repeated reading interventions for students with · pdf filerepeated reading interventions...

20
Vol. 75, No. 3. pp. 263-281. ©2009 Cotmcilfor Exceptional Children. Repeated Reading Interventions for Students With Learning Disabilities: Status of the Evidence DAVID J. CHARD Southern Methodist University LEANNE R. KETTERLIN-GELLER University of Oregon SCOTT K. BAKER Pacific institutes for Research/University of Oregon CHRISTIAN DOABLER CHANISA APICHATABUTRA University of Oregon ABSTRACT: T: For students with or at risk for learning disabilities, developing fluency with reading connected texts remains a formidable challenge. In response, teachers often use repeated reading practices designed to provide students with multiple exposures to the same words. This study exam- ined research focused on determining the efficacy of repeated reading approaches for, improving reading fluency for students with or at risk for learning disabilities. Studies employed experimen- tall quasi-experimental and single-subject research designs. Results suggest that repeated reading is not supported by rigorous research as deflned by the quality indicators used and, therefore, is not an evidence-based practice based on those criteria for students with and at risk for learning disabili- ties. Implications for future research and for practice are discussed. U earning to read remains a hall- well, the occasion is set for them to flourish in mark skill that essentially vocabulary and language development, compre- defines the degree of success hension, and content area learning (Stanovich, students can achieve academi- 1986). Students with or at risk for learning dis- cally throughout their school abilities (LD) are most often identified for special career. For students who learn to read early and education services due to their measurable diffi- Exceptional Children 263

Upload: phungdiep

Post on 08-Mar-2018

217 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: Repeated Reading Interventions for Students With  · PDF fileRepeated Reading Interventions for Students ... supporting the notion that they experience a ... FLUENCY AND ITS ROLE

Vol. 75, No. 3. pp. 263-281.©2009 Cotmcilfor Exceptional Children.

Repeated ReadingInterventions for StudentsWith Learning Disabilities:Status of the Evidence

DAVID J. CHARDSouthern Methodist University

LEANNE R. KETTERLIN-GELLERUniversity of Oregon

SCOTT K. BAKERPacific institutes for Research/University of Oregon

CHRISTIAN DOABLER

CHANISA APICHATABUTRAUniversity of Oregon

ABSTRACT:T: For students with or at risk for learning disabilities, developing fluency with reading

connected texts remains a formidable challenge. In response, teachers often use repeated reading

practices designed to provide students with multiple exposures to the same words. This study exam-

ined research focused on determining the efficacy of repeated reading approaches for, improving

reading fluency for students with or at risk for learning disabilities. Studies employed experimen-

tall quasi-experimental and single-subject research designs. Results suggest that repeated reading is

not supported by rigorous research as deflned by the quality indicators used and, therefore, is not an

evidence-based practice based on those criteria for students with and at risk for learning disabili-

ties. Implications for future research and for practice are discussed.

Uearning to read remains a hall- well, the occasion is set for them to flourish in

mark skill that essentially vocabulary and language development, compre-defines the degree of success hension, and content area learning (Stanovich,students can achieve academi- 1986). Students with or at risk for learning dis-cally throughout their school abilities (LD) are most often identified for special

career. For students who learn to read early and education services due to their measurable diffi-Exceptional Children 263

Page 2: Repeated Reading Interventions for Students With  · PDF fileRepeated Reading Interventions for Students ... supporting the notion that they experience a ... FLUENCY AND ITS ROLE

culties wich early reading acquisition, specificallydifFictilties wich word idencificacion. Basic deficitsin alphabetic coding are the underlying cause ofthese difficulties, with deficits most oftenattributed to deficiencies in reading-related cogni-tive abilities, namely phonological skill deficien-cies (Stanovich; Torgesen et al., 2001). Below-average readers frequently experience difficultieswith metaphonological and metalinguistic tasks,supporting the notion that they experience aphonological core deficit (e.g., Snowling, 2000;Stanovich & Siegel, 1994; Vellutino et al., 1996).

Students who experience a phonological coredeficit are characterized by poor phonologicalawareness and verbal short-term memory as wellas below-average speed of access to phonologicalinformation in long-term memory (Adams, 1990;Lipka, Lesaux, & Siegel, 2006). Decoding diffi-culties limit students' opportunities to read texts,decrease students' exposure to words, limit vocab-ulary learning, and hamper the development ofcontent-area expertise through reading compre-hension. Because of the effect this core deficit canhave on long-term reading achievement, earlyreading interventions generally focus on improv-ing students' phonological awareness, decodingskills, sight word identification, and fluencydevelopment. Evidence suggests that this focuscan be particularly fruitful for many studerits(Mathes, Howard, Allen, & Fuchs, 1998; Mc-Master, Fuchs, Fuchs, & Compton, 2005; Sim-mons et al., 2008; Vellutino, Scanlon, &Tanzman, 1998). Intervention efforts for studentswith LD in reading tend to focus early on the de-velopment of decoding skills followed by fluencybuilding in connected texts. More specifically andimportant for the present review, fluency-buildinginterventions have often featured repeated readingactivities aimed at improving the speed and accu-racy of students' text reading (Chard, Vaughn, &Tyler, 2002; Therrien, 2004). Repeated readingoffers students an opportunity to read and rereadthe same text multiple times and is implementedin a variety of formats including partner reading,reading to an older peer or family member, orreading with an audiotape.

F L U E N C Y A N D I T S R O L E

I N R E A D I N G P R O F I C I E N C Y

Proficient reading can be characterized as a rriulti-faceted process including at least two activities:word identification or decoding and comprehen-sion. The latter, understanding an author's mes-sage, involves making inferences, respondingcritically, and so on, and it always requires atten-tion. In order for a reader to understand what isbeing read, she cannot focus attention on bothword identification and comprehension. A readerwho is not fluent can alternate attention betweenthe two processes; however, this makes reading la-borious. If attention is consumed by decodingwords, little or no capacity is available for the at-tention-demanding process of comprehending.Therefore, automaticity of decoding—a criticalcomponent of fluency—is essential for high levelsof reading achievement (Ehri, 1995; LaBerge &Samuels, 1974).

If attention is consumed by decoding

words, little or no capacity is available

for the attention-demanding process of

comprehending. Therefore, automaticity

of decoding—a critical component

of fluency—is essential for high

levels of reading achievement.

Several theoretical frameworks have beenposited to explain the role of fluency in readingproficiency. Perfetti's (1985) proposed "verbal effi-ciency theory" highlights the importance of lowerlevel lexical skills in reading and explains the im-pact of fluent processing of information to read-ing comprehension. The verbal efficiency theorysuggests that lower level processes (e.g., wordidentification) must reach a particular thresholdlevel before higher level processes (e.g., compre-hension) can be performed simultaneously duringreading. When lower level processes are ineffi-ciently performed, higher order processes arecompromised in an attempt to compensate. Per-fetti's theory assumes that resource demands canbe reduced through learning and practice.

Spring 2009

Page 3: Repeated Reading Interventions for Students With  · PDF fileRepeated Reading Interventions for Students ... supporting the notion that they experience a ... FLUENCY AND ITS ROLE

Logan (1988) proposed an alternative, mem-ory-based theory of fluency, the "instance theoryof automatization," suggesting that automaticityand fluency are based on memory retrieval. Theinstance theory of automatization is based onthree important assumptions: (a) obligatory en-coding, (b) obligatory retrieval, and (c) instancerepresentation (Logan, 1997). Obligatory encodingrefers to focusing attention on a stimulus (e.g., av*ford) and storing details of that stimulus inmemory. Obligatory retrieval suggests that merelyattending to the stimulus is sufficient to retrieveprevious exposures of the stimulus or similarstimuli from memory. Instance representationrefers to the coding and storage of each memorytrace of experiences with a stimulus in memoryInformation recall is automatic when it relies onretrieval of "stored instances," the theoreticalmemory traces laid down in the brain each time atask is executed. Therefore, the strength of thememory trace is increased with the number oftimes the task is performed.

Logan (1997) suggested that the level of au-tomaticity is dependent on the amount of prac-tice, the level of consistency in the taskenvironment, and the number of relevant in-stances of the task recorded in memory. As thereader's knowledge expands and becomes increas-ingly accurate, performance becomes more relianton memory retrieval and less on problem solving(Logan, 1997). Applied to reading fluency, if aword is read frequently, the cumulative practiceresults in an increased likelihood that the wordwill be recognized when encountered later andthe speed will increase. We flnd the combinationof Perfetti's (1985) verbal efficiency theory andLogan's (1988) instance theory to provide intu-itive support for the notion of repeated reading asan intervention for fluency building. As studentsrepeatedly read the same content, it is likely thatthey will practice the same words multiple times,increasing the likelihood they'll be able to auto-matically retrieve those same words in future ex-posures. Simultaneously, they reduce theattention required to read the words and canfocus more intently on the meaning of what theyare reading.

EMPIRICAL RESEARCH ON FLUENCY

INTERVENTIONS

A report of the National Reading Panel (NRP;2000) signiflcantly elevated attention to fluency.NRP's review reflected the position that "fluencydevelops from reading practice" (p. 3-1), and de-voted much of this meta-analysis to research sup-port for two major approaches to providingstudents with reading practice: (a) repeated oralreading practice or guided repeated oral readingpractice, and (b) independent or recreationalreading. NRP identifled 98 articles that met thecriteria for inclusion. The panel determined thatthe mean weighted effect size for guided oral re-peated reading was 0.41, indicating that this pro-cedure "has a moderate impact on the readingachievement of the types of students participatingin these studies" (p. 3-17). It concluded that thereis sufflcient experimental evidence supporting theuse of repeated reading procedures, but insuffl-cient experimental research to suggest that in-creasing independent reading will increase fluencyor reading achievement.

RESEARCH ON FLUENCY

INTERVENTIONS AND STUDENTS

WITH LEARNINC DISABILITIES

Fluency is an essential skill for all students; stu-dents with reading/learning disabilities are mostat risk for presenting difficulties in fluency(Meyer & Felton, 1999). Based on its finding thatrepeated oral reading was associated with im-proved fluency for most students, NRP (2000)recommended teachers begin including repeatedreading activities in their reading instruction.However encouraging this recommendation ap-peared, NRP's findings may not generalize to stu-dents with LD—those with the most significantreading problems. For this reason, researchershave focused on synthesizing the findings fromfluency interventions specifically for students withidentified LD. Many of the approaches to im-proving fluency for students with or at risk forLD could be categorized as focusing on repeatedreading (Meyer & Felton); partner reading (e.g.,Arreaga-Mayer, Terry, & Greenwood, 1998;Fuchs, Fuchs, Máthes, & Simmons, 1997); orother procedures which support repeated reading.

Exceptional Children

Page 4: Repeated Reading Interventions for Students With  · PDF fileRepeated Reading Interventions for Students ... supporting the notion that they experience a ... FLUENCY AND ITS ROLE

WHY FOCUS ON REPEATED

READING?

The purpose of our review was to apply the stan-dards of evidence recommended by Homer et al.(2005) and Gersten et al. (2005) to a body of re-search examining the effects of repeated readingon the reading achievement of students with andat risk for LD. We defme repeated reading as anyintervention procedure that requires students toread passages in connected text or word lists morethan once. We selected repeated reading as ourfocus for several reasons. First, research on theeffects of repeated reading has been well docu-mented in the literature since the 1970s.Researchers have conducted numerous studies onthe effects of this intervention for students in avariety of different contexts and with differingcharacteristics, such as LD, emotional and/or be-havioral disorders, and average achieving stu-dents. In addition, several literature syntheseshave been published documenting the researchsupport for this practice (see Chard et al., 2002;Meyer & Felton, 1999) and meta-analyses havefound positive outcomes on students' readingachievement as a result of implementing this in-tervention (Therrien, 2004). Therrien reportedeffect sizes from four studies oí d = 0.75 on mea-sures of fluency and 0.73 on measures of compre-hension for students with LD. Second, in thefield, repeated reading is widely held as an evi-dence-based practice with documented effective-ness. To many practitioners, there is little doubt asto whether or not repeated reading qualifies as aresearch-based practice that is justifiably imple-mented in support of student learning. Repeatedreading deviates from "craft-based" practices thatsupport reading development in that it is basedon strong theories of learning and cognition(Ehri, 1995; Logan, 1988, 1997; Perfetti, 1985;Stanovich, 1986). Finally, because of the body ofresearch available, we anticipated finding multiplestudies in the literature over the past 30 years doc-umenting the effects of repeated reading. Becausethe focus of this article is limited to students withLD or at risk for LD, we wanted to select a prac-tice backed by research conducted with studentswith this disability.

Although these reasons would suggest thatrepeated reading has already been proven to be an

"evidence-based" practice, as of yet the researchon repeated reading has not been evaluatedagainst the rigorous quality standards needed tojustify the title of "evidence-based." Therefore,the following research question guided our inves-tigation: Is the research base supporting the effec-tiveness of repeated reading based on high-qualitystandards of single-subject andexperimental/quasi-experimental research thatwould lead to the determination that repeatedreading is an evidence-based practice?

METHOD

To address this research question, we focused onfour activities: (a) identification of interventionstudies that examined the impact of repeatedreading for students with or at risk for LD, (b)application of the quality indicators to existingliterature on repeated reading, (c) identification ofstudies that met acceptable quality standards, and(d) determination of whether or not repeatedreading could be considered "evidence-based"given the criteria established by Horner et al.(2005) and Gersten et al. (2005).

IDENTIFICATION OF STUDIES

Prior to conducting the search of studies, weestablished a priori criteria for what would consti-tute a repeated reading intervention. The inter-vention was considered eligible if students wererequired to read connected texts or word listsmore than once with the intention of improvingrate and accuracy, and the intervention did notinclude other instructional components that werefocused on other aspects of reading (e.g., compre-hension, vocabulary development).

The selection of studies for review was a mul-tistep process that began with an exhaustivesearch of the literature. First, we searched Article-First, Educational Resources Information Genter(ERIG), Google Scholar, OVID PsycINFO, andWorldGat electronic databases to locate relevantstudies conducted between January 1975 and De-cember 2006. Our search began with studies con-ducted in 1975 because it corresponded with thepassage of the Education for All HandicappedGhildren Act of 1975 (Public Law 94-142). Weincluded the following literature search terms: re-

Spring 2009

Page 5: Repeated Reading Interventions for Students With  · PDF fileRepeated Reading Interventions for Students ... supporting the notion that they experience a ... FLUENCY AND ITS ROLE

peated reading, reading automaticity, fluency, read-ing speed, reading fluency, reading rate, reading dis-ability, reading strategies, reading improvement,reading aloud, reading practice, reading instruction,assisted reading, oral reading, paired reading, single-subject design, group design, reading difßculties, stu-dents luith learning disabilities, elementaryeducation, secondary education, and primary educa-tion. Electronic searches used multiple combina-tions and sequences of the literature search terms.

Second, we conducted an ancestral searchusing the reference lists from pertinent studiesconducted by Chard et al. (2002); Kuhn andStahl (2003); Meyer and Felton (1999); Swanson,Hoskyn, and Lee (1999); and Therrien (2004).Third, we conducted a manual search of recentliterature in several major journals (from 2004 toDecember 2006) of special, remedial, elementary,and secondary education: American Journal of Ed-ucation; Assessment for Effective Intervention; Be-havioral Disorders; Cognition and Instruction;Educational Psychology; Educational Researcher; El-ementary School Journal; Exceptional Children;Journal of Applied Behavior Analysis; Journal of Ed-ucation and Behavior Statistics; Journal of Educa-tional Psychology; Journal of Educational Research;Journal of Experiential Education; Journal of Exper-imental Education; Journal of Learning Disabilities;Journal of Literacy Research; Journal of Negro Edu-cation; The Journal of Special Education; Journal ofSpecial Education Technology; Journal of Speech,Language, and Hearing Research; Learning Disabil-ities Research & Practice; Learning Disability Quar-terly; Memory and Cognition; Mental Retardation;Peabody Journal of Education; Reading Horizons;Reading Improvement; Reading Research and In-struction; Reading Research Quarterly; ReadingTeacher; Remedial and Special Education; SchoolPsychology Review; and Scientiflc Studies of Read-ing

This process identified 92 articles pertainingto repeated reading. We applied additional criteriaas a secondary screener to these studies: The stud-ies should focus on the effects of repeated readingfor students with or at risk for LD in reading inK to 12 settings. Studies with subjects who wereperforming below average in reading were not in-cluded unless the participants could be identifiedas having LD or being at risk for LD. We re-frained from making any judgments about risk

factors based on the achievement data reportedand relied on the authors' documentation of sta-tus. Additional criteria included the isolation ofthe independent variable. We omitted studies ifrepeated reading was studied in conjunction withother practices and was not isolated as a level ofthe independent variable because it was not possi-ble to examine the effects of repeated reading onstudent performance. This screening process re-sulted in 11 single-subject and experimental orquasi-experimental research studies that met all ofour criteria (see Tables 1 and 2).

APPLICATION OF QUALITY INDICATORS

We limited the examination of repeated readingto single-subject and experimental and quasi-ex-perimental research studies. Focusing on sugges-tions provided by Horner et al. (2005) andGersten et al. (2005), we adapted the quality indi-cators of rigorous research to create rubrics forevaluating the reported methodological rigor ofthe studies. The rubric encompasses the essentialindicators of high-quality research proposed byeach group of authors but provides a continuumfor evaluation. Because most research studies im-plement rigorous methodological standards tovarying degrees, we decided to represent this vari-ance on a Likert scale. As such, instead of consid-ering the indicators as dichotomous qualities of astudy, we created a 4-point scale to refiect thevariable nature of conducting and documentingresearch. We created separate rubrics for single-subject and experimental/quasi-experimental re-search (see Figures 1 and 2).

Interpreting the Quality Indicators. We createdratings for the levels of rigor for each indicator byevaluating the study for each component associ-ated with the quality indicators proposed byHorner et al. (2005) and Cersten et al. (2005). A1-point rating generally refiected a quality indica-tor component that was not documented in thepublished article. For example, an article reportingon an experimental or quasi-experimental studyreceived a component score of 1 for the effect sizecalculation component (included in the data anal-ysis quality indicator) if no effect size was re-ported. In contrast, a 4-point component ratingfor a particular quality indicator indicated that thearticle met the full criteria of the component as

Exceptional Children 2 6 7

Page 6: Repeated Reading Interventions for Students With  · PDF fileRepeated Reading Interventions for Students ... supporting the notion that they experience a ... FLUENCY AND ITS ROLE

TABLE 1

Average Ratings for Single-Subject Research Articles on Repeated Reading

Quality Indicators

StudyParticipants/ Dependent Independent EC/Internal External Social

Setting Variable(s) Variable Baseline Validity Validity Validity

Begeny, Daly, &Valleley (2006) 1.67̂ 3.50 2.33 2.50

Chafouleas, Martens,Dobson, Weinstein,& Gardner (2004) 1.83' 2.90" 2.50' 2.75

Daly & Martens (1994) 2.33" 3.10 2.67 3.50

Freeland, Skinner,Jackson, McDaniel,& Smith (2000) 3.00 3.30» 3.17 2.75»

Strong, Wehby, Falk,& Lane (2004) 2.67 3.40 3.33 3.50

Weinstein & Cooke(1992) 2.33" 3.10" 2.67 3.00

2.67

3.00

2.00"

3.00

3.33

3.17

1.00"

1.00"

2.00

1.00"

1.00"

1.00"

3.00"

3.13"

3.38

3.13

3.13

3.00

Note. Selection based on criteria proposed by Horner et al. (2005). EC = experimental control."Article received a 1-point score for at least one subcomponent in this methodological category.

established by Horner et al. or Gersten et al. Forexample, an article that documented 80% orgreater interobserver agreement for data collectionusing single-subject research methods received 4points for the data collected on the reliability orinterobserver agreement component of the depen-dent variable quality indicator.

We created 2- and 3-point ratings to reflectthe range of reported research fmdings. A 2-pointrating of a component typically reflected a limited

description of the methodological proceduresused in the study. For example, if the descriptionof the participants in an experimental or quasi-ex-perimental study provided some informationabout the interventionists or teachers but no in-formation about their comparability across treat-ment groups, the article received a 2-point ratingon the information about interventionists orteachers component included in the descriptionof participants quality indicator. Because some

TABLE 2

Average Ratings for Experimental and Quasi-Experimental Research Articles on Repeated Reading

Essential Quality Indicators

StudyDescription

of Participants

3.67

2.50"

2.50"

2.33

2.33"

Intervention/ComparisonConditions

3.00

2.33"

2.33"

1.83"

2.00"

Outcome

Measures

3.75

2.752.50

2.75

3.00

Data

Analysis

3.75

2.25"1.50"

1.75"

2.25"

MathesSi Fuchs (1993)

O'Shea Sindelar, & O'Shea (1987)

Rashotte & Torgesen (1985)

Sindelar, Monda, & O'Shea (1990)

Young, Bowers, &c MacKinnon, (1996)

Note. Selection based on criteria proposed by Horner et al. (2005)."Article received a 1-point score for at least one subcomponent in this methodological category.

2 6 8Spring 2009

Page 7: Repeated Reading Interventions for Students With  · PDF fileRepeated Reading Interventions for Students ... supporting the notion that they experience a ... FLUENCY AND ITS ROLE

F I G U R E 1

Quality Indicators of Single-Subject Research Articles and Reports

Participants and Setting

Sample characteristics(e.g., age, gender,disability, diagnosis)

Process for selectingparticipants

Critical features of thephysical setting

Dependent Variable

Description ofdependent variable

Measurementprocedure

Measurement validityand description

Measurementfrequency

Data collected onreliability (minimalstandards: IOA = 80%;Kappa = 60%)

Independent Variable

Description ofindependent variable

IV manipulation

Fidelity ofimplementation

1

No detailprovided

No descriptionof selectionprocess

No descriptionprovided

1

No descriptionprovided

No procedureprovided or noquantifiablevariables

No validmeasures anddescription notreplicable

No repeatedmeasures

No reliabilitydata reported

1

Only name orvague descrip-tion of IV pro-vided

IV is providedwith no control

No measure offidelity

2

Limited detailprovided

Proceduresdescribed butnot appropriateand/or withlimited detail

Litniteddescriptionprovided

2

Limiteddescriptionprovided

Procedureprovided butno quantiftablevariables

No validmeasures ordescription notreplicable

Measurement re-peated but veryinfrequently

Reliability dataincorrectlycollected oranalyzed

2

IV is describedwith little detail

Little controlexercised (e.g.,monitor, scripts)

Fidelity ismonitored butnot directly

3

Some detailprovided

Proceduresdescribed areappropriate butminimallydescribed

Some descriptionprovided

3

Some descriptionprovided but notoperational

Procedure pro-vided but onlysome variablesquantiftable

Some measuresvalid; descriptionis replicable

Measurementrepeated butinfrequently

Reliability datareported butminimal stan-dards not met

3

Majorcomponents ofIV provided withsome detail (e.g.,scripts provided)

Conditionassignmentis planned

Fidelity ismonitoreddirectly, but atlarge componentlevel

4

Ample detailprovided

Procedures wereappropriate andadequatelydescribed

Detailed descrip-tion provided toallow replication

4

Operationaldescriptionprovided

Procedureprovided andall variablesquantifiable

Measures arevalid anddescription isreplicable

Measurementrepeatedfrequently

Reliability datareported andminimalstandards met

4

All componentsof IV describedin detail with ef-forts to commu-nicate precision

Randomassignment tocondition

Fidelity ismonitored indetail with cor-rections providedwhen necessary

Score

Score

Score

continues

Exceptional Children 2 6 9

Page 8: Repeated Reading Interventions for Students With  · PDF fileRepeated Reading Interventions for Students ... supporting the notion that they experience a ... FLUENCY AND ITS ROLE

F I G U R E 1 (Continued)

Baseline

DV measurement

Description of baselinecondition

Experimental Control/Internal Validity

Design demonstratesexperimental effect

Design controls for commonthreats to internal validity(e.g., elimination of rivalhypotheses)

Pattern of results

Extemal Validity

Replication of effects (e.g..across participants, settings.or materials to establishexternal validity)

Social Validity

Importance of DV

Importance of magnitudeof change in DV

Practicality and costeffectiveness ofimplementation of IV

Typical nature ofimplementation of IV

1

DVnotmeasuredobjectively

No descriptionof baseline

1

No demon-stration ofexperimentaleffect

No control forthreats tovalidity

Results donot suggestexperimentalcontrol

1

No effors toreplicate efforts

1

No importance

No importance

Impracticaland not costeffective

IVimplementedin atypicalmanner

2

DV measuredinfrequently;data is missingor not stable

Vaguedescriptionof baseline

2

Only onedemonstrationof experimentaleffect

Few threatscontrolled

Results suggesta change intrend, level, orvariability

2

Few replicationsattempted

2

Somewhatimportant

Either practicalor cost effective.but not both

IV imple-mented eitherin typical con-text or by typi-cal agent, not30th

3

DV measuredfrequently butnot stable

Baselinedescriptiondetailedbut limited

3

More than onedemonstrationof experimentaleffect

Most threatscontrolled

Results docu-ment a changein trend, level.or variability

3

Somereplicationattempted

3

Important

Some evidenceof practicalityand costeffectiveness

Implementa-tion extendedin somewhattypical contextsand with asomewhattypical agent(e.g., certifiedteacher)

4

DV measuredfrequently andis stable beforeintervention

Baselinedescriptiondetailed andextensive

4

Three or moredemonstrationsof experimentaleffect

All threatscontrolled

Results docu-ment a patternof experimentalcontrol

4

Multiplereplicationsacross variables

4

Important

Very important

Practical andcost effective

Implementa-tion extendedin typicalcontexts withtypical agents(e.g., thecertifiedteacher)

Score

Score

Score

Score

Note. Based on quality indicators proposed by Horner et al. (2005) and Gersten et al. (2005). IOA = interobserveragreement; IV = independent variable; DV = dependent variable.

2 7 0 Spring 2009

Page 9: Repeated Reading Interventions for Students With  · PDF fileRepeated Reading Interventions for Students ... supporting the notion that they experience a ... FLUENCY AND ITS ROLE

F I G U R E 2

Quality Indicators of Experimental and Quasi- Experimental Research Articles and Reports

Description of Participants

Information aboutdiagnosis of disabilityor difficulty

Samples are comparableacross conditions onrelevant characteristics

Information aboutinterventionists orteachers; comparabilityacross conditions

Description andImplementation ofIntervention andComparison Conditions

Description ofintervention andimplementationprocedures

Description of fidelityof implementationprocedures

Description ofcomparison conditionactivities

1

No evidenceand/ordescription

No proceduresfor comparabilitydescribed

No informationor descriptionprovided; noinformationabout com-parability acrossgroups

1

Minimaldescriptionprovided; nodetails

No informationprovided

Minimaldescriptionprovided; nodetails

2

Little evidenceand/ordescription

Proceduresdescribed bubnot appropriate

Someinformadon ordescriptionprovided; noInformationabout com-parability acrossgroups

2

Some descriptionprovided; limiteddetails

Someinformationprovided;evaluation andeffects onintervention im-pact notdescribed

Some descriptionprovided; limiteddetails

3

Some evidenceand/ordescription

Proceduresappropriate butminimallydescribed

Some informa-tion or descrip-tion provided;some informa-tion about com-parability acrossgroups

3

Some descriptionprovided; generaldetails lackingspecificity forreplication

Some informa-tion provided;evaluation andeffects on inter-vention impactminimallydescribed

Some descriptionprovided; generaldetails lackingspecificity forreplication

4

Ample evidenceand/ordescription

Proceduresappropriate andadequatelydescribed

Sufficientinformation ordescriptionabout interven-tionists provided;comparableacross groups

4

Description cleatand specific forreplication

Sufficientinformationprovided;evaluation andeffects oninterventionimpact described

Description clearand specific forreplication

Score

Score

continues

Exceptional Children 27 f

Page 10: Repeated Reading Interventions for Students With  · PDF fileRepeated Reading Interventions for Students ... supporting the notion that they experience a ... FLUENCY AND ITS ROLE

F I G U R E 2 (Continued)

Outcome Measures

Multiple measures ormeasures of generalizedperformance (Weremultiple measures used tomeasure the DV? Weremeasures of generalizedperformance used?)

Appropriateness of datacollection times

Data Analysis

Data analysis linked toresearch questions/hypotheses; consideredthe unit of analysis

EfFect size calculation

1

Only usedmeasure tightlyaligned tointervention

No informationabout timingprovided

1

No informationabout dataanalysis provided

Effect sizes notreported

2

Used measuretightly alignedto interventionalong with oneother measure

Timing ofadministrationof outcomemeasures notappropriate

2

Data analysistechniques notappropriategiven the re-searchquestions/hypotheses; unitof analysis mayor may not havebeen appropriate

EfFect sizesreported butnot interpreted

3

Used measureof generalizedperformance

Timing ofadministrationof outcomemeasures some-what appropriate

3

Data analysistechniquesmostly appropri-ate given re-search questions/hypotheses(alternate meth-ods could beused that weremore elegant);used appropriateunit analysis

Effect sizesreported butnot accuratelyinterpreted

4

Used multiplemeasures andmeasure ofgeneralizedperformance

Timing ofadministrationof outcomemeasuresappropriate

4

Data analysistechniquesappropriategiven theresearchquestions/hypotheses; usedappropriate unitof analysis

Effect sizesreported andaccuratelyinterpreted

Score

Score

Note. Based on quality indicators proposed by Horner et al. (2005) and Gersten et al. (2005). IOA = interobserveragreement; DV = dependent variable.

information was provided, the component shouldscore higher than 1 point, but limited details stillmake this description unacceptable. Conversely, ifmore details were provided but lacked specificityfor replication or generalization, the articlereceived 3 points for that component. For exam-ple, a single-subject research study that reportedmonitoring the fidelity of implementation, acomponent of the independent variable qualityindicator, might receive 3 points because althoughdetailed information was provided about moni-toring, the authors may not have documented

correction procedures. The study should not re-ceive 4 points for the independent variable qualityindicator because fidelity of implementation in-formation was not provided at the level of preci-sion proposed by Horner et al. (2005). However,because much of the necessary information waspresent, we considered this acceptable.

Members of the author team not involved increating the rubrics evaluated them to ensurealignment with the quality indicators proposed byHorner et al. (2005) and Gersten et al. (2005). Inaddition, we applied the rubric to a sample article

2 7 2 Spring 2009

Page 11: Repeated Reading Interventions for Students With  · PDF fileRepeated Reading Interventions for Students ... supporting the notion that they experience a ... FLUENCY AND ITS ROLE

to calibrate the raters' interpretations of the scale.After independently reviewing the rubric and ap-plying it to a sample article, we made several ad-justments to clarify the expectations of the levelsof rigor associated with each score.

IDENTIFICATION OF STUDIES THAT MET

ACCEPTABLE QUALITY STANDARDS

Two raters using the corresponding rubric inde-pendently evaluated single-subject and experi-mental/quasi-experimental research articles. Theraters included a researcher with 4 years experi-ence teaching and researching in the field of ap-plied measurement and assessment and asecond-year doctoral student studying special ed-ucation. Initial ratings were entered into a spread-sheet and evaluated for correspondence. Ininstances where raters disagreed by 2 or morepoints and/or one rater assigned a componentscore of 1, an independent third rater evaluatedthe quality of the component and provided an ar-bitration score. As such, the third rater evaluatedany indicator that received a component score ofL The arbitration score supplanted the initial rat-ing. The independent rater was a third-year doc-toral student in special education who wasfamiliar with the rubric but not the literature inrepeated reading.

Scores on the quality indicators were derivedfrom the totals of each component associatedwith each quality indicator. We aggregated thescores for the quality indicators across raters basedon the methodological category (e.g., descriptionof participants, dependent variable). Combiningscores at the broader level accounted for theminor variance across raters present at the compo-nent level. Any category that did not receive anaverage acceptable score (3 points or greater)across indicators was considered unacceptable.Moreover, any indicator that earned a componentscore of 1 by two or more raters (the two initialraters and/or the independent rater) was consid-ered unacceptable.

Interrater reliability was calculated by divid-ing the number of exact matches on ratings at thecomponent level by the total number of exactmatches and disagreements. This resulted in relia-bilities of 0.36 for single-subject studies and 0.53for experimental/quasi-experimental studies.

When calculating interrater reliability by consid-ering agreements to include exact matches and1-point discrepancies, reliabilities were 0.96 forsingle-subject studies and 0.91 forexperimental/quasi-experimental studies.

DETERMINATION OF EVIDENCE-BASED

PRACTICE

We used the criteria proposed by Horner et al.(2005) and Gersten et al. (2005) to make a fmalevaluation of repeated reading as an evidence-based practice. For single-subject research, Horneret al. noted that a practice can be labeled "evi-dence-based" if five studies that meet minimallyacceptable methodological criteria document pos-itive effects from implementing the practice. Ofthese studies, the research needs to be conductedby three different researchers from three regionsof the country. In addition, the number of sub-jects across the five studies needs to total at least20. According to Gersten et al., research con-ducted using experimental or quasi-experimentaldesigns needs to meet all but one of the essentialquality indicators to be considered high quality.These studies must also have one to four desirableindicators to be acceptable. Four acceptable-qual-ity studies or two high-quality studies document-ing the effectiveness of the practice along with aweighted effect size significantly greater than zeroare needed in order to label the practice as "evi-dence-based." A "promising practice" requires thesame numbers of studies but can have a 20% con-fidence interval for the weighted effect sizes that isgreater than zero. We applied these criteria to therepeated reading studies we examined to deter-mine if repeated reading could be labeled as "evi-dence-based" for students with or at risk for LD.

R ESU LTS

QUALITY OF THF SINGLE-SUBJECT

RESEARCH STUDIES

We analyzed six single-subject research studies forcorrespondence with the quality indicators identi-fied by Horner et al. (2005), organizing 21 com-ponents into seven methodological categories orquality indicators: description of participants andsetting, dependent variable, independent variable.

Exceptiorml Children 2 7 3

Page 12: Repeated Reading Interventions for Students With  · PDF fileRepeated Reading Interventions for Students ... supporting the notion that they experience a ... FLUENCY AND ITS ROLE

baseline, experimental control/internal validity,external validity, and social validity (see Table 1).

We evaluated the description of the partici-pants and setting by reviewing the articles' discus-sion of the sample characteristics, process forselecting participants, and description of criticalfeatures of the physical setting. Only one studymet the minimum requirements of acceptabilityfor this category (i.e., mean rating of 3 or greaterfor each quality indicator and no componentrated as 1 by two or more raters). Most studiesprovided adequate information about the sampleparticipants; however, the process for selecting theparticipants was either absent or superficially dis-cussed. Moreover, the description of the settinglacked sufficient details to interpret the quality ofthe research design.

We considered documentation of the depen-dent variable, measurement procedures includingvalidity and technical consideration, and fre-quency and reliability of data collection for eachmanuscript. Of the six single-subject researchstudies reviewed, all but one of the studies re-ceived an average score of 3 on this quality indica-tor. However, two studies received a score of 1based on the lack of description of measurementvalidity and otie study received a 1-point score be-cause it did not discuss reliabihty information fordata collection. Because the minimurh criteria es-tablished for our review also required each studyto have adequate representation of the subcompo-nents in the category, these three studies pre-sented unacceptable evidence about the quality ofthe research for this category.

Reports of the overall description, manipula-tion, and fidehty of implementation of the inde-pendent variable determined if the researchstudies presented acceptable evidence to docu-ment the rigor of the research with regard to thequality indicator on the independent variable.Four of the six research studies did not meet theminimutn criteria for acceptability in this cate-gory. Two of the studies that did not provide ade-quate documentation failed to provide detailsabout the fidelity of implementation of the inter-vention. Other studies provided limited informa-tion about the independent variable includingprecise implementation techniques or scripts.

Three studies did not meet the minimumcriteria needed to establish rigorous research

methodology for documentation of baseline pro-cedures primarily due to a limited description ofthe baseline condition. Two studies did not meetthe minimum criteria for their documentation ofthe experiniental control/internal validity. Mostmanuscripts reported on studies that were welldesigned to demonstrate experimental effects ofrepeated reading including observing a pattern ofresults due to the implementation of the interven-tion that documents variability, trend changes, ordifferences in level of performance. However, sev-eral studies provided limited information on theexperittiental controls employed by the re-searchers to address common threats to internalvalidity.

Only one subcomponent evaluated the ade-quacy of the external validity of the documentedresearch. No study we reviewed met the minimumrequirfenifents needed to document the potentialgeneralizability of results by replicating the studyacross participants, settings, or materials. It shouldbe noted that we only considered participants withLD in this evaluation. As such, some studies mayhave included measures to ensure external validityacross the range of participants, but did not repli-cate with students with or at risk for LD.

Four of the six research articles reviewed pro-vided adequate documentation to justify the so-cial validity of the research. We evaluated researchstudies on four dimensions: importance of the de-pendent variable, importance of the magnitude ofchange in the dependent variable, practicality andcost effectiveness of the independent variable, andnature of implementation of the independentvariable. Two studies received 1-point scores be-cause they used procedures and/or resources riottyjjically available to classroom teachers (such asspecially trained interventionists).

QUALITY OF THE EXPERIMENTAL AND

QUASI-EXPERIMENTAL RESEARCH STUDIES

We evaluated five research studies that used ex-perimental or quasi-experimental research meth-ods, categorizing nine components into the foui-methodological categories of quality indicator de-fined by Gersten et al. (2005): description of par-ticipants, description and implementation of theintervention and comparison groups, outcome

2 7 4 Spring 2009

Page 13: Repeated Reading Interventions for Students With  · PDF fileRepeated Reading Interventions for Students ... supporting the notion that they experience a ... FLUENCY AND ITS ROLE

measures used, and data analytic techniques (seeTable 2).

We evaluated each research study for its de-scription of the participants—the quality of thedocumentation of students' demographic infor-mation including disability diagnoses proceduresand comparability of students across conditions.In addition, we evaluated information about theinterventionists or teachers implementing the in-tervention along with the comparability of theircharacteristics and/or credentials across condi-tions. One study met the minimum criteria ofearning an average score of 3 across all compo-nents with no 1-point scores. The remainingstudies did not adequately report informationabout the interventionists and/or teachers. In ad-dition, several studies provided limited informa-tion about the diagnoses procedures used todocument the participants' disabilities.

To evaluate the description and implementa-tion of the intervention and comparison condi-tions, we reviewed the research studies for theiroverall description of the intervention and com-parison conditions as well as the fidelity of imple-mentation procedures. Four of the five studiesreviewed did not meet the minimum require-ments for rigorous research in this methodologicalcategory primarily due to the lack of informationabout the fidelity of implementation of the inter-vention. Most studies reviewed did not providesufficient information to determine if the inter-vention was instituted as proposed by the re-searchers.

Three studies did not meet the minimumcriteria needed to establish rigorous researchmethodology for documentation of the outcomemeasures. In most cases, the articles documentedappropriate data collection times given the re-search clesign used. However, several studies didnot use multiple dependent measures or measuresof generalized performance to document the ef-fects of the independent variable. Similarly, fourstudies did not meet the minimum requirementsfor data analysis. Although most studies used dataanalysis techniques that aligned with the researchquestion and were applied at the appropriate unitof analysis, only one study reported efFect sizes. Assuch, all other studies received a 1-point score be-cause this information was missing.

D I S C U S S I O N

EVALUATION OF EVIDENCE BASE

To determine if repeated reading could be classi-fied as an evidence-based practice, we applied thecriteria established by Horner et al. (2005) andGersten et al. (2005) to the single-subject and ex-perimental/quasi-experimental research studies,respectively.

Single-Subject Research Studies. We examinedsix single-subject research studies for correspon-dence with the expectations of rigorous researchin seven methodological categories. No studiesmet the minimum requirements for rigorous re-search in all seven categories. Strong, Wehby,Falk, and Lane (2004) provided acceptable evi-dence in five of the seven categories and Freeland,Skinner, Jackson, McDaniel, and Smith (2000)provided adequate documentation in four of theseven categories. The remaining four studies pro-vided satisfactory evidence in three or fewermethodological categories. Applying the criteriaproposed by Horner et al. (2005), these studiesdo not provide ample evidence that the researchwas conducted following rigorous methodologicalstandards. To be considered an evidence-basedpractice, at least five high-quality research studiesare needed that document the effects of the inter-vention. Because no studies reviewed in the inves-tigation qualified as high-quality single-subjectresearch, repeated reading does not meet the re-quirements for being an evidence-based practicefor students with or at risk for LD.

Experimental and Quasi-Experimental Re-search Studies. We evaluated five experimental orquasi-experimental research studies across fourmethodological categories to determine if the arti-cles documented high-quality reports of research.Four of the five studies provided acceptable docu-mentation in only one or fewer categories. How-ever, Mathes and Fuchs (1993) provided ampleevidence of rigorous research across the fourmethodological categories. Follow-up analysisdocumenting the presence of desirable indicatorsof quality research was necessary to determine ifthis article reports acceptable-quality or high-quality research as described by Gersten et al.(2005). Mathes and Fuchs provided adequate evi-dence for four desirable indicators by document-ing the concurrent related evidence for validity of

Exceptional Children 2 7 S

Page 14: Repeated Reading Interventions for Students With  · PDF fileRepeated Reading Interventions for Students ... supporting the notion that they experience a ... FLUENCY AND ITS ROLE

the reading acbievement measures, monitoring fi-delity of implementation beyond "surface" fea-tures, documenting instructional practices for thecomparison groups, and presenting the results ina clear and coherent manner. As such, the re-search reported in this article would be classifiedas high quality.

To be classified as an evidence-based practice,four acceptable-quality or two high-quality exper-imental or quasi-experimental research studies areneeded as well as a weighted effect size that is sig-nificantly greater than zero. Because the articlewritten by Mathes and Fuchs (1993) was classi-fied as high quality, one additional high-qualitystudy with a weighted effect size significantlygreater than zero would be needed to classify re-peated reading as an evidence-based practice.Without this documentation, our results indicatethat repeated reading does not qualify as an evi-dence-based or promising practice for studentswith or at risk for learning disabilities accordingto the requirements proposed by Gersten et al.(2005). It is possible to interpret the criteria foran evidence-based practice as stating that the ef-fect size across high-quality and acceptable studieshad to be significantly different from zero. Ourinterpretation, however, was that the effect size ineach experimental/quasi-experimental study hadto be significantly greater than zero. We chose thismore conservative interpretation as we could en-vision the possibility that in some high-qualitystudies there could be a finding of no differencebetween treatment and comparison but, if aggre-gated with the effect sizes ftom other studies withsignificantly large effects, the result could misleadthe field.

IMPLICATIONS FOR RESEARCH

Most special education researchers would likelyhave identified repeated reading as an evidence-based practice for use in building reading fluencyfor students with learning disabilities. However,the results of our review—based on the standardsfor rigorous research established by Horner et al.(2005) and Gersten et al. (2005)—suggest other-wise. We believe these results have implicationsfor the design and implementation of specificmethodologies as well as for research and researchfunding in general.

Implications for Single-Subject Research. Re-flecting on the findings from those studies em-ploying single-subject research designs, we foundseveral consistent problems. Many of the single-subject studies either did not describe or only su-perficially discussed the process for selectingparticipants. Additionally, the description of thesetting often lacked sufficient detail for replicabil-ity. The single-subject studies also often omitteddescriptions of measurement validity, and in somecases reliability. Moreover, many studies providedlittle detail about how fidelity of implementationwas monitored, if at all. In lieu of measuring fi-delity of implementation, we expected that tbeinterventions would have been sufficiently de-tailed to allow replicability. However, this wasoften not the case. Frequently, the studies failed intheir designs to address common threats to valid-ity either by failing to describe the baseline condi-tion thoroughly, not achieving stability in thebaseline before implementing the treatment, ornot having a control or comparison condition(e.g., some studies examined growth from pretestto posttest but did not include a second groupthat did not receive the experimental treatment).Finally, a common difficulty with the studies wasthe lack of replication across participants, settings,or materials. In fact, the set of studies includedfor criteria did not include any studies that repli-cated research on a specific treatment. In all cases,the treatments were so sufficiently different fromtreatments in all other studies that it was impossi-ble to generalize the findings to a particular ap-proach to repeated reading.

Our results indicate that repeated readingdoes not qualify as an evidence-based or

promising practice for students with or at

risk for learning disabilities according to therequirements proposed by Gersten et al

These common problems suggest that thesingle-subject research on repeated reading needsto be more metbodologically rigorous. Horner etal. (2005) required that in order for a study to beconsidered of acceptable rigor, it needed to meetall the established criteria. Although this standardis set high, it was not the case that many studies

2 7 6 Spring 2009

Page 15: Repeated Reading Interventions for Students With  · PDF fileRepeated Reading Interventions for Students ... supporting the notion that they experience a ... FLUENCY AND ITS ROLE

failed to meet the standard by only one or twocriteria; several studies failed to demonstrate high-quality research techniques in multiple method-ological categories. As such, the research is notclearly presented to allow the authors of thesestudies to rule out alternative hypotheses for theirfindings.

Implications for Experimental and Quasi-Experimental Research. Several problems were per-sistent across the experimental or quasi-experi-mental research studies examined, includingproviding very limited information about the in-terventionists or teachers implementing the treat-ment. This is particularly problematic as it makesit difficult to attribute the outcomes of the inter-vention to the intervention itself rather than possi-bly to the teachers' specific qualifications orcharacteristics. In addition, studies frequentlylacked detail regarding the diagnosis of students'disabilities. This is mostly a problem in terms ofgeneralizing the effects of the interventions topopulations of students with learning or readingdisabilities. Moreover, several studies lacked suffi-cient detail about the fidelity of implementationboth in terms of how it was measured, if at all, andhow it was ensured. Many early studies failed toinclude any information on fidelity of implemen-tation; more contemporary studies have includedmore detailed information in this area as it has be-come standard to expect fidelity data as well as in-formation regarding how fidelity was maintained.In a similar fashion, many of the older studies didnot report effect sizes. Until the American Psycho-logical Association (2001) made it an expectationto teport the magnitude of effect for treatments,they were not commonly included in studies. Fi-nally, it was common for experimental and quasi-expetimental studies to use a single measure ofperformance and in many cases they did not in-clude a generalized measure of performance.Under these circumstances it is difficult to deter-mine the efficacy of the repeated reading interven-tions on general measures of reading fluency.

In recent years, expectations for rigor haveincreased in experimental and quasi-experimentalresearch designs. One could argue that theseheightened expectations are both a result of theevolution of educational research in general aswell as increased rigor related to funded researchcompetitions. The fact that only one experimental

or quasi-experimental study of repeated readingconducted with students with or at risk for learn-ing disabilities met the criteria for high-quality re-search is disconcerting. On its face, this findingwarrants attention and should encourage re-searchers in this area to more carefully considerthe criteria for high-quality research. Additionally,these findings should alert funding agencies toconsider supporting studies that focus on replica-tion of research aimed at answering fundamentalresearch questions about interventions that areoften assumed effective. Unfortunately, there islittle evidence in the studies we reviewed of effortsto replicate findings, a basic tenet of scientific re-search.

Beyond encouraging researchers to strive tomeet the criteria, we believe it is important toconsider different interpretations for the ratherdismal results of our reviews of both sets of stud-ies. One possible reason for the poor outcome ofboth our reviews may be that we applied the crite-ria inappropriately. As noted in the Method sec-tion, we created a process for our review thatallowed reviewers to provide a score for eachstudy across all indicators that reflected more thanjust a dichotomous "present" or "absent" score.However, we believe this approach actually re-laxed the criteria and more likely overestimatedthe quality of the research rather than underesti-mated it. In addition, we implemented an ap-proach that we felt resulted in relative consensusamong our reviewers while allowing for varyingviews. Our approach was designed to refiect aprocess that would more closely represent the waythese criteria will be applied generally.

A second possible reason for our results maybe that the criteria established by Horner et al.(2005) and Gersten et al. (2005) are simply toorigorous. This explanation seems untenable to us.Although the criteria established by these twogroups of researchers are indeed rigorous, the cri-teria represent fundamental features of researchstudies that we convey to doctoral students in in-troductory courses every year. Moreover, these cri-teria serve as clear targets that researchers shouldconsider both in designing their research as wellas in describing the results of their research fordissemination. In fact, we would suggest that sev-eral additional criteria should be considered. Forexample, a "high-quality" research study should

Exceptional Children 2 7 7

Page 16: Repeated Reading Interventions for Students With  · PDF fileRepeated Reading Interventions for Students ... supporting the notion that they experience a ... FLUENCY AND ITS ROLE

include a strong theoretical or conceptual frame-work for the intervention under review. Althoughthese frameworks are often assumed, authorsshould be encouraged to ensure that the researchquestions they propose are logically linked toexisting theories so that their answers contributemeaningfully to the literature. We also feel thatseveral of the desirable indicators proposed byGersten et al. should be included as essential qual-ity indicators. For example, we would prefer tosee the addition of construct and criterion-relatedevidence for validity and quality of implementa-tion as essential quality indicators. In brief, we donot believe that the criteria were too rigorous; wefelt the rigorous nature of the criteria should in-spire research designs that more definitively ad-dress the focal research question(s) and encourageresearchers to strive to accurately and precisely re-port their research findings.

These criteria serve as clear targetsthat researchers should consider both

in designing their research as well asin describing the results of their

research for dissemination.

Although the criteria as established byHorner et al. (2005) and Gersten et al. (2005) re-sulted in disappointing findings, our findings il-lustrate how difficult it is to conduct high-qualityresearch. Rather than reducing the criteria forquality, our research community should carefullyconsider how to reinforce efforts to improve thequality of research. These considerations mayinclude such steps as encouraging replicationstudies in dissertations, providing replicationcompetitions through funding agencies, andencouraging research collaborations that wouldresult in replicating studies in multiple regions ofthe country. To achieve the goal of increasing thequality of our research, we may need to considercriteria for tenure and promotion that are basedon quality of research as much as quantity of re-search published. Perhaps, if application of thesecriteria across numerous intervention areas resultsin similarly poor outcomes, we should rethink theway we support the implementation of efficacystudies in the future.

IMPLICATIONS FOR PRACTICE

We are reluctant to draw too many conclusionsabout the implementation of repeated readingpractices with students with LD based on thefindings of our review. Glearly, we are unable todescribe repeated reading as an "evidence-based"practice according to the criteria set forth byHorner et al. (2005) and Gersten et al. (2005).Despite this finding, we are loath to suggest thatteachers stop implementing repeated readingpractices for two reasons. First, as described insome detail in the introduction, repeated readingis a logical extension of multiple theoreticalframeworks that suggest its use in supporting stu-dents who need fluency development. Second,though few of the studies we reviewed met thecriteria for high-quality research, the effect sizespublished in meta-analyses suggest that repeatedreading is likely to positively affect fluency out-comes for students who are building fluency. Thecombination of theoretical support and positiveeffect sizes suggest to us that in the absence of in-novations with documented effectiveness in im-proving fluency, repeated reading practices shouldbe continued.

CHALLENGES IN REVIEWING THE

RESEARCH

As a team of researchers, we found the process ofthis review to be both humbling and daunting.Humbling because we found ourselves surprisedby our own assumptions of what would be con-sidered high-quality or acceptable quality re-search. We recognize that in most cases, our ownresearch would not have met the criteria for "highquality." Daunting as we considered the prospectof evaluating critically the research published byour respected colleagues. We want to acknowl-edge, therefore, that we are acutely aware of howcomplex it is to conceive, conduct, and report re-search that will meet the criteria described hereand contribute meaningfully to the research liter-ature.

Our greatest challenge in this review was cre-ating a process that took advantage of the range ofknowledge of our reviewers about interventionsand intervention research while accurately reflect-ing the quality of each study as it related to thecomponents and indicators of quality. For us.

2 7 8 Spring 2009

Page 17: Repeated Reading Interventions for Students With  · PDF fileRepeated Reading Interventions for Students ... supporting the notion that they experience a ... FLUENCY AND ITS ROLE

overcoming this challenge required us to pursuethe rubric-based review we offered in this report.This was clearly an extension of the initial criteriafor which we are responsible. We felt our rubricapproach allowed us to inject variance in the re-view of studies that resulted in a more accurate re-flection of the quality of each study. The rubricwas difficult to apply consistently across reviewersas is illustrated in our interrater reliabilities. Webelieve this inconsistency was due to the subjec-tivity of language used in the rubric. For example,reviewers often did not agree about whether therewas "little" information provided about partici-pants or "some" information provided. However,we were encouraged when we relaxed the reliabil-ity criteria to include ratings foi- each componentthat were within one point of each other. In thiscase, the reliability increased significantly. With-out the range of ratings, we feared that morestudies would have failed to meet the standards.In the case of repeated reading, we anticipatedthat this approach would result in our findingthat the practice was promising, if not evidence-based. This was not the case. However, we feelthat our process warrants further use and critiqueand we welcome both. In addition, we feel it ispremature to suggest changes to the quality indi-cators and standards for evidence-based practicesbased on our review alone.

Horner et al. (2005) and Gersten et al.(2005) should be applauded for their efforts topropel special education research forward throughthe establishment of rigorous standards for ourwork. We hope that in our effort to fairly evaluatethe quality of the research, we haven't overlookedany particular feature that would have resulted ina different outcome. Research on repeated readingis very valuable and we implore special educationresearchers to consider further work in this areathat will result in enhancing our understanding ofrepeated reading's efficacy for building readingflueiicy in students with or at risk for learning dis-abilities.

REFERENCES

Adams, M. (1990). Beginning to read: Thinking andlearning about print. Cambridge, MA: MIT Press.

American Psychological Association. (2001). Publica-tion manual of the American Psychological Association(5th ed.). Washington, DC: Author.

Arreaga-Mayer, C , Terry, B., & Greenwood, C.(1998). Classwide peer tutoring. In K. Topping & S.Ehly (Eds.), Peer-assisted learning (pp. 105-120). Mah-Wah, NJ: Erlbaum.

Begeny, J. C , Daly E. J., Ill, & Valleley, R. J. (2006).Improving oral reading fluency through response op-portunities: A comparison of phrase drill error correc-tion with repeated readings, journal of BehaviorEducation, 15, 229-235.

Chafouleas, S. M., Martens, B. K., Dobson, R. L, We-instein, K. S., & Gardner, K. B. (2004). Fluent readingas the improvement of stimulus control: Additive ef-fects of performance-based interventions to repeatedreading on students' reading and error rates, journal ofBehavioral Education, 13, 67-81.

Chard, D. J., Vaughn, S., &C Tyler, B. J. (2002). A syn-thesis of research on efFective interventions for buildingfluency with elementary students with learning disabili-ties./o«r««/o/Z-iflr«/«^ £)¿ítí¿///«'iíi, 35, 386-406.

Daly E. J., Ill, & Martens, B. K. (1994). A comparisonof three interventions for increasing oral reading perfor-mance: Application of the instructional hierarchy, jour-nal of Applied Behavior Analysis, 27, 459-469.

Education of All Handicapped Children Act, 20U.S.C. § 1400 et seq. (1975).

Ehri, L. C. (1995). Stages of development in learningto read words by sight, journal of Research in Reading,18, 116-125.

Freeland, J. T, Skinner, C. H., Jackson, B., McDaniel,C. E., & Smith, S. (2000). Measuring and increasingsilent reading comprehension rates: Empirically validat-ing a repeated reading intervention. Psychology in theSchools, 37, 415-429.

Fuchs, D., Fuchs, L. S., Mathes, P G., & Simmons, D.C. (1997). Peer-assisted learning strategies: Makingclassrooms more responsive to diversity. American Edu-cational Research journal, 34, 174-206.

Gersten, R., Fuchs, L. S., Compton, D., Coyne, M.,Greenwood, C , & Innocenti. M. S. (2005). Qualityindicators for group experimental and quasi-experi-mental research in special education. Exceptional Chil-dren, 71, 149-164.

Horner, R. H., Carr, E. G., Halle, J., McGee, G.,Odom, S., & Wolery, M. (2005). The use of single-subject research to identify evidence-based practice inspecial education. Exceptional Children, 71, 165-179.

Exceptional Children 2 7 9

Page 18: Repeated Reading Interventions for Students With  · PDF fileRepeated Reading Interventions for Students ... supporting the notion that they experience a ... FLUENCY AND ITS ROLE

Kuhn, M. R., & Stahl, S. (2003). Fluency: A review ofdevelopmental and remedial strategies. Journal of Edu-cational Psychology, 95, 3-21.

LaBerge, D., & Samuels, S. J. (1974). Towards a theoryof automatic information processing in reading. Cogni-tive Psychology, 6, 293-323.

Lipka, O., Lesaux, N. K., & Siegel, L. S. (2006). Ret-rospective analyses of the reading development of grade4 students with reading disabilities: Risk status andprofiles over 5 years. Journal of Learning Disabilities,39, 364-378.

Logan, G. D. (1988). Toward an instance theory of au-tomatization. Psychological Review, 95, 492-527.

Logan, G. D. (1997). Automaticity and reading: Per-spectives from the instance theory of automatization.Reading and Writing (¿uarterly, 13, 123-146.

Mathes, R G., & Fuchs, L. S. (1993). Peer-mediatedreading instruction in special education resourcerooms. Learning Disabilities Research & Practice, 8,233-243.

Mathes, P G., Howard, J. K., Allen, S. H., & Fuchs,D. (1998). Peer-assisted learning strategies for first-grade readers: Responding to the needs of diverse learn-ers. Reading Research Quarterly, 33, 62-94.

McMaster, K. N., Fuchs, D., Fuchs, L. S., & Comp-ton, D. L. (2005). Responding to nonresponders: Anexperirhental field trial of identification and interven-tion methods. Exceptional Children, 71, 445—463.

Meyer, M. S., & Felton, R. H. (1999). Repeated read-ing to enhance fluency: Old approaches and new direc-tions. AnnaL of Dyslexia, 49, 283-306.

National Reading Panel. (2000). Teaching children toread: An evidence-based assessment of the scientific re-search literature on reading and its implications for read-ing instruction. Washington, DC: National Institute ofChild Health and Human Development.

O'Shea, L. J., Sindelar, P T, & O'Shea, D. J. (1987).The effects of repeated readings and attentional cues onthe reading fluency and comprehension of learning dis-abled readers. Learning Disabilities Research, 2,103-109.

Perfetti, C. A. (1985). Reading ability. New York: Ox-ford University Press.

Rashotte, C. A., ÔC Torgesen, J. K. (1985). Repeatedreading and reading fluency in learning disabled chil-dren. Reading Research Quarterly 20, 180-188.

Simmons, D. C , Coyne, M. D., Kwok, O., McDon-agh, S., Harn, B., & Kame'enui, E. J. (2008). Indexingresponse to intervention: A longitudinal study of read-ing risk from kindergarten through third grade. Joumalof Learning Disabilities, 41, 158-173.

Sindelar, P T., Monda, L. E., ôc O'Shea, L. J. (1990).Effects of repeated readings on instructional- and mas-tery-level readers. Journal of Educational Research, 83,220-226.

Snowling, M. J. (2000). From language and reading todyslexia. Dyslexia, 7, 37-46.

Stanovich, K. E. (1986). Matthew effects in reading:Some consequences in individual differences in the ac-quisition of literacy. Reading Research Quarterly, 21,360-407.

Stanovich, K. E., & Siegel, L. S. (1994). Phenotypicperformance profile of children with reading disabili-ties: A regression-based tests of the phonological-corevariable-difference model. Journal of Educational Psy-chology, 86, 24-55.

Strong, A. C , Wehby, J. H., Falk, K. B., & Lane, K. L.(2004). The impact of a structured reading curriculumand repeated reading on the performance of junior highstudents with emotional and behavioral disorders.Schoot Psychology Review, 33, 561-581.

Swanson, H. L, Hoskyn, M. W, &i Lee, C. (1999). In-terventions for students with learning disabilities: A meta-analysis of treatment outcomes. New York: Guilford.

Therrien, W J. (2004). Fluency and comprehensiongains as a result of repeated reading: A meta-analysis.Remedial and Special Education, 25, 252-261.

Torgesen, J. K., Alexander, A. W, Wagner, R. K.,Rashotte, C. A., Voeller, K., Conway, T, et al. (2001).Intensive remedial instruction for children with severereading disabilities: Immediate and long-term out-comes from two instructional approaches. Journal ofLearning Disabilities, 34, 33-58.

Vellutino, E R., Scanlon, D. M., Sipay, E. R., Small, S.G., Pratt, A., Chen, R., et al. (1996). Cognitive profilesof difficult-to-remediate and readily remediated poorreaders: Early intervention as a vehicle for distinguish-ing between cognitive and experiential deficits as basiccauses of specific reading disability. Journal of Educa-tional Psychology, 88, 601-638.

Vellutino, F. R., Scanlon, D. M., & Tanzman, M. S.(1998). The case for early intervention in diagnosingspeciflc reading disability. Journal of School Psychology,36, 5G7-?,^7.

Weinstein, G., & Cooke, N. L. (1992). The effects oftwo repeated reading interventions on generalization offluency. Learning Disability Quarterly, 15, 21—28.

Young, A. R., Bowers, P G., & MacKinnon, G. E.(1996). Effects of prosodie modeling and repeatedreading on poor readers' fluency and comprehension.Applied Psycholinguistics, 17, 59-84.

2 8 O Spring 2009

Page 19: Repeated Reading Interventions for Students With  · PDF fileRepeated Reading Interventions for Students ... supporting the notion that they experience a ... FLUENCY AND ITS ROLE

ABOUT THE AUTHORS

DAVID J. CHARD (CEC TX Federation), LeonSimmons Endowed Dean, Annette Caldwell Sim-mons School of Education and Human Develop-ment, Southern Methodist University, Dallas,Texas, L E A N N E R. K E T T E R L I N - G E L L E R

(CEC OR Federation), Assistant Professor, Edu-cational Leadership, College of Education, Uni-versity of Oregon, Eugene. SCOTT K. BAKER(CEC OR Federation), Director, Pacific Institutesfor Research, and Associate Director of the Cen-ter on Teaching and Learning at the University ofOregon, Eugene. CHRISTIAN DOABLER (CECOR Federation), Doctoral Candidate; andCHAN ISA APICHATABUTRA, Doctoral Candi-date, Department of Special Education and Clini-cal Sciences, College of Education, University ofOregon, Eugene.

Address correspondence to David Chard, AnnetteCaldwell Simmons School of Education andHuman Development, Southern Methodist Uni-versity, P.O. Box 750382, Dallas, TX 75275-0382(e-mail: [email protected]).

Manuscript submitted September 2007; acceptedApril 2008.

Exceptional Children 281

Page 20: Repeated Reading Interventions for Students With  · PDF fileRepeated Reading Interventions for Students ... supporting the notion that they experience a ... FLUENCY AND ITS ROLE