20). Lembke et al. [150] had students differentiate
elementary schools learn addition, subtraction, multiplica- quantity (name larger of 2 numbers), name missing numbers,
and name randomly ordered numbers. Martinez et al. [151]
tion, and division in a relatively straightforward manner
had students count aloud from 1, name or identify digits from
but often without considering the conceptual nature of the
1 to 20, discriminate the larger of 2 numbers, and identify the
number system. This issue is even more pronounced in
missing number in a sequence.
middle schools with introduction of geometry, algebra, and Preschool and kindergarten CBM probes also have in-
statistics/probability. As Fuchs notes, “CBM differs from cluded choosing a number (20 items each with four options),
other forms of classroom-based assessment that rely on counting objects (of specific pictured things with up to 20
mastery measurement. With mastery measurement, teachers pictures), counting (from 1 to the highest number possible up
assess mastery on a problem type, such as adding 2-digit to 30), and visually discriminating among four unique objects
numbers with regrouping and after mastery is demonstrated, (with 30 such trials). All of these tasks have sufficient reliabil-
move on to the next skill. By contrast, with CBM, every skill ity. Three kindergarten probes used for criterion-referenced
in the annual curriculum is represented on each weekly test. evidence included counting circles, counting letters (sounded
So, like a norm-referenced test, CBM samples a broad range out), and discriminating the unique item from four pictured
of skills, and each repeated measurement is an alternate form [152].
of equivalent difficulty, assessing the same constructs” [139, Some of the research includes both computational and
p. 34]. As noted earlier, a mastery monitoring approach is requisite skills. For example, Fuchs et al. [144], first grade
confined to assessing skills from a limited and instruction- measures, included fact retrieval (25 addition fact problems
ally delivered domain. In mathematics CBM research, this with answers from 0 to 12 as well as 25 subtraction fact prob-
approach has been labeled as a “subskill mastery measure lems with answers from 0 to 1), computation (problems with
(SMM)” versus a “general outcomes measure (GOM)” [140]. two single-digit numbers requiring adding or subtracting),
Another labeling system used by Foegen et al. [141] refers problems with three single-digit numbers requiring adding,
to the domain as “curriculum samples” (from skills taught) and problems with two double-digit numbers (requiring
or “robust indicators” (reflecting core competence with high adding or subtracting without regrouping), number identi-
likelihood of being related to important curricula). In all of fication and counting (four-item tests that present students
these labels, the key issue is the number of skills represented with number sequences, the last two of which are shown
in the domains (and often hidden in the GOM perspective). as blanks, for example, 4, 5, 6, . . .), and finally concepts
and application (25 items including numeration, concepts,
6.1. Domains Sampled in Elementary Schools. In mathematics geometry, measurement, applied computation, charts and
research, both the domain sampling used in specific studies graphs, and word problems). Another example of broad
and the study purposes have been extremely varied, making domain sampling includes Leh et al. [153]. On occasion, the
it difficult to bring consistency to the outcomes. For the most tasks vary whether the problem is fully presented (and the
part, this body of research is similar in many ways to that student is directed to supply the answer) or the problem is
done with any of the other CBM subject areas. A typical incompletely rendered but with a solution (and the student
range of skills includes computation, applications, and (real must create the conditions for the solution shown).
life) problem solving [139]. Given the focus of elementary Sensitivity to instructional programs has been studied,
school curriculum on basic operations, it is not surprising often on the efficacy of tutoring. For example, on first
that computation has been the focus of considerable CBM grade math (math concepts, procedural calculations, word
research with students in grades 1–5 [142–147]. In a study by problems, and math facts), an effect size of .40 has been
Burns et al. [142], single and multiple skills were compared reported with follow-up research addressing math fluency in
with students in grades three to five. They reported sufficient first grade and word problems in third grade that showed an
reliability and moderate predictive validity for older grade “effect size compared to the control condition was 0.27 for
band students using three levels of placement (frustration, math facts tutoring but almost double that for word-problem
instructional, and mastery). tutoring (0.53)” [154, p. 261]. A final study showed the effects
12 ISRN Education

of tutoring with “validated instruction” resulting in a very structured on identifying reliable and valid measures for
large ES of 1.34. screening and progress monitoring, using progress monitor-
ing for students with special needs, analyzing specific skills,
6.2. Domains Sampled in Middle Schools. Ann Foegen has and using the measures as part of class wide peer tutoring and
been the most dominant researcher in middle school math- consulting with teaching.
ematics and has addressed a number of problem types. Christ et al. [140] review computation CBMs and use
Foegen’s [155] early work on middle school math focused on Messick’s validity framework [1] relating content and criteri-
both a fact probe and an estimation probe (both computation on-related evidence while also noting the lack of research
only and as word problems). Using teacher ratings and a addressing consequential validity. In particular, two domain
standardized achievement test as criterion measures, she
samples are considered: sub-skill mastery measure and gen-
reported sufficient levels of criterion-referenced correlations
eral outcomes measures as well as stimulus sampling (cur-
over 10 weeks. She also noted that reliability increases when
riculum and robust indicators); clearly the two features are
successive probes were aggregated but slope of improve-
ment remained moderate. She extended this work [137] by related. Most reliability studies address internal consistency
addressing low achieving students, administering a basic and alternate forms; no research is presented on test-retest
math operations task, a basic estimation task (both com- reliability. A key issue is the relative stability of single skill
putational and word problems), a modified estimation task measures versus the greater variance associated with multiple
(with changes in the number types), along with ratings skill measures, particularly when randomly placed on the
from teachers, grades from students, and performance on a measure. Finally, duration of administration is summarized.
standardized test. Multiple types of reliability were computed “A single brief assessment of 1 to 2 min is probably sufficient
(and found to be adequate); in addition, criterion-related to assess most single skills. Longer and more numerous
evidence (correlations) among the CBMs and teacher ratings assessments are necessary to assess multiple skills” (p. 203).
were moderate. For Fuchs et al. [144], the focus of their review of four
Finally, Foegen [156] used The Monitoring Basic Skills studies was on the capability of a response to intention (RTI)
Progress (MBSP) measures for computation [157] and for system in mathematics to reduce the need for intensive, long-
concepts and applications [158] (reflecting skills and concepts term service. The studies address first grade tutoring, math
from the Tennessee state mathematics curriculum); in addi- facts fluency in first grade, math facts tutoring in third grade,
tion, she used basic facts, estimation, complex quantity dis- and finally tutoring with word problems. Although they
crimination, and missing number. She reported moderately describe significant effects from the first grade tutoring (effect
high levels of alternate form reliability (.80–.91) in grades 6– size of .40), they also noted the low relative performance level
8 (fall, winter, and spring) and test-retest reliability (.70–.85). of the students (compared to students not at risk); therefore,
Using teacher ratings and a standardized achievement test as another study was implemented with an emphasis on tutoring
criterion measures, she reported sufficient levels of criterion- for fluency. In the third study, transfer from math facts to
referenced correlations and moderate rates of growth. word problems was the focus for third grade student with
effect sizes reported for math facts of .27 and word problems
6.3. Summary and Reflection of Mathematics Domains with .53, suggesting lack of transfer. Finally, in this third study, the
CBM. Four reviews have been published on mathematics interaction of tutoring with validated classroom instruction
CBM. In this chronological summary, the highlights of these was the focus that was found to be significantly more effective
reviews are presented. than when implemented with teacher-designed instruction.
Foegen et al. [141] comprehensively review two approach- Following the initial research on reading and writing at
es to developing CBMs in math: curriculum sampling versus the University of Minnesota, there has been a very significant
robust indicators. Rather than a mastery approach, cur- increase in the research on mathematics. Quite likely, the fed-
riculum sampling reflects items from within a year “given eral legislation (NCLB) that began in 2001 had an influence
the high level of curriculum specificity in mathematics” as the two areas in which districts were held accountable
(p. 122); in contrast, robust indicators reflect problems that was reading and mathematics. And, like the early literacy
“identify aspects of core competence in mathematics [that] skills, waiting until grade 3 made little sense so much of this
enables students’ growth to be modeled over multiple years research spanned the entire grade range of K-8. Also, like
of learning” (p. 122). This review analyzes results from 32 the early literacy skills research, mathematics is comprised of
reports on progress monitoring summarizing reliability and multiple domains that eventually need to be stitched together.
validity research (stage 1) as well as student growth (stage 2).
However, an important difference is that these domains
Results are summarized for early, elementary, and secondary
successively build upon each other. Whereas learning to
mathematics. The seven studies on using data to improve
student achievement have all been conducted by Fuchs and read involves generalizations with many exceptions to the
colleagues with six of them focused on computation. rules (of sounds and syllabication), learning to compute
Lembke and Stecker [159] summarize procedural issues and solve mathematics problems is lawful with successive
in implementing math CBMs, including steps in data-based algorithms used. For example, an early skill in counting
decision making (screening and progress monitoring), met- (using objects and finger counting) turns into addition and
rics for summarizing performance (correct digits, problems, then multiplication. A number of principles are applied in
or responses), and establishing goals. The review is then solving problems so that students can learn more general case
ISRN Education 13

strategies (e.g., distributive and associative principles). There- demands made on students and is likely part of measurement
fore, the field of CBM research addressed domains that had to systems that are “curriculum-based.”
be integrated into long-range goal measurement, a feat more Initially, Espin’s focus was on reading aloud from content
similar to the development of general outcome measurement area text books [163]. In this study, 124 students in grade
in secondary settings that is addressed next. That is, to avoid 10 were administered a background knowledge vocabulary
a mastery-monitoring model (with domains of assessment test (matching a word with its definition with 10 words
tightly associated with instruction), a more general case sampled from the content text), reading passages (both pre
sampling was developed, integrating a range of skills. and post), and a classroom study task (with a multiple choice
test that included literal and inferential questions) in English
7. Secondary Content CBM and Science. They reported high reliability (both alternate
form and test-retest) as well as a pattern of performance for
Tindal and Espin have contributed the most research for most students with disabilities (5 of 7) that was consistent
CBM in secondary settings, with the essential difference with a general deficit in both content areas (below the
between them being the source for ensuring measurement 40th percentile rank) with only 28 of 102 general education
comparability and applicability to content area instruction. students in this group; no differences were found in special
A comprehensive comparison of these two approaches is education comparing general deficit versus content-specific
published by Espin and Tindal [160] addressing critical issues deficit performance. However, when comparing these two
such as the curriculum in secondary classrooms, the skills groups following a period of study in the text, the two groups
were different (with higher performance by content-specific
to be assessed in CBM, research in reading, writing, and
deficit students in the number of words read correctly per
mathematics, and finally a comparison of the two approaches.
minute). Finally, they reported differences between these two
In general, Espin has addressed vocabulary as the key to
groups on background knowledge.
understanding measurement in secondary settings while The study was completed by Espin and Deno [164] with
Tindal has addressed content concepts. 10th grade students (𝑛 = 121) who were given English and
Science texts to read (for one minute) and study (for 30
7.1. Vocabulary Assessment in Secondary Schools. Much of
minutes in preparation for completing a 25-item multiple
Espin’s research laid the groundwork for this area (correlating
choice test). They used grade point average and performance
reading aloud from content text, completing a maze test, and on a standardized achievement test as criterion measures.
matching vocabulary words with definitions); consistently, They reported reliable and positive correlations between
low moderate correlations have been reported among these reading measures and all performance measures (in the range
measures, often highlighting the importance of vocabulary. of 40 s to 70 s), accounting for 11% to 28% of the variances. The
Quirk [161] notes that words in secondary settings are correlations with reading fluency and study questions and
Graeco-Latin in origin and distinguishes them from English published achievement tests, as well as grade point average,
words having an Anglo-Saxon origin, which are typically were about the same value.
short (monosyllabic), learned early in life, and form the basis In a similar study, Espin and Deno [165] measured 120
of everyday discourse. When students move from elementary tenth grade students with four tasks in both English and
schools to middle schools, the nature of reading changes Science: vocabulary matching (10 items from the study pas-
drastically. The words of secondary content are typically mul- sage), prestudy reading aloud, content area study task (900-
tisyllabic, sound foreign, contain morphophonemic struc- word passages with 25 multiple choice (literal or inferential)
tures that are only somewhat transparent, and are learned questions to answer), and poststudy reading aloud (which
later in life [162]. They comprise the primary language of were not analyzed). They reported moderate correlations
the secondary content classroom and pose a more serious among the measures. Text-based reading added 13% of the
problem for students at risk of failure, “the component parts variance to performance on the content area study task with
of Graeco-Latin words no longer carry much meaning for vocabulary subsequently adding an additional 9% of the
most uses of the words in English. The words have lost variance; this finding was true for both English and Science
their semantic transparency” [162, p. 689]. Consequently, (though with slightly more positive results for English).
they typically are abstract, do not appear frequently in the When vocabulary was entered into the regression analysis
English language, and are not easy to understand. “When first and subsequent analyses included text-based reading,
these features combine in words, they interfere with word use no additional variance was accounted. Separate analyses for
and with word learning” [162, p. 696], and as a result, meaning students from both ends of the distribution showed the test-
is derived through whole word search which is slower. Fur- based reading passages to function as well as vocabulary
thermore, in secondary settings, the lens for learning is not measures. In a follow-up study, they also reported that the
spoken language or conversation. “Academic Graeco-Latin vocabulary test was slightly higher in its correlation with a
words are mainly literary in their use. Most native speakers content study task (and accounted for more of the variance)
of English begin to encounter these words in quantity in their than oral reading of the content text (counting number of
upper primary school reading and in the formal secondary words correct per minute).
school setting. So, the word’s introduction in literature or Espin and Foegen [166] added a maze test to the reading
textbooks, rather than conversation, restricts people’s access aloud from text and vocabulary test as secondary measures;
to them” [162, p. 677]. This change has significance in the in this study, 184 students in grades six to eight participated
14 ISRN Education

by taking these three measures and then taking three criterion the Direct Instruction approach in which the focus is on the
measures: a 10-item multiple choice comprehension test, daily juxtaposition of examples and nonexamples as the basis for
tests (from five timed reading passages with 10 multiple defining concepts [173].
choice questions), and a posttest comprised of a compilation The reason for taking this route to curriculum-based
of multiple choice questions from the daily tests. The results measurement in secondary content areas is threefold. First,
showed moderately high correlations among the three CBM content (information) becomes much more subject specific
tasks and the three criterion measures though vocabulary was in middle and high school grades and less generalizable
the best predictor measures. Again, in a series of stepwise outside of the subject area. Second, this content specificity is
regression analyses, vocabulary appeared to contribute the likely to be noteworthy even within a subject area, favoring a
most unique variance (when it was added first, no more mastery monitoring approach rather than a general outcome
variance was accounted for by reading aloud or maze). measurement (GOM) methodology. Third, students with
Espin et al. [167] in a later study that directly addressed disabilities or those at risk are likely to have difficulty reading
the vocabulary measure as a generalized indicator of perfor- in addition to learning subject matter content, requiring a
mance in content area; in this study, text from the classroom measurement system to do basically double duty, in serving
of instruction was used. They administered a number of as a barometer of reading AND content learning. The subject
measures to 58 seventh grade students (five of whom were specificity (both across and within content areas), however,
receiving special education services); a number of vocabulary is the most troubling issue to address in generating a GOM
matching measures (with 49 terms and definitions used in approach. An artifact of this issue is that content information
each area of sociology, psychology, and geography), both stu- becomes difficult to traverse; every new content area is new
dent read and administrator read, were administered along with little opportunity to preview and review of information.
with a standardized achievement test and a classroom knowl-
edge test (36 items using factual and applied knowledge). 7.2.1. General Description of CBA. In the concept-based as-
Alternate form reliability was moderate (but higher when sessment (CBA), three to four primary attributes per concept
multiple probes were aggregated); the correlation was mod- are identified in a content area from content experts who
erately high between the vocabulary probes (particularly the use facts in the curriculum to inductively derive them. “A
administrator read measures) and the published test as well as concept-based focus provides the teacher with a template for
the classroom post-test, though moderate-low with grades. specifying the domain-specific conceptual knowledge with
The study conducted by Espin et al. [168] continued corresponding explicitly identified attributes and examples
this research on vocabulary matching in social studies with and nonexamples” [174, p. 335]. Attributes provide critical
more focus on progress and administration format (student rules for organizing the domain of facts with generalizability
versus administrator). Using the same population as used to novel content in the same subject area. For example,
earlier by Espin et al. [167], they analyzed the data using Twyman and Tindal [175] define “nationalism as a concept
hierarchical linear modeling to compare growth for the two [with] three attributes as: (a) has an ideology based on
administration formats. They found more growth in correct devotion and loyalty to a nation, (b) primary emphasis on
matches with student (.65) than administrator (.22) read promotion of its culture and interests, and (c) identifies
tests. In a comparison of three criterion measures as level with the power of historical tradition to oppose those of
two predictors, they found all to contribute significantly to other nations or supranational groups” (p. 5). As another
vocabulary knowledge. illustration of a major concept in social studies, Twyman et
In the most recent study, [169] presented the results al. [176] use four critical attributes to characterize civilization:
from two studies. In the first study, two measures were (a) having religious beliefs, (b) forming social groups, (c)
administered to 236 students in grade eight (a read aloud generating support systems and activities, and (d) employing
task administered with timing noted at one, two, and three writing as a primary form of communication. In both
minutes and a maze task with timing noted at two, three, illustrations, the concepts and their corresponding attributes
and four minutes); both tasks were based on human interest can be applied to novel content with new examples.
stories in the newspaper. Performance on these tasks was then
correlated with a state test. Alternate form reliability was quite 7.2.2. Critical Effects from CBA. One of the important side
high and both correlated with the state test. effects from this approach is the immediate reduction in com-
plexity of content information with clear explication of crit-
7.2. Concept-Based Assessment in Secondary Schools. As a ical information. Furthermore, the concept-based approach
basis for framing a concept-based approach, Nolet, and leads to assessments that take advantage of constructed
Tindal [170] divide content into three knowledge forms [171]: responses with explicit guidelines for a partial credit-scoring
facts (single dimensional declarative statements), concepts algorithm [177]. For example, a student’s response can be
(with a label, attributes, and both examples and nonexam- scored for framing an argument, as well as providing both
ples), and principles (if-then relations that present cause- a rational and details that support the rationale with up to
effect relations). With concepts highlighted as the key knowl- seven points possible [174]; alternatively, a multistep flow
edge form, attributes and subsequently examples and nonex- chart, with binary decisions, can provide a score of 0 to 4
amples are then articulated for each concept [172]. Attributes points [172, 175]. In either partial credit system, reliability
are essential for ensuring comparability of alternate forms, is enhanced by using an explicit scoring guide [178]. A
a critical feature of CBM. This approach is in contrast to concept-based approach also provides an explicit link to
ISRN Education 15

instruction through the use of semantic maps [179] or the a number of training manuals are available in middle school
verbal interactions between teachers and students during science [186], high school science [187], mathematics [188],
instruction [180, 181]. Furthermore, the approach provides a language arts [189], and social science [187].
mechanism for special education teachers to collaborate with
general education content teachers in an effective manner 7.3. Summary and Reflection of Access and Target Skills for
[182]. Finally, both teachers and researchers can begin to CBM. Assessment in secondary settings did not develop
integrate curriculum content, instructional delivery, student until well after the initial research at the University of Min-
perception, and formative test performance that provides nesota IRLD. And the research that did eventually emerge was
essential feedback [183]. The approach is particularly useful considerably different. No longer looking at measurement
in linking assessment with instruction and is universal across of access skills, the focus directly targeted the curriculum
content areas with application in the social and physical content. For both researchers and programs of research,
sciences [181]. this target reflected a challenge to a GOM perspective as
well as other principles underlying the initial CBM research
7.2.3. Empirical Outcomes from CBA. The approach has platform. To maintain a time series (and long-range goal
empirical support from several studies. For example, Twy- sampling) design, target content had to be approached from a
man et al. [176] showed that by anchoring both instruction broad perspective that nevertheless maintained relevance for
and assessment on concepts, students form a deeper teachers to use in evaluating instructional programs. Espin
understanding of content. Although students using a concept and colleagues used vocabulary (knowledge) as their content
basis showed no difference on tests of facts (declarative medium while Tindal and colleagues addressed concepts
knowledge), they were significantly better on tests of vocab- (attributes and examples-nonexamples). Both approaches
ulary and problem solving. This approach has been used in a distilled measurement to a command of academic words
case study with a seventh grade student for whom English was and helped remove reading as a necessary access skill which
his second language [176]. In a unit on four Mesoamerican would have presented a tautology. Students could not reach
civilizations with a cloze technique using examples and proficiency on content unless they learned to read, and
nonexamples of religion, social groups, support systems and they would not be able to read (content) unless they knew
activities, and writing, a control group of students taught in the academic words. However, neither provided quite the
a traditional factual approach performed more poorly on a elegance attained with other features of CBM (like easy to
writing task than students in the experimental CBI group. create and implement as well as administer and score).
In a similar study on history, Twyman et al. [174] used three
types of measures (fact test, vocabulary knowledge, and 8. Analysis of Current Practice and a Proposal
problem-solving essay) to compare two groups of students for an Alternative Conception
taught with CBI and traditionally (as factual information)
over 21 days (with each group for 46 minutes). Although The most dominant use of CBM has been for screening and
both groups improved significantly on the facts test, students progress monitoring. Even before response-to-intervention
in CBI improved differentially more than the traditional (RTI) was common parlance among educators, CBM offered
group for the other two measures. a system for monitoring students attainment of goals on their
IEPs and evaluating instructional programs. As noted in the
7.2.4. Research to Practice. Two training modules provide early history, the measures were validated and used to make a
an overview of concept-based instruction and assessment in variety of decisions. Of late, however, their use in a formal
middle-high schools (Tindal et al. [184] and Nolet et al. [185]). system such as RTI has become widespread among school
In the former, a complete description is provided for under- districts in the United States.
standing content information in terms of knowledge forms The National Research Center on Learning Disabili-
(facts, concepts, and principles) as well as designing instruc- ties (NRCLD) defines RTI as “student-centered assessment
tion for higher order intellectual operations that addresses models that use problem-solving and research-based meth-
curriculum analysis, instructional planning, and interactive ods to identify and address learning difficulties in chil-
teaching. An appendix is included with black line masters dren [through] high-quality classroom instruction, universal
for various semantic maps to organize concepts. The latter screening, continuous progress monitoring, research-based
publication provides companion training in development of interventions, and fidelity of instructional interventions”
assessments for understanding students’ ability to reiterate, [190, p. 86]. RTI depends upon the coordinate relation
summarize, illustrate, evaluate, predict, and explain. In addi- between implementation of interventions and change on
tion, a number of important assessment issues are addressed measures over time so that the interventions are either
including sampling planning and indexing reliability, both of vindicated or modified for students with disabilities or who
which are necessary for development of assessment problem- are at risk of failure. For Wanzek and Cavanaugh [191], fully
solving tasks. An appendix provides black line masters for developed RTI models also integrate general and special
a decision-logic scoring guide using constructed problem- education to identify and integrate school resources in pro-
solving essays. These initial training materials were later viding effective instruction and intervention. Furthermore,
adapted with empirical references and applied to a consul- scientifically based evidence is used to develop effective
tation [182] and learning assessment system [172] for special instruction including the intensity of instruction (time, fre-
education teachers in middle schools settings. In addition, quency, duration, and instructional group size).
16 ISRN Education

8.1. Empirical Results and Needed Research. Systemic research on the effects of instructional modifications, they found
on RTI is more conceptual than actual. For example, Barnett that teachers with the self-monitoring added to CBM made
et al. [192] conceptualized technical adequacy of RTI from more significant modification and had significantly greater
Messick’s [193] evidential (efficacy and effectiveness) and improvement in their students’ performance (pre-post) in
consequential perspective. They present a technical adequacy math than teachers with CBM alone or no CBM.
checklist for each of three instructional tiers “emphasizing Within 20 years, the research on data-based decision
technical checks and iterative use, via a fictitious example” (p. making was sufficient to arrive at some generalizations (with-
26). More recently, VanDerHeyden [194] added the following out and with achievement results) [200]. In particular, five
to considerations of technical adequacy of RTI implemen- features appeared critical for the use of CBM for elementary
tation in which “assessments must be technically adequate or middle school students with mild to moderate disabilities.
for the purpose for which it is used [and] decision rules
must be correctly applied with correct actions occurring in (1) Progress monitoring alone is insufficient; rather,
sequence” (p. 336). In the end, her focus was on sensitivity instruction should be tailored to student needs.
and specificity of decisions (and subsequent errors) using a (2) Both types of changes are needed: (a) raising goals
receiver operating curve (ROC) analysis. when progress is higher than expected and (b) chang-
The problem with both previous analyses (and a number ing instruction when progress is less than expected.
of similar publications) is not the researchers’ perspectives
(3) Decision making is facilitated (made more efficient
but the need to follow up on their perspectives. Or, as
and with greater satisfaction) with computerized data
Gersten et al. [195] note, “fully determining the validity of an
collection, storage, management, and analysis.
assessment process transcends what any one researcher can
accomplish. It is a task for a community of researchers and (4) Student strengths and weaknesses can be identified
practitioners to consider meanings and utility of assessment with skills analysis in conjunction with consultation.
procedures in relation to current thinking about how to (5) Meaningful programmatic changes can be facilitated
improve instructional practice and issues raised by studies of with consultation.
implementation” (p. 512).
This research on teacher skills in the use of CBM also
8.1.1. What We Know. Two factors appear particularly influ- is based on the use of various decision rules for identifying
ential in how well teachers implement CBM: adequacy of risk. For example, four dual discrepancy (DD) cutoffs were
planning time and teachers’ degrees of personal and/or compared on DIBELS (winter and spring) in which student
teaching efficacy [196]. Even though five indices of imple- growth (in first grade) was set below the 25th percentile, at
mentation were examined (number of measurement points, the 33rd percentile, at the 50th percentile, and less than one
ambitiousness of the goal set for the student, number of standard deviation below the mean [201]. On an end of year
times student goal was raised, number of times teachers made reading measure, the percentile rank measures consistently
instructional changes, and timing of changes), she reported (and increasingly by rank) reflected reading risk; “both the
that “teachers with high personal efficacy and high teaching 25th and 33rd percentile criteria moderately differentiated
efficacy increased the end-of-year goal for students more reading skills of responsive and non-responsive students
often than their counterparts with low degrees of efficacy; who were at risk, but the 50th percentile and one standard
teachers with high teaching efficacy set goals that were overall deviation criteria both led to small effect sizes” (p. 402). It
more ambitious than those of teachers with low teaching appears that teachers need guidance to ensure that students
efficacy” (p. 5). In a related and later study, Allinder [197] also with the greatest need are identified.
reported that when teachers implemented CBM more accu- And this aspect is important at the systems level. As
rately, their students made significantly greater math gains Mellard et al. [202] determined, in one of the few studies
than the students with teachers who either implemented done on systems use of RTI. They addressed four compo-
CBM less accurately or who did not use CBM. nents as “a framework that includes (a) universal screening,
This research on teacher use of CBM is based, in part, (b) tiered levels of high-quality interventions, (c) progress
on measurement of, and intervention on, teacher behavior. monitoring, and (d) data-based curricular decisions” (p. 186).
For example, Allinder and BeckBest [198] used an accuracy of In articulating the actual practice of RTI, they concluded
implementation scale to investigate teachers’ CBM structure that schools were screening students in various ways, using
(measuring baseline, graphing data, writing goals, and draw- norms or percentage of population as cut points for risk
ing goal lines), measurement (test administration, scoring, assessment, placing students into instructional tiers (in vary-
and frequency of measurement), and evaluation (describing ing proportions), and monitoring progress in tiers two and
instructional strategies and the changes to them both in three. Finally, they noted that “good recordkeeping systems
terms of content and timing) with each item rated on a 5- was a recurring theme” (p. 192). In addition, school personnel
point scale. Their intervention included an initial training and heeded the need to make screening and progress monitoring
then either university-based consultation or self-monitoring. results accessible and to share data from year to year.
Though significant gains in math achievement were made, no Ironically, none of this research on data-based decision
differences were found between the two treatment conditions. making included training on single subject methodology.
Again, in a related and later study, Allinder et al. [199] studied In fact, during the first 20 years of CBM, only one study
self-monitoring with CBM versus CBM alone. In their focus investigated the effects of decision making using a single
ISRN Education 17

subject design [203]. These researchers reported that, when organized into tiers, time, and groups, as well as specific
teachers are trained in a single subject design, they can obtain instructional emphases being used (along with curriculum
significant results that transfer well beyond the content of materials and strategies), and how well decisions have been
instruction. This same argument appears in consideration made with subsequent changes in level, variability, or slope.
of the curriculum in CBM. In terms of content-related To get teachers with these resources and information, training
evidence and the relation between instructional planning, the needs to be based on student reports that can be accessed to
curriculum has not been found to be a critical feature that evaluate not only the effectiveness of interventions but also
identifies CBM as an effective tool in monitoring progress the systematic use of data-based decision making. Training
[204]. That is, the use of CBM does not appear to be dictated needs to focus not only on how to implement best practice but
by the relation of measurement content with instruction. The also how to interpret information on student performance
studies Fuchs and Deno reference were from some of the
and progress. Most importantly, professional development is
early research on curriculum differences conducted at the
needed in how to use the information in a systematic manner.
IRLD in the 1980s. They also reference within curriculum
differences in sampling passages being as great as between 8.2. A Nomological Net Supporting a Theory of Change. To
curriculum differences and the lack of (or potential for) support this “next generation” of training on data use, future
generalization. Rather than limiting sampling plans to the
research first needs to be based on a nomological net that
curriculum, they suggest three other criteria as instrumental
references three warrants.
for instructional utility: (a) comparable difficulty, (b) valid
outcome indicators, and (c) qualitative feedback. Therefore, Assumption 1 (measurement sufficiency). Students are ap-
in developing teachers skills for using CBM, these important propriately placed in long-range goal material to ensure the
contexts should be considered so that teachers know how measures are sensitive to change. What is the type of measure,
to generalize from measurement results to more broadly grade level of measure, and the time interval (density of meas-
considered instructional changes. ures) used during the year?
8.1.2. What We Do Not Know. Presently, few researchers Assumption 2 (instructional adequacy). Instruction is
have directly investigated training systems. For example, detailed and explicit, allowing a team of teachers to coordi-
in a training manual for reading decision making, such nate various elements such as providing an instructional tier
topics include benchmark measures, expected growth rates, (1–3), allocating sufficient time to teach, grouping students
goal establishment, decision rules, instructional strategies, appropriately, deciding on the instructional emphasis using
effective reading instruction, and exemplar case studies [205]. specific curriculum (core and supplemental) materials, and
Yet, no information exists on the effectiveness of this training. determining what instructional strategies to use.
And, as noted by Barnett et al. [192, p. 20], implementation of
RTI requires practitioners to consider prevention, screening, Assumption 3 (decision making). Interventions need to be
and successive instructional programs that are scientifically introduced for low performing students when data warrant
based and changed using decision rules. It is this combination change. Are interventions provided at the right time and in
of multiple variables that needs to be investigated concur- accordance with specific data features (e.g., level, variability,
rently and at the systems level. and slope)?
In summary, more research and development is needed
for training teachers on systematic use of data, consideration The theory of change can be viewed as the interlocking
of goals, skills analysis, and data management systems; it union of the three components in a chain: measurement
also should include use of single subject designs in practice. sufficiency, instructional adequacy, and decision making. See
Yet, most CBM training systems have little data on the Figure 1. It is not each link itself that is critical, but it is the
effectiveness of the training, even though they are premised intersection of the link with subsequent links that is critical
upon the collection of student performance and progress. (the time series is important in making causal arguments in
Another problem is that data are collected only at the which time proceeds from left to right). As teachers collect
individual student level and not on teachers. Therefore, data (from benchmark to decisions of risk and monitoring
professional development cannot be tailored to the areas in of progress), the data used to inform them needs to be
which teachers need the most assistance, whether it is about sufficient, directed toward instruction, and adjusted as
how to effectively progress monitor students, how to develop needed. Furthermore, this information needs to be collected
effective instructional programs, or how to make decisions into a single database for teachers to monitor their application
for maintaining or changing these programs. of RTI as well as policy makers and administrators to use
Next generation training needs to allow information the information in making system decisions. The theory
to be used from teachers and students through relational is driven by accessibility as a key ingredient to change. If
databases. Teachers need systematic models of assessment information is not easily accessible and tractable, then it is
practices with classroom vignettes, exemplary practices, and unlikely to result in use. This theoretical approach also needs
resources that can be immediately used to develop effective to be holistic. Changing individual components as separate
progress monitoring. Teachers also need to determine how events is unlikely to change systems. Rather, the whole needs
many students are being monitored on specific measures, to be reflected in the parts that in turn need to connect
grade levels, and time intervals, how students are being teachers and administrators.
18 ISRN Education

Teacher
professional
development
Progress monitoring

Instructional
adequacy
• Tier
Measurement • Time Decision
sufficiency • Group making
Teacher • Type • Emphasis • Level
practice • Curriculum • Variability
• Grade level
• Strategy • Slope
• No. (density)

Proximal
variables

Student Distal Fall Winter Spring

learning benchmark benchmark benchmark
variables

Figure 1: Teacher professional development → Teacher practice → Student learning.

The combined effects from all three components (prox- and progress monitoring data with state test scores in the
imal variables) are critical as well as the relation between Oregon Assessment of Knowledge and Skills (OAKS) to doc-
them and the outcome (distal variables), which is within- ument the effects for students with varying characteristics.
year growth (on benchmark measures) to document change Most of the research examples in this section are based in
relative to peers. It is not enough to have only one of the reading, but are likely to apply equally well in mathematics.
proximal variables—the right measures, targeted interven-
tions based on best evidence, and decisions tied to their affect 9.1. Measurement Sufficiency. The focus of this component is
on students. All three are needed. However, they need to to provide teachers and administrators training on student
work synergistically. And even then, changing these three measurement information to improve measurement suffi-
components is not enough either. Rather, the effect needs to ciency (i.e., increase or decrease the frequency of measure-
close the achievement gap in which students with disabilities ment, change the grade level of measurement, or modify
are catching up to their peers on grade level performance the type or domain of measure being used). These three
measures (e.g., benchmarks). Finally, for systemic change measurement components would allow policy to reflect an
data need to be collected on proximal variables for developing empirical basis for practice.
reports on use, allowing for professional development to be Many suggestions have been published in the literature
tailored and specific. on how often a student should be progress monitored, but
these reflect few actual studies of different types or schedules
9. Empirical Research on Three of measurement. We know little about the full range of
Assumptions (Warrants) reading measures, as most of the research in reading has
focused on oral reading fluency. Yet, teachers need to consider
With these assumptions to guide training, and given what we the full range of skills for students, including very early
know and do not know about teacher practice, the research reading (phoneme segmentation and alphabetic principles),
agenda for the next 30 years is organized around three critical as well as vocabulary and comprehension. As Gersten and
warrants supporting claims and guiding collection of evi- Dimino [207] note, “equally important is research on RTI that
dence. In this last section, these three warrants are explicated addresses students with problems with vocabulary, listening
and used to critique the empirical literature that has been col- comprehension, and reading comprehension” (p. 105).
lected in the previous 30 years. However, when the appropri-
ate literature is absent, the analysis is based on easy CBM data 9.1.1. What (Domain) to Measure? Very little research has
extracted from the 2010-2011 school year, with approximately been completed on the sensitivity of various measures for
10,000 students having progress monitoring in kindergarten monitoring reading development. For example, should first
through grade five. A team of researchers from Behavioral grade students with low reading skills be monitored in letter
Research and Teaching (BRT) has presented the initial results names, letter sounds, phoneme segmentation, or passage
from these analyses at two national conferences, respectively: reading fluency? At what point in skill development should
Pacific Coast Research Conference and National Council on measures be changed (e.g., from monitoring letter sounds
Measurement in Education with [206] or Tindal and Nese to word reading fluency)? Hintze’s work may be the most
[68]. Data continue to be harvested on annual benchmarking revealing, but it is limited primarily to oral reading fluency
ISRN Education 19

[82, 208, 209]. In our initial analysis of progress monitoring, Probably the most significant limitation of prior research in
we have found considerable variation in the type of reading this area is that benchmark measures, not progress measures,
measure administered over time; see Tindal et al. [206]. are often used to document growth [68]. The general reason
In grades three to five, teachers sometimes measured for this is the completeness of the data set for large groups
passage reading fluency one week and then either word of students and the regularity of data collection (all students
reading fluency or reading comprehension a few weeks later, in the district take measures in the fall, winter, and spring).
followed by word reading fluency. We found little pattern in Two research exceptions reflecting different schedules report
this cycle of different reading measures being successively generally similar results to the earlier research on growth:
administered over time. Jenkins and Terjeson [76] and Jenkins et al. [75].
We also know, however, that growth is increasingly being
9.1.2. What Grade Level to Measure? Recommendations on documented as nonlinear, with more growth from fall to
the grade or domain level to measure are much less promi- winter than from winter to spring [73, 77]. These findings
nent than how often to measure. Early research documented have several implications for policy and practice. If indeed
differential sensitivity as a function of breadth of the reading growth is nonlinear, teachers should not use a single goal as
domain [47, 210]. Otherwise, very few studies have addressed part of an Individualized Education Plan (IEP) to create an
grade level of the measures but rather consider passage vari- aim line. This practice would result in teachers underpredict-
ation [88] or difficulty [82]; again, most of this research is on ing early progress and overpredicting later progress. From
oral reading fluency. The general finding is that passages differ our own (easyCBM) data sets on progress monitoring, we
considerably unless explicitly controlled [83]. Again, our have found that teachers are quite haphazard in the way they
initial analyses of progress monitoring indicate considerable monitor students over time. For example, the average length
variation in the grade level of the measures used to monitor of time between September 1 and the first passage reading
progress (from 1st to 8th grade). In 3rd to 5th grade, teachers fluency (PRF) progress measure in 3rd through 5th grades is
generally (most frequently) monitored progress with grade- more than 10 weeks with a standard deviation greater than
level passages or passages that were one grade level below. seven weeks and students predominantly measured every
two to five weeks thereafter. Finally, much of this research
9.1.3. How Often to Measure Progress? Frequency of measure- fails to include full demographic information for students,
ment is a function of likelihood of change over time, and for particularly those being progress monitored (e.g., students at
that, the best literature to reference is research on within-year risk of failing to learn to read) [68].
growth for oral reading fluency. As noted earlier, Fuchs et
al. [69] documented the average slope of students in grades
9.2. Instructional Adequacy. Two primary sources are useful
one to six using a least-squares regression between scores
as sources for training on instructional systems: (a) What
and calendar days with slope converted to a weekly time
Works Clearinghouse [211–213] and (b) the recently published
frame. For the 103 special and general education students,
edition of the Handbook for Reading Interventions [214] as
respectively, the weekly growth in words correct per minute
well as Wanzek and Cavanaugh [191]. These sources provide a
was 1.5 and 2.0 in first and second grades, 1.0 and 1.5 in
wealth of information that can be used by teachers to organize
third grade, 0.85 and 1.1 in fourth grade, 0.5 and 0.8 in fifth
their instruction, particularly the former two documents. We
grade, and 0.3 and 0.63 in sixth grade. Eight years later, Deno
also use the results from a recent publication on grouping
et al. [70] conducted a far more reaching study in terms of
by Schwartz et al. [215] as well as writing from the National
geographic sampling plan and sample size with nearly 3,000
Center to Improve the Tools of Educators (NCITE).
students tested from four regions of the country. For 2,675
students in general education and 324 in special education
classes, they reported 2.0 words correct per minute growth 9.2.1. Instructional Tiers (1–3), Time, and Grouping. Most
per week until students achieved 30 WRCM; thereafter, RTI systems use a 3-tier system [211]. In these three tiers, the
students in the general education population improved at following levels of evidence are associated with each tier: (a)
least 1.0 word correct per minute/week. In the past decade tier 1 (differentiated instruction for all students using current
years, seven studies have been published on slope of progress reading performance) has low levels of evidence, (b) tier two
in oral reading fluency (Deno et al. [70]). In summary, a (intensive, systematic instruction on multiple reading skills
number of variables have been considered in studying growth that is delivered in small groups three to five times per week
of oral reading fluency over the past 15 years. Probably for 20–40 minutes for students scoring below benchmarks
the most remarkable finding is the general consistency in on universal screening) receives strong levels of evidence,
outcomes with very inconsistent sampling plans and research and (c) tier three (daily intensive instruction on a number
methods. of different reading skills for students exhibiting minimum
Note that most of these studies have focused on oral progress in tier two after a reasonable time) receives low
reading fluency with few other reading measures being inves- levels of evidence. An important caveat for tiers two and
tigated. The field also needs to address other reading skills three is the recommendation for small group instruction to
(in addition to fluency) that are critical in learning to read be homogenously configured though the use of a certified
for young students in the early elementary grades (e.g., letter teacher or paraprofessional is not deemed critical. Rather, it is
naming, letter sounding, and phoneme segmenting) as well the focus on how time is spent and apportioned that is critical,
as later elementary grades (vocabulary and comprehension). not the total amount of time.
20 ISRN Education

In the easyCBM research, considerable variation exists Soar to Success, Study Island, Step Up to Writing, various
across districts in their allocations of time in tiers 2 and 3 trade consumable workbooks, or various leveled texts. Less
(from 30 minutes to 60 and 120 minutes extra) though most than 3% were from the core curriculum: Houghton Mifflin or
(51%) of the students receive instruction 4 days per week and Treasures.
another 25% receive instruction five days per week for their “These instructional factors are encouraging high student
first intervention. Fully 22% receive this instruction for fewer engagement in learning, having a strong academic focus
than 30 minutes per day, 39% received it for 30–59 minutes and clarity of content coverage, facilitating moderate-to-high
per day, another 14% for an hour each day, and only 4% more success rates, and performance monitoring and informative
than an hour every day [216]. feedback” [219, p. 582]. In this study, a specific mathematics
The most clear research on grouping comes from program (Accelerated Math) was implemented with CBM.
Schwartz et al. [215] who used an experimental design to The results of the analyses for both the NALT and STAR
document the effects of student-teacher ratios in teasing out Math exams indicated that students who participated in
the very significant effects found with 1 : 1 (versus small group AM demonstrated more growth than students who did not
instruction) along with professional development (primarily participate (p. 532).
driven by the Reading Recovery model). With teachers In general, and consistent with the What Works Clear-
randomly assigned to small groups of two, three, and five inghouse, Wanzek and Cavanaugh reported that tier two
students in grade two, teachers taught for 20 weeks in daily involved small group instruction and “more frequent pro-
30-minute lessons in each condition. Using several pre-post grams monitoring (weekly or biweekly) to ensure the effec-
measures, 1 : 1 was compared to all other groups combined tiveness of instruction for the students” (p. 193). They found
and found to be significantly different (students scored group sizes of three to four students, with increasing amounts
higher on the post-test after covarying on the pre-test); of time over the grades (though in smaller bites in the lower
pairwise comparisons of 1 : 1 with each group were made, grades), increasing intensity by offering additional time or
and no significant differences were found among the small smaller groups with instruction more explicit, practice more
group comparisons (1 : 2, 1 : 3, and 1 : 5). Finally, post hoc repeated, feedback more frequent, and more emphasis on
comparisons showed the 1 : 1 condition scoring significantly high priority skills (also see Saez [216]).
higher on most of the measures. The problem, then, is simply
the resources needed to implement 1 : 1 programs, which is 9.3. Decision Making. CBM has generally been viewed as
an expensive venture that may be best reserved for students ideographic not nomothetic with students serving as their
in tier 3. Otherwise, the meta-analysis by Elbaum et al. [217] own control. In this view, the critical currency is how students
can be referenced to justify small group instruction, given change over time rather than how they compare to each
that 1 : 1 instruction is very resource intensive. other. The purpose is to make an individual difference in
measurement and evaluation not to document individual
9.2.2. Instructional Emphasis. In the What Works Clearing- differences [220].
house [211], the recommendation is made to “use a curricu- Single subject research designs are used to present data
lum that addresses the components of reading instruction on a line graph and analyzed to make data-based decisions
(phonemic awareness, phonics, vocabulary, comprehension, [221]. Typically, such designs include a baseline phase prior to
and fluency) and relates to students’ needs and development implementation of intervention; this phase is then compared
level” (p. 20). In the easyCBM 2001–2011 analysis by Saez to postintervention trajectory [222]. Visual analysis is used
[216], four areas were reportedly emphasized as the first inter- to evaluate data of individuals or small groups [221], and
vention (in decreasing use) targeting: fluency (38%), word decisions rules are then used to signal when instructional
identification (32%), comprehension (21%), and vocabulary changes are needed (not just who should be labeled with a
(8%). In addition, we add phonemic awareness and alphabetic disability). With appropriate labels and particularly timely
principles noted by O’ Connor [218]. well-depicted interventions, “cause-effect” relations can be
inferred. Using one of three references, teachers can deter-
9.2.3. Curriculum Materials and Instructional Strategies. A mine who is benefitting and who is not benefitting from the
myriad of materials and strategies are available for read- instructional program using (a) the normative distribution
ing instruction. Suffice it to say that “reading instruction of student performance based on all those in the group or a
should also be explicit. Explicit instruction involves a high subset of students similar to the targeted student or group,
level of teacher student interaction that includes frequent (b) a standard that corresponds with a successful outcome, or
opportunities for students to practice the skill with specific (c) individual reference using a comparison to an individual
corrective feedback” [211, p. 22]. According to the analysis student’s prior data [223]. Once sufficient data are collected
by Saez [216], approximately 66% were intensive interven- (e.g., four to five occasions), four indices are summarized to
tions: CARS/STARS (Comprehensive Assessment of Reading, evaluate instructional phases, the first two of which are within
Strategies/Strategies to Achieve Reading Success), Corrective phases, and the last two are between phases: (a) slope, (b)
Reading, Explode the Code, Phonics for Reading, Harcourt variability, (c) level, and (d) overlap (which incorporates all
Intervention Station, Horizons, Voyager Passport, Open three of these indices).
Court Intervention Kit, Triumphs, Read Naturally, Reading Decisions can also be based on goal lines or aim lines.
Mastery, Rewards, Score4Reading, or SLANT (Structured Decision rules for intervention change, including service
Language Training). About 31% were strategic interventions: eligibility, are based upon these goal lines in concert with
ISRN Education 21

an empirically sound instructional and decision making or intensity duration (19%); relatively few targeted changes in
sequence [192, 224]. The aim line or goal line refers to the tier (6%) or group size (6%).
projected amount of weekly growth across time that teachers The recommendation is made to monitor progress at
establish as a minimum for adequate progress [76]. Aim lines least eight times per year by the What Works Clearinghouse
are established in research using such techniques as norm [211]. “Although no studies have experimentally tested the
references or ordinary least squares (OLS) regression. Deno et impact of progress monitoring on outcomes in reading, we
al. [70] suggested normative growth rates; others considered still encourage schools to monitor the progress of these
slope of improvement relative to aim lines by fitting an OLS students so that personnel possess information on how a
regression line to each student’s scores, computed the weekly student is doing in general education reading proficiency
increase based on the obtained parameter, and averaged and improving in specific skills” [211, p. 7]. This information
these weekly increases (across students within grade levels would allow decision making to be based on changes in
and within general or special education programs). Ardoin level (average performance), variability (standard deviation
[225] developed aim lines by using each student’s median of values), and slope (amount of improvement) as indices that
baseline score as the aim line start point and calculated the can be used to make decisions to change programs. Generally,
end point of the aim line by multiplying the number of slope is used most frequently, though it is interpreted relative
weeks of intervention implementation by the desired gain to the amount of variability.
per week and adding the product to the start point (using
an OLS solution). The beginning point of the goal line 10. Summary and Parting Comments
is placed at the beginning of intervention, the end point
is placed at the predicted end of data collection, and a Curriculum-based measurement research has a solid 30-year
line is drawn to connect the two points. VanDerHeyden et track record that supports its technical adequacy. It represents
al. [226] developed a local norm in a manner similar to a standardized form of assessment so that comparability
the technique used by Ardoin [225] using students in the of measures (and outcomes) can be ensured, whether it is
“instructional range” as the comparison group. They then comparing a student’s performance to other students (norm
used level of performance and trend over time to estimate referenced) or comparing their previous performance to later
students’ growth by subtracting each student’s fall score from performance (individual referenced). The former provides
his/her spring score and then dividing this difference by the effective screening for risk and resource allocation, while
number of intervention weeks. The average of the students’ the latter provides a system for progress monitoring and
weekly growth estimates was then considered to represent evaluation of instruction.
the growth rate of students considered not to be at risk (i.e., For any measurement system, the first order of business
those students who scored in the instructional range during is about technical adequacy, which has been the essential
the spring administration). focus throughout the 30 years. The key areas for this research
Presently, most decision rules are conceptually not empir- began in reading (fluency), but it was only after major
ically based. Furthermore, a number of methodological political events that it expanded to early literacy or the
problems are inherent in their use: (a) as noted earlier, the foundation of disparate reading skills. Likewise, the initial
assumption of linear growth, which may not be accurate research on writing was limited, primarily in the populations
[73, 77], (b) use of gain scores [227], and (c) difficulties being monitored (elementary age students but later becoming
in combining statistical and graphic representations of data secondary students), which gave rise to systems for scoring
[192]. Although aim lines may improve decisions, they also student work (products). Attention to mathematics also was
have their limitations: disagreement in estimates, distraction not part of the original research but quickly gained stride
from other data qualities, ambiguous results that can lead to across the age ranges of students; given the unique nature
misinterpretations [192], and confounding growth with level of mathematics, the issue mostly addressed domains for
of baseline performance [194, 224]. sampling items. Finally, in middle and high schools, the focus
In the analysis by Saez [216], three districts were com- was on the target skills of content trying to “essentialize” the
pared in their response to nonresponders with the following access skills needed to get to the content. Obviously, more
options being promulgated: (a) four to six points below aim research is needed in all of these subject areas.
line or slope is flat/decreasing, (b) measured achievement However, and perhaps more importantly, further research
falls below aim line or flat progress trend, or (c) after four and development needs to occur in the training of teachers
data points to “make an instructional change or continue on the use of data and on the reporting systems that teachers
to monitor.” Unfortunately, the vast majority of teachers are using. Although early research targeted decision making,
were implementing few changes, with 355 students (65%) it was under well-controlled studies being conducted with
receiving only one intervention, 138 students (25%) receiving university support. Little attention was given to full-scale
two changes, 25 (4.6%) receiving three, and only 29 (5.3%) adoption and systematic incorporation into the decision
receiving four or more instructional interventions. Although making of the classrooms. Essentially, the research had little
some of the second interventions were introduced in Septem- verisimilitude. Now, however, it is important to begin this
ber (𝑛 = 15) or October (𝑛 = 13), many were implemented research in typical classrooms that are part of not apart
in the late fall (48 in November and December) and winter from teachers and students participating in a response-to-
(21 in January and February). And of the changes made, most intervention system. In order to begin this research, it is best
were targeted toward instructional program curricula (50%) to consider training and reporting. The training can address
22 ISRN Education

the use of norm- or individual-referenced data as well as [11] S. Deno, P. Mirkin, and M. Shinn, Behavioral Perspectives on the
on smart reports that are electronic and provide prompts Assessment of Learning Disabled Children, Monograph, Institute
for decision making, much like many software applications for Research on Learning Disabilities (IRLD), University of
being developed around scheduling flights, meetings, and Minnesota, Minneapolis, Minn, USA, 1979.
weight loss. Given the ubiquitous nature of technology in [12] J. R. Jenkins, S. Deno, and P. Mirkin, Measuring Pupil Progress
the classroom, there is no better time for a new generation Toward the Least Restrictive Environment, Monograph, Institute
of curriculum-based measurement embedded into a digital for Research on Learning Disabilities (IRLD), University of
Minnesota, Minneapolis, Minn, USA, 1979.
environment.
[13] L. Fuchs and S. Deno, “The relationship between curriculum-
based mastery measures and standardized achievement tests in
Acknowledgments reading,” Research Report, Institute for Research on Learning
Disabilities (IRLD), University of Minnesota, Minneapolis,
This paper would not have been possible without others at Minn, USA, 1981.
Behavioral Research and Teaching (BRT) Denise Swanson, [14] L. Fuchs, G. Tindal, D. Fuchs, M. Shinn, S. Deno, and G. Ger-
helped organize the many references cited in this compila- mann, “The technical adequacy of a basal reading mastery test:
tion. A Steffani Mast, provided exceptional editing in moving the Holt reading series,” Research Report, Institute for Research
the paper to a final draft. Finally, Dr. Joseph Nese helped craft on Learning Disabilities (IRLD), University of Minnesota,
the logic of the decision making argument at the end of the Minneapolis, Minn, USA, 1983.
paper as part of a grant application to fund this line of work. [15] L. Fuchs, G. Tindal, M. Shinn, D. Fuchs, and G. Germann,
“Technical adequacy of basal readers’ mastery tests: the Ginn
References 720 series,” Research Report, Institute for Research on Learning
Disabilities (IRLD), University of Minnesota, Minneapolis,
[1] S. Messick, “Standards of validity and the validity of standards Minn, USA, 1983.
in performance assessment,” Educational Measurement, vol. 14, [16] G. Tindal, L. Fuchs, D. Fuchs, M. Shinn, S. Deno, and G. Ger-
no. 4, pp. 5–8, 1995. mann, “The technical adequacy of a basal series mastery test: the
[2] M. Kane, “Validation,” in Educational Measurement, R. Bren- Scott-Foresman reading program,” Research Report, Institute
nan, Ed., pp. 17–64, American Council on Education and Prae- for Research on Learning Disabilities (IRLD), University of
ger, Westport, Conn, USA, 4th edition, 2006. Minnesota, Minneapolis, Minn, USA, 1983.
[3] L. S. Fuchs, “The past, present, and future of curriculum-based [17] G. Tindal, M. Shinn, L. Fuchs, D. Fuchs, S. Deno, and G. Ger-
measurement research,” School Psychology Review, vol. 33, no. 2, mann, “The technical adequacy of a base reading series mastery
pp. 188–192, 2004. test,” Research Report, Institute for Research on Learning Dis-
abilities (IRLD), University of Minnesota, Minneapolis, Minn,
[4] American Educational Research Association, American Psy-
USA, 1983.
chological Association, & National Council on Measurement in
Education, Standards for Educational and Psychological Testing, [18] D. Fuchs and S. Deno, “Reliability and validity of curriculum-
Amer Psychological Association, Washington, DC, USA, 1999. based informal reading inventories,” Research Report, Institute
for Research on Learning Disabilities (IRLD), University of
[5] S. Deno and P. Mirkin, Data Based Program Modification: A Minnesota, Minneapolis, Minn, USA, 1981.
Manual, Leadership Training Institute for Special Education,
[19] L. Fuchs, D. Fuchs, and S. Deno, “The nature of inaccuracy
Minneapolis, Minn, USA, 1977.
among readability formulas,” Research Report, Institute for
[6] G. Tindal and J. F. T. Nese, “Applications of curriculum-based Research on Learning Disabilities (IRLD), University of Min-
measures in making multiple decisions with multiple reference nesota, Minneapolis, Minn, USA, 1983.
points,” in Assessment and Intervention: Advances in Learning [20] L. Fuchs and S. Deno, “A comparison of reading placement
and Behavioral Disabilities, M. M. T. Scruggs, Ed., vol. 24, pp. based on teacher judgment, standardized testing, and curricu-
31–58, Emerald, Bingley, UK, 2011. lum-based assessment,” Research Report, Institute for Research
[7] P. Mirkin, S. Deno, G. Tindal, and K. Kuehnle, “Formative eval- on Learning Disabilities (IRLD), University of Minnesota, Min-
uation: continued development of data utilization systems,” neapolis, Minn, USA, 1981.
Research Report, Institute for Research on Learning Disabilities [21] L. Fuchs, S. Deno, and P. Mirkin, “Direct and frequent mea-
(IRLD), University of Minnesota, Minneapolis, Minn, USA, surement and evaluation: effects on instruction and estimates of
1980. student progress,” Research Report, Minneapolis, Minn, USA,
[8] B. Meyers, J. Meyers, and S. Deno, “Formative evaluation and 1982.
teacher decision-making: a follow-up investigation,” Research [22] L. Fuchs, S. Deno, and P. Mirkin, “Effects of frequent curricu-
Report, Institute for Research on Learning Disabilities (IRLD), lum-based measurement and evaluation on student achieve-
University of Minnesota, Minneapolis, Minn, USA, 1980. ment and knowledge of performance: an experimental study,”
[9] S. Deno and P. Mirkin, Data-Based IEP Development: An Research Report, Institute for Research on Learning Disabilities
Approach to Substantive Compliance, Monograph, Institute for (IRLD), University of Minnesota, Minneapolis, Minn, USA,
Research on Learning Disabilities (IRLD), University of Minne- 1982.
sota, Minneapolis, Minn, USA, 1979. [23] L. Fuchs, S. Deno, and A. Roettger, “The effect of alterna-
[10] G. Tindal, L. Fuchs, S. Christenson, P. Mirkin, and S. Deno, “The tive data-utilization rules on spelling achievement,” Research
relationship between student achievement and teacher assess- Report, Institute for Research on Learning Disabilities (IRLD),
ment of short or long-term goals,” Research Report, Institute University of Minnesota, Minneapolis, Minn, USA, 1983.
for Research on Learning Disabilities (IRLD), University of [24] L. Fuchs, C. Wesson, G. Tindal, P. Mirkin, and S. Deno, “Instruc-
Minnesota, Minneapolis, Minn, USA, 1981. tional changes, student performances, and teacher preferences:
ISRN Education 23

the effects of specific measurement and evaluation procedures,” [37] R. Skiba, C. Wesson, and S. Deno, “The effects of training
Research Report, Institute for Research on Learning Disabilities teachers in the use of formative evaluation in reading: an exper-
(IRLD), University of Minnesota, Minneapolis, Minn, USA, imental control comparison,” Research Report, Institute for
1982. Research on Learning Disabilities (IRLD), University of Min-
[25] P. Mirkin and S. Deno, “Formative evaluation in the classroom: nesota, Minneapolis, Minn, USA, 1982.
an approach to improving instruction,” Research Report, Insti- [38] L. Fuchs, C. Wesson, G. Tindal, P. Mirkin, and S. Deno, “Teacher
tute for Research on Learning Disabilities (IRLD), University of efficiency in continuous evaluation of IEP goals,” Research
Minnesota, Minneapolis, Minn, USA, 1979. Report, Institute for Research on Learning Disabilities (IRLD),
University of Minnesota, Minneapolis, Minn, USA, 1981.
[26] P. Mirkin, L. Fuchs, G. Tindal, S. Christenson, and S. Deno,
“The effect of IEP monitoring strategies on teacher behavior,” [39] R. King, C. Wesson, and S. Deno, “Direct and frequent mea-
Research Report, Institute for Research on Learning Disabilities surement of student performance: does it take too much time?”
(IRLD), University of Minnesota, Minneapolis, Minn, USA, Research Report, Institute for Research on Learning Disabilities
1981. (IRLD), University of Minnesota, Minneapolis, Minn, USA,
1982.
[27] B. Sevcik, R. Skiba, G. Tindal et al., “Curriculum-based mea-
[40] D. Marston and S. Deno, “Implementation of direct and repeat-
surement: effects on instruction, teacher estimates of student
ed measurement in the school setting,” Research Report, Insti-
progress and student knowledge of performance,” Research
tute for Research on Learning Disabilities (IRLD), University of
Report, Institute for Research on Learning Disabilities (IRLD),
Minnesota, Minneapolis, Minn, USA, 1982.
University of Minnesota, Minneapolis, Minn, USA, 1983.
[41] D. Marston, G. Tindal, and S. Deno, “Predictive efficiency of
[28] C. Wesson, S. Deno, P. Mirkin et al., “Teaching structure and direct, repeated measurement: an analysis of cost and accuracy
student achievement effects of curriculum-based measurement: in classification,” Research Report, Institute for Research on
a casual (structural) analysis,” Research Report, Institute for Learning Disabilities (IRLD), University of Minnesota, Min-
Research on Learning Disabilities (IRLD), University of Min- neapolis, Minn, USA, 1982.
nesota, Minneapolis, Minn, USA, 1982.
[42] P. Mirkin, D. Marston, and S. Deno, “Direct and repeated
[29] C. Wesson, P. Mirkin, and S. Deno, “Teachers’ use of self-in- measurement of academic skills: an alternative to traditional
structional materials for learning procedures for developing and screening referral, and identification of learning disabled stu-
monitoring progress on IEP goals,” Research Report, Institute dents,” Research Reports, Institute for Research on Learning
for Research on Learning Disabilities (IRLD), University of Disabilities (IRLD), University of Minnesota, Minneapolis,
Minnesota, Minneapolis, Minn, USA, 1982. Minn, USA, 1982.
[30] C. Wesson, R. Skiba, B. Sevcik et al., “The impact of the structure [43] G. Tindal, G. Germann, and S. Deno, “Descriptive research on
of instruction and the use of technically adequate instructional the Pine County norms: a compilation of findings,” Research
data on reading improvement,” Research Report, Institute Report, Institute for Research on Learning Disabilities (IRLD),
for Research on Learning Disabilities (IRLD), University of University of Minnesota, Minneapolis, Minn, USA, 1983.
Minnesota, Minneapolis, Minn, USA, 1983. [44] G. Tindal, G. Germann, D. Marston, and S. Deno, “The
[31] M. Shinn, R. Good, and S. Stein, “Summarizing trend in student effectiveness of special education,” Research Report, Institute
achievement: a comparison of methods,” School Psychology for Research on Learning Disabilities (IRLD), University of
Review, vol. 18, no. 3, pp. 356–370, 1989. Minnesota, Minneapolis, Minn, USA, 1983.
[45] L. Fuchs, G. Tindal, and S. Deno, “Effects of varying item
[32] R. Skiba and S. Deno, “A correlational analysis of the statistical
domain and sample duration on technical characteristics of
properties of time-series data and their relationship to student
daily measures in reading,” Research Report, Institute for
achievement in resource classrooms,” Research Report, Institute
Research on Learning Disabilities (IRLD), University of Min-
for Research on Learning Disabilities (IRLD), University of
nesota, Minneapolis, Minn, USA, 1981.
Minnesota, Minneapolis, Minn, USA, 1983.
[46] M. Shinn, M. Gleason, and G. Tindal, “Varying the difficulty of
[33] R. Skiba, D. Marston, C. Wesson, B. Sevcik, and S. Deno, “Char- testing materials: implications for curriculum-based measure-
acteristics of the time-series data collected through curriculum- ment,” The Journal of Special Education, vol. 23, pp. 223–233,
based reading measurement,” Research Report, Institute for 1989.
Research on Learning Disabilities (IRLD), University of Min-
[47] G. Tindal and S. Deno, “Daily measurement of reading: effects
nesota, Minneapolis, Minn, USA, 1983.
of varying the size of the item pool,” Research Report, Institute
[34] G. Tindal, S. Deno, and J. Ysseldyke, “Visual analysis of time for Research on Learning Disabilities (IRLD), University of
series data: factors of influence and level reliability,” Research Minnesota, Minneapolis, Minn, USA, 1981.
Report, Institute for Research on Learning Disabilities (IRLD), [48] G. Tindal, D. Marston, S. Deno, and G. Germann, “Curriculum
University of Minnesota, Minneapolis, Minn, USA, 1983. differences in direct repeated measures of reading,” Research
[35] D. Marston, L. Lowry, S. Deno, and P. Mirkin, “An analysis of Report, Institute for Research on Learning Disabilities (IRLD),
learning trends in simple measures of reading, spelling, and University of Minnesota, Minneapolis, Minn, USA, 1982.
written expression: a longitudinal study,” Research Report, [49] S. Deno, D. Marston, P. Mirkin, L. Lowry, P. Sindelar, and J. R.
Institute for Research on Learning Disabilities (IRLD), Univer- Jenkins, The Use of Standard Tasks to Measure Achievement in
sity of Minnesota, Minneapolis, Minn, USA, 1981. Reading, Spelling, and Written Expression: A Normative and
[36] R. King, S. Deno, P. Mirkin, and C. Wesson, “The effects of Developmental Study, Institute for Research on Learning Dis-
training in the use of formative evaluation in reading: an exper- abilities, Minneapolis, Minn, USA, 1982.
imental control comparison,” Research Report, Institute for [50] S. Deno, P. Mirkin, B. Chiang, and L. Lowry, “Relationships
Research on Learning Disabilities (IRLD), University of Min- among simple measures of reading and performance on stan-
nesota, Minneapolis, Minn, USA, 1983. dardized achievement tests,” Research Report, Institute for
24 ISRN Education

Research on Learning Disabilities (IRLD), University of Min- [67] C. A. Espin, K. L. McMaster, S. Rose, and M. M. Wayman, A
nesota, Minneapolis, Minn, USA, 1980. Measure of Success: How Curriculum-Based Measurement has
[51] S. Deno, P. Mirkin, L. Lowry, and K. Kuehnle, “Relation- Influenced Education and Learning, University of Minnesota
ships among simple measures of spelling and performance on Press, Minneapolis, Minn, USA, 2012.
standardized achievement tests,” Research Report, Institute for [68] G. Tindal and J. F. T. Nese, “Within year achievement growth
Research on Learning Disabilities (IRLD), University of Min- using curriculum based measurement,” in Proceedings of the
nesota, Minneapolis, Minn, USA, 1980. National Council on Measurement in Education, Vancouver,
[52] S. Deno, P. Mirkin, and D. Marston, “Relationships among Canada, April 2012.
simple measures of written expression and performance on [69] L. Fuchs, D. Fuchs, C. Hamlett, L. Walz, and G. Germann,
standardized achievement tests,” Research Report, Institute for “Formative evaluation of academic progress: how much growth
Research on Learning Disabilities (IRLD), University of Min- can we expect?” School Psychology Review, vol. 22, pp. 27–48,
nesota, Minneapolis, Minn, USA, 1980. 1993.
[53] L. Fuchs, S. Deno, and D. Marston, “Use of aggregation to [70] S. L. Deno, L. S. Fuchs, D. Marston, and J. Shin, “Using curric-
improve the reliability of simple direct measures of academic ulum-based measurement to establish growth standards for
performance,” Research Report, Institute for Research on students with learning disabilities,” School Psychology Review,
Learning Disabilities (IRLD), University of Minnesota, Min- vol. 30, no. 4, pp. 507–524, 2001.
neapolis, Minn, USA, 1982. [71] S. P. Ardoin and T. J. Christ, “Evaluating curriculum-based mea-
[54] D. Marston and S. Deno, “The reliability of simple, direct surement slope estimates using data from triannual universal
measures of written expression,” Research Report, Institute for screenings,” School Psychology Review, vol. 37, no. 1, pp. 109–125,
Research on Learning Disabilities (IRLD), University of Minne- 2008.
sota, Minneapolis, Minn, USA, 1981. [72] T. J. Christ, “Short-term estimates of growth using curriculum-
[55] G. Tindal, D. Marston, and S. Deno, “The reliability of direct and based measurement of oral reading fluency: estimating stan-
repeated measurement,” Research Report, Institute for Research dard error of the slope to construct confidence intervals,” School
on Learning Disabilities (IRLD), University of Minnesota, Psychology Review, vol. 35, no. 1, pp. 128–133, 2006.
Minneapolis, Minn, USA, 1983. [73] T. J. Christ, B. Silberglitt, S. Yeo, and D. Cormier, “Curriculum-
[56] J. Videen, S. Deno, and D. Marston, “Correct word sequences: based measurement of oral reading: an evaluation of growth
a valid indicator of proficiency in written expression,” Research rates and seasonal effects among students served in general and
Reports, Institute for Research on Learning Disabilities (IRLD), special education,” School Psychology Review, vol. 39, no. 3, pp.
University of Minnesota, Minneapolis, Minn, USA, 1982. 447–462, 2010.
[57] C. Wesson, S. Deno, and P. Mirkin, Research on Developing [74] S. B. Graney, K. N. Missall, R. S. Martı́nez, and M. Bergstrom,
and Monitoring Progress on IEP Goals: Current Findings and “A preliminary investigation of within-year growth patterns in
Implications for Practice, Monograph, Institute for Research on reading and mathematics curriculum-based measures,” Journal
Learning Disabilities (IRLD), University of Minnesota, Minne- of School Psychology, vol. 47, no. 2, pp. 121–142, 2009.
apolis, Minn, USA, 1982. [75] J. R. Jenkins, J. J. Graff, and D. L. Miglioretti, “Estimating
[58] G. Tindal, G. Germann, S. Deno, and P. Mirkin, The Pine County reading growth using intermittent CBM progress monitoring,”
Model for Special Education Delivery: A Data-Based System, Exceptional Children, vol. 75, no. 2, pp. 151–163, 2009.
Monograph, Institute for Research on Learning Disabilities [76] J. Jenkins and K. Terjeson, “Monitoring reading growth: goal
(IRLD), University of Minnesota, Minneapolis, Minn, USA, setting, measurement frequency, and methods of evaluation,”
1982. Learning Disabilities Research & Practice, vol. 26, no. 1, pp. 28–
[59] P. K. Mirkin, L. S. Fuchs, and S. L. Deno, Consideration for De- 35, 2011.
signing a Continuous Evaluation : An Interpretive Review, Mono- [77] J. F. T. Nese, G. Biancarosa, D. Anderson, C. F. Lai, J. Alonzo,
graph, Institute for Research on Learning Disabilities (IRLD), and G. Tindal, “Within-year oral reading fluency with CBM: a
University of Minnesota, Minneapolis, Minn, USA, 1982. comparison of models,” Reading and Writing, vol. 25, no. 4, pp.
[60] M. R. Shinn, Curriculum-Based Measurement: Assessing Special 887–915, 2012.
Children, The Guilford Press, New York, NY, USA, 1989. [78] S. Raudenbush, A. Bryk, Y. Cheong, and R. Congdon, HLM 6:
[61] M. R. Shinn, Advanced Applications of Curriculum-Based Mea- Hierarchical Linear and NonLinear Modeling, Scientific Software
surement, Guilford Press, New York, NY, USA, 1998. International, Lincolnwood, Ill, USA, 2004.
[62] G. Tindal and D. Marston, Classroom-Based Assessment: Eval- [79] S. Raudenbush and A. Bryk, Hierarchical Linear Models: Appli-
uating Instructional Outcomes, Merrill, Columbus, Ohio, USA, cations and Data Analysis Methods, Sage Publications, Thou-
1990. sand Oaks, Calif, USA, 2nd edition, 2002.
[63] E. E. Gickling and V. P. Thompson, “A personal view of curricu- [80] R. Brennan, Generalizability Theory, Springer, New York, NY,
lum-based assessment,” Exceptional Children, vol. 52, no. 3, pp. USA, 2001.
205–218, 1985. [81] R. Linn and E. Burton, “Performance-based assessment: impli-
[64] K. Howell, Curriculum-Based Evaluation for Special and Reme- cations of task specificity,” Educational Measurement, vol. 13, pp.
dial Education: A Handbook for Deciding What to Teach, Merrill 5–8, 1994.
Publishing, Columbus, Ohio, USA, 1987. [82] J. M. Hintze, E. J. Daly, and E. S. Shapiro, “An investigation of
[65] C. S. Blankenship, “Using curriculum-based assessment data to the effects of passage difficulty level on outcomes of oral reading
make instructional decisions,” Exceptional Children, vol. 52, no. fluency progress monitoring,” School Psychology Review, vol. 27,
3, pp. 233–238, 1985. no. 3, pp. 433–445, 1998.
[66] M. R. Shinn, S. Rosenfield, and N. Knutson, “Curriculum-based [83] J. M. Hintze and T. J. Christ, “An examination of variability as
assessment: a comparison of models,” School Psychology Review, a function of passage variance in CBM progress monitoring,”
vol. 18, no. 3, pp. 299–316, 1989. School Psychology Review, vol. 33, no. 2, pp. 204–217, 2004.
ISRN Education 25

[84] J. M. Hintze, S. V. Owen, E. S. Shapiro, and E. J. Daly, “Research [100] D. Starch, “The measurement of efficiency in reading,” The
design and methodology section—generalizability of oral read- Journal of Educational Psychology, vol. 6, no. 1, pp. 1–24, 1915.
ing fluency measures: application of g theory to curriculum- [101] V. L. Anderson and M. A. Tinker, “The speed factor in reading
based measurement,” School Psychology Quarterly, vol. 15, no. performance,” Journal of Educational Psychology, vol. 27, no. 8,
1, pp. 52–68, 2000. pp. 621–624, 1936.
[85] B. C. Poncy, C. H. Skinner, and P. K. Axtell, “An investigation of [102] D. LaBerge and S. J. Samuels, “Toward a theory of automatic
the reliability and standard error of measurement of words read information processing in reading,” Cognitive Psychology, vol.
correctly per minute using curriculum-based measurement,” 6, no. 2, pp. 293–323, 1974.
Journal of Psychoeducational Assessment, vol. 23, no. 4, pp. 326–
[103] R. Good, D. Simmons, and E. Kame’enui, “The importance and
338, 2005.
decision-making utility of a continuum of fluency-based indica-
[86] T. J. Christ and S. P. Ardoin, “Curriculum-based measurement tors of foundational reading skills for third-grade high-stakes
of oral reading: passage equivalence and probe-set develop- outcomes,” Scientific Studies of Reading, vol. 5, no. 3, pp. 257–
ment,” Journal of School Psychology, vol. 47, no. 1, pp. 55–75, 288, 2001.
2009.
[104] R. H. Good, J. Gruba, and R. A. Kaminski, “Best practices
[87] S. P. Ardoin and T. J. Christ, “Curriculum-based measurement
in using Dynamic Indicators of Basic Early Literacy Skills
of oral reading: standard errors associated with progress moni-
(DIBELS) in an outcomes-driven model,” in Best Practices in
toring outcomes from DIBELS, AIMSweb, and an experimental
School Psychology IV, A. Thomas and J. Grimes, Eds., pp. 679–
passage set,” School Psychology Review, vol. 38, no. 2, pp. 266–
700, National Association of School Psychologists, Washington,
283, 2009.
DC, USA, 2001.
[88] D. J. Francis, K. L. Santi, C. Barr, J. M. Fletcher, A. Varisco, and
[105] R. H. Good and R. A. Kaminski, “Assessment for instructional
B. R. Foorman, “Form effects on the estimation of students’ oral
decisions: toward a proactive/prevention model of decision-
reading fluency using DIBELS,” Journal of School Psychology,
making for early literacy skills,” School Psychology Quarterly,
vol. 46, no. 3, pp. 315–342, 2008.
vol. 11, no. 4, pp. 326–336, 1996.
[89] T. Christ, “Curriculum-based measurement of oral reading:
multi-study evaluation of schedule, duration and dataset quality [106] H. L. Rouse and J. W. Fantuzzo, “Validity of the dynamic
on progress monitoring outcomes,” Journal of School Psychology, indicators for basic early literacy skills as an indicator of early
vol. 78, no. 3, 2012. literacy for urban kindergarten children,” School Psychology
Review, vol. 35, no. 3, pp. 341–355, 2006.
[90] D. C. Briggs, “Synthesizing causal inferences,” Educational
Researcher, vol. 37, no. 1, pp. 15–22, 2008. [107] N. Clemens, E. Shapiro, and F. Thoemmes, “Improving the
efficacy of first grade reading screening: an investigation of
[91] National Institute for Literacy, Developing Early Literacy: Report
word identification fluency with other early literacy indicators,”
of the National Early Literacy Panel (A Scientific Synthesis of
School Psychology Quarterly, vol. 26, no. 3, pp. 231–244, 2011.
Early Literacy Development and Implications for Intervention),
Washington, DC, USA, 2008. [108] S. Hagan-Burke, M. Burke, and C. Crowder, “The convergent
[92] R. C. Anderson, E. H. Hiebert, J. A. Scott, and I. A. G. Wilkin- validity of the dynamic indicators of basic and early literacy
son, Becoming a Nation of Readers: The Report of the Commis- skills and the test of word reading efficiency for the beginning
sion on Reading, National Institute of Education, Washington, of first grade,” Assessment for Effective Intervention, vol. 31, no.
DC, USA, 1985. 4, pp. 1–15, 2006.
[93] National Institutes of Child Health and Human Development, [109] Y. Y. Lo, C. Wang, and S. Haskell, “Examining the impacts of
“Report of national reading panel: teaching children to read: early reading intervention on the growth rates in basic literacy
an evidence-based assessment of the scientific literature on skills of at-risk urban kindergarteners,” The Journal of Special
reading and its implications for reading instruction,” Report of Education, vol. 43, no. 1, pp. 12–28, 2009.
the Subgroups, Washington, DC, USA, 2000. [110] U. Yesil-Dagli Ummuhan, “Predicting ELL students’ beginning
[94] No Child Left Behind, Committee on Education and Labor, first grade English oral reading fluency from initial kinder-
Government Printing Office, Washington, DC, USA, 1st edition, garten vocabulary, letter naming, and phonological awareness
2001. skills,” Early Childhood Research Quarterly, vol. 26, no. 1, pp. 15–
[95] M. J. Adams, Beginning to Read: Thinking and Learning about 29, 2011.
Print, MIT Press, Cambridge, Mass, USA, 1990. [111] L. C. Ehri, S. R. Nunes, D. M. Willows, B. V. Schuster,
[96] L. S. Fuchs, D. Fuchs, and D. L. Compton, “Monitoring early Z. Yaghoub-Zadeh, and T. Shanahan, “Phonemic awareness
reading development in first grade: word identification fluency instruction helps children learn to read: evidence from the
versus nonsense word fluency,” Exceptional Children, vol. 71, no. National Reading Panels meta-analysis,” Reading Research
1, pp. 7–21, 2004. Quarterly, vol. 36, no. 3, pp. 250–287, 2001.
[97] K. D. Ritchey and D. L. Speece, “From letter names to word [112] S. Stage, J. Sheppard, M. Davidson, and M. Browning, “Pre-
reading: the nascent role of sublexical fluency,” Contemporary diction of first-graders’ growth in oral reading fluency using
Educational Psychology, vol. 31, no. 3, pp. 301–327, 2006. kindergarten letter fluency,” Journal of School Psychology, vol.
[98] K. D. Ritchey, “Assessing letter sound knowledge: a comparison 39, pp. 225–237, 2001.
of letter sound fluency and nonsense word fluency,” Exceptional [113] H. Yopp, “A test for assessing phonemic awareness in young
Children, vol. 74, no. 4, pp. 487–506, 2008. children,” The Reading Teacher, vol. 49, no. 1, pp. 20–29, 1995.
[99] P. G. Aaron, R. Malatesha Joshi, R. Gooden, and K. E. Bentum, [114] D. L. Linklater, R. E. O’Connor, and G. J. Palardy, “Kindergarten
“Diagnosis and treatment of reading disabilities based on the literacy assessment of English Only and English language
component model of reading: an alternative to the discrepancy learner students: an examination of the predictive validity of
model of LD,” Journal of Learning Disabilities, vol. 41, no. 1, pp. three phonemic awareness measures,” Journal of School Psychol-
67–84, 2008. ogy, vol. 47, no. 6, pp. 369–394, 2009.
 26 ISRN Education

[115] D. L. Speece, K. D. Ritchey, D. H. Cooper, F. P. Roth, and C. [132] B. Diercks-Gransee, J. W. Weissenburger, C. L. Johnson, and P.
Schatschneider, “Growth in early reading skills from kinder- Christensen, “Curriculum-based measures of writing for high
garten to third grade,” Contemporary Educational Psychology, school students,” Remedial and Special Education, vol. 30, no. 6,
vol. 29, no. 3, pp. 312–332, 2004. pp. 360–371, 2009.
[116] C. Juel, “Learning to read and write: a longitudinal study of 54 [133] C. A. Espin, S. De La Paz, B. J. Scierka, and L. Roelofs, “The
children from first through fourth grades,” Journal of Education- relationship between curriculum-based measures in written
al Psychology, vol. 80, no. 4, pp. 437–447, 1988. expression and quality and completeness of expository writing
[117] J. M. T. Vloedgraven and L. Verhoeven, “Screening of phonolog- for middle school students,” The Journal of Special Education,
ical awareness in the early elementary grades: an IRT approach,” vol. 38, no. 4, pp. 208–217, 2005.
Annals of Dyslexia, vol. 57, no. 1, pp. 33–50, 2007. [134] S. Fewster and P. D. Macmillan, “School-based evidence for
[118] M. Twain, “Simplified spelling,” in Letters from the Earth: the validity of curriculum-based measurement of reading and
Uncensored Writings, B. DeVoto, Ed., HarperCollins, New York, writing,” Remedial and Special Education, vol. 23, no. 3, pp. 149–
NY, USA, 1942. 156, 2002.
[119] L. Fuchs and D. Fuchs, “Determining adequate yearly progress [135] J. M. Amato and M. W. Watkins, “The predictive validity of
from kindergarten through grade 6 with curriculum-based CBM writing indices for eighth-grade students,” The Journal of
measurement,” Assessment for Effective Intervention, vol. 29, no. Special Education, vol. 44, no. 4, pp. 195–204, 2011.
4, pp. 25–37, 2004. [136] K. McMaster and C. Espin, “Technical features of curriculum-
[120] L. S. Fuchs, D. Fuchs, and D. L. Compton, “Monitoring early based measurement in writing: a literature review,” The Journal
reading development in first grade: word identification fluency of Special Education, vol. 41, no. 2, pp. 68–84, 2007.
versus nonsense word fluency,” Exceptional Children, vol. 71, no. [137] A. Foegen and S. L. Deno, “Identifying growth indicators for
1, pp. 7–21, 2004. low-achieving students in middle school mathematics,” The
[121] R. Parker, G. Tindal, and J. Hasbrouck, “Countable indices of Journal of Special Education, vol. 35, no. 1, pp. 4–16, 2001.
writing quality: their suitability for screening-eligibility deci-
[138] M. Calhoon, “Curriculum-based measurement for mathemat-
sions,” Exceptionality, vol. 2, pp. 1–17, 1991.
ics at the high school level: what we do not know and what we
[122] R. I. Parker, G. Tindal, and J. Hasbrouck, “Progress monitoring need to know,” Assessment for Effective Intervention, vol. 33, pp.
with objective measures of writing performance for students 234–239, 2008.
with mild disabilities,” Exceptional Children, vol. 58, no. 1, pp.
[139] L. Fuchs, D. Fuch, and S. Courey, “Curriculum-based measure-
61–73, 1991.
ment of mathematics competence: from competence to con-
[123] G. Tindal and R. Parker, “Assessment of written expression for cepts and applications to real life problem solving,” Assessment
students in compensatory and special education programs,” The for Effective Intervention, vol. 30, no. 2, pp. 33–46, 2005.
Journal of Special Education, vol. 23, no. 2, pp. 169–183, 1989.
[140] T. Christ, S. Sculin, A. Tolbize, and C. Jiban, “Implication of
[124] G. Tindal and R. Parker, “Identifying measures for evaluating
recent research: curriculum-based measurement of math com-
written expression,” Learning Disabilities Research and Practice,
putation,” Assessment for Effective Intervention, vol. 33, pp. 198–
vol. 6, pp. 211–218, 1991.
205, 2008.
[125] K. A. Gansle, A. M. VanDerHeyden, G. H. Noell, J. L. Resetar,
[141] A. Foegen, C. Jiban, and S. Deno, “Progress monitoring mea-
and K. L. Williams, “The technical adequacy of curriculum-
sures in mathematics. A review of the literature,” The Journal of
based and rating-based measures of written expression for
Special Education, vol. 41, no. 2, pp. 121–139, 2007.
elementary school students,” School Psychology Review, vol. 35,
no. 3, pp. 435–450, 2006. [142] M. K. Burns, A. M. VanDerHeyden, and C. L. Jiban, “Assessing
[126] K. L. McMaster and H. Campbell, “New and existing curricu- the instructional level for mathematics: a comparison of meth-
lum-based writing measures: technical features within and ods,” School Psychology Review, vol. 35, no. 3, pp. 401–418, 2006.
across grades,” School Psychology Review, vol. 37, no. 4, pp. 550– [143] T. J. Christ and O. Vining, “Curriculum-based measurement
566, 2008. procedures to develop multiple-skill mathematics computa-
[127] K. A. Gansle, G. H. Noell, A. M. VanDerHeyden et al., “An tion probes: evaluation of random and stratified stimulus-set
examination of the criterion validity and sensitivity to brief arrangements,” School Psychology Review, vol. 35, no. 3, pp. 387–
intervention of alternate curriculum-based measures of writing 400, 2006.
skill,” Psychology in the Schools, vol. 41, no. 3, pp. 291–300, 2004. [144] L. S. Fuchs, D. Fuchs, D. L. Compton, J. D. Bryant, C. L. Hamlett,
[128] C. Espin, B. Scierka, and S. Skare, “Criterion-related validity and P. M. Seethaler, “Mathematics screening and progress mon-
of curriculum-based measures in writing for secondary school itoring at first grade: implications for responsiveness to inter-
students,” Reading & Writing Quarterly, vol. 15, no. 1, pp. 5–27, vention,” Exceptional Children, vol. 73, no. 3, pp. 311–330, 2007.
1999. [145] C. Jiban and S. Deno, “Using math and reading curriculum-
[129] C. Espin, J. Shin, S. L. Deno, S. Skare, S. Robinson, and B. Ben- based measurements to predict state mathematics test perfor-
ner, “Identifying indicators of written expression proficiency for mance: are simple one-minute measures technically adequate?”
middle school students,” The Journal of Special Education, vol. Assessment for Effective Intervention, vol. 32, pp. 78–89, 2007.
34, no. 3, pp. 140–153, 2000. [146] P. Seethaler and L. Fuchs, “Using curriculum-based measure-
[130] J. W. Weissenburger and C. A. Espin, “Curriculum-based meas- ment to monitor kindergarteners’ mathematics development,”
ures of writing across grade levels,” Journal of School Psychology, Assessment for Effective Intervention, vol. 36, no. 4, pp. 219–229,
vol. 43, no. 2, pp. 153–169, 2005. 2011.
[131] C. Espin, T. Wallace, H. Campbell, E. S. Lembke, J. D. Long, and [147] E. Shapiro, L. Edwards, and N. Zigmond, “Progress monitoring
R. Ticha, “Curriculum-based measurement in writing: predict- of mathematics among students with learning disabilities,”
ing the success of high-school students on state standards tests,” Assessment for Effective Intervention, vol. 30, no. 2, pp. 15–32,
Exceptional Children, vol. 74, no. 2, pp. 174–193, 2008. 2005.
 ISRN Education 27

[148] B. Clarke, S. Baker, K. Smolkowski, and D. J. Chard, “An analysis [166] C. A. Espin and A. Foegen, “Validity of general outcome meas-
of early numeracy curriculum-based measurement: examining ures for predicting secondary students performance on content-
the role of growth in student outcomes,” Remedial and Special area tasks,” Exceptional Children, vol. 62, no. 6, pp. 497–514,
Education, vol. 29, no. 1, pp. 46–57, 2008. 1996.
[149] D. Chard, B. Clarke, S. Baker, J. Otterstedt, D. Braun, and R. [167] C. Espin, T. Busch, J. Shin, and R. Kruschwitz, “Curriculum-
Katz, “Using measures of number sense to screen for difficulties based measurement in the content areas: validity of vocabulary
in mathematics: preliminary findings,” Assessment for Effective matching as an indicator of performance in social studies,”
Intervention, vol. 30, no. 2, pp. 3–14, 2005. Learning Disabilities, vol. 16, pp. 142–151, 2001.
[150] E. Lembke, A. Foegen, T. Whittaker, and D. Hampton, “Estab- [168] C. A. Espin, J. Shin, and T. W. Busch, “Curriculum-based
lishing technically adequate measures of progress in early measurement in the content areas: vocabulary matching as
numeracy,” Assessment for Effective Intervention, vol. 33, no. 4, an indicator of progress in social studies learning,” Journal of
pp. 206–214, 2008. Learning Disabilities, vol. 38, no. 4, pp. 353–363, 2005.
[151] R. Martinez, K. Missal, S. Graney, O. Aricak, and B. Clarke, [169] C. Espin, T. Wallace, E. Lembke, H. Campbell, and J. Long,
“Technical adequacy of early numeracy curriculum-based mea- “Creating a progress-monitoring system in reading for middle
surement in kindergarten,” Assessment for Effective Intervention, school students: tracking progress toward meeting high-stakes
vol. 34, pp. 116–125, 2009. standards,” Learning Disabilities Research and Practice, vol. 25,
[152] A. M. VanDerHeyden, C. Broussard, and A. Cooley, “Further pp. 60–75, 2010.
development of measures of early math performance for pre-
[170] V. Nolet and G. Tindal, “Special education in content area class-
schoolers,” Journal of School Psychology, vol. 44, no. 6, pp. 533–
es: development of a model and practical procedures,” Remedial
553, 2006.
and Special Education, vol. 14, no. 1, pp. 36–48, 1993.
[153] J. Leh, A. Jitendra, G. Caskie, and C. Griffin, “An evaluation of
curriculum-based measurement of mathematics word problem [171] G. Roid and T. M. Haladyna, A Technology for Test-Item Writing,
solving measures monitoring third-grade students’ mathemat- Academic Press, Orlando, Fla, USA, 1982.
ics competence,” Assessment for Effective Intervention, vol. 32, [172] G. Tindal and V. Nolet, “Curriculum-based measurement in
pp. 90–99, 2007. middle and high schools: critical thinking skills in content
[154] L. Fuchs, D. Fuchs, and D. Compton, “The early prevention of areas,” Focus on Exceptional Children, vol. 27, no. 7, pp. 1–22,
mathematics difficulty: its power and limitations,” Journal of 1995.
Learning Disabilities, vol. 45, no. 3, Article ID 269, p. 257, 2012. [173] S. Engelmann and D. Carnine, Theory of Instruction: Principles
[155] A. Foegen, “Technical adequacy of general outcome measures and Applications, Irvington Publishers, New York, NY, USA,
for middle school mathematics,” Assessment for Effective Inter- 1982.
vention, vol. 25, pp. 175–203, 2000. [174] T. Twyman, J. McCleery, and G. Tindal, “Using concepts to
[156] A. Foegen, “Progress monitoring in middle school mathemat- frame history content,” Journal of Experimental Education, vol.
ics: options and issues,” Remedial and Special Education, vol. 29, 74, no. 4, pp. 331–349, 2006.
no. 4, pp. 195–207, 2008. [175] T. Twyman and G. Tindal, “Reaching all of your students in
[157] L. S. Fuchs, C. L. Hamlett, and D. Fuchs, Monitoring Basic Skills social studies,” Teaching Exceptional Children, vol. 1, no. 5,
Progress: Basic Math Computation, 2nd edition, 1998. article 1, 2005.
[158] L. S. Fuchs, C. L. Hamlett, and D. Fuchs, Monitoring Basic Skills [176] T. Twyman, L. R. Ketterlin-Geller, J. D. McCoy, and G. Tindal,
Progress: Basic Math Concepts and Applications, 1999. “Effects of concept-based instruction on an English language
[159] E. Lembke and P. Stecker, Curriculum-Based Measurement in learner in a rural school: a descriptive case study,” Bilingual
Mathematics: An Evidence-Based Formative Assessment Pro- Research Journal, vol. 27, no. 2, pp. 259–274, 2003.
cedure, RMC Research Corporation, Center on Instruction, [177] S. Embretson and S. Reise, Item Response Theory for Psycholo-
Portsmouth, UK, 2007. gists, Lawrence Erlbaum Associates, Mahwah, NJ, USA, 2000.
[160] C. Espin and G. Tindal, “Curriculum-based measurement for [178] T. Twyman and G. Tindal, “Extending curriculum-based meas-
secondary students,” in Advanced Applications of Curriculum- urement into middle/secondary schools: the technical adequacy
Based Measurement, M. R. Shinn, Ed., Guilford Press, New of the concept maze,” Journal of Applied School Psychology, vol.
York, NY, USA, 1998. 24, no. 1, pp. 49–67, 2007.
[161] R. Quirk, The Linguist and the English Language, Arnold, Lon-
[179] J. E. Heimlich and S. D. Pittelman, Semantic Mapping: Class-
don, UK, 1974.
room Applications, International Reading Association, Newark,
[162] D. Corson, “The learning and use of academic english words,” Del, USA, 1986.
Language Learning, vol. 47, no. 4, pp. 671–718, 1997.
[180] F. P. Hunkins, Teaching Thinking Through Effective Questioning,
[163] C. Espin and S. Deno, “Content-specific and general reading
Christopher-Gordon Publisher, Boston, Mass, USA, 1989.
disabilities of secondary-level students: identification and edu-
cational relevance,” The Journal of Special Education, vol. 27, pp. [181] G. Tindal and V. Nolet, “Serving students in middle school
321–337, 1993. content classes: a heuristic study of critical variables linking
[164] C. Espin and S. Deno, “Performance in reading from content instruction and assessment,” The Journal of Special Education,
area text as an indicator of achievement,” Remedial and Special vol. 29, no. 4, pp. 414–432, 1996.
Education, vol. 14, no. 6, pp. 47–59, 1993. [182] V. Nolet and G. Tindal, “Curriculum-based collaboration,”
[165] C. Espin and S. Deno, “Curriculum-based measures for sec- Focus on Exceptional Children, vol. 27, no. 3, pp. 1–12, 1994.
ondary students: utility and task specificty of text-based read- [183] V. Nolet and G. Tindal, “Instruction and learning in middle
ing and vocabulary measures for predicting performance on school science classes: implications for students with disabili-
content-area tasks,” Diagnostique, vol. 20, no. 1–4, pp. 121–142, ties,” The Journal of Special Education, vol. 28, no. 2, pp. 166–187,
1994. 1994.
 28 ISRN Education

[184] G. Tindal, V. Nolet, and G. Blake, Focus on Teaching and [202] D. Mellard, M. McKnight, and K. Woods, “Response to inter-
Learning in Content Classes, University of Oregon Behavioral vention screening and progress-monitoring practices in 41 local
Research and Teaching, Eugene, Ore, USA, 1992. schools,” Learning Disabilities Research & Practice, vol. 24, no. 4,
[185] V. Nolet, G. Tindal, and G. Blake, Focus on Assessment Learning pp. 186–195, 2009.
in Content Classes, University of Oregon Behavioral Research [203] C. H. Hofstetter, “Contextual and mathematics accommodation
and Teaching, Eugene, Ore, USA, 1993. test effects for English-language learners,” Applied Measurement
[186] L. Ketterlin-Geller and G. Tindal, Concept-Based Instruction: in Education, vol. 16, no. 2, pp. 159–188, 2003.
Science, University of Oregon Behavioral Research and Teach- [204] L. Fuchs and S. Deno, “Must instructionally useful performance
ing, Eugene, Ore, USA, 2002. assessment be based in the curriculum?” Exceptional Children,
[187] T. Twyman, L. Ketterlin-Geller, and G. Tindal, Concept-Based vol. 61, no. 1, pp. 15–24, 1994.
Instruction: Social Science, University of Oregon Behavioral [205] P. Stecker and E. Lembke, Advanced Applications of CBM in
Research and Teaching, Eugene, Ore, USA, 2002. Reading (K-6): Instructional Decision-making Strategies Manual,
[188] M. McDonald, L. Ketterlin-Geller, and G. Tindal, Concept- National Center on Student Progress Monitoring, Washington,
Based Instruction: Mathematics, University of Oregon Behav- DC, USA, 2011.
ioral Research and Teaching, Eugene, Ore, USA, 2002. [206] G. Tindal, J. Alonzo, J. F. T. Nese, and L. Saez, “Validating
[189] G. Tindal, J. Alonzo, and L. Ketterlin-Geller, Concept-Based progress monitoring in the context of RTI,” in Pacific Coast
Instruction: Language Arts, University of Oregon Behavioral Research Conference (PCRC), Coronado, Calif, USA, February
Research and Teaching, Eugene, Ore, USA, 2002. 2012.
[190] S. Berkeley, W. N. Bender, L. Gregg Peaster, and L. Saunders, [207] R. Gersten and J. A. Dimino, “RTI (Response to Intervention):
“Implementation of response to intervention: a snapshot of rethinking special education for students with reading difficul-
progress,” Journal of Learning Disabilities, vol. 42, no. 1, pp. 85– ties (yet again),” Reading Research Quarterly, vol. 41, no. 1, pp.
95, 2009. 99–108, 2006.
[191] J. Wanzek and C. Cavanaugh, “Characteristics of general educa- [208] J. M. Hintze and E. S. Shapiro, “Curriculum-based measurement
tion reading interventions implemented in elementary schools and literature-based reading: is curriculum-based measure-
with reading difficulties,” Remedial and Special Education, vol. ment meeting the needs of changing reading curricula?” Journal
33, no. 3, pp. 192–202, 2012. of School Psychology, vol. 35, no. 4, pp. 351–375, 1997.
[192] D. Barnett, N. Elliott, J. Graden et al., “Technical adequacy [209] J. Hintze, E. Shapiro, and J. Lutz, “The effects of curriculum on
for response to intervention practices,” Assessment for Effective the sensitivity of curriculum-based measurement in reading,”
Intervention, vol. 32, no. 1, pp. 20–31, 2006. The Journal of Special Education, vol. 28, pp. 188–202, 1994.
[193] S. Messick, “Validity,” in Educational Measurement, R. Linn, Ed., [210] L. Fuchs, G. Tindal, and S. Deno, Effects of Varying Item
pp. 13–103, Macmillan Publishing Company, New York, NY, Domains and Sample Duration on Technical Characteristics of
USA, 3rd edition, 1989. Daily Measures in Reading, University of Minnesota Institute
[194] A. VanDerHeyden, “Technical adequacy of response to inter- for Research on Learning Disabilities, Minneapolis, Minn, USA,
vention decisions,” Council for Exceptional Children, vol. 77, no. 1982.
3, pp. 335–350, 2011.
[211] Department of Education, Assisting Students Struggling With
[195] R. Gersten, T. Keating, and L. K. Irvin, “The burden of proof: Reading: Response to Intervention and Multi-Tier Intervention in
validity as improvement of instructional practice,” Exceptional the Primary Grades, Institute of Education Sciences, Washing-
Children, vol. 61, no. 5, pp. 510–519, 1995. ton, DC, USA, 2009.
[196] R. Allinder, “An examination of the relationship between [212] Department of Education, Improving Reading Comprehension
teacher efficacy and curriculum-based measurement and stu- in Kindergarten Through 3rd Grade, Institute of Education
dent achievement,” Remedial and Special Education, vol. 16, pp. Sciences, Washington, DC, USA, 2010.
247–254, 1995.
[213] Department of Education, WWC Evidence Review Protocol for
[197] R. M. Allinder, “When some is not better than none: effects
K-12 Students with Learning Disabilities, Institute of Education
of differential implementation of curriculum-based measure-
Sciences, Washington, DC, USA.
ment,” Exceptional Children, vol. 62, no. 6, pp. 525–535, 1996.
[214] E. R. O’Connor and P. Vadasy, The Handbook of Reading Inter-
[198] R. Allinder and M. BeckBest, “Differential effects of two
ventions, Guilford Press, New York, NY, USA, 2011.
approaches to supporting teachers’ use of curriculum-based
measurement,” School Psychology Review, vol. 24, pp. 287–298, [215] R. M. Schwartz, M. C. Schmitt, and M. K. Lose, “Effects of teach-
1995. er-student ratio in response to intervention approaches,” The
[199] R. M. Allinder, R. M. Bolling, R. G. Oats, and W. A. Gagnon, Elementary School Journal, vol. 112, no. 4, pp. 547–567, 2012.
“Effects of teacher self-monitoring on implementation of [216] L. Saez, “Instructional responsiveness: what are teachers
curriculum-based measurement and mathematics computation doing?” in Proceedings of the Pacific Coast Research Conference,
achievement of students with disabilities,” Remedial and Special Coronado, Calif, USA, Februrary 2012.
Education, vol. 21, no. 4, pp. 219–226, 2000. [217] B. Elbaum, S. Vaughn, M. T. Hughes, and S. W. Moody, “How
[200] P. M. Stecker, L. S. Fuchs, and D. Fuchs, “Using curriculum- effective are one-to-one tutoring programs in reading for ele-
based measurement to improve student achievement: review of mentary students at risk for reading failure? A meta-analysis
research,” Psychology in the Schools, vol. 42, no. 8, pp. 795–819, of the intervention research,” Journal of Educational Psychology,
2005. vol. 92, no. 4, pp. 605–619, 2000.
[201] M. K. Burns and B. V. Senesac, “Comparison of dual discrep- [218] R. E. O’Connor, “Phoneme awareness and the alphabetic princi-
ancy criteria to assess response to intervention,” Journal of ple,” in The Handbook of Reading Interventions, R. E. O’Connor
School Psychology, vol. 43, no. 5, pp. 393–406, 2005. and P. Vadasy, Eds., Guilford Press, New York, NY, USA, 2011.
 ISRN Education 29

[219] R. Spicuzza, J. Ysseldyke, A. Lemkuil, S. Kosciolek, C. Boys, and

E. Teelucksingh, “Effects of curriculum-based monitoring on
classroom instruction and math achievement,” Journal of School
Psychology, vol. 39, no. 6, pp. 521–542, 2001.
[220] S. Deno, “Individual differences and individual difference: the
essential difference of special education,” The Journal of Special
Education, vol. 24, pp. 160–173, 1990.
[221] D. L. Gast, Single-Subject Research Methodology in Behavioral
Science, Routledge, New York, NY, USA, 2010.
[222] R. H. Horner, E. G. Carr, J. Halle, G. Mcgee, S. Odom, and M.
Wolery, “The use of single-subject research to identify evidence-
based practice in special education,” Exceptional Children, vol.
71, no. 2, pp. 165–179, 2005.
[223] L. Fuchs, “Assessing intervention responsiveness: conceptual
and technical issues,” Learning Disabilities Research & Practice,
vol. 18, no. 3, pp. 172–186, 2003.
[224] M. K. Burns, S. E. Scholin, S. Kosciolek, and J. Livingston,
“Reliability of decision-making frameworks for response to
intervention for reading,” Journal of Psychoeducational Assess-
ment, vol. 28, no. 2, pp. 102–114, 2010.
[225] S. P. Ardoin, “The response in response to intervention: evaluat-
ing the utility of assessing maintenance of intervention effects,”
Psychology in the Schools, vol. 43, no. 6, pp. 713–725, 2006.
[226] A. M. VanDerHeyden, J. C. Witt, and D. W. Barnett, “The emer-
gence and possible futures of response to intervention,” Journal
of Psychoeducational Assessment, vol. 23, no. 4, pp. 339–361,
2005.
[227] J. B. Willet, “Measuring change more effectively by modeling
individual change over time,” in The International Encyclopedia
of Education, T. Husen and T. N. Postlethwaite, Eds., Pergamon
Press, Elmsford, NY, USA, 2nd edition, 1994.
Journal of
Criminology

Depression Research Journal of Journal of Journal of

Hindawi Publishing Corporation

and Treatment
Hindawi Publishing Corporation
Anthropology
Hindawi Publishing Corporation
Addiction
Hindawi Publishing Corporation
Archaeology
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014

Journal of Economics
Biomedical Education Research International
Hindawi Publishing Corporation Volume 2014 Hindawi Publishing Corporation Volume 2014
http://www.hindawi.com http://www.hindawi.com

Submit your manuscripts at

http://www.hindawi.com

International Journal of Education

Population Research Research International
Hindawi Publishing Corporation Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014

Child Development
Research

Sleep Disorders
Autism
Research and Treatment
Nursing
Research and Practice
Schizophrenia
Research and Treatment
Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014

Geography Psychiatry
Journal Journal
Current Gerontology
& Geriatrics Research

Journal of Urban Studies

Hindawi Publishing Corporation
Aging Research
Hindawi Publishing Corporation Hindawi Publishing Corporation
Research
Hindawi Publishing Corporation
Hindawi Publishing Corporation Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014
http://www.hindawi.com