A Synthesis of Peer-Mediated Academic Interventions for Secondary Struggling Learners

Abstract

A synthesis of the extant research on peer-mediated reading and math interventions for students in regular or alternative education settings with academic difficulties and disabilities in Grades 6 to 12 (ages 11–18) is presented. Interventions conducted between 2001 and 2012 targeting reading and math were included if they measured effects on at least one academic outcome measure. A total of 13 intervention studies were synthesized in which 10 studies employed an experimental or quasi-experimental design and three studies used a single-case design. Findings from the 13 studies revealed mostly moderate to high effects favoring peer mediation, particularly when implementing a peer-mediated feedback component. In addition, findings suggest such interventions have social validity among adolescents and teachers. More rigorous research on secondary peer-mediated math interventions, peer-mediated interventions in alternative settings, and effective ways to pair dyads to incorporate a structured feedback component is warranted. Implications for peer-mediated instruction for academically struggling adolescents are discussed.

Keywords

peer-mediated instruction academic interventions adolescents synthesis

The Common Core State Standards (CCSS) Initiative (National Governors Association Center for Best Practices & Council of Chief State School Officers, 2010) targets students’ college and career readiness skills and suggests more rigorous academic preparation is needed in U.S. middle and high schools (National Association of Secondary School Principals, n.d.). Although there is a growing body of research on the kinds of targeted interventions capable of helping students meet these rising standards (e.g., Kamil et al., 2008; Siegler et al., 2010), several characteristics of the upper grade levels can make it difficult to implement scientifically based reading and math interventions (Bulgren, Graner, & Deshler, 2013).

First, teachers feel pressured to cover a large amount of content in a limited amount of time and, therefore, tend to resist instructional adaptations meant to support students of lower ability levels if the teachers perceive those adaptations as deemphasizing the content (Siebert & Draper, 2008). In addition, secondary classes tend to be larger than elementary classes and represent a wider range of ability levels as well as a variety of behavioral issues (Blatchford, Bassett, & Brown, 2011). It is also common for schools with the greatest need to have limited resources to assist with providing evidence-based instruction in the classroom (Carr, Gray, & Holley, 2007). Finally, many adolescents exhibit low motivation and engagement after having struggled academically for many years (Casillas, Robbins, Allen, & Kuo, 2012; Docan, 2006).

As a result of these complicated factors, middle and high school teachers often opt to design lessons to circumvent addressing students’ poor skills such as avoiding the use of text altogether (McCulley, Katz, & Vaughn, 2013). Many consequences can result from such a decision. For example, avoiding or limiting exposure to text reading during the school day can result in students not acquiring the knowledge and skills they need to be successful in upper level content area classes or postsecondary settings (Mol & Bus, 2011). Poor academic achievement has led to soaring school dropout rates (Glennie, Bonneau, Vandellen, & Dodge, 2012), prompting educational leaders to describe the situation as nothing less than “a public health crisis” (Fuchs, Fuchs, & Compton, 2010, p. 26). Evidence of this crisis is also documented in the numbers of adolescents with academic difficulties who are in incarcerated settings (Bloomberg, Bales, Mann, Piquero, & Berk, 2011). Hence, student outcomes often hinge on the decisions teachers make about the design of their instruction. Given the challenges of secondary schools just described, it is important to consider the research on instructional delivery practices.

Effective Instructional Delivery Practices

Besides identifying the important content of reading and mathematics instruction (e.g., Kamil et al., 2008; Siegler et al., 2010), previous research has also begun to identify the features influencing the effective delivery of that instruction (Pashler et al., 2007). It is recommended that, despite the content focus, instruction for all students include the following features: modeling, explicit instruction, opportunities for students to respond, immediate corrective feedback, increased time on task, and systematic instruction (Archer & Hughes, 2011; Vaughn, Wanzek, Murray, & Roberts, 2012). All students benefit from these features of effective instruction, but they are essential for students who struggle (Rupley, Blair, & Nichols, 2009). Results of a synthesis support this notion. Scammacca et al. (2007) concluded interventions for adolescents are more effective when designed and delivered by researchers, perhaps because they maintain higher fidelity to the content and features of delivery than teachers do. The authors also found improved outcomes when trained peers delivered interventions.

Peer-Mediated Instruction

Having students of the same age tutor each other or work together as partners or in small groups to complete assignments has been referred to as peer-mediated instruction (Maheady, Harper, & Sacca, 1988). Numerous studies have been conducted using peer-mediated instruction in core academic classrooms and supplemental reading or math intervention settings to improve student outcomes, particularly at the elementary level (e.g., Fuchs et al., 1997; Fuchs, Fuchs, Hamlett, Phillips, & Karns, 1995; Greenwood, Delquadri, & Carta, 1988; Klingner & Vaughn, 1996). Potential benefits of peer-mediated instruction include the ability to provide ongoing corrective feedback in a timely manner and increased opportunities to respond and practice (Hattie & Timperley, 2007).

Previous Research

Although we have a great deal of positive evidence supporting the use of peer-mediated interventions at the elementary level, less is known about the effectiveness of these interventions for secondary students. Three previously published reviews examined, in whole or part, the effects of peer-mediated interventions on academic outcomes of secondary students with academic difficulties and disabilities (Kunsch, Jitendra, & Sood, 2007; Okilwa & Shelby, 2010; Stenhoff & Lignugaris/Kraft, 2007). Below we highlight the findings of each of these reviews and identify the aspects or applications warranting further examination.

Kunsch et al. (2007) reviewed 17 studies published between 1978 and May 2006 that investigated the effectiveness of a peer-mediated learning on the mathematics performance of students in grades K–12 identified as or at risk of having a learning disability. Single-case research design (SCRD) studies were excluded, as well as any studies that lacked sufficient quantitative data to calculate effect sizes. Within the corpus, only three studies were conducted with students in Grades 6–12 (Bahr & Rieth, 1991; Calhoon & Fuchs, 2003; Roach, Paolucci-Whitcomb, Meyers, & Duncan, 1983). The seven effect sizes from these studies ranged from d = −0.31to 0.53, and findings were most favorable for students at risk for mathematics disabilities and interventions that targeted math computations. Secondary students did not demonstrate improvements as robust as those for elementary participants. The authors did not report social validity findings. Given more recent changes to mathematics instruction with the advent of the Principles and Standards for School Mathematics (National Council of Teachers of Mathematics, 2000) and the CCSS (National Governors Association Center for Best Practices & Council of Chief State School Officers, 2010), the findings for the limited number of studies identified in secondary grades need updating.

Stenhoff and Lignugaris/Kraft (2007) reviewed 20 studies published between 1980 and 2005 that investigated peer tutoring research conducted with students with mild disabilities in Grades 7–12 in general and special education classrooms and one correctional facility school. The authors excluded any studies that lacked sufficient quantitative data to calculate effect sizes or percentage of nonoverlapping data (PND). They found that peer tutoring in secondary settings resulted in improved academic performance of students with mild disabilities. Moreover, training tutors to implement instruction prior to beginning peer tutoring sessions produced a large effect on tutee outcomes (d = 0.25 to 1.15; PND = 50–98%). Eight studies implemented reciprocal tutoring, nine studies implemented heterogeneous grouping, two studies implemented homogenous grouping, and one study implemented cross-age. The authors identified heterogeneous peer tutoring as the only type of evidence-based practice for secondary students based on the SCRD studies. Monitoring peers during the tutoring sessions was also considered effective based on findings of the SCRD studies reviewed. As with the Kunsch et al. (2007) synthesis, Stenhoff and Lignugaris/Kraft did not examine any social validity results reported in the studies. Moreover, they excluded students in Grade 6 and those who were low achieving or at risk for but not yet identified as having a disability.

Most recently, Okilwa and Shelby (2010) reviewed 12 studies published between 1997 and 2007 that investigated the effects of peer tutoring on academic performance of students with disabilities in Grades 6–12. Each of the 12 studies implemented peer tutoring (defined as classwide peer tutoring [CWPT], peer-assisted learning strategies [PALS], or reciprocal peer tutoring) in at least one core content area: English language arts (ELA), mathematics, science, and/or social studies. Peer tutoring was reported as effective, regardless of content area, at improving academic outcomes for students receiving special education services in both general education and special education settings. With the exception of two studies (Mastropieri et al., 2001; Mastropieri, Scruggs, Spencer, & Fontana, 2003), the authors did not independently calculate effect sizes or PND for SCRD studies. Rather, they reported the effects as provided in the original studies. There was a wide range in effects across all measures (d = −0.19 to 2.20), with more robust outcomes reported on researcher-developed measures specific to the content taught. Furthermore, the authors did not synthesize social validity findings or evaluate the overall quality of the studies in their corpus.

Rationale and Research Questions

We extend the findings of these previously conducted syntheses in several ways. We include more recently published studies and allow for implementation of peer-mediated learning in math and reading instruction, as well as studies that include peer-mediated instructional support for reading content area text. Because many adolescents who continue to struggle academically are some of the same students who end up in some type of alternative or incarcerated setting, it is essential to examine the effectiveness of these practices with adolescents who struggle to learn across a variety of settings (Allensworth & Easton, 2005; Balfanz, Herzog, & MacIver, 2007). We, therefore, extended our search to include studies implemented in alternative settings such as juvenile justice settings and residential treatment facilities. In addition to including studies that use either treatment-comparison or SCRD designs, we synthesize social validity findings across studies. Finally, we independently calculate the effect sizes for treatment-comparison or PND for SCRD studies, and we analyze the studies against accepted quality indicators (Gersten et al., 2005; Raudenbush, 2005; Shadish, 2002; What Works Clearinghouse, 2010). Our purpose in conducting this synthesis was to examine the effects of peer-mediated interventions for students with reading and math difficulties in Grades 6–12 to better understand effective practices for adolescents who struggle with learning. We addressed the following research questions:

How effective is peer-mediated intervention implemented in the regular or alternative school settings on academic outcomes for low-achieving adolescents with learning difficulties in Grades 6–12?

What are the features (e.g., type, duration, grouping, feedback) of peer-mediated intervention for low-achieving adolescents with learning difficulties in Grades 6–12?

What evidence of the social validity of peer-mediated intervention exists in studies conducted with low-achieving adolescents with learning difficulties in Grades 6–12?

Method

Search Procedure and Criteria

We identified studies through a multistep process that included an exhaustive search of the literature using a combination of search methods including electronic searches, manual searches of several journals, and an ancestral search of the literature. The methods used to conduct the search and procedures used during coding were developed by several of the authors through prior syntheses (Reed, Sorrells, Cole, & Takakawa, 2013; Wexler, Pyle, Flower, Williams, & Cole, 2013). First, we conducted the electronic search using the ERIC, PsycINFO, and Academic Search Complete databases for studies published between 2001 and July 2012 that spans more than 10 years and represents the most current research. The following search terms were used in various combinations in the electronic search: at-risk, low achiev*, struggling, below grade level, learning dis*, peer partner*, peer tutor*, peer mentor*, peer mediat*, reciprocal tutor*, reciprocal teaching, collaborat* group*, small group*, cooperative learning, peer-assist*.

Second, to identify any articles that may not have appeared in our electronic search, an additional hand search was conducted of the following 16 journals for articles published between the years of 2010 and July 2012: Annals of Dyslexia, Exceptional Children, Journal of Educational Psychology, Journal of Learning Disabilities, Journal of Special Education, Learning Disability Quarterly, Learning Disabilities Research and Practice, Reading Research Quarterly, Reading and Writing Quarterly, Remedial and Special Education, Scientific Studies of Reading, Journal of Correctional Education, Behavioral Disorders, Journal of Offender Rehabilitation, Journal of Emotional and Behavioral Disorders, Journal of Offender Counseling, Services & Rehabilitation. These journals were selected for manual search because their aims include publishing studies relevant to the field of remedial and special education (the critical contributors of our target population as described in the criteria), and/or they were identified in the electronic search as including studies specifically about peer-mediated learning.

More than 4,000 abstracts were identified and evaluated for inclusion in the synthesis based on the following criteria:

Studies appeared in a peer-reviewed journal.

The articles were published in English.

Participants were identified as at risk, low achieving, struggling, below grade level, or diagnosed with a learning disability (LD). The designation could pertain to students’ reading, literacy, and/or math performance. Studies including students without these designations were included if the data were disaggregated for the population of interest.

Participants were enrolled in regular or alternative education settings, including juvenile justice or residential treatment facilities.

Participants were in Grades 6–12 or ages 11–18. Studies with older or younger students were included as long if one of two conditions were met: (a) the majority were in the grades or of the ages pertinent to this synthesis, or (b) data were disaggregated for the specified grades or ages of interest.

Studies targeted math or reading instruction, including reading in content areas such as social studies, using a structured peer-mediated format with either partners or small groups of students. Studies were excluded if they only described allowing students to work together cooperatively or collaboratively without teaching them particular tutoring steps to follow in doing so (Ridlon, 2009).

The research design was treatment-comparison, SCRD, or single-group design.

Outcomes included at least one academic measure assessing students’ literacy, math, or content learning and were reported with sufficient data for calculating effect sizes.

A total of 13 studies met all selection criteria for the synthesis. We then conducted an ancestral search of the reference lists in these publications to determine if any relevant studies might have been missed in the electronic and manual searches. No new studies were identified that had not already been evaluated for inclusion.

Coding Procedures

An extensive coding sheet was developed to examine the essential elements of each study. The coding sheet contained the following information: (a) participants, (b) methodology, (c) intervention and comparison information, (d) clarity of causal inference, (e) measures, (f) findings, (g) implications and limitations, and (h) social validity information. Four raters were trained on the use of the coding sheet. Given the variety of studies included in the corpus, we believed that establishing interrater reliability on a select article or a small subset of articles would be insufficient for ensuring accuracy. Therefore, we planned to have multiple raters for each of the 13 studies.

One rater initially scored all the single case design studies (n = 3), that were then double scored by a second rater. Disagreements were resolved in consultation with an expert in SCRD who served as a collaborating author on this work. All group design studies (n = 10) were independently scored by two of the four reviewers. Some disagreements, such as the classification of an intervention type, were resolved in discussion between the two coders. Other disagreements, such as an error in the calculation of an effect size, were resolved by a third rater.

Effect Size Calculation

For treatment-comparison design studies, the effect size, Hedges’s g, was calculated as the difference in means between the treatment and comparison condition divided by the pooled standard deviation with a correction for small sample sizes (Hedges, 1981). Effect sizes were interpreted as a small effect if g = 0.20–0.49, medium effect if g = 0.50–0.79, and large effect if g = 0.80 or greater (Cohen, 1988). Outcomes from SCRD studies were calculated as PND (Scruggs, Mastropieri, & Casto, 1987). PND is a computational method for quantitatively synthesizing data from single case studies. The benefit of using PND is that it is not compromised by serial dependency in the data. However, it is compromised by variability in baseline data, a baseline datum point at the floor or ceiling, and trends in data series (Wolery, Busick, Reichow, & Barton, 2010); thus, results should be interpreted with caution. PND is calculated as the percentage of data points during the treatment phase that are higher than the highest datum point from the baseline phase (Scruggs et al., 1987).

Results

Data Analysis Plan

A range of study designs and peer-mediated intervention types were represented in this synthesis. We first present synthesized study features (e.g., study design and quality) to highlight similarities and differences across the corpus of studies. We then present effects on reading and math outcomes by peer-mediated intervention type (e.g., CWPT) across all studies, including treatment-comparison and SCRD studies. Finally, we present synthesized social validity findings.

Study Features

A total of 13 studies, reported in peer-reviewed journals between the years of 2001 and July 2012, met the criteria for inclusion in this synthesis. Studies were distributed across the years of interest with the largest number of studies published in 2003 (n = 3) and 2006 (n = 3). A summary of the study characteristics is reported in Table 1.

Table 1.

Study Characteristics.

Study and Design	Description of Conditions	Sessions	Effect Sizes
Treatment-comparison designs
1. Calhoon & Fuchs (2003)Quasi-experimental (classes randomly assigned to condition)	T: Both students in pair serve as tutor and tutee in each session, peer tutor used verbal statements or questions to guide the tutee through concrete knowledge of the math skill. On the first problem set, the tutor provided positive or corrective feedback. The tutee completed the second problem set independently and explained the work to the tutor. Then the pair switched roles for a third and fourth problem set.C: Students worked independently on the Buckle Down! on Tennessee Mathematics workbook to practice skills on the state assessment.	Number of sessions: 30Frequency of sessions: two times a week × 15 weeksLength of sessions: 30 min	Math Operations Test–Revised^a: T v C = 0.40Math Concepts and Applications Test^a: T v C = 0.26Math portion of the Tennessee Comprehensive Achievement Test^a: T v C = 0.07
2. Calhoon (2005)Quasi-experimental (teachers randomly assigned to condition)	T: Three times per week, the teacher began with scripted phonics lessons and directed practice. Peer partners then practiced the skills using a linguistic coding system. On alternating days, students engaged in partner reading. The peer tutor read first to serve as a model. After both partners have read, one student retold the story sequence for 2 min. Then the students summarized the text by composing a main idea statement before doing a prediction activity.C: Teachers implemented Saxon Phonics Interventions (explicit, systematic instruction) using scripted lessons in a whole-class format. Paired with SRA Skills Acquisition.	Number of sessions: NRFrequency of sessions: Saxon phonics 3 days a week × 31 weeks; LST/PALS and SRA Skills Acquisition 2 days a week × 31 weeksLength of sessions: Saxon phonics and LST/PALS 40 min; SRA Skills Acquisition 35 min	WJ-III letter word identification^a: T v C = 0.95WJ-III passage comprehension^a: T v C = 0.84WJ-III word attack^a: T v C = 1.01WJ-III reading fluency^a: T v C = 0.09
3. Gersten, Baker, Smith-Johnson, Dimino, & Peterson (2006)Experimental (random assignment of matched LD/non-LD)	T: Students watched short segments of the Eyes on the Prize documentary (with teacher clarifications and questions), completed four compare–contrast text structure activities, and were taught narrative text structures to generate discussion. Students worked in pairs to complete tasks, alternate reading aloud to each other, and discuss using a Think–Pair–Share.C: Students used the same curriculum content as the treatment students but answered questions and completed tasks individually.	Number of sessions: school 1 = NR; school 2 = 25Frequency of sessions:school 1 = NR; school 2 = dailyLength of sessions: school 1 = 52 min; school 2 = 41 min	Vocabulary matching task: T v C = 0.60Written exam: T v C = 1.12Content interview: T v C = 0.75
4. Kim et al. (2006)Quasi-experimental (class sections randomly assigned to condition)	T: Students learned interactive comprehension strategies (preview, click and clunk, gist, and wrap-up) and worked in pairs to partner read, discuss all questions, and answer questions about the passages on the computer assisted program.C: Teacher provided fluency instruction, timed readings with student partners, vocabulary instruction, and comprehension instruction.	Number of sessions: 17–23Frequency of sessions: two times a weekLength of sessions: 50 min	WRMT-R Passage Comp^a: T v C = 0.48CSR 4th grade gist: T v C = 0.92CSR 4th grade comprehension questions: T v C = 1.14CSR instructional-level gist = 0.76CSR instructional-level comprehension questions = 0.82
5. Mastropieri et al. (2001)Quasi-experimental (teachers randomly assigned to condition)	T: Students used partner reading with corrective feedback of oral reading errors. Students took turns restating the story and generating a gist statement to summarize what they read.C: Traditional whole-class instruction that included teacher questioning, oral and silent reading, worksheet activities, activating prior knowledge, making predictions, and vocabulary development with the same materials as in the treatment condition.	Number of sessions: NRFrequency of sessions: daily × 5 weeksLength of sessions: 50 min	Reading comprehension posttest: T v C = 1.12
6. Mastropieri, Scruggs, Spencer, & Fontana (2003)Quasi-experimental (classes randomly assigned to condition)	T: Students used partner reading with corrective feedback of oral reading errors. Tutoring pairs summarized the passage and responded to questions. Teachers awarded points for correctly following procedures.C: Teachers provided daily review and introduction of new information. Students then took turns reading orally. After reading, students independently used guided study notes with teachers’ assistance. Teachers awarded points for correctly following procedures.	Number of sessions: approximately 18 including pre- and posttestingFrequency of sessions: every other day × 9 weeksLength of sessions: 90 min	Reading fluency rate: T v C = −0.27Reading fluency errors: T v C = −0.36Reading comprehension (generate summary sentence) : T v C = 1.95World history content area knowledge: T v C = 0.56Content Chapter 1: T v C = 1.02Content Chapter 2: T v C = 1.04Content Chapter 3: T v C = 1.70Unit open-ended test: T v C = 1.29Unit multiple choice: T v C = 2.03End of year not from unit: T v C = 0.05End of year from unit: T v C = 1.14
7. Scruggs, Mastropieri, & Marshak (2012)Quasi-experimental (classes randomly assigned to condition)	T: Students worked in pairs to discuss content and alternated asking each other questions. They worked sequentially through fact sheets in a self-paced fashion and took pre- and posttests as proxies for high-stakes tests. Students recorded their performance.C: Teachers provided daily review and introduction of new information. Students then took turns reading orally. Students participated in answering teacher questioning of content, taking notes independently, and in completing relevant seatwork.	Number of sessions: NRFrequency of sessions: alternating between three times a week and two times a week (on A-B block schedule) × 18 weeksLength of sessions: 50 min if two times a week and 90 min if three times a week (alternating A-B schedule)	Measure of target items (direct effects)ES = 0.78Measure of nontarget items (indirect effect)ES = 0.67
8. Spencer, Scruggs, & Mastropieri (2003) Quasi-experimental (classes randomly assigned to condition; within subject crossover)	T: Students implemented partner reading with corrective feedback of errors. They took turns guiding each other in summarizing the content with a gist statement and in asking each other review questions. C: Teachers began with a whole-class discussion. Then, students took turns reading the chapter out loud before independently completing guided notes or worksheets. The teacher occasionally used supplemental videos and assigned projects or worksheets as homework.	Number of sessions: NR Frequency of sessions: daily × 2 weeks Length of sessions: 35 min	Weekly quizzes Peer tutoring first ES = 0.74 Traditional first ES = 0.91 Multiple-choice chapter tests Peer tutoring first ES = 0.61 Traditional first ES = 1.04 Open-ended chapter tests Peer tutoring first ES = 0.43 Traditional first ES = 0.40 Percentage on-task behavior^b Peer tutoring first ES = 0.95 Traditional first ES = 0.37
9. Vaughn et al. (2011) Experimental (randomized block design; students randomized to classes and classes randomized to conditions)	T: Teachers modeled each comprehension strategy and implemented short lessons to provide practice and review. Students assigned to roles within small groups and read through the text applying the strategies (use learning logs to record previews, gists, clunks, and summaries). C: Used the typical English language arts curriculum.	Number of sessions: 24 to 48 sessions reported by teachers; mean number of sessions = 34.62 (7.37 SD) Frequency of sessions: two times a week × approximately 18 weeks Length of sessions: 50 min	Gates-MacGinitie reading comprehension^a: T v C (low on TOWRE) = 2.73 T v C (low on state test) = 2.34 AIMSweb Maze^a: T v C (low on TOWRE) = −0.10 T v C (low on state test) = 0.08 TOSREC^a: T v C (low on TOWRE) = 0.46 T v C (low on state test) = 1.66
10. Wexler, Vaughn, Roberts, & Denton (2010) Experimental (students were paired within class and then randomly assigned by pair to condition)	T1: In the repeated reading condition, students worked in pairs to read the same texts three times each while providing error correction for each other. They then summarized the text and graphed their words read correctly. T2: In the wide reading condition, students worked in pairs to read three different texts each. They provided similar error correction and also summarized the texts and graphed words read correctly as in the repeated reading condition. C: Students participated in the typical practice reading instruction, which emphasized preparation for the reading portion of the state accountability test by reading and answering comprehension questions independently with minimal feedback.	Number of sessions: 50 Frequency of sessions: daily × 10 weeks Length of sessions: 15–20 min	WJ-III letter word identification^a: T1 v C = −0.05 T2 v C = −0.11 T1 v T2 = 0.06 WJ-III passage comprehension^a: T1 v C = −0.10 T2 v C = −0.19 T1 v T2 = 0.10 TOSRE^a: T1 v C = −0.31 T2 v C = −0.21 T1 v T2 = 0.10 AIMSWeb^a: T1 v C = −0.08 T2 v C = −0.26 T1 v T2 = 0.18 TOSCRF^a: T1 v C = 0.04 T2 v C = −0.23 T1 v T2 = 0.28
Single-case designs
11. Dufrene et al. (2010) Multiple baseline across participants, AB	T: Peer tutor modeled oral reading of the passage while tutee followed along. Then, tutee read the passage aloud two times in a row to build fluency. No error correction was provided. The third time the tutee read the passage, the tutor recorded the words read correctly and gave tickets for positive behavior support.	Number of sessions: baseline = 6, intervention = 20 Frequency of sessions: one student received three times per week; others received an average 2.1, 1.7, or 2 per week. Length of sessions: 9.5 min on average	DIBELS-ORF Grade 6 passages^a: Jan PND = 13/20 (65%); minimally effective Donna PND = 8/20 (40%); minimally effective Darrian PND = 8/20 (40%); minimally effective Tanya PND = 3/20 n(15%); ineffective
12. Hughes & Fredrick (2006) Multiple probe across behaviors and replicated across students	T: Teacher pronounced and defined words. Students then practiced orally reading the words in unison three times. Using a time delay procedure, peer tutors read definitions from index cards. Tutee was to say the word before the printed card was presented. Tutor delivered positive or corrective feedback and had tutee write the word. They reshuffled the deck of word cards and continued practicing for 8 min before switching roles. The teacher awarded points for correctly following the procedures.	Number of sessions: Set A = 6–7 sessions (median 4.0 accuracy, 5.0 rate) and 175–280 trials; Set B = 3–6 sessions (median 3.0 accuracy, 4.0 rate) and 135–161 trials; Set C = 3–6 sessions (median 3.5 accuracy, 3.5 rate) and N/A - 265 trials. Frequency of sessions: Daily Length of sessions: approximately 16 min (8 min per partner)	Measure of trained vocabulary words in sets of 5 (similar to matching). Graded on accuracy and rate (number completed in a minute): Set A = N/A for Student 1, 100% for Student 2 (highly effective), 100% for Student 3 (highly effective) Set B = N/A for Student 1, 100% for Student 2 (highly effective), 90% for Student 3 (highly effective) Set C = 71% for Student 1 (moderately effective), 100% for Student 2 (highly effective), 89% for Student 3 (moderately effective)
13. Veerkamp, Kamps, & Cooper (2007) Reversal ABAB design modified to accommodate a 3rd “C” condition (A = CWPT, B = CWPT plus lottery, C = student academic achievement as measured by weekly tests). Class A: ABABCACACA Classes B and C: ABCACACA	T: Teacher read the entire text aloud while students followed along. Then, partners took turns orally reading the text. The student acting as tutor awarded points for correctly read sentences and provided error correction. In between partners’ turns, students asked each other comprehension questions. Initially the questions were scripted, but students eventually students generated their own using question starter cards. The student acting as tutor provided corrective feedback and awarded points for correct responses. After 5 min, the students reversed roles. During vocabulary instruction, teachers read the list of words for the day before tutors quizzed each other by holding up the word card for the tutee to read and provide a definition. The tutor delivered corrective feedback and awarded points for correct responses. In the treatment plus lottery condition, the instruction was the same except teachers awarded tickets for good tutoring and on-task behavior. Tickets were drawn for prizes (i.e., the lottery).	Number of sessions: NR Frequency of sessions: daily × 19 weeks Length of sessions: 43 min × four days a week and 25 min once a week	PND for weekly tests during CWPT phase 1: 4/7 (57%) Thomas = minimally effective, 0/7 (0%) Cathy and Michelle = ineffective. PND for weekly tests during CWPT phase 2: 0/2 (0%) Thomas = ineffective, 1/2 (50%) Cathy and Michelle = minimally effective. PND for weekly tests during CWPT+ Lottery phase 1: 3/4 (75%) all students = moderately effective. PND for weekly tests during CWPT+ Lottery phase 2: 2/3 (67%) all students = minimally effective. PND for ORF-WCPM during CWPT novel 1: 2/4 (50%) Thomas and Michelle = minimally effective, 0/4 (0%) Cathy = ineffective PND for ORF-WCPM during CWPT novel 2: 3/4 (75%) Thomas = moderately effective, 1/4 (25%) Cathy = ineffective, 0/4 (0%) Michelle = ineffective PND for ORF-WCPM during CWPT novel 3: 2/3 (67%) Thomas and Cathy = minimally effective, 3/4 (75%) Michelle = moderately effective.

Note. C = comparison; CSR = collaborative strategic reading; CTD = constant time delay; CWPT = classwide peer tutoring; DIBELS = Dynamic Indicators of Basic Early Literacy Skills; ES = effect size; LD = learning disability; LST = linguistics skills training; min = minutes; N/A = not applicable; NR = not reported; ORF = oral reading fluency; PALS = peer-assisted learning strategies; PND = percentage of nonoverlapping data; reciprocal = alternated roles as tutor and tutee; T = treatment; TOSCRF = Test of Silent Contextual Reading Fluency; TOSRE = Test of Silent Reading Efficiency; TOSREC = Test of Silent Reading Efficiency and Comprehension; TOWRE = Test of Word Reading Efficiency; WJ-III = Woodcock–Johnson 3rd Edition; WRMT-R = Woodcock Reading Mastery Test–Revised.

Standardized measure. ^bObservational measure.

Study design and quality

A total of 10 studies used a treatment-comparison design and three studies used a SCRD. Several articles and documents have proposed quality indicators and standards for experimental and quasi-experimental studies (Gersten et al., 2005; Raudenbush, 2005; Shadish, 2002). Because of their contribution to the validity of the study findings, we elected to highlight several characteristics of quality experimental and quasi-experimental studies that we derived from the quality indicators and standards set forth by the What Works Clearinghouse (WWC, 2010) and Gersten et al. (2005). An analysis of the quality indicators of the treatment-comparison studies in this synthesis is reported in Figure 1. These characteristics include (a) using random assignment to form two groups of study participants, (b) reporting a replicable description of the independent and dependent variable, (c) reporting fidelity of implementation, (d) using multiple outcome measures to provide an appropriate balance between measures closely aligned with the intervention and measures of generalized performance, (e) reporting the reliability and validity of outcome measures used, (f) reporting enough data to compute effect sizes, and (g) accounting for attrition rates within acceptable limits (i.e., less than 30%).

Figure 1.

Quality indicators for treatment-control studies.

As expected from our delimiting criteria, all studies used random assignment of students, classes, or teachers to condition. Two of the 10 experimental treatment-comparison studies used random assignment of students to conditions (Gersten, Baker, Smith-Johnson, Dimino, & Peterson, 2006; Wexler, Vaughn, Roberts, & Denton, 2010), and 6 of the studies used random assignment of classes to condition (Calhoon & Fuchs, 2003; Kim et al., 2006; Mastropieri et al., 2003; Scruggs, Mastropieri, & Marshak, 2012; Spencer, Scruggs, & Mastropieri, 2003; Vaughn et al., 2011). The remaining 2 of the 10 studies used random assignment of teacher to condition (Calhoon, 2005; Mastropieri et al., 2001). Authors of all 10 studies sufficiently contrasted their treatment and comparison conditions and addressed fidelity of implementation. However, some studies either did not specify how treatment checks were conducted (Mastropieri et al., 2001; Spencer et al., 2003) or did not evaluate fidelity more than one time during implementation (Calhoon, 2005; Calhoon & Fuchs, 2003).

Figure 1 also identifies those studies that reported the use of multiple outcome measures (n = 7; Calhoon & Fuchs, 2003; Gersten et al., 2006; Kim et al., 2006; Mastropieri et al., 2003; Spencer et al., 2003; Vaughn et al., 2011; Wexler et al., 2010) or multiple subtests from one outcome measure (n = 1; Calhoon, 2005). Only two studies did not use multiple outcome measures or subtests (Mastropieri et al., 2001; Scruggs et al., 2012). Of the 10 treatment-comparison studies, 5 used standardized outcomes measures (Calhoon, 2005; Calhoon & Fuchs, 2003; Kim et al., 2006; Vaughn et al., 2011; Wexler et al., 2010), and the remaining 5 studies exclusively used measures with insufficient evidence of reliability and validity (Gersten et al., 2006; Mastropieri et al., 2001; Mastropieri et al., 2003; Scruggs et al., 2012; Spencer et al., 2003). Kim et al. (2006) used researcher-developed measures in addition to standardized measures. All 10 studies provided enough data to compute effect sizes on outcome measures and reported acceptable attrition rates. Summarizing across Figure 1, 5 of the 10 treatment-comparison studies met all quality indicators for treatment-comparison studies (Calhoon, 2005; Calhoon & Fuchs, 2003; Kim et al., 2006; Vaughn et al., 2011; Wexler et al., 2010).

SCRD studies are able to demonstrate experimental control and document functional relations, or casual relations where the value of the dependent variable changes as a function of the independent variable. Thus, they can be used to identify evidenced based practices (Kazdin, 1982). Moreover, SCRD studies help identify what works for whom under what conditions (Horner et al., 2005). However, the rigor and quality of SCRD must be considered when determining if experimental control was established and when synthesizing across studies to establish evidence-based practices. Two sets of quality indicators for SCRD studies were used to analyze the rigor of those included in this review. First, the Horner et al. (2005) quality indicators have been used in recent literature syntheses on a variety of topic areas (e.g., Barton & Wolery, 2008). The indicators were used to create seven criteria for documenting the quality and rigor of each study as well as determining whether the studies established experimental control and demonstrated a functional relation. The text and graphed data from each of the three SCRD studies reviewed here were analyzed for the following seven criteria, that also serve as the descriptors in Figure 2: (a) presence of a stable baseline, (b) amount of overlap across conditions, (c) consistency of change across tiers (i.e., replications), (d) an adequate number of attempts at a treatment effect (i.e., at least one demonstration of the treatment effect and two replications of the treatment effect), (e) reporting of interobserver agreement (IOA) across at least 20% of participants and conditions with a minimum average of 80%, (f) adequate descriptions of the independent variable and the baseline condition with replicable precision, and (g) reporting of intervention fidelity.

Figure 2.

Evidence of experimental control by study (Horner et al., 2005).

All three studies had at least three attempts to demonstrate a treatment effect and described their independent variable with replicable precision. However, one study did not have a stable baseline (Veerkamp, Kamps, & Cooper, 2007), two studies had significant overlap between conditions, had inconsistent changes across tiers, and did not adequately report IOA (Dufrene et al., 2010; Veerkamp et al., 2007), and one did not report intervention fidelity (Veerkamp et al., 2007). Therefore, of the three SCRD studies, only Hughes and Fredrick (2006) met all quality indicators.

Second, we analyzed these three studies using criteria designed in previous school-based research syntheses (e.g., Maggin & Chafouleas, 2011; Maggin, Chafouleas, Mosely, & Johnson, 2011; Maggin, Johnson, Chafouleas, Ruberto, & Berggren, 2012) based on the recommendations of the WWC (Kratochwill et al., 2011). The initial five criteria presented in Figure 3 were scored using a dichotomous scale (i.e., present, not present) at the study level, and the final one design standards were coded at the case level. The final criteria identifies the number of data points per condition. According to the recommendations of the WWC, at least five data points per condition are ideal, and fewer than three data points per condition would suggest the SCRD study cannot contribute to establishing a functional relation (Kratochwill et al., 2011).

Figure 3.

Application of WWC single-case research design standards.

All three studies met four of the six criteria in Figure 3. That is, all the SCRD studies included in this review systematically manipulated the independent variable, measured the dependent variable repeatedly, reported IOA at higher than 80%, and had three attempts at demonstrating a treatment effect. However, only Hughes and Fredrick (2006) met all six criteria, although the study only had three data points in multiple conditions. The strength of evidence of Hughes and Fredrick (2006) was further analyzed using the WWC SCRD procedures for visual analysis (Maggin & Chafouleas, 2011). The visual analysis clearly indicated the presence of a functional relation. Because the other two studies did not meet all quality indicators and Hughes and Fredrick (2006) had only three data points, interpretations of the peer-mediated intervention literature are limited.

Sample characteristics

Data on study participants and settings are reported in Table 2. The 13 studies included 653 total students identified as at risk, low achieving, struggling, below grade level, or LD. Sample sizes of the targeted students in the studies ranged from 3 to 253, with an average of 64 participants (Mdn = 35) for treatment-comparison studies and an average of three subjects for SCRD studies. A total of 643 students participated in treatment-comparison designs, and 10 participated in SCRD studies. Twelve studies reported the grade levels of their participants. Ninth and eleventh graders were included least often (n = 2 studies), and seventh grade students were included most frequently (n = 6 studies). Six studies (Calhoon, 2005; Kim et al., 2006; Mastropieri et al., 2001; Mastropieri et al., 2003; Spencer et al., 2003; Veerkamp et al., 2007) identified the mean ages of the target participants with a combined mean age of 13.19 years. The remaining seven studies did not specify age of participants (Calhoun & Fuchs, 2003; Gersten et al., 2006), specified an age range only (Dufrene et al., 2010; Hughes & Fredrick, 2006; Wexler et al., 2010), or specified age for the whole sample, but did not disaggregate for the target students (Scruggs et al., 2012; Vaughn et al., 2011).

Table 2.

Study Participants and Settings.

Study	Number of Participants	Grade/Age	Student Descriptors	Disability Category	Setting
Treatment-Comparison Designs
Calhoon & Fuchs (2003)	N = 92 (45 T; 47 C)	Grade 9: 29 T; 28 C Grade 10: 6 T; 9 C Grade 11: 8 T; 8 C Grade 12: 2 T; 2 C	African American: 29 T; 18 C Caucasian: 16 T; 29 C	LD: 35 T; 33 C BD: 2 T; 2 C MR: 6 T; 6 C Other: 2 T; 6 C	3 urban high schools in the Southeast Self-contained math resource classes
Calhoon (2005)	N = 38 (18 T; 20 C)	Grade 6 = 32 Grade 7 = 5 Grade 8 = 1 Mean Age: 12.11 T; 11.71 C	African American: 2 T; 0 C Caucasian: 7 T; 4 C Hispanic: 9 T; 16 C	LD: 18 T; 20 C	Middle school in the Southwest Self-contained language arts resource classes
Gersten, Baker, Smith-Johnson, Dimino, & Peterson (2006)	N = 76, n = 36 (18 T; 18 C)	Grade 7 Grade 8	NR	LD = 33 OHI = 3	Two middle schools in a moderate-size city in the Northwest
Kim et al. (2006)	N = 34 (16 T; 18 C)	Grade 6: 2 T; 3 C Grade 7: 7 T; 9 C Grade 8: 7 T; 6 C Mean age: 13.23 T; 13.34 C	African American: 4 T; 3 C Caucasian: 5 T; 10 C Hispanic: 7 T; 5 C Low Income: 8 T; 9 C	LD: 13 T; 15 C Other: 3 T; 3 C	Urban middle school Reading resource class or language arts class for students with reading difficulties
Mastropieri et al. (2001)	N = 24 (12 T; 12 C)	Grade 7 Mean age LD = 12.9 MR = 13	African American = 1 Caucasian = 21 Hispanic = 2	LD = 20 MR = 4 Spent 50% of school day in self-contained	Midwestern state Special education English class
Mastropieri et al. (2003)	N = 16 (7 T; 9 C)	Grade 10 = 15 Grade 12 = 1 Average age = 15.8	African American = 4 Asian = 2 Caucasian = 6 Hispanic = 4	LD = 14 LD and ED = 1 MR = 1 Spent 80% of day in self-contained	High school in major metro area in a mid-Atlantic region Special education world history class
Scruggs et al. (2012)	N = 157, n = 24 (12 T; 12 C)	Middle school	NR for target students	LD = 21 ED = 3	Middle school in a large mid-Atlantic school district 10 social studies classes
Spencer et al. (2003)	N = 30 (12 T1; 18 T2)	Average age: Grade 7 = 13.15 Grade 8 = 14.05	African American = 7 Asian = 1 Caucasian = 20 Hispanic = 2	ED = 30	Alternative middle school for students with BD and/or ED in a suburban mid-Atlantic state Social studies or civics classes
Vaughn et al. (2011)	N = 782 (400 T; 382 C) n (low on TOWRE) = 95 (49 T; 46 C) n (low on state test) = 253 (125 T; 128 C)	Grade 7 Grade 8	NR for target students	Sped: 56 T; 28 C All with reading difficulty (i.e., failing state test and scoring 1+ SD below mean on TOWRE	Three school districts (two near urban and one urban) Regularly scheduled 7th or 8th grade English language arts/reading class
Wexler et al. (2010)	N = 96 (33 T1; 34 T2; 29 C)	Grade 9 = 33 Grade 10 = 24 Grade 11 = 27 Grade 12 = 12	ELL = 19	LD = 76 ED = 3 MR = 4 OHI = 8 Autism = 1 AI = 1 Unknown = 3	Metropolitan area in Southwest Special education English language arts or reading classes
Single-case designs
Dufrene et al. (2010)	N = 4	Grade 6 = 4	NR	Receiving Tier II intervention = 4	Classroom at school Noninstructional time (e.g., study hall) selected by regular classroom teacher
Hughes & Fredrick (2006)	N = 3	Grade 6 = 3	NR for target students	LD = 3	Middle school in satellite community near large metropolitan area in Southeast Inclusive general education classroom
Veerkamp et al. (2007)	N = 71, n = 3	Grade 6 = 3 Average age = 11.6	African American = 2 Caucasian = 1	“Low achievers”	Urban middle school in Midwest General education reading classes

Note. AI = auditory impairment; BD = behavioral disorder; C = comparison; ED = emotional disorder; ELL = English language learners; LD = learning disability; MR = mentally retarded; N = full sample; n = subgroup of full sample who qualified for this review; NR = not reported; OHI = other health impairment; SD = standard deviation; Sped = special education; T = treatment; TOWRE = Test of Word Reading Efficiency.

Ethnicity was reported for the targeted students (i.e., those identified as at risk, low achieving, struggling, below grade level, or LD) in seven studies (Calhoon, 2005; Calhoon & Fuchs, 2003; Kim et al., 2006; Mastropieri et al., 2001; Mastropieri et al., 2003; Spencer et al., 2003; Veerkamp et al., 2007). Of the 237 participants with reported ethnicity, just more than half were Caucasian (n = 119). Fewer were African American (n = 70) or Hispanic (n = 45), and only 3 were Asian. One study did not report the ethnicities of participants but did identify 19 students as English language learners (Wexler et al., 2010).

Disability subtype was reported and disaggregated in 10 studies (Calhoon, 2005; Calhoon & Fuchs, 2003; Gersten et al., 2006; Hughes & Fredrick, 2006; Kim et al., 2006; Mastropieri et al., 2001; Mastropieri et al., 2003; Scruggs et al., 2012; Spencer et al., 2003; Wexler et al., 2010). Of the total 390 participants identified with disabilities, 77% (n = 301) were identified as students with LD, 9% (n = 36) were students with an emotional disorder (ED), 5% (n = 21) were identified with an intellectual disability (described in the studies as mental retardation), 4% (n = 14) were identified generically as “other” disability, and 3% (n = 11) were identified with other health impairment. Only 1% (n = 4) were identified with a behavioral disorder. Finally, fewer than 1% (n = 1) were identified with comorbid LD and ED, autism, and auditory impairment.

Intervention duration and session length

The corpus of studies included a range of reported information on the number of sessions, frequency of the intervention, and length of intervention sessions (see Table 1). Of the eight studies explicitly reporting the total amount or a range of sessions implemented, the fewest held was three (Hughes & Fredrick, 2006) and the most was 50 (Wexler et al., 2010). Based on the frequency of sessions reported, it is possible that more sessions were implemented in other studies. However, we declined to infer a number not directly stated because it is possible that holidays or other disruptions could have reduced what might have been calculated using the number of days per week and total number of weeks reported.

All 13 studies provided information on the frequency of intervention implementation. Six studies implemented interventions daily (Gersten et al., 2006; Hughes & Fredrick, 2006; Mastropieri et al., 2001; Spencer et al., 2003; Veerkamp et al., 2007; Wexler et al., 2010). The remaining studies implemented interventions two to three times per week. All studies reported the length of individual intervention sessions that ranged from an average 9.5 min (Dufrene et al., 2010) to 90 min per session (Mastropieri et al., 2003). Most commonly (n = 4), sessions were held from 40 to 50 min (Gersten et al., 2006; Kim et al., 2006; Mastropieri et al., 2001; Vaughn et al., 2011). Three studies also reported differing lengths of sessions for different components (35–40 min; Calhoon, 2005), different days of the week (25–43 min; Veerkamp et al., 2007), or alternating weekly schedules (50–90 min; Scruggs et al., 2012). The extraordinary variety of duration, frequency, and session length is notable as there does not seem to be a pattern emerging in these interventions from which to draw conclusions about the necessary amount of time needed for peer-mediated interventions to be effective.

Features of the peer mediation

We examined the common and unique features of the peer-mediation implemented in the 13 studies identified for this synthesis (see Figure 4). Of the studies, 10 implemented a formal peer-mediated learning strategy including CWPT (two treatment-comparison and two SCRD studies), PALS (four treatment-comparison designs), and CSR (two treatment-comparison designs). Of these 10 studies, 4 were focused on a content area other than ELA and reading. Two studies of CWPT (Scruggs et al., 2012; Spencer et al., 2003) and one study of PALS (Mastropieri et al., 2003) were designed for teaching social studies content, and another study of PALS was designed for teaching mathematics (Calhoon & Fuchs, 2003). The remaining three studies implemented a peer-mediated intervention that did not use a formal strategy: two focused on ELA/reading (Dufrene et al., 2010; Wexler et al., 2010), and one study focused on teaching social studies content (Gersten et al., 2006).

Figure 4.

Features of peer-mediated instruction implemented.

Regardless of whether or not a formal strategy was used, the studies incorporated similar features such as reciprocal tutoring (i.e., having students alternate roles as tutor and tutee), that was implemented in three of the studies using CWPT (Hughes & Fredrick, 2006; Scruggs et al., 2012; Veerkamp et al., 2007), all four of the studies using PALS (Calhoon, 2005; Calhoon & Fuchs, 2003; Mastropieri et al., 2001; Mastropieri et al., 2003), and two of the studies using no formal peer-mediation strategy (Gersten et al., 2006; Wexler et al., 2010).

Similarly, partner reading (i.e., having students alternate reading and rereading portions of text to support fluency and/or comprehension) was implemented in two studies using CWPT (Spencer et al., 2003; Veerkamp et al., 2007), three studies using PALS (Calhoon, 2005; Mastropieri et al., 2001; Mastropieri et al., 2003), and one of the studies using CSR (Kim et al., 2006). Partner reading was also implemented in three studies that did not identify a formal peer-mediation strategy (Dufrene et al., 2010; Gersten et al., 2006; Wexler et al., 2010). Besides partner reading and reciprocal tutoring, peer discussion explicitly was described as a feature of just one treatment-comparison design study (Gersten et al., 2006).

Pairing more- and less-abled peers was sometimes implemented within CWPT (Hughes & Fredrick, 2006; Spencer et al., 2003; Veerkamp et al., 2007) or PALS (Mastropieri et al., 2001; Mastropieri et al., 2003), but was also implemented without a formal strategy (Dufrene et al., 2010; Wexler et al., 2010). It should be noted that Wexler et al. (2010) aimed to pair a higher and lower ability peer together, but reported this was compromised because the majority of the sample was functioning at a low level. Only Dufrene et al. (2010) designated one student as the peer tutor rather than having the more- and less-abled partners alternate roles as tutor/tutee. Finally, one study had small groups rather than dyads of peers implement the CSR strategy (Vaughn et al., 2011), and Kim et al. (2006) uniquely utilized computer-assisted instruction with pairs to implement the CSR strategy.

Finally, most studies included a feedback procedure. The only one that did not was also the SCRD study that did not pair higher and lower ability students (Dufrene et al., 2010). Where there was a rather uniformly low-functioning sample, the authors commented the intended feedback component was potentially compromised (Wexler et al., 2010). The other anomaly in the provision of feedback during peer-mediated instruction was reported by Gersten et al. (2006), who noted feedback was delivered by teachers and expressed during peer discussion rather than within the peer tutoring component.

Study Findings

Findings are summarized by the type of peer-mediated intervention, starting with the formal strategies (i.e., CWPT, PALS, and CSR) and then grouping together the studies that did not use a formal strategy. Within each section, we first report effect sizes from the treatment-comparison design studies followed by findings from SCRD studies, if applicable. From the data reported in our corpus of studies, we calculated a total of 56 unique effect sizes for the former and 32 PND computations for the latter. We end this section by summarizing the social validity findings from the studies that provided enough information.

Classwide peer tutoring

Of the 13 studies, 4 implemented CWPT: 2 treatment-comparison designs (Scruggs et al., 2012; Spencer et al., 2003) and 2 SCRD studies (Hughes & Fredrick, 2006; Veerkamp et al., 2007). The 2 treatment-comparison CWPT studies administered researcher-developed measures of social studies content and one also observed students’ time on task (Spencer et al., 2003). There was a wide range in effect sizes (ES = 0.37 to 1.04) across all outcomes reported. Scruggs et al. (2012) randomly assigned 10 inclusive social studies classrooms to a an experimental CWPT condition or a typical instruction comparison condition (e.g., teacher lecture). Students in the treatment condition worked in pairs to discuss content and alternated asking each other questions and providing feedback. Stronger effects were reported for students in the CWPT condition on researcher-developed measures of target content items (ES = 0.78) and nontarget items (ES = 0.67).

Spencer et al. (2003) also randomly assigned social studies classes to condition but used a within-subjects crossover design so that each class participated for an equal amount of time in the CWPT treatment condition and the comparison condition. The latter typically consisted of whole-class discussion, read-alouds by individual students, and independent note taking. Similar to the Scruggs et al. (2012) study, students in the Spencer et al. CWPT treatment condition implemented partner reading and took turns delivering corrective feedback and summarizing the content. After only 4 weeks in the tutoring condition, students scored higher on weekly quizzes (ES range = 0.74 to 0.91), multiple-choice chapter tests (ES range = 0.61 to 1.04), open-ended chapter tests (ES range = 0.40 to 0.43), and percentage of on-task behavior (ES range = 0.37 to 0.95) than when in the typical instruction comparison condition.

As with the treatment-comparison studies, the outcome measures administered in the two CWPT singe-case design studies were all researcher-developed. Both studies addressed ELA. Using a multiple-probe design, Hughes and Fredrick (2006) intervened with three sixth grade students with LD using a CWPT time delay procedure. Instruction was designed to help students read and define vocabulary words with the help of a peer who provided immediate and corrective feedback. Veerkamp et al. (2007) implemented vocabulary instruction and oral text reading practice. Partners delivered error correction and awarded points for correctly read sentences and correctly defined words. PND estimates indicated the peer-mediated vocabulary interventions were moderately to highly effective in the Hughes and Fredrick (2006) study, but ineffective to moderately effective for the three low-achieving students in the Veerkamp et al. (2007) study. However, given the known limitations with PND, these calculations should be interpreted with caution.

Peer-assisted learning strategies

All four studies implementing PALS used treatment-comparison designs. Calhoon and Fuchs (2003) implemented a peer-mediated mathematics intervention with tutors alternating roles as tutor/tutee and providing feedback. This demonstrated mostly small effects (ES = 0.26 to 0.40), with no discernible effect on students’ state mathematics test performance (ES = 0.07).

Stronger effects were realized in the two studies focused on students’ reading outcomes (Calhoon, 2005; Mastropieri et al., 2001). In the Calhoon (2005) study, teachers randomly assigned to the treatment condition implemented scripted phonics lessons three times per week followed by peer partner practice on the phonics concepts. On alternating days, students engaged in partner reading practice. Peers alternated roles as tutor/tutee and delivered positive and corrective feedback to each other during practice. Mastropieri et al. (2001) implemented a combined partner reading with corrective feedback on oral reading errors and main idea generation intervention, that was compared to traditional whole class instruction. Large effects were found on standardized measures of word reading skills (ES = 0.95 to 1.01) and comprehension (ES = 0.84) in the Calhoon (2005) study, and on researcher-developed measures of comprehension (ES = 1.12) in the Mastropieri et al. (2001) study. However, the peer-mediated intervention had no discernible effect (ES = 0.09) on students’ performance on a standardized measure of fluency (Calhoon, 2005).

Although the final study of PALS was focused on social studies content, the researchers assessed students’ fluency and comprehension outcomes in addition to their content learning (Mastropieri et al., 2003). The treatment, which incorporated feedback delivered by a peer, had a small effect on reducing students’ reading rate and oral reading errors (ES = −0.27 and −0.36, respectively) and a large effect on a researcher-developed measure of comprehension (ES = 1.95). There was a wide range in effect sizes on researcher-developed measures of students’ social studies content learning (ES = 0.05 to 2.03).

Collaborative strategic reading

Both studies implementing CSR used treatment-comparison designs and focused on students’ reading outcomes (Kim et al., 2006; Vaughn et al., 2011). In the Kim et al. (2006) study, the students in classes randomly assigned to the treatment condition were paired by assigning a more- to a less-abled peer. As partners, they learned and practiced the CSR strategies using a computer-assisted CSR program that incorporated computer-driven supports (e.g., requesting a definition or review of CSR). Vaughn et al. (2011) used a randomized block design with students randomly scheduled into classes and classes randomly assigned to condition. In the treatment condition, teachers modeled each comprehension strategy and implemented short lessons to provide practice in the CSR strategies. Students were assigned roles (e.g., clunk expert) within small groups to guide reading and application of the strategies. Teachers in the business as usual comparison group were taught using the typical ELA curriculum. Mostly large effects were found on researcher-developed measures of reading comprehension (ES = 0.76 to 1.14) in the Kim et al. (2006) study, with a smaller effect on the standardized measure of comprehension administered (ES = 0.48). However, a wide range of effects (ES = −0.10 to 2.73) were found on standardized measures of comprehension in the Vaughn et al. (2011) study.

No formal peer-mediation strategy

Two treatment-comparison design studies (Gersten et al., 2006; Wexler et al., 2010) and one single case design (Dufrene et al., 2010) implemented features of peer-mediated learning outside of the context of a formal strategy. Gersten et al. (2006) utilized two approaches for teaching a history unit on the civil rights movement to middle schools students with LD in the general education setting. In the treatment condition, teachers used brief video clips and peer partners who alternated roles as tutor/tutee to support learning of content through activities, including partner reading. Feedback was provided by the teachers and through peer discussion. Moderate to large effects were found on researcher-developed measures of students’ social studies content learning, including key vocabulary (g = 0.60 to 1.12).

Wexler et al. (2010) examined the efficacy of two different 20-min reading fluency interventions (i.e., repeated reading and wide reading) compared to a typical instruction condition on the word reading, fluency, and comprehension outcomes of high school students with severe reading disabilities. Students alternated roles as tutor/tutee and were paired based on placing a more- and less-abled reader together; however, this procedure was compromised because the majority of the sample was functioning at a low level. With the exception of a small effect (ES = 0.28) found on a standardized fluency measure in the Wexler et al. (2010) study, the peer-mediated instruction produced no discernible effects (ES = −0.31 to 0.28) on standardized assessments of students’ comprehension, fluency, and word identification.

Dufrene et al. (2010) implemented a multiple baseline study using peer tutoring to build reading fluency with modeling and practice, but the intervention did not incorporate error correction. PND calculated on a standardized measure of students’ oral reading fluency revealed peer-mediation was minimally effective for three students and ineffective for one (Dufrene et al., 2010). No other SCRD studies assessed the effects of peer-mediated learning on students’ fluency outcomes.

Social validity

Of the 13 studies, 8 provided social validity findings (Calhoon & Fuchs, 2003; Hughes & Fredrick, 2006; Kim et al., 2006; Mastropieri et al., 2001; Mastropieri et al., 2003; Scruggs et al., 2012; Spencer et al., 2003; Veerkamp et al., 2007). Converging evidence from these studies, as included in Figure 5, suggests students believed peer-mediated instruction was beneficial to their learning. Authors from 7 of the 8 studies reported that, regardless of the type of peer-mediated instruction, students agreed that they liked participating in this type of instruction, particularly due to the feedback opportunities. Only Hughes and Fredrick (2006) did not report whether the students liked the peer-mediated instruction, but they stated the students in their study agreed that they would participate in peer-mediated instruction again or in another class. A similar sentiment among students was reported in 4 additional studies (Kim et al., 2006; Mastropieri et al., 2001; Mastropieri et al., 2003; Veerkamp et al., 2007). Authors from only one study reported that students expressed not wanting to participate in peer-mediated instruction again or in another class (Spencer et al., 2003). Overall, feedback from students was positive.

Figure 5.

Social validity data.

In all eight studies that reported social validity data, teachers also expressed that they felt the peer-mediated instruction was beneficial to student learning. With the exception of the Veerkamp et al. (2007) study, authors stated that the teachers would likely use peer-mediated instruction again. In five of these eight studies, teachers reportedly believed discipline problems decreased and engagement increased due to the peer-mediated instruction (Hughes & Fredrick, 2006; Mastropieri et al., 2001; Mastropieri et al., 2003; Spencer et al., 2003; Veerkamp et al., 2007). Despite teachers’ generally positive view of peer-mediated instruction, a challenge noted in three studies was that the teachers found implementation of peer-mediated instruction with low achievers to be difficult (Mastropieri et al., 2001; Mastropieri et al., 2003; Spencer et al., 2003). Finally, one study reported that both teachers and their secondary students were unsure about the benefits of the extrinsic reward system used within the peer-mediated instruction (Calhoon & Fuchs, 2003).

Discussion

The primary purpose of this synthesis was to determine the effectiveness of peer-mediated reading and math interventions on academic outcomes for struggling learners in the secondary grades in regular and alternative settings. We also sought to describe the various features of peer-mediated instruction, as well as the social validity of these interventions as reported by students and teachers. We focused on Grades 6 to 12 in regular and special education settings because previous syntheses have more extensively examined the effectiveness of these practices with students in the elementary grades (Elbaum, Vaughn, Hughes, & Moody, 1999; Rohrbeck, Ginsburg-Block, Fantuzzo, & Miller, 2003), students from the elementary grades to postsecondary education (Heron, Welsch, & Goddard, 2003), and students with emotional and behavioral disorders (Ryan, Reid, & Epstein, 2004) than with secondary students experiencing academic difficulties (Kunsch et al., 2007; Okilwa & Shelby, 2010; Stenhoff & Lignugaris/Kraft, 2007). In addition, many secondary teachers struggle with finding effective ways to differentiate and intensify instruction as well as increase engagement for secondary struggling learners (Kosanovich, Reed, & Miller, 2010). Therefore, determining research-based methods of intervening with this challenging population is essential.

Overall, there were relatively few treatment-comparison studies (n = 10), including only one study that focused on math (Calhoon & Fuchs, 2003). No studies conducted in juvenile justice facilities or residential treatment centers met our search criteria. Thus, questions remain about the generalizability of peer-mediated intervention to all content areas and alternative educational settings that contradicts findings by Okilwa and Shelby (2010), who reported peer tutoring as effective regardless of content area.

Although only half of the 10 studies (Calhoon, 2005; Calhoon & Fuchs, 2003; Kim et al., 2006; Vaughn et al., 2011; Wexler et al., 2010) met all accepted quality indicators for experimental and quasi-experimental studies (Gersten et al., 2005; WWC, 2010), the remaining 5 studies met all but one or two indicators. Given the range of effect sizes across the studies, no pattern linking study quality to outcomes can be discerned beyond the tendency for higher effect sizes to be associated with researcher-developed measures as discussed in the next paragraph. There are many challenges to implementing research in secondary schools (Matto & Vercellotti, 2012), so we acknowledge that not meeting all quality indicators may be a function of the compromises that must be made to carry out a treatment-comparison study with adolescents. For example, only two studies randomly assigned students to conditions (Gersten et al., 2006; Wexler et al., 2010). This is the most desirable approach but is subject to scheduling restrictions by class availability and student credit requirements.

Outcomes for the treatment-comparison studies ranged from no effects to large effect sizes on measures of comprehension, content acquisition, fluency, vocabulary, word reading, and math operations and concepts. It is not uncommon to use researcher-developed proximal measures to measure content acquisition, as several studies in this synthesis did (Gersten et al., 2006; Kim et al., 2006; Mastropieri et al., 2001; Mastropieri et al., 2003; Scruggs et al., 2012; Spencer et al., 2003), and we note the majority of large effect sizes for content knowledge and reading skill improvement were found with researcher-developed measures from which inflated effect sizes are typical (Swanson, Hoskyn, & Lee, 1999). This finding aligns with findings from previous research on peer-tutoring (Okilwa & Shelby, 2010; Stenhoff & Lignugaris/Kraft, 2007). Future studies could increase our confidence in the findings by employing more standardized measures.

Three studies used SCRD to examine the effectiveness of peer-mediated reading intervention in the regular education setting. Findings from these studies were mixed, and only one study met the quality standards set forth by Horner et al. (2005). None of the three studies met the quality indicators set forth by the WWC (Horner et al., 2005; Kratochwill et al., 2011), decreasing our confidence in conclusions that may be drawn from these SCRD studies.

For two of the four reading intervention studies with smaller effect sizes, we identified anomalies in the features and/or implementation of the peer-mediated instruction. The treatment-comparison study reporting small to no effects on standardized measures of reading had a uniformly low-functioning sample of students and, therefore, had difficulty with the provision of peer-delivered feedback (Wexler et al., 2010). The SCRD study with ineffective to minimally effective results did not incorporate peer feedback at all and did not pair higher and lower ability students (Dufrene et al., 2010). It is possible that the method of pairing and the delivery of quality feedback contributed to the poor outcomes in these studies as compared to others in the corpus. Findings from the review conducted by Stenhoff and Lignugaris/Kraft (2007) help explain these findings as they reported heterogeneous tutoring to be the only type of evidence based practice for secondary students. The only other reading intervention studies with no discernible effects on an outcome measure did incorporate peer feedback, but also had large effects on other standardized measures of reading that were administered (Calhoon, 2005; Vaughn et al., 2011). The lack of effects found in these two studies was limited to a single outcome: fluency or a maze measure, respectively. By comparison, the studies reporting no or problematic feedback reported no or small effects across all students (Dufrene et al., 2010) or measures administered (Wexler et al., 2010).

Most studies, despite research design, represented a variety of formal peer-mediated interventions (n = 10) including CWPT, PALS, and CSR. Only three studies did not use a formal strategy. It is unknown whether or not this is consistent with the ratio of formal and ad hoc peer-mediation practices being implemented by practitioners in secondary classrooms. Therefore, we note that “peer mediation” must be carefully defined when attempting to draw conclusions that will be meaningful to teachers. To that end, we identified the common features of instruction across all studies. These included reciprocal tutoring (n = 9), partner reading (n = 9), and pairing more- and less-abled peers (n = 7). Peer discussion (n = 1), the use of small groups (n = 1), and the use of computers to facilitate peer-mediated instruction (n = 1) have isolated but promising support as components of peer-mediated instruction.

There is more evidence, overall, supporting formal approaches but limited evidence when disaggregated by type of formal approach. The common patterns are that peer-mediated instruction has more history of use with literacy instruction and demonstrates more robust outcomes on intervention specific outcomes. There are moderate to large effects on standardized measures, particularly when the intervention incorporates feedback, with the exception of fluency outcomes that were minimal to nonexistent. This is consistent with studies finding limited growth/improvement in adolescence (Denton, Wexler, Vaughn, & Bryan, 2008; Vaughn, Wexler, et al., 2012; Wexler et al., 2010).

Taken together, results from the 13 studies included in this review suggest secondary struggling learners can improve in reading comprehension and/or content acquisition with peer-mediated interventions, particularly when incorporating some type of feedback component. Almost all studies included in this synthesis that resulted in at least small effects across researcher and standardized measures incorporated some sort of structured feedback component, usually from a peer and less often from a peer and teacher (Gersten et al., 2006). Interventions that did not include a reciprocal feedback component (Dufrene et al., 2010) or used a feedback component that was compromised due to poor pairing techniques (Wexler et al., 2010) resulted in minimal to no effects. With only one study in our corpus focusing on math, we are unable to draw conclusions about the overall effectiveness of peer-mediated learning in this subject and how the results align with previous research (e.g., Kunsch et al., 2007). However, the positive impact on reading outcomes supports findings from syntheses of peer-mediated interventions at the early grades (Elbaum et al., 1999; Rohrbeck et al., 2003) and previous research from findings from syntheses at the secondary grades (Okilwa & Shelby, 2010; Stenhoff & Lignugaris/Kraft, 2007).

Finally, the social validity data reported in eight of the studies suggests most students and teachers believed peer-mediated instruction was beneficial and would want to implement it again. This seems to reflect the social needs of adolescents and their desire for more autonomy in their work (Kamil et al., 2008; Reeve & Jang, 2006). The cautions noted about working with low-achieving students (Mastropieri et al., 2001; Mastropieri et al., 2003; Spencer et al., 2003; Wexler et al., 2010) indicate it may be beneficial to determine how to pair students in the most efficacious way and perhaps more flexibly at this level. Doing so also might help reduce behavior problems and increase engagement.

Limitations

There are several limitations to this synthesis. First, the overall number of studies included in the corpus was fairly limited: only 10 treatment-comparison and 3 SCRD studies. In addition, most had rather small sample sizes of our target students. Although other studies were available, they did not disaggregate findings for secondary struggling learners. In addition to these sampling issues, interpretations of the findings must take into consideration the fact that several studies did not meet accepted quality standards (Gersten et al., 2005; Raudenbush, 2005; Shadish, 2002; WWC, 2010). With respect to the kinds of peer mediation implemented, the presence of particular features (see Figure 4) was known only if stated by the authors. For example, peer discussion was specifically mentioned in only one study, but is a common element in the formal strategies—none of which specifically described discussion. Finally, as is common in intervention research (Swanson et al., 1999), many of the larger effect sizes were found when using researcher-developed measures. These limitations notwithstanding, peer-mediated intervention can be interpreted as a promising instructional method to address the needs of adolescents who struggle with reading and learning from content area texts.

Summary of Implications and Future Research

Based on these synthesized findings, we conclude that further secondary-level peer-mediated instruction research, designed to meet experimental and SCRD quality standards, is warranted. Studies using standardized outcome measures are especially needed to improve the generalizability of effects. Future research also might examine ways to effectively pair students to increase learning, while taking into account the emotional and behavioral challenges of the secondary struggling learner.

It is important to remember we located only one peer-mediated math intervention study and no studies conducted in an alternative setting (e.g., correctional facility). Secondary-level students are being held to more rigorous math standards as preparation for college and careers (National Council of Teachers of Mathematics, 2000; National Governors Association Center for Best Practices & Council of Chief State School Officers, 2010). Therefore, more intervention research implementing peer-mediated math interventions is necessary to determine its effectiveness with this content. Considering the benefits of the feedback component found across studies in this synthesis, incorporating this feature when implementing peer-mediated instruction in the content areas may be beneficial and an efficient means of increasing the overall amount of explicit feedback provided to adolescents experiencing academic difficulty.

Finally, struggling academically is one of the biggest predictors of school disengagement, dropout, and even incarceration (Allensworth & Easton, 2005; Balfanz et al., 2007; Fall & Roberts, 2012). Because instruction in adjudicated and treatment-based facilities may be adolescents’ last chance to make gains before transitioning back to the community, school, and work (Malmgren & Leone, 2000), it is essential that we examine the effectiveness of the same promising practices for secondary struggling learners in the regular school setting with those students who are expected to learn in an alternative setting. In light of the unique challenges associated with teaching in an alternative setting (Coulter, 2004; Georgetown Law Human Rights Institute, 2012), it is crucial to confirm evidence-based practices that are realistic and beneficial for teachers to implement in these settings. For these reasons, research on the efficacy of peer-mediated instruction in the alternative setting is an area ripe for research.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Allensworth

Easton

(2005). The on-track indicator as a predictor of high school graduation. Chicago, IL: Consortium on Chicago School Research.

Archer

A. L.

Hughes

C. A.

(2011). Explicit instruction: Effective and efficient teaching. New York, NY: Guildford.

Bahr

C. M.

Rieth

H. J.

(1991). Effects of cooperative, competitive, and individualistic goals on student achievement using computer-based drill-and-practice. Journal of Special Education Technology, 11, 33–48.

Balfanz

Herzog

MacIver

D. J.

(2007). Preventing student disengagement and keeping students on the graduation path in urban middle-grades schools: Early identification and effective interventions. Educational Psychologist, 42, 223–235.

Barton

E. E.

Wolery

(2008). Teaching pretend play to children with disabilities: A review of the literature. Topics in Early Childhood Special Education, 28, 109–125. doi:10.1177/0271121408318799

Blatchford

Bassett

Brown

(2011). Examining the effect of class size on classroom engagement and teacher-pupil interaction: Differences in relation to pupil prior attainment and primary vs. secondary schools. Learning and Instruction, 21, 715–730. doi:10.1016/j.learninstruc.2011.04.001

Bloomberg

T. G.

Bales

W. D.

Mann

Piquero

A. R.

Berk

R. A.

(2011). Incarceration, education, and transition from delinquency. Journal of Criminal Justice, 39, 355–365. doi:10.1016/j.jcrimjus.2011.04.003

Bulgren

J. A.

Graner

P. S.

Deshler

D. D.

(2013). Literacy challenges and opportunities for students with learning disabilities in social studies and history. Learning Disabilities Research and Practice, 28, 17–27.

*Calhoon

M. B.

(2005). Effects of a peer-mediated phonological skill and reading comprehension program on reading skill acquisition for middle school students with reading disabilities. Journal of Learning Disabilities, 38, 424–433.

10.

*Calhoon

M. B.

Fuchs

L. S.

(2003). The effects of peer-assisted learning strategies and curriculum-based measurement on the mathematics performance of secondary students with disabilities. Remedial and Special Education, 24, 235–245.

11.

Carr

M. J.

Gray

N. L.

Holley

M. J.

(2007). Shortchanging disadvantaged students: An analysis of intra-district spending patterns in Ohio. Journal of Educational Research & Policy Studies, 7, 36–53. Retrieved from http://www.eric.ed.gov/PDFS/EJ809432.pdf

12.

Casillas

Robbins

Allen

Kuo

Y. L.

(2012). Predicting early academic failure in high school from prior academic achievement, psychosocial characteristics, and behavior. Journal of Educational Psychology, 104, 407–420. doi:10.1037/a0027180

13.

Cohen

(1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum.

14.

Coulter

(2004). Using one-to-one tutoring and proven reading strategies to improve reading performance with adjudicated youth. Journal of Correctional Education, 55, 321–333.

15.

Denton

C. A.

Wexler

Vaughn

Bryan

(2008). Intervention provided to linguistically diverse middle school students with severe reading difficulties. Learning Disabilities Research and Practice, 23, 79–89.

16.

Docan

T. N.

(2006). Positive and negative incentives in the classroom: An analysis of grading systems and student motivation. Journal of Scholarship of Teaching and Learning, 6, 21–40. Retrieved from http://www.eric.ed.gov/PDFS/EJ854925.pdf

17.

*Dufrene

B. A.

Reisener

C. D.

Olmi

D. J.

Zoder-Martell

McNutt

M. R.

Horn

D. R.

(2010). Peer tutoring for reading fluency as a feasible and effective alternative in response to intervention systems. Journal of Behavior Education, 19, 239–256. doi:10.1007/s10864-010-9111-8

18.

Elbaum

Vaughn

Hughes

Moody

S. W.

(1999). Grouping practices and reading outcomes for students with disabilities. Exceptional Children, 65, 399–415.

19.

Fall

A. M.

Roberts

(2012). High school dropouts: Interactions between social context, self-perceptions, school engagement, and student dropout. Journal of Adolescence, 35, 787–798. doi:10.1016/j.adolescence.2011.11.004

20.

Fuchs

L. S.

Fuchs

Compton

D. L.

(2010). Rethinking response to intervention at middle and high school. School Psychology Review, 39, 22–28.

21.

Fuchs

L. S.

Fuchs

Hamlett

C. L.

Phillips

N. B.

Karns

(1995). General educators’ specialized adaptation for students with learning disabilities. Exceptional Children, 61, 440–459. doi:10.2307/1511370

22.

Fuchs

L. S.

Fuchs

Hamlet

C. L.

Phillips

N. B.

Karns

Dutka

(1997). Enhancing students helping behaviors during peer-mediated instruction with conceptual mathematical explanations. Elementary School Journal, 101, 495–510. doi:10.1086/461863

23.

Georgetown Law Human Rights Institute. (2012). Kept out: Barriers to meaningful education in the school-to-prison pipeline. Washington, DC: Author.

24.

*Gersten

Baker

Smith-Johnson

Dimino

Peterson

(2006). Eyes on the prize: Teaching complex historical content to middle school students with learning disabilities. Exceptional Children, 72, 264–280.

25.

Gersten

Fuchs

L. S.

Compton

Coyne

Greenwood

Innocenti

M. S.

(2005). Quality indicators for group experimental and quasi-experimental research in special education. Exceptional Children, 71, 149–164.

26.

Glennie

Bonneau

Vandellen

Dodge

K. A.

(2012). Addition by subtraction: The relationship between dropout rates and school-level academic achievement. Teachers College Record, 114, 1–26.

27.

Greenwood

C. R.

Delquadri

Carta

J. J.

(1988). Classwide peer tutoring. Seattle, WA: Educational Achievement Systems.

28.

Hattie

Timperley

(2007). The power of feedback. Review of Educational Research, 77, 81–112. doi:10.3102/003465430298487

29.

Hedges

L. V.

(1981). Distribution theory for Glass’s estimator of effect size and related estimators. Journal of Educational Statistics, 6, 107–128.

30.

Heron

T. E.

Welsch

R. G.

Goddard

Y. L.

(2003). Applications of tutoring systems in specialized subject areas. Remedial and Special Education, 24, 288–300.

31.

Horner

R. H.

Carr

E. G.

Halle

McGee

Odom

S. L.

Wolery

(2005). The use of single-subject research to identify evidence-based practices in special education. Exceptional Children, 71, 165–179.

32.

*Hughes

T. A.

Fredrick

L. D.

(2006). Teaching vocabulary with students with learning disabilities using classwide peer tutoring and constant time delay. Journal of Behavior Education, 15, 1–23.

33.

Kamil

M. L.

Borman

G. D.

Dole

Kral

C. C.

Salinger

Torgesen

(2008). Improving adolescent literacy: Effective classroom and intervention practices: A practice guide (NCEE 2008-4027). Washington, DC: National Center for Education Evaluation and Regional Assistance, Institute of Education Sciences, U.S. Department of Education. Retrieved from http://ies.ed.gov/ncee/wwc

34.

Kazdin

A. E.

(1982). Single-case experimental designs in clinical research and practice. New Directions for Methodology of Social and Behavioral Science, 13, 33–47.

35.

*Kim

A. H.

Vaughn

Klingner

J. K.

Woodruff

A. L.

Klein

Kouzekanani

(2006). Improving the reading comprehension of middle school students with disabilities through computer-assisted collaborative strategic reading (CACSR). Remedial and Special Education, 27, 235–248.

36.

Klingner

J. K.

Vaughn

(1996). Reciprocal teaching of reading comprehension strategies for students with learning disabilities who use English as a second language. Elementary School Journal, 96, 275–293.

37.

Kosanovich

M. L.

Reed

D. K.

Miller

D. H.

(2010). Bringing literacy strategies into content instruction: Professional learning for secondary-level teachers. Portsmouth, NH: RMC Research Corporation, Center on Instruction. Retrieved from http://www.centeroninstruction.org/files/Bringing%20Literacy%20Strategies%20into%20Content%20Instruction.pdf

38.

Kratochwill

T. R.

Hitchcock

Horner

R. H.

Levin

J. R.

Odom

S. L.

Rindskopf

D. M.

Shadish

W. R.

(2011). Single case designs technical documentation. In What Works Clearinghouse: Procedures and standards handbook (Version 2.0). Retrieved from http://ies.ed.gov/ncee/wwc/pdf/wwc_procedures_v2_standards_handbook.pdf

39.

Kunsch

C. A.

Jitendra

A. K.

Sood

(2007). The effects of peer-mediated instruction in mathematics for students with learning problems: A research synthesis. Learning Disabilities Research and Practice, 22, 1–12.

40.

Maggin

D. M.

Chafouleas

S. M.

(2011). Protocol for applying the What Works Clearinghouse single-case standards. Unpublished research instrument.

41.

Maggin

D. M.

Chafouleas

S. M.

Mosely

K. M.

Johnson

A. J.

(2011). A systematic evaluation of token economies as a classroom management tool for students with challenging behavior. Journal of School Psychology, 49, 529–544.

42.

Maggin

D. M.

Johnson

A. H.

Chafouleas

S. M.

Ruberto

L. M.

Berggren

(2012). A systematic evidence review of school-based group contingency interventions for students with challenging behavior. Journal of School Psychology, 50, 625–654.

43.

Maheady

Harper

G. F.

Sacca

M. K.

(1988). Peer-mediated instruction: A promising approach to meeting the diverse needs of LD adolescents. Learning Disability Quarterly, 11, 108–113.

44.

Malmgren

K. W.

Leone

P. E.

(2000). Effects of a short-term auxiliary reading program on the reading skills of incarcerated youth. Education and Treatment of Children, 23, 239–247.

45.

*Mastropieri

M. A.

Scruggs

T. E.

Mohler

Beranek

Spencer

Boon

R. T.

Talbott

(2001). Can middle school students with serious reading difficulties help each other learn anything? Learning Disabilities Research and Practice, 16, 18–27.

46.

*Mastropieri

M. A.

Scruggs

T. E.

Spencer

Fontana

(2003). Promoting success in high school history: Peer tutoring versus guided notes. Learning Disabilities Research and Practice, 18, 52–65.

47.

Matto

E. C.

Vercellotti

(2012). Methodological lessons learned from conducting civic education research in high schools. PS: Political Science & Politics, 45, 728–735. doi:10.1017/S1049096512000820

48.

McCulley

L. V.

Katz

Vaughn

(2013). Reading instruction and students with learning disabilities. In Bakken

J. P.

Obiakor

F. E.

Rotatori

A. F.

(Eds.), Learning disabilities: Practice concerns and students with LD. Advances in special education (Vol. 25, pp. 19–43). Bingley, UK: Emerald.

49.

Mol

S. E.

Bus

A. G.

(2011). To read or not to read: A meta-analysis of print exposure from infancy to early adulthood. Psychological Bulletin, 137, 267–296. doi:10.1037/a0021890

50.

National Association of Secondary School Principals. (n.d.). Implementing the Common Core State Standards: The role of the secondary school leader. Retrieved from http://www.nassp.org/Content/158/SecondaryActionBrief_Final.pdf

51.

National Council of Teachers of Mathematics. (2000). Principles and standards for school mathematics (Vol. 1). Reston, VA: Author. Retrieved from http://www.nctm.org/standards

52.

National Governors Association Center for Best Practices & Council of Chief State School Officers. (2010). Common Core State Standards. Retrieved from http://www.corestandards.org/the-standards

53.

Okilwa

N. S. A.

Shelby

(2010). The effects of peer tutoring on academic performance of students with disabilities in Grades 6 through 12: A synthesis of the literature. Remedial and Special Education, 31, 450–463. doi:10.1177/0741932509355991

54.

Pashler

Bain

Bottge

Graesser

Koedinger

McDaniel

Metcalfe

(2007). Organizing instruction and study to improve student learning (NCER 2007-2004). Washington, DC: U.S. Department of Education, National Center for Education Research, Institute of Education Sciences. Retrieved from http://ncer.ed.gov.

55.

Raudenbush

S. W.

(2005). Learning from attempts to improve schooling: The contribution of methodology. Educational Researcher, 34(5), 25–31.

56.

Reed

D. K.

Sorrells

A. M.

Cole

H. A.

Takakawa

N. N.

(2013). The ecological and population validity of reading interventions for adolescents: Can effectiveness be generalized? Learning Disability Quarterly, 36, 131–144. doi: 10.1177/0731948712451976.

57.

Reeve

Jang

(2006). What teachers say and do to support students’ autonomy during a learning activity. Journal of Educational Psychology, 98, 209–218. doi:10.1037/0022-0663.98.1.209

58.

Ridlon

C. L.

(2009). Learning mathematics via a problem-centered approach: A two-year study. Mathematical Teaching and Learning, 11, 188–225. doi:10.1080/10986060903225614

59.

Roach

J. C.

Paolucci-Whitcomb

Meyers

H. W.

Duncan

D. A.

(1983). The comparative effects of peer tutoring in math by and for secondary special needs students. Pointer, 27, 20–24.

60.

Rohrbeck

C. A.

Ginsburg-Block

M. D.

Fantuzzo

J. W.

Miller

T. R.

(2003). Peer-assisted learning interventions with elementary school students: A meta-analytic review. Journal of Educational Psychology, 95, 240–257. doi:10.1037/0022-0663.95.2.240

61.

Rupley

W. H.

Blair

T. R.

Nichols

W. D.

(2009). Effective reading instruction for struggling readers: The role of direct/explicit teaching. Reading and Writing Quarterly, 25, 125–138. doi:10.1080/10573560802683523

62.

Ryan

J. B.

Reid

Epstein

M. H.

(2004). Peer-mediated intervention studies on academic achievement for student with EBD. Remedial and Special Education, 25, 330–341.

63.

Scammacca

Roberts

Vaughn

Edmonds

Wexler

Reutebuch

C. K.

Torgesen

J. K.

(2007). Intervention for adolescent struggling readers: A meta-analysis with implication for practice. Portsmouth, NH: RMC Research Corporation, Center on Instruction.

64.

Scruggs

T. E.

Mastropieri

M. A.

Casto

(1987). The quantitative synthesis of single subject research. Remedial and Special Education, 8, 24–33.

65.

*Scruggs

Mastropieri

M. A.

Marshak

(2012). Peer-mediated instruction in inclusive secondary social studies learning: Direct and indirect learning effects. Learning Disabilities Research and Practice, 27, 12–20.

66.

Shadish

W. R.

(2002). Revisiting field experimentation: Field notes for the future. Psychological Methods, 7, 3–18. doi:10.1037/1082-989x.7.1.3

67.

Siebert

Draper

(2008). Why content area literacy messages do not speak to mathematics teachers: A critical content analysis. Literacy Research and Instruction, 47, 229–245.

68.

Siegler

Carpenter

Fennell

Geary

Lewis

Okamoto

. . . Wray

(2010). Developing effective fractions instruction for kindergarten through 8th grade: A practice guide (NCEE #2010–4039). Washington, DC: U.S. Department of Education, National Center for Education Evaluation and Regional Assistance, Institute of Education Sciences. Retrieved from http://ies.ed.gov/ncee/wwc

69.

*Spencer

V. G.

Scruggs

T. E.

Mastropieri

M. A.

(2003). Content area learning in middle school social studies classrooms and students with emotional or behavioral disorders: A comparison of strategies. Behavioral Disorders, 28, 77–93.

70.

Stenhoff

D. M.

Lignugaris/Kraft

(2007). A review of the effects of peer tutoring on students with mild disabilities in secondary settings. Exceptional Children, 74, 8–30.

71.

Swanson

H. L.

Hoskyn

Lee

(1999). Intervention for students with learning disabilities: A meta-analysis of treatment outcomes. New York, NY: Guilford.

72.

*Vaughn

Klingner

J. K.

Swanson

E. A.

Boardman

A. G.

Roberts

Mohammed

S. S.

Stillman-Spisak

S. J.

(2011). Efficacy of collaborative strategic reading with middle school students. American Educational Research Journal, 48, 938–964.

73.

Vaughn

Wanzek

Murray

C. S.

Roberts

(2012). Intensive interventions for students struggling in reading and mathematics: A practice guide. Portsmouth, NH: RMC Research Corporation, Center on Instruction.

74.

Vaughn

Wexler

Leroux

Roberts

Denton

C. A.

Barth

Fletcher

(2012). Effects of an intensive reading intervention for eighth grade students with persistently inadequate response to intervention. Journal of Learning Disabilities, 45, 515–525.

75.

*Veerkamp

M. B.

Kamps

D. M.

Cooper

(2007). The effects of classwide peer tutoring on the reading achievement of urban middle school students. Education and Treatment of Children, 30(2), 21–51. doi:10.1353/etc.2007.0010

76.

Wexler

Pyle

Flower

Williams

Cole

(2013). A synthesis of academic interventions for incarcerated adolescents. Review of Educational Research. Advance online publication. doi:10.3102/0034654313499410

77.

*Wexler

Vaughn

Roberts

Denton

C. A.

(2010). The efficacy of repeated reading and wide reading practice for high school students with severe reading disabilities. Learning Disabilities Research and Practice, 25, 2–10.

78.

What Works Clearinghouse. (2010). WWC procedures and standards handbook (Version 2.1). Retrieved from http://ies.ed.gov/ncee/wwc/documentsum.aspx?sid=19

79.

Wolery

Busick

Reichow

B. R.

Barton

E. E.

(2010). Comparison of overlap methods for quantitatively synthesizing single subject data. Journal of Special Education, 44, 18–28.