a synthesis of research on effective interventions for building … · 2017-11-29 · effective...

JOURNAL OF LEARNING DISABILITIESVOLUME 35, NUMBER 5, SEPTEMBER/OCTOBER 2002, PAGES 386–406

A Synthesis of Research on Effective Interventions for Building Reading Fluency with Elementary Students with Learning Disabilities

David J. Chard, Sharon Vaughn, and Brenda-Jean Tyler

Abstract

Fluent reading, often defined as speed and accuracy, is an important skill for all readers to develop. Students with learning disabilities(LD) often struggle to read fluently, leading to difficulties in reading comprehension. Despite recent attention to reading fluency andways to improve fluency, it is not clear which features of interventions that are designed to enhance fluency are beneficial for the moststruggling readers. The purpose of this study is to synthesize research on interventions that are designed primarily to build reading flu-ency for students with LD. The search yielded 24 published and unpublished studies that reported findings on intervention features, in-cluding repeated reading with and without a model, sustained reading, number of repetitions, text difficulty, and specific improvementcriteria. Our findings suggest that effective interventions for building fluency include an explicit model of fluent reading, multiple op-portunities to repeatedly read familiar text independently and with corrective feedback, and established performance criteria for in-creasing text difficulty.

The development of reading flu-ency has been linked to success-ful reading since the early re-

search on the psychology of reading. Inhis classic review of 19th-century read-ing research, Huey (1908) likened thedevelopment of fluent reading to thedevelopment of other psychomotorskills such as playing tennis, remark-ing that both skills benefited frompractice. “Repetition progressivelyfrees the mind from attention to de-tails, and makes facile the total act,shortens the time, and reduces the ex-tent to which consciousness must con-cern itself with the process” (p. 104).Despite this attention afforded to flu-ency in the early 1900s, it was not until1974, when LaBerge and Samuels of-fered their theory of automatic pro-cessing, that reading fluency took amore prominent role in our under-standing of reading development.

LaBerge and Samuels (1974) pro-posed that learning to read involves in-creasing automaticity in processingword units (e.g., letter–sound corre-spondences), processing these unitsinto recognizable words, and connect-ing the words while reading a passage.In effect, improvement in the process-ing of units, words, and connected textcognitively releases the reader to thinkabout the meaning of the text. Thistheory of automatic information process-ing resulted in research that focused onimproving the speed at which studentsrecognized words (e.g., Ehri & Wilce,1983) and on repeated reading (Sam-uels, 1997).

Similarly, Perfetti’s (1977, 1985) ver-bal efficiency model suggested thatslow word processing speed interfereswith automaticity of reading and,therefore, with comprehension. How-ever, Perfetti extended this explanation

to suggest that slow word reading isalso debilitating because it consumesworking memory and, therefore, pre-vents the individual from thinkingabout the text while reading. Slowword reading clogs working memorywith the processing of word-levelreading so as to prevent understandingat the content level. Thus, both rapidreading of high-frequency words andrapid decoding as a means to enhancetext understanding appear critical fortypical reading development (L. S.Fuchs, Fuchs, Hosp, & Jenkins, 2001;Kuhn & Stahl, 2000; Meyer & Felton,1999).

Students with learning or reading dis-abilities demonstrate difficulties in thearea of fluency. A common core prob-lem is the ability to read sight words,decode words, and read phrases andsentences automatically and rapidly.Thus, reading fluency is an essential

jld-v35n5-p386 8/22/02 2:34 PM Page 386

VOLUME 35, NUMBER 5, SEPTEMBER/OCTOBER 2002 387

skill for all students. However, stu-dents with reading or learning disabil-ities are most at risk for presenting dif-ficulties in fluency (Meyer & Felton,1999). The U.S. National ResearchCouncil’s Committee for the Preven-tion of Reading Failure noted that

because the ability to obtain meaningfrom print depends so strongly on the de-velopment of word recognition accuracyand reading fluency, both the lattershould be regularly assessed in the class-room, permitting timely and effective in-structional response when difficulty ordelay is apparent. (Snow, Burns, & Grif-fin, 1998, p. 7)

Contemporary research has demon-strated that some students with learn-ing disabilities (LD) can be charac-terized as having a specific deficit innaming speed that distinguishes themfrom students with learning disabili-ties that stem from phonological pro-cessing deficits (Wolf & Katzir-Cohen,2001). Wolf and Bowers (1999, 2000)have referred to this distinction as adouble deficit model of reading disabil-ity. This model has led to the conjecturethat interventions should be specifi-cally tailored to address fluency defi-cits for those students who are phono-logically aware and able to decodeaccurately but remain dysfluent. Thesespecific intervention practices warrantfurther investigation and validation(Wolf & Katzir-Cohen, 2001).

Fluency is a critical but often ne-glected element of reading programs(Allington, 1983; Kameénui & Sim-mons, 2001). This is likely due in partto the fact that effective interventionsfor improving fluency are not widelyknown. For example, whereas there isa consensus that fluency is dependenton adequate word recognition skills,there is also an understanding thatword recognition proficiency may notalways yield fluent readers (NationalReading Panel, 2000). Many of the ap-proaches to improving fluency couldbe categorized as focusing on repeatedreading (Meyer & Felton, 1999),whether through partner reading (e.g.,Arreaga-Mayer, Terry, & Greenwood,

1998; D. Fuchs, Fuchs, Mathes, & Sim-mons, 1997) or through other proce-dures that support repeated reading(see Topping & Ehly, 1998, for a re-view).

Due to the perceived importance offluency development to reading suc-cess and because of its apparent ne-glected status in classroom instruction,the U.S. National Reading Panel (NRP;2000) has provided a meta-analysis ofrecent research on fluency building.Sufficient research articles were lo-cated for the NRP meta-analysis to in-form two areas of fluency building,namely, guided repeated oral readingand formal efforts to increase students’independent or recreational reading.NRP criteria for studies to be includedin their meta-analysis were that they(a) examined the impact of repeatedreading or some other form of guidedoral reading instruction on readingachievement; (b) provided reading in-struction in English with students(K–12); and (c) appeared in a refereedjournal. The NRP identified 98 articleson guided repeated oral reading thatmet all of these criteria. The NRP de-termined that the mean weighted ef-fect size for guided oral repeatedreading was .41, indicating that thisprocedure “has a moderate impact onthe reading achievement of the types ofstudents participating in these studies”(p. 3-17).

When distinguishing effect sizes forbelow-average versus average andabove-average readers, the NRP (2000)reported that in general, guided oralrepeated reading was more beneficialto average readers than to poor read-ers. The NRP conjectured that these re-sults were probably less importantthan the individual findings of somestudies that suggested that readers atdifferent levels of proficiency benefitedfrom different aspects of the interven-tions. For example, poor readers mayhave benefited more in terms of wordreading accuracy, whereas more profi-cient readers may have benefited in de-veloping prosody. Unfortunately, theNRP was unable to distinguish differ-ences in specific intervention features

because of the small sample of studiesthat addressed each specific interven-tion feature.

Based on its synthesis of guided re-peated oral reading and its associationwith improved fluency for most stu-dents, the NRP (2000) recommendedthat teachers begin including such ac-tivities in their classroom routines tosupplement their reading instruction.Despite these promising recommenda-tions for most readers in the classroom,these findings may not generalize tothe students with the most significantreading problems. For this reason, wewere interested in synthesizing the find-ings from fluency interventions specif-ically intended for students with LD.Because most students with LD havereading disabilities, we thought itwould be beneficial to specifically de-scribe those fluency intervention stud-ies in which students with learningdisabilities were participants.

Furthermore, considering timelines,resources, and purpose, it simply wasnot possible for the NRP to locate andcode the broadest array of research.Thus, dissertations, single group de-signs, and single-subject/case studydesigns were not included in the NRPstudy. The purpose of this study was toexamine all intervention studies acrossa complete array of sources that iden-tified students with LD as the targetgroup for reading fluency intervention.

Method

Operational Definitions

For the purpose of this study, a learningdisability refers to any researcher- orschool-identified learning disability. Inmany of the included studies, the cri-teria for identifying a learning disabil-ity were explicitly outlined. In somecases, however, the authors did not re-port the procedures used for identify-ing the learning disability but statedthat participating students met the cri-teria for state identification. In stillother cases, no specific criteria weredescribed, but it was stated that stu-

jld-v35n5-p386 8/22/02 2:34 PM Page 387

JOURNAL OF LEARNING DISABILITIES388

dents had been identified as having alearning disability.

The term fluency refers to the speedand accuracy with which a studentreads connected text orally. Measuresof fluency included oral reading flu-ency, in which both rate and accuracywere measured, and measures of oralreading that gauged accuracy, fluency,or prosody separately.

Fluency intervention refers to any in-structional intervention designed spe-cifically to increase students’ readingfluency in connected text. Fluency in-terventions could focus at the wordlevel or the connected-text level.

Locating and Selecting Studies

To identify a broad range of studies,computer searches were conducted ofthe ERIC, PsycINFO, and ArticleFirstdatabases; the first two of these data-bases include dissertations in theirsearch returns. Key words and termsused to locate studies included fluency,reading fluency, reading aloud, readingrate, repeated reading, reading practice, as-sisted reading, oral reading, and pairedreading. Moreover, hand searches werecarried out on the most relevant jour-nals, including Annals of Dyslexia; Edu-cation and Treatment of Children; Excep-tional Children; Journal of EducationalResearch; Journal of Experimental Psy-chology; Journal of Learning Disabilities;Journal of Reading Behavior; Learning,Memory, and Cognition; Learning Dis-ability Quarterly; Learning DisabilitiesResearch & Practice; Psychology in theSchools; Reading Horizons; Reading Re-search and Instruction; Reading ResearchQuarterly; Remedial and Special Educa-tion; and School Psychology Review. Ref-erences from studies that met our cri-teria and from article reviews offluency research were checked forpromising studies that might also fitour criteria. Our broad-based searchyielded a pool of 104 studies in a vari-ety of formats, including journal arti-cles, book chapters, reports, and dis-sertations.

The titles and abstracts of prospec-tive studies were carefully reviewed.

Three sets of criteria guided our selec-tion of studies:

1. the students targeted for the inter-vention were elementary-age stu-dents with LD,

2. the purpose of the study specifi-cally targeted reading fluency, and

3. the study was published in the lastquarter of the 20th century.

We considered for inclusion only thosestudies wherein the sample was de-scribed as students with LD. In thosestudies where students with LD wereincluded in general education class-rooms, we accepted the study only ifthe findings for the students with dis-abilities were reported separately, or ina way that made it possible to disag-gregate results, or when the percentageof students with LD was at least 60% ofthe total sample. In terms of gradelevel, we restricted our synthesis tothose studies wherein at least 50% ofthe participants were identified as stu-dents between the first and fifth grade.Studies were excluded if they includedstudents within our focal grade rangeas well as outside that range but didnot provide sufficient information todetermine the number of students ineach grade.

The second set of criteria addressedaspects of the studies’ purpose. First,we included only studies that reportedon the implementation of an inter-vention that expressly targeted theimprovement of reading fluency inconnected text. Studies in which inter-ventions focused on improved wordanalysis and where improved fluencymight be a by-product were excludedfrom our analysis. Second, althoughfluency development has been studiedin many languages, we restricted ourselection to those studies that focusedon reading in English. Finally, we lim-ited our corpus of studies to those pub-lished between January 1975 and De-cember 2000 inclusive. In cases whereit was difficult to determine from thetitle and abstract whether a particularstudy would be appropriate for inclu-sion, the work was reviewed to clarify

ambiguous or incomplete information.In cases of remaining uncertainty, wediscussed the work, and a decision wasreached. Following these criteria, weestablished a final pool of 24 studies.

Coding Studies

Studies that met all criteria were dis-tributed among the authors and atrained research assistant, who codedthem using a comprehensive codingprotocol designed to facilitate therecording of, and ease of access to, keyaspects of the included studies. Thecoding protocol focused on the follow-ing major categories: (a) study infor-mation, (b) comparison and treatmentgroup information, (c) treatment andcomparison descriptions, (d) measureinformation, and (e) specific findings.

Coders participated in a 3-hourtraining session in which the specificcategories and codes were describedand discussed and coding was prac-ticed on sample studies. As in thesearch process, confusing aspects ofstudies discovered during coding werediscussed at meetings between the au-thors, and decisions were reached byconsensus. During the study, 20% of allstudies were coded twice and inter-rater agreement was calculated. Over-all interrater agreement ranged be-tween .85 and .97, with a mean of .92.

Instruction for comparison groupswas coded similar to the treatmentgroups unless no intervention wasspecified. Five studies meeting ourcriteria included a no-treatment com-parison. In all five studies, this no-treatment comparison was describedas either traditional reading instruc-tion or using a commercial or basalreading program. A summary of alltreatment and comparison groups isincluded in Tables 1 through 3.

Results

Tables 1 through 5 provide an over-view of the studies included in thissynthesis according to the focus of theintervention. For example, Table 1 de-

jld-v35n5-p386 8/22/02 2:34 PM 88


scribes studies that examined repeatedreading without a model. All tables usethe following descriptions of quasi-experimental and experimental re-search designs (Campbell & Stanley,1966): single group, single grouppretest–posttest design, treatment–control, multiple group comparison,nonequivalent control group design.Furthermore, tables include studiesthat employed single-subject or casestudy methodology. In some instances,information about a particular studymay be reported on more than onetable because samples within the studyreceived treatments that included mul-tiple features.

In each table, Column 1 includes au-thor names, publication date, partici-pant age range, and design categoryfor each study. Column 2 describes thetreatments administered to partici-pants and reports the sample sizes.Furthermore, the duration of the treat-ment in minutes is listed if this in-formation was provided in the study.Column 3 describes the dependentmeasures for each study. Although thefocus of this synthesis is on reading flu-ency, all dependent measures relatedto reading are included. Column 4summarizes the results reported ineach study. The summary of resultsreflects the significant differences re-ported between samples. For example,in Table 3, PALs > No PALs indicatesthat there was a statistically significantdifference between the PALs (peer-assisted learning strategies) sampleand the No PALs sample, favoringPALs. Those studies that reported nosignificant differences and did not pro-vide means and standard deviationswere coded as such.

Column 4 also reports effect sizes(ES) for comparisons between treat-ment samples or between treatmentsand comparison samples on all depen-dent measures. Cohen’s d was theindex of effect size (Hedges & Olkin,1985). Where mean scores and stan-dard deviations were reported, d wascalculated as the difference betweentreatment and comparison posttestmean scores (or adjusted posttest

means if that information was pro-vided) divided by the pooled standarddeviations of the posttest. In somecases, when neither mean scores norstandard deviations were reported, ef-fect sizes were estimated from F val-ues. In some cases, effect sizes were notreported because neither means, stan-dard deviations, nor F values were re-ported.

When a study included only two in-dependent samples, the effect sizes inColumn 4 reflect the magnitude of theeffect between two treatments or be-tween the treatment and the compari-son condition. When more than two in-dependent samples were included inthe study, the effect sizes in Column 4are preceded by the relevant compari-son, usually with the treatment dem-onstrating superior performance listedfirst. For example, in Table 3, “PALs vs.No PALs: d = .63” indicates that thePALs sample outperformed the NoPALs sample with an effect size of .63. For the purpose of interpretation, J. Cohen’s (1988) distinctions on themagnitude of effect were used, with d = .20 reflecting a small effect size, d =.60 a moderate effect size, and d = .80 alarge effect size. This corpus of studiesrevealed that variations on two inter-vention types were used to enhancefluency for students with LD: repeatedreading interventions and word practiceinterventions.

Repeated Reading

Repeated Reading Without a Model.Twenty-one studies (N = 128) ad-dressed the question of whether re-peatedly reading text is an effectiveway to improve the reading fluency ofstudents with LD. In all samples, thestudents independently read con-nected text repeatedly a minimum oftwo times and a maximum of seventimes. The effect sizes on measures offluency (including rate and accuracy)ranged from d = .02 to d = 3.02, with anaverage of d = .68. These studies aresummarized in Table 1.

A. L. Cohen (1988) compared thefour-times-repeated reading of text to a

no-treatment condition and found nomain effects on measures of oral read-ing fluency, accuracy, or passage com-prehension. However, the mean effectsize for repeated readings on fluencyand accuracy was d = 1.98 for studentswith LD. Furthermore, in anotherstudy, the repeated reading of textcompared with reading text just onceyielded significantly better scores forfluency and comprehension (O’Shea,Sindelar, & O’Shea, 1987) and for accu-racy and fluency when compared withan audiotaped model of word readingin yet another study (Daly & Martens,1994).

Monda (1989) compared both silentand oral repeated reading without amodel to reading that followed a read-aloud text by the teacher. The mean ef-fect size for the repeated reading inter-ventions compared to the listeningintervention yielded no differences (d = –.05). However, although Mondafound no significant differences be-tween groups on any measure of flu-ency, she found moderate effect sizes (d = .57 and d = .46) on measures ofreading accuracy, favoring oral re-peated reading over modeled readingas treatments.

Rashotte and Torgesen (1985) com-pared the reading fluency of studentswith LD under two different repeatedreading conditions (one with overlap-ping words and one without) with asustained reading condition. The meaneffect size for the repeated reading con-ditions relative to the sustained read-ing condition was d = .34. Students inboth repeated reading conditions per-formed significantly better than thosein the sustained reading condition on ameasure of reading rate. When the re-peated reading condition with no over-lapping words was compared to thesustained reading condition, the effectsize was moderate (d = .65). When therepeated reading condition with a highpercentage of overlapping words wascompared to sustained reading, the ef-fect size was somewhat smaller (d =.35).

In an alternating treatment designacross six students, Rose (1984) com-

jld-v35n5-p386 8/22/02 2:34 PM Page 389

390

TABLE 1Studies Examining Repeated Reading Without a Model

Author/participant Treatment description/sampleage/design size/treatment duration Dependent measures Results/effect sizes (d )

Cohen, 19888 years 7 months–13 years

2 months Multiple group comparison

Daly & Martens, 19948 years 10 months–11 years

11 monthsMulti-element design

Marston, Deno, Dongil, Diment,& Rogers, 1995

Age not reportedTreatment–Comparisona

Monda, 19899.0–13.5 yearsMultiple group comparison

• Processing power (PP; n = 16): Re-peated reading (4 times) with text pre-sented 3–5 words at a time.

• Repetitive reading (RR; n = 16): Re-peated reading (4 times) with studentcontrol of text amount.

• No-treatment comparison (C; n = 15).Duration: 195–202 minutes

N = 4• All conditions followed a model, drill,

and train to criterion lesson structure.• Subject Passage Preview (SP): Read

passage silently and reread for assess-ment.

• Listening Passage Preview (LP): Lis-tened to audiotaped (130 wpm) pas-sage, read list of unknown words, andreread passage for assessment.

• Taped Words (TW): Read aloud alongwith audiotaped word list (80 wpm).Reread list and read passage for as-sessment.

Length and duration: Session length notspecified. Treatments lasted 21 days.

• Peer tutoring (PT; n = 27): Repeatedreading of text with peer partners tak-ing turns reading.

• Comparison (C; n = 23): not describedDuration: 2,250 minutes

• Oral Repeated Rereading (OR; n =20): Read, reread with timing and nofeedback.

• Silent Repeated Reading (SR; n = 20):Read, reread silently with timing.

• Listening Repeated Reading (LR; n =20): Read, reread aloud to student byteacher.

Duration: All students reread the passagea third time before responding to com-prehension questions. Procedure ad-ministered once to each condition.

Paragraph reading speed RR vs. PP: d = .19(practiced)

Paragraph reading accuracy PP vs. RR: d = .56(practiced)

Paragraph reading speed PP vs. RR: d = .26(unpracticed)

Paragraph reading accuracy PP vs. RR: d = .19(unpracticed)

Reading fluency, final text PP vs. RR: d = .30; PP vs. C: d = 1.58; RR vs. C: d = 1.25

Reading accuracy, final text PP vs. RR: d = .89; PP vs. C: d = 3.02; RR vs. C: d = 2.09

Word reading speed PP > C(single syllable, practiced)

Word reading speed PP > C(multisyllable, practiced)

Word reading speed no significant difference(unpracticed) between groups

Passage comprehension no significant differencebetween groups

Passage reading accuracy • LP > SP > TW on Passage reading fluency accuracy and fluency Word reading accuracy of passage reading.Word reading fluency • TW condition outper-

formed SP and LPon measures of wordreading accuracyand fluency.

Oral reading fluency PT vs. C: d = –.14

Words read per minute (exp). SR vs. OR: d = .13;OR vs. LR: d = .02; SR vs. LR: d = .15

Words read per minute (transfer)b SR vs. OR: d = .27; OR vs. LR: d = –.06; SR vs. LR: d = .20

Errors per minute (exp.)b OR vs. SR: d = .70; OR vs. LR: d = .57; SR vs. LR: d = –.17

Errors per minute (transfer)b,c OR vs. SR: d = .76; OR vs. LR: d = .46; SR vs. LR: d = –.46

(Table continues)

jld-v35n5-p386 8/22/02 2:34 PM Page 390

391

(Table 1 continued)


O’Shea, Sindelar, & O’Shea, 1987

11.3–13.6 years2 (Focus) × 3 (No. of

reading) factorial design

Rashotte & Torgesen, 1985 (based on dissertation by

Rashotte, 1983)8.6–12 yearsMultiple treatment design

Rose, 19849.5–13.1 yearsAlternating treatment design

Stout, 19978 years–9 years 3 monthsMultiple baseline across

participants with an additional prebaseline phase

N = 29• Repeated Reading 7 (R7): Read text 7

times.• Repeated Reading 3 (R3): Read text 3

times.• Reading (R): Read text once.Duration: Each treatment administered

once.

N = 12• Repeated reading, no overlapping

words (NO): Read stories 4 times withabout 20 words in common acrossstories.

• Repeated reading with high overlap(HO): Same as NO but with about 60words in common.

• Sustained Reading (SR): Read 4 differ-ent stories each day.

Duration: 7 days

N = 6• Baseline: Oral reading only, no pre-

viewing. • Silent previewing: Student read pas-

sage silently before reading it aloud tothe teacher.

• Listening previewing: Teacher readpassage aloud while student followedalong. Then student read passagealoud to teacher.

Length and duration: Approximately 25days of intervention. Length of daily in-tervention not specified.

N = 4• Baseline: Student read a passage with

feedback provided after reading. Thenstudents were allowed to study thepassage. Comprehension questionswere asked. (at least 3 sessions).

• Treatment: Students read randomly se-lected passages 3 times, using thesame procedures used in baseline.After 3 readings, students answeredwritten multiple choice comprehensionquestions. (At least 3 sessions).

Length and duration: Not specified.

Oral reading fluency R7 > R3 > R

Story retell R7, R3 > R

Reading rate (slope of progress) Both interventions withrepeated readingperformed signifi-cantly better thansustained readingon reading rate.

NO vs. HO: d = .18NO vs. SR: d = .65HO vs. SR: d = .35HO slope, SR slope > 0

Reading accuracy (slope of NO vs. HO: d = –.26progress)b NO vs. SR: d = –.52

HO vs. SR: d = –.24HO slope, SR slope > 0

Passage comprehension no significant differ-(slope of progress) ences between

groups.NO vs. HO: d = .10NO vs. SR: d = .08HO vs. SR: d = .17

Words read correctly per minute • Listening previewingled to faster readingrates.

• Both silent and lis-tening previewing ap-peared to be morebeneficial than baseline.

Words read correctly per minute • Statistically signifi-Errors per minute cant increases in % comprehension questions fluency across allanswered correctly four students.

• Students did notdemonstrate statisti-cally significant im-provements in errorsper minute.

• Statistically signifi-cant difference incomprehension only for one student.

(Table continues)

jld-v35n5-p386 8/22/02 2:34 PM Page 391


pared silent repeated reading withouta model to a no-treatment baseline andto a teacher-modeled repeated readingcondition. Whereas the modeled con-dition led to higher reading rates thanrepeated reading alone, both repeatedreading conditions again led to higherreading rates than the baseline condi-tion.

In a multiple baseline study acrossfour participants, Stout (1997) reportedsignificant increases in reading fluencyfrom baseline to treatment for all fourstudents with LD. However, the differ-ences in reading accuracy and writtencomprehension were not significant.

Swain and Allinder (1996) studiedthe effectiveness of a repeated readingintervention that provided studentsthe opportunity to read and reread textto multiple people, including theteacher, an instructional assistant, aparent, and a younger student. The in-tervention was relatively brief—24, 12,and 6 days, respectively, for three stu-dents. Still, two of the three studentsdemonstrated improved oral readingfluency. However, only one studentdemonstrated improvement on a read-ing maze measure. In general, the find-ings for repeated reading of text with-out a model revealed that repeatedreading was associated with improvedoutcomes in accuracy and fluency.

Repeated Reading With a Model.Several types of studies can be distin-guished based on the method used for

modeling. All studies using repeatedreading with a model are listed in Ta-ble 2.

Modeling by an adult. Fourteen sam-ples (one group [N = 10] and 13 singlecases) were studied to address theissue of repeated reading with an adultmodel. Monda (1989) compared stu-dent repeated reading of text to a lis-tening preview condition in which theteacher read the passage to the studenttwice before the student reread thepassage alone. Although there were nostatistically significant differences be-tween groups, the effect sizes for read-ing accuracy (errors per minute) weremoderate (d = .46 to d = .57), favoringoral repeated reading without a model.In contrast, the effect size for accuracyon a transfer measure compared tosilent repeated reading was moderate,favoring the listening preview or adultmodel (d = .46). On a measure of com-prehension, students in the listeningpreview condition outperformed stu-dents in the oral repeated reading con-dition, although effect sizes were small(d = .25 on cued recall, and d = .34 onfree recall).

In a multiple baseline design, Smith(1979) presented students with a pas-sage of text read by the teacher at 100words per minute. At the end of thereading by the teacher, the student con-tinued reading from where the teacherstopped. For the two students with LDincluded in the sample, reading flu-ency increased during and after the

modeling phase from approximately 6to 14 correct words per minute. This in-crease in fluency over the baseline wasmaintained during a follow-up phase.Smith (1979) replicated this study with a student with much lower read-ing fluency at baseline than the stu-dents in her first study and found sim-ilar results.

Rose and Beattie (1986) comparedthree interventions for improving flu-ency—a teacher-modeled condition, abaseline condition that included oralreading and skill instruction, and ataped previewing condition. For threeof the four participating students inthis alternating treatment design, theteacher model resulted in higher oralreading fluency than either the base-line or the taped previewing condition.However, reading accuracy was not af-fected by the various previewing pro-cedures. In a similar study, Rose (1984)found that a listening preview (teachermodel) promoted faster reading ratesthan either a no-preview baseline con-dition or a repeated reading alone con-dition.

Modeling by a more proficient peer. Thefindings from three samples (N = 89)addressed using a more proficient peerfor modeling fluent reading. Mathesand Fuchs (1993) compared the use ofrepeated reading with a partner, sus-tained reading of text with a partner,and a comparison condition. The sam-ple in the sustained reading conditionperformed significantly better on flu-

(Table 1 continued)


Swain & Allinder, 19967 years 8 months–8 years

1 monthMultiple baseline across

individuals

aSeveral other interventions were conducted in this study, but they were not focused on fluency development. bNegative d reflects positive outcome favoring treat-ment listed first. cAdditional measures were reported in this study that were not directly related to reading fluency.

N = 3Baseline: No additional instructional activ-

ities.Repeated Reading (RR): Read passage

four times over 4 days to the teacher,instructional assistant, parent, andyounger student.

Length and duration: Session length notspecified; treatment lasted 24, 12, and6 days for Students 1–3, respectively.

Oral reading fluency • Intervention im-Curriculum-based maze measure proved oral reading

fluency for 2 of 3 stu-dents.

• Maze performance decreased for 2 of 3 students during inter-vention phase.

jld-v35n5-p386 8/22/02 2:34 PM Page 392

393

TABLE 2Studies Examining Repeated Reading with a Model


Daly & Martens, 19948 years 10 months–

11 years 11 monthsMulti-element design

Gilbert, Williams, & McLaughlin, 1986

7 years 1 month–7 years 4 months

Multiple baseline design across participants

Mathes & Fuchs, 19934th–6th grade, no age

information reportedMultiple group comparison

Moseley, 199311 yearsIndividual case study

N = 4• All conditions followed a model, drill,

and train to criterion lesson structure.• Subject Passage Preview (SP): Read


• Listening Passage Preview (LP): Lis-tened to audiotaped (130 wpm) pas-sage, read list of unknown words andreread passage for assessment.

• Taped Words (TW): Read aloud alongwith audiotaped word list (80 wpm).Reread list and read passage for as-sessment.


N = 3• Baseline (45-minute session; duration:

1 day): Teacher introduction to vocabu-lary and phonics; students practicedpassage silently before reading aloudfor 4 min. No feedback provided.

• Intervention (session length not pro-vided; duration: 4 days): Students lis-tened once and read along 3 timeswith a tape-recorded reading beforereading for 4 minutes.

All three conditions were divided ran-domly into groups reading instructionaland independent texts.

• Sustained Reading (SR; n = 23): Part-ner reading from same text for 9 min.

• Repeated Reading (RR; n = 22): Part-ner reading of same text 3 times (re-peated once by partner)

• Control (C; n = 22): Traditional readinginstruction.

Duration: 1,350 minutes

N = 1Intervention (one 30-min. session weekly;Duration: 300 minutes): 1. Passage (one year above instructional

level) read to student by a speechsynthesizer.

2. Passage presented to student visuallyin phrases and in speech.

3. Passage read by student either inphrases or learner-paced sentence bysentence.

4. Student read and reread passagesuntil criterion of 120 wpm with 2 orfewer errors was met.

Passage reading accuracy • LP > SP > TW on ac-Passage reading fluency curacy and fluency of Word reading accuracy passage reading.Word reading fluency • TW condition outper-

formed SP and LPon measures of word reading accuracy and fluency.

Oral reading fluency • Intervention led to Oral reading accuracy increased fluency

and increased accu-racy for all 3 stu-dents.

Average number of words read SR > C; SR vs. C: orally in 3 min. d = .17; RR vs. C:

d = .10; SR vs. RR: d = .07

Average number of correct SR vs. C: d = .04; responses to 10 comprehension RR vs. C: d = .01;questions SR vs. RR: d = .04

Number of maze items correcta SR vs. C: d = .15; RR vs. C: d = –.24; SR vs. RR: d = .38

Schonell Word Recognition Test • Student improved in Reading accuracy word recognition, Reading rate reading accuracy

from pretest to follow-up.

• Student’s reading rate decreased from pretest to posttest but increased consid-erably at follow-up.

(Table continues)

jld-v35n5-p386 8/22/02 2:34 PM Page 393

394

(Table 2 continued)


Rose, 19849.5–13.1 yearsAlternating treatment design

Rose & Beattie, 19868.7–11.6 yearsAlternating treatment design

Smith, 1979 (Study 1);8 yearsMultiple baseline,

multi-element design

Smith, 1979 (Study 2)12 yearsMulti-element single-

subject design

N = 6• Baseline: Oral reading only, no pre-

viewing. • Silent previewing: Student read pas-

sage silently before reading it aloud tothe teacher.

• Listening previewing: Teacher readpassage aloud while student followedalong. Then student read passagealoud to teacher.

Length and duration: Approximately 25 days of intervention. Length of dailyintervention not specified.

N = 4• Baseline: Daily individual oral reading,

introduction to new words, practice ofnew words via flash cards, sentenceconstruction using new words, andworksheet practice.

• Listening previewing: Teacher readsassigned passage orally at relativelyslow conversational rate (approxi-mately 130–160 wpm) as students fol-low along. All other instructional proce-dures same as baseline.

• Taped previewing: Identical to listening,except teacher had prerecorded thetaped passage and students followedalong with the tape.

Length and duration: 3–4-min. sessionsdaily. Approximately 30 days of inter-vention alternating between 3 condi-tions.

N = 2• Baseline: Children read a passage at

instructional level without support. • Modeling: Teacher read child’s pas-

sage for 1 minute at 100 wpm. Childcontinued reading from where teacherstopped.

• Follow-up: Students read passage ad-ditional time.

Length and duration: Not specified, allconditions administered once.

N = 1• Baseline: Child read a passage at in-

structional level without support.• Modeling: Teacher read passage for

1 minute at 100 wpm. Child continuedreading from where teacher stopped.

• Modeling with correction: Same asmodeling, plus corrections providedduring student reading.

Words read correctly per minute • Listening previewing led to faster reading rates.

• Both silent and lis-tening previewing ap-peared to be more beneficial than baseline.

Oral reading fluency • Listening and taped Reading accuracy previewing resulted

in increased oral reading rates relative to baseline.

• Listening previewing procedure was more beneficial than taped previewing for 3 of 4 participants.

• Error rates were not effected by the pre-viewing procedures.

Words read correctly per minute • Intervention resulted Errors per minute in increased speed

and accuracy for both students.

Words read correctly per minute • Speed and accuracy Errors per minute increased with each

additional interven-tion component.

• Maximum speed and accuracy achieved under the modeling with correction and previewing condition.

(Table continues)

jld-v35n5-p386 8/22/02 2:34 PM Page 394


ency than the control sample (d = .17),but there were no significant differ-ences between the repeated readingcondition and the control condition onmeasures of fluency (d = .10) or com-prehension (d = .01). Moreover, on amaze task, effect sizes were small butsignificant when the repeated readingwith partners condition was comparedwith the control condition (d = .24) andslightly higher, though still moderate,when the sustained reading conditionwas compared with the control condi-tion (d = .38).

Modeling by audiotape or computer.Results from four samples (N = 12) ad-dressed the question of whether an au-diotaped or computer model or pre-view of the text to be read by thestudents in the sample improved thereading fluency of students with LD.Of the four samples, one used a casestudy design (Moseley, 1993) and threeused a single-subject design (Daly &Martens, 1994; Gilbert, Williams, & Mc-Laughlin, 1986; Rose & Beattie, 1986).In the case study sample, the modelwas provided through a speech syn-thesizer, with the pace controlled bythe student. In this case, the student’s

reading fluency decreased from 76words correct per minute at pretest to63 words correct per minute during theintervention. However, at the follow-up test of fluency, the student read at112 words correct per minute, suggest-ing that the overall impact of the inter-vention may have been positive. In oneof the single-subject designs, Rose andBeattie (1986) compared listening pre-viewing, in which the teacher modeledreading of the text, with a taped pre-view, in which the student controlledthe tape and followed along readingwith the tape. For three of the four stu-dents, the teacher-modeled reading ofthe text was more effective than thetaped model.

Daly and Martens (1994) compared ataped model of passage reading withrepeated reading without a model andwith audiotaped reading of a relatedword list. On measures of passagereading accuracy and fluency, thetaped reading model resulted in con-sistently better performance than re-peated reading without a model andtaped words. The taped words condi-tion resulted in better performance ona measure of word reading accuracy

for three of the four participants in thestudy.

Gilbert et al. (1986) compared ataped model of fluent reading fol-lowed by three repeated readings to abaseline condition in which the teacherintroduced vocabulary and importantphonics elements and the studentsilently read the passage once. Fluencyand accuracy improved for all threestudents in both conditions.

Repeated Reading Interventionswith Multiple Features. Three groupsamples and four single-case samples(N = 52) involved interventions that in-cluded repeated reading as one of sev-eral instructional features. The meaneffect size across interventions on mea-sures of fluency was d = .71 and rangedfrom d = .20 to d = 1.17. These studiesare listed in Table 3.

Simmons, Fuchs, Fuchs, Mathes, andHodge (1995) compared an interven-tion that combined an effective teach-ing component and peer-mediated re-peated reading to traditional readinginstruction. On a measure of oral read-ing fluency, the students who receivedthe combination of effective teaching

(Table 2 continued)


Smith (continued)

Vaughn et al., 20008.5–8.8 yearsQuasi-experimental

pretest–posttest comparison design

aNegative d reflects positive outcome favoring treatment listed first. bStatistical comparisons were not possible because of the small sample size in each group.

• Modeling with correction and preview-ing: After modeling, student reread themodeled portion and continued to readfor 5 minutes.

• Follow-up: Student read passage addi-tional time.


Partner Reading (PR; n = 7): Partnerstook turns reading (3 minutes each)with the more proficient reader readingfirst. 1-minute, timed reading followed.

Collaborative Strategic Reading (CSR; n = 9): Partners used four-strategy ap-proach to reading text

Length and duration: 2–3 sessionsweekly for 12 weeks; approximately 25miuntes per session for PR and 45minutes per session for CSR.

Gray Oral Reading Test, rateb PR vs. CSR: d = .69Gray Oral Reading Test, accuracyb PR vs. CSR: d = .65Gray Oral Reading Test, PR vs. CSR: d = .30

comprehensionb

Test of Reading Fluency, words PR vs. CSR: d = .16correct per minuteb

jld-v35n5-p386 8/22/02 2:34 PM Page 395

396

TABLE 3Studies Examining Repeated Reading with Multiple Features


Simmons, Fuchs, Fuchs, Mathes, & Hodge, 1995

9.47–9.91 yearsTreatment–Comparison

D. Fuchs, Fuchs, Mathes, & Simmons, 1996

Mean age = 9.87 years (PALs group); 10.09 years (No PALs group)

Treatment–Comparison

Sutton, 1991Ages not providedPre–Posttest design

Weinstein & Cooke, 19928 years 1 month–10 years

2 monthsMulti-treatment, single-

subject design (ABACA)

• Effective teaching plus peer tutoring(ET+PT; n = 11): Effective instructionalprinciples and peer tutoring using re-peated reading (3 readings/passagefor first 4 weeks, 2 readings/passagefor second 4 weeks), story retells, andparagraph summarization.

• Comparison (C; n = 29): Traditionalreading instruction.

Duration: 800 minutes

• Peer Assisted Learning (PALs; n = 20):Partner reading with retell (one re-peated reading), paragraph summary,and prediction relay.

• No PALs (n = 20): Traditional readinginstruction.

Duration: 1,350 minutes

N = 17Four-element condition:1. Teacher modeled reading of story.2. Target students read to tutor partners.3. Partner reading.4. Target student read story to teacher.

N = 4Baseline: Each student read first set of

3 passages for first baseline phaseand the intervention conditions. Sameprocedure was used for the second setof passages. Third set of 3 passageswas used for final baseline.

Intervention (10 min./day):1. Students listened to taped model at

100 wpm.2. Students asked to read passage

quickly and accurately.

No. of words read in 3 min. ET + PT > C; ET + PT vs. C: d = .73

No. of words read in 3 min. No significant differences(delayed) between groups;

ET + PT vs. C: d = .53No. comprehension questions ET + PT > C;

correct ET + PT vs. C: d = .82No. comprehension questions No significant differences

correct (delayed) between groups; ET + PT vs. C: d = .36

No. maze items correct in 2 min. ET + PT > C; ET + PT vs. C: d = 1.00

Matched words in recall summaries No significant differences between groups; ET + PT vs. C: d = 1.05

Total words in recall summaries No significant differences between groups; ET + PT vs. C: d = .78

SAT Comprehension No significant differences between groups; ET + PT vs. C: d = .56;

Mean effect size (ET + PT vs. C): d = .73

Average number of words read PALs vs. No PALs: orally in 3 min. d = .20

Average number of correct PALs vs. No PALs: responses to 10 comprehension d = .63questions

Number of maze items correct PALs vs. No PALs: d = .49

Mean effect size (PALs v. No PALs): d = .44

Brigance Test of Oral Reading Posttest vs. Pretest: (words per minute) d = 1.17

Brigance Test of Oral Reading Posttest vs. Pretest: (errors per minute)a d = .91;

Mean effect size (Posttest vs. Pretest): d = 1.04

Oral reading fluency • All students made progress over base-line; mean gains rang-ing from 16.1 to 39.4 words correct per minute.

• Mean gain for the fixed-rate phase = 62%

• Mean gain for the im-provement phase = 58%

(Table continues)

jld-v35n5-p386 8/22/02 2:34 PM Page 396


and repeated reading performed sig-nificantly better than the comparisonsample. Similar significant differenceswere found for a measure of compre-hension and for a maze measure. Al-though no other significant differenceswere noted, the mean effect size for thecombination intervention versus thecontrol sample was moderate to largeat d = 73. D. Fuchs et al. (1997) com-pared a partner reading interventionthat included repeated readings of textand comprehension activities (para-graph summarization and predictionactivities) to a traditional reading pro-gram, yielding a low to moderatemean effect size of d = .44.

A four-element intervention imple-mented by Sutton (1991), which in-cluded a combination of teacher-modeled reading, target students’rereading to a tutor, peer-paired read-ing, and target students’ rereading tothe teacher, resulted in a considerableincrease in reading rate and a decreasein reading errors, with a large mean ef-fect size of d = 1.04. Weinstein andCooke (1992) used a similar interven-tion in a single-subject design in whichstudents listened to a taped model be-fore rereading the passage to a partic-ular criterion and then examining theirprogress as it was plotted on a graph.The four participating students all ex-perienced increased fluency as a resultof this intervention.

Other Elements That Influence Flu-ency Performance in Repeated Read-ing Interventions. Various other ele-

ments of interventions may affect read-ing fluency. Studies that addressedthese elements are listed in Table 4.

Amount of text. A. L. Cohen (1988)compared the amount of text pre-sented to students as they repeatedlyread passages from a computer screen.One sample (N = 16) was presented apassage at a rate of three to five wordsat a time, whereas a second sample (N = 16) had control over the amountof text that was presented. Both sam-ples were compared to a control sam-ple. No significant differences werenoted between the repeated readingsamples. The sample that receivedonly three to five words per screenscored significantly higher on mea-sures of single- and multisyllabic wordreading accuracy. Both repeated read-ing samples demonstrated improvedfluency over the control condition, witha large mean effect size of d = 1.98. Ef-fect size comparisons of the two re-peated reading groups were small,with the exception of reading accuracy(ranging from d = .56 to d = .89), favor-ing the controlled presentation of threeto five words per screen.

Text difficulty. Three samples (N =37) were studied to better understandthe influence of text difficulty in re-peated reading interventions. Sindelar,Monda, and O’Shea (1990) comparedrepeated reading of instructional-leveltexts (defined as text that could be readat 50–100 words per minute with twoor fewer errors) to repeated reading ofmastery-level texts (defined as text thatcould be read at more than 100 words

per minute). Statistically significantdifferences on a measure of oral read-ing fluency favored the mastery-leveltext sample (d = 1.57). However, theinstructional-level sample significantlyoutperformed the mastery-level sam-ple on accuracy (d = .61). No significantdifferences were identified on the mea-sure of comprehension.

In a related study, Rashotte and Tor-gesen (1985) compared repeated read-ing of text that included a high pro-portion of overlapping words withrepeated reading of text in which therewas a low degree of overlap. Therewere no significant differences betweengroups. However, the condition inwhich there were few overlappingwords performed better than the con-dition with a high degree of overlap onall measures.

Number of repetitions. To determinethe number of times that studentsshould repeatedly read text for themost fluency benefit, the findings fromtwo samples (N = 54) are relevant.O’Shea et al. (1987) used a factorial de-sign to study the relative influence ofthe number of repetitions on fluency.They used three intervention levels: asingle reading, three repeated read-ings, and seven repeated readings. Ona measure of oral reading fluency,main effects were identified with sig-nificant differences between all groups.Seven readings resulted in higher per-formance than three readings, whichwas significantly better than a singlereading. On a measure of story retell-ing, there were no differences between

(Table 3 continued)


Weinstein (continued)

aThese effect sizes were changed to positive numbers to reflect a growth in student accuracy rather than a decrease in errors.

3. For fixed criterion phase, each studentreread the passage twice daily until heor she met the specified criterion of 90 wcpm.

4. For improvement phase, studentsreread a passage until they achieved 3 successive improvements.

5. Results were plotted and shared withstudent immediately.

• Generalization im-proved after improve-ment phase from 5% to 89% but was mixed for fixed-rate phase, ranging from –25% to 56%.

jld-v35n5-p386 8/22/02 2:34 PM Page 397

398

TABLE 4Studies That Examined Other Elements of Repeated Reading Interventions


A. L. Cohen, 19888 years 7 months–13 years

2 monthsMultiple-group comparison

Lovitt, T. W., & Hansen, C. L., 1976

8–12 yearsOne-group pretest–posttest

Sindelar, Monda, & O’Shea, 1990Age not reportedTreatment–Comparison

• Processing Power (PP; n = 16): Re-peated reading (4 times) with text pre-sented 3–5 words at a time.

• Repetitive Reading (RR; n = 16): Re-peated reading (4 times) with studentcontrol of text amount.

• No-treatment comparison (C; n = 15).Duration: 195–202 minutes

• Baseline (B) (N = 7): Students readaloud to the teacher. Teacher suppliesmissed or mispronounced words.Teacher records responses and stu-dents respond to written comprehen-sion questions after reading.

• Treatment (T): Students reread or skiplevelled passages contingent on theircorrect word rate and comprehensionscores. Drill was provided on portionsof reading that were problematic.

Duration: 800 minutes

• Repeated reading–Instructional level (I; n = 17): Reread text 3 times at50–100 wpm.

• Repeated reading–Mastery level (M; n = 8): Reread text 3 times at 100 wpmor faster.

Paragraph reading speed RR vs. PP: d = .19(practiced)

Paragraph reading accuracy PP vs. RR: d = .56(practiced)

Paragraph reading speed PP vs. RR: d = .26(unpracticed)

Paragraph reading accuracy PP vs. RR: d = .19(unpracticed)

Reading fluency, final text PP vs. RR: d = .30; PP vs. C: d = 1.58; RR vs. C: d = 1.25

Reading accuracy, final text PP vs. RR: d = .89; PP vs. C: d = 3.02; RR vs. C: d = 2.09

Word reading speed (single PP > C; insufficient in-syllable, practiced) formation for d

Word reading speed (multisyllable, PP > C; insufficient in-practiced) formation for d

Word reading speed (unpracticed) No significant differ-ences between groups; insufficient information for d

Passage comprehension No significant differ-ences between groups; insufficient information for d;

Mean effect size for re-peated reading inter-ventions: d = 1.98

Correct word rate (cwpm) All students improved in correct rate. Mean gain of 9.3 cwpm. In-creases ranged from 2.4 cwpm to 15.3 cwpm.

Orral error rate (epm) Four of the seven par-ticipants decreased their error rates. The mean error rate im-proved from 31 epm to 2.9 epm. De-creases ranged from .5 to –.9 epm.

Percentage of comprehension All students improved in questions correct their comprehension

responses, with a mean increase of 11.9% and a range of 5.7% to 16.1%.

Oral reading fluency M > I; M vs. I (1 reading): d = 2.31;M vs. I (3 readings): d = 1.57

(Table continues)

jld-v35n5-p386 8/22/02 2:34 PM Page 398

399

(Table 4 continued)


Sindelar (continued)

Rashotte & Torgesen, 1985 (based on dissertation by Rashotte, 1984)

8.6–12 yearsMultiple treatment design

O’Shea, Sindelar, & O’Shea, 1987

11.3–13.6 years2 (Focus) × 3 (No. of read-ings) factorial design

Sindelar, Monda, & O’Shea, 1990

Age not reported2 (Reading level) × 2 (No.

of readings) factorial design

Smith, 1979 (Study 2)12 years

Duration: Each condition administeredonce.

N = 12• Repeated reading, no overlapping

words (NO): Read stories 4 times withabout 20 words in common acrossstories.

• Repeated reading with high overlap(HO): Same as NO but with about 60words in common.

• Sustained Reading (SR): Read 4 differ-ent stories each day.

Duration: 7 days

N = 29• Repeated Reading 7 (R7): Read text 7

times.• Repeated Reading 3 (R3): Read text 3

times.• Reading (R): Read text once.Duration: Each treatment administered

once.

N = 25• Reading text once (R1).• Repeated reading (R3): Reread text 3

times.Duration: Each condition administered

once.

N = 1• Baseline: Child read a passage at in-

structional level without support.

Errors per minutea I > M; M vs. I (1 read-ing): d = .88;M vs. I (3 readings): d = .61

Number of propositions retold No significant differ-ences between groups; M vs. I (1 reading): d = .78;M vs. I (3 readings): d = .34

Reading rate (slope of progress) Both interventions with repeated reading performed signifi-cantly better than sustained reading on reading rate.

NO vs. HO: d = .18; NO vs. SR: d = .65;HO vs. SR: d = .35; HO slope, SR slope > 0

Reading accuracy (slope of NO vs. HO: d = .26; progress)a NO vs. SR: d = .52;

HO vs. SR: d = .24; HO slope, SR slope > 0

Passage comprehension (slope No significant differ-of progress) ences between

groups; NO vs. HO: d = .10;NO vs. SR: d = .08; HO vs. SR: d = .17;

Mean effect size, re-peated reading vs. sustained reading: d = .34

Oral reading fluency R7 > R3 > RInsufficient information

to calculate ES.Story retell R7, R3 > R

Insufficient information to calculate ES.

Oral reading fluency R3 > R1, R3 vs. R1: d = 2.70

Oral reading accuracy R1 > R3, R3 vs. R1: d = –.62

Number of propositions retold R3 > R1, R3 vs. R1: d = 1.67;

Mean effect size (R3 vs.R1): d = 1.25

Words read correctly per minute • Fluency and accu-Errors per minute racy increased with

(Table continues)

jld-v35n5-p386 8/22/02 2:34 PM Page 399


the repeated reading conditions, butboth repeated reading conditions re-sulted in significantly higher scoresthan the single reading condition.

Similarly, Sindelar et al. (1990) foundthat rereading text three times resultedin significantly better performance ona measure of oral reading fluency thanreading the text once. Similar differ-ences were noted on a comprehensionmeasure.

Types of feedback. Findings from onesingle-subject sample relate to the in-fluence of feedback during repeatedreading. Smith (1979) found that fol-lowing teacher modeling of fluent

reading, providing the correct wordswhen the student read words incor-rectly during oral reading resulted inan increase of more than 20 words cor-rect per minute over baseline, and er-rors decreased from 13.6 to 9.4 errorsper minute.

Criteria for repeated reading. Elevensingle-subject samples studied the in-fluence of establishing particular crite-ria for repeated-reading interventions.Using an alternative treatment design,Weinstein and Cooke (1992) comparedrepeated reading using a criterion of 90 words per minute (fixed rate) witha criterion based on individual im-

provement as a basis for increasing thedifficulty of the text. They found thefixed-rate criterion more effective thanthe individual improvement criterion,although the individual improvementseemed to facilitate generalization tounpracticed text.

Similar to the fixed criterion condi-tion studied by Weinstein and Cooke(1992), Lovitt and Hansen (1976) de-signed an intervention that requiredstudents to meet a particular set of cri-teria (> 75 words read correctly perminute, < 4.5 errors per minute, and87% comprehension questions correct)in order to skip a difficulty level of text.

(Table 4 continued)


Smith (continued)Multi-element single- subject design.

Weinstein & Cooke, 19928 years 1 month–10 years

2 monthsMulti-treatment, single-

subject design (ABACA)

aNegative d reflects positive outcome favoring treatment listed first.

• Modeling: Teacher read passage for 1 minute at 100 wpm. Child continuedreading from where teacher stopped.

• Modeling with correction: Same asmodeling, plus corrections providedduring student reading.

• Modeling with correction and preview-ing: After modeling, student reread themodeled portion and continued to readfor 5 minutes.

• Follow-up: Student read passage addi-tional time.


N = 4Baseline: Each student read first set of

3 passages for first baseline phase.The same passages were used in theintervention conditions. Same proce-dure was used for the second set ofpassages. Third set of 3 passageswere used for final baseline.

Intervention (10 min./day):1. Students listened to taped model at

100 wpm.2. Students asked to read passage

quickly and accurately.3. For fixed criterion phase, each student

reread the passage twice daily until heor she met the specified criterion of 90wcpm.

4. For improvement phase, studentsreread a passage until they achieved3 successive improvements.

5. Results were plotted and shared withstudent immediately.

each additional inter-vention component.

• Maximum fluency and accuracy achieved under the modeling with correc-tion and previewing condition.

Oral reading fluency • All students made progress over base-line; mean gains ranging from 16.1 to 39.4 words correct per minute.

• Mean gain for the fixed-rate phase = 62%

• Mean gain for the im-provement phase = 58%

• Generalization im-proved after improve-ment phase from 5% to 89% but was mixed for fixed-rate phase, ranging from –25% to 56%.

jld-v35n5-p386 8/22/02 2:34 PM Page 400


If the three criteria were not met in 7 days, the reader received drill andpractice on the difficult words andphrases in the text. The researchers re-ported that all participating studentsimproved in reading fluency duringthe performance-based advancementcondition by 10 words correct perminute and answered almost 12%more comprehension questions accu-rately. These results were maintainedat follow-up.

Word Practice Interventions

Nine single-subject samples were stud-ied to examine the effectiveness of de-coding interventions on improvingreading fluency (see Table 5). O’Shea,Munson, and O’Shea (1984) applied an

alternating treatment design to com-pare the relative effectiveness of aword drill procedure and a phrase drillprocedure. Each procedure was basedon words missed during an initial oralreading baseline phase. Words missedafter oral reading were placed on flashcards, and students were given the op-portunity to practice on half of thewords in isolation (word drill). Alter-natively, in the phrase drill condition,the error words were practiced in con-textual phrases. There were negligibledifferences between alternating treat-ment conditions for isolated wordreading and reading fluency. Signifi-cant differences between conditionswere demonstrated only on a measureof passage reading accuracy, favoringthe phrase drill condition.

Employing a multi-element design,Daly and Martens (1994) comparedtaped previewing of words read in alist to participating students to re-peated reading conditions. Studentsdemonstrated greater word readingaccuracy in this taped words conditionthan in the baseline or repeated read-ing conditions. However, on measuresof passage reading fluency and accu-racy, the repeated reading conditionwas consistently more effective.

Discussion

Fluency has been identified as an es-sential link between word analysis andcomprehension of text and is consid-ered a necessary tool for learning from

TABLE 5Studies Examining Fluency Practice at the Word Level

Number of students/Author/participant treatment description/

age/design length and duration Dependent measures Results

Daly & Martens, 19948 years 10 months–11 years

11 monthsMulti-element design

O’Shea, Munson, & O’Shea, 1984

7–11 yearsAlternating treatment

design

N = 4* All conditions followed a model, drill,

and train to criterion lesson structure.* Subject Passage Preview (SP): Read


* Listening Passage Preview (LP): Lis-tened to audiotaped (130 wpm) pas-sage, read list of unknown words, andreread passage for assessment.

* Taped Words (TW): Read aloud alongwith audiotaped word list (80 wpm).Reread list and read passage for as-sessment.


N = 5* Baseline (Word supply): When an error

was made during oral reading, theteacher supplied the word.

* Word drill: Words were supplied byteacher during oral reading. After read-ing, students were drilled on half oferror words using 5" × 8" flash cards.

* Phrase drill: Same procedure as con-trol. After reading, words were drilledusing phrases in which they occurredin text.

Length and duration: 30 min. sessions/day for 10 days; 300 min.

Passage reading accuracy * LP > SP > TW on Passage reading fluency accuracy and fluency Word reading accuracy of passage reading.Word reading fluency * TW condition outper-

formed SP and LPon measures of word reading accuracy and fluency.

Error words read correctly No significant differ-in isolation ences between drill

Error words read correctly conditions and in passage context control.

Words read correctly per minute Significant effects favor-in daily reading passage ing phrase drill.

Negligible differences between conditions.

jld-v35n5-p386 8/22/02 2:34 PM Page 401


reading (Chall, 1983). The relationshipof fluent oral reading and overall read-ing ability is supported by both em-pirical and clinical evidence (Meyer &Felton, 1999; Rasinski, Padak, Linek, &Sturtevant, 1994; Reutzel & Hollings-worth, 1993). Despite the importanceof fluency, its essential role in buildingoverall reading ability has only re-cently been highlighted (NRP, 2000).

Fluency appears to be particularlyimportant for students with significantreading problems, because they oftenhave labored reading with manypauses, which results in slow and dis-connected oral reading. This effortfulreading is problematic because it fo-cuses reading at the decoding andword level, which makes comprehen-sion virtually impossible. Chall (1979)described these readers as “glued toprint” (p. 41) and unable to delight inthe reading of text. The students withLD who were the target group for thissynthesis represent a large subgroup ofdysfluent readers.

The National Reading Panel (2000)summarized findings about guided re-peated oral reading as a means to im-prove fluency and indicated that theoverall weighted effect size produced amoderate effect for repeated oral read-ing. The NRP presented the case thatinstruction in guided oral reading is animportant part of a reading programand is associated with gains in fluencyand comprehension. Oral reading in-terventions were found to be superiorto instruction encouraging students toread silently. Furthermore, the NRPreported that good and poor readersboth benefited from the repeatedguided reading, although they maybenefit differentially from different as-pects of the treatment (Faulkner &Levy, 1999). However, the NRP readingfluency synthesis did not address theextent to which individuals with LDmight benefit from fluency interven-tions or the extent to which other typesof fluency interventions (other thanoral repeated reading) might be associ-ated with improvements in fluencyand other aspects of reading.

The purpose of this study was toprovide a synthesis of the research onfluency interventions conducted withstudents with LD. Our goal was to lo-cate all intervention studies publishedand all dissertations conducted withinthe past 25 years that evaluated the ef-fects of fluency training on elementarystudents with LD. The comprehensivesearch yielded 24 studies: 8 multiplegroup, 5 single group, and 11 casestudies or single-subject design stud-ies. Two of the single-group studieswere part of factorial designs that alsoincluded other samples.

Before interpreting the findings ofthe present synthesis, it is important tonote that effect sizes can be consideredonly within the context of the compar-isons with which treatment groupswere contrasted. Because effect sizesare largely dependent on the nature ofthe comparison groups, it is criticalthat a synthesis include detailed infor-mation regarding the comparisons. Forthis reason, Tables 1 through 5 includedetails of both treatment and compari-son conditions. However, comparisongroups differ considerably across sam-ples, complicating the interpretation ofthe findings. The interpretations thatfollow were developed with this limi-tation in mind.

In general, the findings from thissynthesis suggested that repeatedreading interventions for studentswith LD are associated with improve-ments in reading rate, accuracy, andcomprehension. This would providesupport for the theory of automaticityproposed by LaBerge and Samuels(1974) and extended as a verbal effi-ciency model by Perfetti (1977, 1985).These studies, and the theory support-ing them, provide evidence that thefocus on developing students’ rapidprocessing of print by reading targetpassages more than once is often effec-tive as a means to improve accuracyand speed, and ultimately leads to bet-ter understanding of text.

One procedure for enhancing flu-ency is for teachers to model reading of text by reading aloud to students

(Dowhower, 1987; Hoffman, 1987;Smith, 1979). Repeated reading with amodel seems to be more effective thanrepeated reading with no model, par-ticularly for students with low fluency(e.g., Rose & Beattie, 1986; Smith, 1979).Tape- or computer-modeled readingseems more effective than having nomodel but may not be as effective asteacher modeling (Daly & Martens,1994; Rose & Beattie, 1986). Further-more, having text read initially by amodel promoted comprehension, per-haps because it allowed students tofocus initially on the content of the pas-sage before they read it themselves(Monda, 1989; Rose & Beattie, 1986).

Asking peers, who are often betterreaders, to serve as the model for stu-dents with LD was investigated in sev-eral studies reported here and re-viewed in separate syntheses (Elbaum,Vaughn, Hughes, & Moody, 1999;Mathes & Fuchs, 1994). Repeated read-ing with a partner as a means to im-proving fluency has yielded somewhatequivocal results (e.g., Marston, Deno,Dongil, Diment, & Rogers, 1995), al-though there are few studies docu-menting its effectiveness alone (Mar-ston et al., 1995; Mathes & Fuchs, 1994).In a separate analysis of the effects ofpeer tutoring on broad reading out-comes, cross-age tutoring was associ-ated with higher mean weighted effectsizes (.50) than cooperative partners(.00), and with peer tutoring, the role ofthe student within the pair had a sig-nificant effect on outcomes, with recip-rocal tutor–tutee roles demonstratinglow mean weighted effect sizes (.09; El-baum, Vaughn, Hughes, & Moody,2000).

Speed and accuracy have tradition-ally been considered the hallmarks ormost essential features of fluency(LaBerge & Samuels, 1974; Samuels,1997). Most researchers agree that ac-curacy in itself is insufficient and thatstudents need to read rapidly if theyare going to understand the connec-tions that need to be made betweenideas in print (Nathan & Stanovich,1991). Variables associated with effects

jld-v35n5-p386 8/22/02 2:34 PM Page 402


for fluency include controlling the dif-ficulty of text and providing feedbackfor words missed. Advancing studentsthrough progressively more difficulttext based on their performance seemsto enhance their overall fluency (Lovitt& Hansen, 1976; Weinstein & Cooke,1992), as does correction and feedbackfor words read incorrectly (Smith,1979). Rereading text many times andto many different people and provid-ing progressively more difficult textwith feedback and correction formissed words may be the componentsessential to improving fluency.

Another approach to fluency build-ing is to provide struggling readerswith text chunked in words or phrasesas a means of improving fluency andcomprehension (Young & Bowers, 1995).The research in this review reveals thatvarying the amounts of text presentedin repeated reading does not seem tochange the outcome. However, con-trolling the amount of text presentedmay be beneficial for students who areexperiencing difficulty with readingaccuracy, as it may force them to focuson the words for a longer period oftime (A. L. Cohen, 1988).

Several researchers have argued thatfluency is enhanced when reading ad-dresses the meaning of the text (An-derson, Wilkinson, & Mason, 1991). Atleast for struggling readers and stu-dents with dyslexia in the third grade,a fluency intervention (repeated read-ing) and a comprehension intervention(collaborative strategic reading) wereboth associated with gains in fluencyand comprehension (Vaughn et al.,2000). In this synthesis we found thatalthough comprehension was not typ-ically the focus of the intervention, inmany cases fluency growth was asso-ciated with growth in comprehension(e.g., Simmons et al., 1995; Sindelar et al., 1990). In the lone study wherecomprehension instruction was com-bined with repeated reading (D. Fuchset al., 1997), the effect sizes for growthin comprehension were moderate andexceeded the effect sizes for fluency.Additional research focusing on the re-

lationship between fluency and com-prehension for students with LD iswarranted.

Rereading text or repeated oral read-ing is perhaps the best documented ap-proach to improving fluency (NPR,2000; Rashotte & Torgesen, 1985) andhas been associated with improvedoutcomes for young students (e.g.,O’Shea et al., 1987) and for college stu-dents (Carver & Hoffman, 1981). Gen-erally, intervention research on fluencydevelopment for students with LD hasbeen dominated by research on re-peated reading. This likely reflects theapplication of the theory that fluentreading is promoted by frequent op-portunities to practice with familiartext and to increase exposure to words.It may also be influenced by the find-ing that repeated readings do improvefluency for students with LD. Further-more, rereading the same text moretimes is better than fewer.

Future Research

Like most research syntheses, thisstudy both answers questions andraises new ones. Questions that wouldbe valuable to address in future re-search include,

1. What aspects of guided oral read-ing are associated with positiveoutcomes in fluency? To what ex-tent do these aspects differ basedon the reading level of the student?What about the decoding ability ofthe student?

2. When students are reading, atwhat level is repeated reading as-sociated with the greatest gains influency? What about comprehen-sion?

3. Most research questions haveasked about the extent to whichfluency intervention, particularlyrepeated reading, influences com-prehension. What about the extentto which comprehension instruc-tion influences outcomes in flu-ency? This question is promptedby the strong correlation between

fluent reading and comprehension(Dowhower, 1987; Shinn, Good,Knutson, Tilly, & Collins, 1992).

4. How much text needs to be in-cluded in the repeated reading in-tervention to most effectively influ-ence fluency? Does this vary byage and reading level of students?Should reading material bechunked and repeated at thephrase level, sentence level, ormultiple sentence level? Some re-search has suggested that word-level reading (repeated), such aswith flash cards, is associated withimproved outcomes in comprehen-sion (Tan & Nicholson, 1997).

5. Is there a small number of studentsfor whom automaticity is not pos-sible (e.g., due to neurological diffi-culties), in which case fluencybuilding may be exceedingly diffi-cult? What is the best way to buildcomprehension for these students?

6. Are the effects of fluency-buildingactivities sustainable? Only onestudy in the present corpus (Sim-mons et al., 1995) included a follow-up measure of fluency.Although Simmons et al. (1995)reported no significant differenceson this measure, further researchshould focus on this issue.

Implications for Practice

For many struggling readers, particu-larly students with LD, becoming a flu-ent reader is a challenge that must beovercome in order to progress from de-coding to understanding what is read.Despite the integral role of fluency inreading development, fluency has notplayed a prominent role in reading in-struction (Allington, 1983). Because ofthis apparent lack of attention to flu-ency development, recent syntheses ofresearch on reading have highlightedthe importance of including fluencybuilding as part of daily instruction(Chard, Simmons, & Kameénui, 1998;NRP, 2000; Snow, Burns, & Griffin,1998). The importance of becoming afluent reader warrants careful atten-

jld-v35n5-p386 8/22/02 2:34 PM Page 403


tion to the evidence documentingwhich interventions are most effectiveat promoting reading fluency. To date,research on fluency interventions forstudents with LD has focused almostexclusively on repeated reading inter-ventions. The results of the presentsynthesis have provided a more de-tailed look at which features of inter-ventions make them more effective orless effective.

Generally, the findings of this syn-thesis suggest that students with LDwho are experiencing difficulties withfluent reading would benefit from in-terventions that have multiple compo-nents focusing attention on increasingthe rate and accuracy of reading. Thefindings of the present synthesis sup-port earlier findings suggesting thatopportunities to practice reading andrereading familiar text is one way forstudents with LD to enhance theirreading fluency. Although silent read-ing has become a popular feature ofreading instruction nationwide, thereis little evidence to suggest that it is aneffective way to build students’ flu-ency. Equal attention should be paid torepeated reading of text for studentswho continue to struggle with readingfluency.

Another salient finding that has im-plications for classroom instruction isthat students benefit from having amodel of fluent reading. As repeatedreading is implemented, it will be im-portant for teachers to consider thebest way to model fluency before stu-dents engage in repeated reading.Seemingly, the most effective way todo this is by having an adult providethat model. Realistically, however, re-sources are not always available for anadult to model fluent reading. In thesecases, an audiotaped or computer-generated model is an effective substi-tute. Moreover, the findings of thissynthesis support earlier findings (El-baum et al., 1999) suggesting that us-ing grouping practices that allow moreproficient readers to guide less ablereaders is also an effective way to buildfluency.

Several other intervention featuresshould be considered as teachers de-velop instructional activities for flu-ency development. In instances wherecorrective feedback was combinedwith repeated reading, students weremore successful at boosting their flu-ency, primarily by decreasing theirreading errors. Moreover, fluency ap-pears to develop more quickly if delib-erate attention is given to setting crite-ria and adjusting the difficulty level oftext as students progress.

Although more research is needed tobetter understand how reading flu-ency and comprehension are related,the results of this synthesis support thecombination of instructional compo-nents that focus students’ attentionboth on increasing their fluency and onimproving their understanding ofwhat they read. It is important to notethat modeling of fluent reading alsoseems to boost students’ comprehen-sion, as they not only hear how askilled reader reads but are able to un-derstand the text rather than focusingall their attention on decoding. More-over, repeated reading interventionsthat were combined with comprehen-sion activities enhanced both fluencyand comprehension. Thus, it wouldseem to confirm the importance of in-cluding both these elements in dailyinstruction.

The findings of the present reviewprovide strong support for the imple-mentation of fluency-building activi-ties for students with learning disabil-ities. Given the narrowly definedpopulation of students considered inthis review, further research is neededto understand the extent to which re-peated reading and other fluency-building activities can enhance aca-demic outcomes for all readers.

ABOUT THE AUTHORS

David J. Chard, PhD, is an assistant professorof special education at the University of Oregon.His current interests include teacher develop-ment and effective instruction in early literacyand numeracy. Sharon Vaughn, PhD, is the

Mollie Villeret Davis Professor of Reading andLearning Disabilities and director of the TexasCenter for Reading and Language Arts in theCollege of Education at The University of Texasat Austin. Brenda-Jean Tyler, MEd, is a doc-toral candidate in the Department of Special Ed-ucation at The University of Texas at Austin.Her areas of interest include sound reading in-struction for English language learners withlearning disabilities and teacher preparation fordiverse students with learning disabilities. Ad-dress: David J. Chard, 5261 University of Ore-gon, Special Education Area, Eugene, OR97403; e-mail: [email protected]

REFERENCES

Note. Asterisks denote articles that were in-cluded in the synthesis.

Allington, R. E. (1983). The reading instruc-tion provided readers of differing ability.Elementary School Journal, 83, 548–559.

Anderson, R. C., Wilkinson, I. A. G., &Mason, J. A. (1991). A microanalysis ofsmall-group, guided reading lessons: Ef-fects of an emphasis on global storymeaning. Reading Research Quarterly, 26,417–441.

Arreaga-Mayer, C., Terry, B. J., & Green-wood, C. R. (1998). Classwide peer tutor-ing. In K. Topping & S. Ehly (Eds.), Peer-assisted learning (pp. 105–120). Mahwah,NJ: Erlbaum.

Campbell, D. T., & Stanley, J. C. (1966). Ex-perimental and quasi-experimental designsfor research. Chicago: Rand McNally.

Carver, R. P., & Hoffman, J. V. (1981). Theeffect of practice through repeated read-ing on gain in reading ability using acomputer-based instructional system.Reading Research Quarterly, 16, 374–390.

Chall, J. S. (1979). The great debate: Tenyears later, with a modest proposal forreading stages. In L. B. Resnick & P. A.Weaver (Eds.), Theory and practice of earlyreading (Vol. 1, pp. 29–55). Hillsdale, NJ:Erlbaum.

Chall, J. S. (1983). Stages of reading develop-ment. New York: McGraw-Hill.

Chard, D. J., Simmons, D. C., & Kameénui,E. J. (1998). Word recognition: Instruc-tional and curricular basics and implica-tions. In D. C. Simmons & E. J. Kameénui(Eds.), What reading research tells us aboutchildren with diverse learning needs: Thebases and the basics (pp. 169–182). Hills-dale, NJ: Erlbaum.

jld-v35n5-p386 8/22/02 2:34 PM Page 404


*Cohen, A. L. (1988). An evaluation of the ef-fectiveness of two methods for providing com-puter-assisted repeated reading training toreading disabled students. Unpublisheddoctoral dissertation, Florida State Uni-versity, Tallahassee.

Cohen, J. (1988). Statistical power analysis forthe behavioral sciences (2nd ed.). Hillsdale,NJ: Erlbaum.

*Daly, E. J., & Martens, B. K. (1994). A com-parison of three interventions for increas-ing oral reading performance: Applica-tion of the instructional hierarchy. Journalof Applied Behavior Analysis, 27, 459–469.

Dowhower, S. L. (1987). Effects of repeatedreadings on selected second grade transi-tional readers’ fluency and comprehen-sion. Reading Research Quarterly, 22, 389–406.

Ehri, L. C., & Wilce, L. S. (1983). Develop-ment of word identification speed inskilled and less skilled beginning readers.Journal of Educational Psychology, 75, 3–18.

Elbaum, B., Vaughn, S., Hughes, M., &Moody, S. W. (1999). Grouping practicesand reading outcomes for students withdisabilities. Exceptional Children, 65, 399–415.

Elbaum, B. E., Vaughn, S., Hughes, M. T., &Moody, S. W. (2000). How effective areone-to-one tutoring programs in readingfor elementary students at risk for read-ing failure? Journal of Educational Psychol-ogy, 92, 605–619.

Faulkner, H. J., & Levy, B. A. (1999). Fluentand nonfluent forms of transfer in read-ing: Words and their message. Psycho-nomic Bulletin and Review, 6, 111–116.

*Fuchs, D., Fuchs, L. S., Mathes, P. G., &Simmons, D. C. (1997). Peer-assistedlearning strategies: Making classroomsmore responsive to diversity. AmericanEducational Research Journal, 34, 174–206.

Fuchs, L. S., Fuchs, D., Hosp, M. K., & Jen-kins, J. R. (2001). Oral reading fluency asan indicator of reading competence: Atheoretical, empirical, and historical anal-ysis. Scientific Studies of Reading, 5, 239–256.

*Gilbert, L. M., Williams, R. L., & McLaugh-lin, T. F. (1986). Use of assisted reading toincrease correct reading rates and de-crease error rates of students with learn-ing disabilities. Journal of Applied BehaviorAnalysis, 29, 255–257.

Hedges, L. V., & Olkin, I. (1985). Statisticalmethods for meta-analysis. Orlando, FL:Academic Press.

Hoffman, J. V. (1987). Rethinking the role oforal reading in basal instruction. Elemen-tary School Journal, 87, 367–374.

Huey, E. B. (1908). The psychology and peda-gogy of reading. Cambridge, MA: MIT Press.

Kameénui, E. J., & Simmons, D. C. (2001).Introduction to this special issue: TheDNA of reading fluency. Scientific Studiesof Reading, 5, 203–210.

Kuhn, M. R., & Stahl, S. A. (2000). Fluency:A review of developmental and remedial prac-tices (Technical Report No. 2-008). AnnArbor: University of Michigan, Center for the Improvement of Early ReadingAchievement.

LaBerge, D., & Samuels, S. J. (1974). Towarda theory of automatic information pro-cessing in reading. Cognitive Psychology,6, 293–323.

*Lovitt, T. W., & Hansen, C. L. (1976). Theuse of contingent skipping and drilling to improve oral reading and compre-hension. Journal of Learning Disabilities, 9,20–26.

*Marston, D., Deno, S. L., Dongil, K., Di-ment, K., & Rogers, D. (1995). Compari-son of reading intervention approachesfor students with mild disabilities. Excep-tional Children, 62, 20–37.

*Mathes, P. G., & Fuchs, L. S. (1993). Peer-mediated reading instruction in specialeducation resource rooms. Learning Dis-abilities Research & Practice, 8, 233–243.

Mathes, P. G., & Fuchs, L. S. (1994). Theefficacy of peer tutoring in reading forstudents with mild disabilities: A best-evidence synthesis. School Psychology Re-view, 23, 59–80.

Meyer, M. S., & Felton, R. H. (1999). Re-peated reading to enhance fluency: Oldapproaches and new direction. Annals ofDyslexia, 49, 283–306.

*Monda, L. E. (1989). The effects of oral, silent,and listening repetitive reading on the flu-ency and comprehension of learning disabledstudents. Unpublished doctoral disserta-tion, Florida State University, Tallahas-see.

*Moseley, D. (1993). Visual and linguisticdeterminants of reading fluency indyslexics: A classroom study with talkingcomputers. In S. F. Wright & R. Groner(Eds.), Facets of dyslexia and its remediation(pp. 567–584). London: Elsevier.

Nathan, R. G., & Stanovich, K. E. (1991).The causes and consequences of differ-ences in reading fluency. Theory Into Prac-tice, 30(3), 176–184.

National Reading Panel. (2000). Teachingchildren to read. Washington, DC: NationalInstitutes of Health.

*O’Shea, L. J., Munson, S. M., & O’Shea, D. J. (1984). Error correction in oral read-ing: Evaluating the effectiveness of threeprocedures. Education and Treatment ofChildren, 7, 203–214.

*O’Shea, L. J., Sindelar, P. T., & O’Shea, D. J.(1987). The effects of repeated readingsand attentional cues on the reading flu-ency and comprehension of learning dis-abled readers. Learning Disabilities Re-search, 2, 103–109.

Perfetti, C. A. (1977). Language compre-hension and fast decoding: Some psycho-linguistic prerequisites for skilled read-ing comprehension. In J. T. Guthrie (Ed.),Cognition, curriculum, and comprehension(pp. 20–41). Newark, DE: InternationalReading Association.

Perfetti, C. A. (1985). Reading ability. NewYork: Oxford University Press.

Rashotte, C. A. (1984). Reported reading andreading fluency in learning disabled children.Unpublished doctoral dissertation, Flo-rida State University.

*Rashotte, C. A., & Torgesen, J. K. (1985).Repeated reading and reading fluency inlearning disabled children. Reading Re-search Quarterly, 20, 180–188.

Rasinski, T. V., Padak, N., Linek, W., &Sturtevant, E. (1994). Effects of fluencydevelopment on urban second-gradereaders. Journal of Educational Research, 87,158–165.

Reutzel, D. R., & Hollingsworth, P. M.(1993). Effects of fluency training on sec-ond graders’ reading comprehension.Journal of Educational Research, 86, 325–331.

*Rose, T. L. (1984). The effects of twoprepractice procedures on oral reading.Journal of Learning Disabilities, 17, 544–548.

*Rose, T. L., & Beattie, J. R. (1986). Relativeeffects of teacher-directed and taped pre-viewing on oral reading. Learning Disabil-ity Quarterly, 9, 193–199.

Samuels, S. J. (1997). The method of re-peated readings. The Reading Teacher, 50,376–381.

Shinn, M. R., Good, R. H., Knutson, N.,Tilly, W. D., & Collins, V. (1992). Curriculum-based measurement of oralreading fluency: A confirmatory analysisof its relation to reading. School Psychol-ogy Review, 21, 459–479.

jld-v35n5-p386 8/22/02 2:34 PM Page 405


*Simmons, D. C., Fuchs, L. S., Fuchs, D.,Mathes, P., & Hodge, J. P. (1995). Effectsof explicit teaching and peer tutoring onthe reading achievement of learning-disabled and low-performing students inthe classroom. The Elementary School Jour-nal, 95, 387–408.

*Sindelar, P. T., Monda, L. E., & O’Shea, L. J.(1990). Effects of repeated readings on in-structional- and mastery-level readers.Journal of Educational Research, 83, 220–226.

*Smith, D. D. (1979). The improvement ofchildren’s oral reading through the use ofteacher modeling. Journal of Learning Dis-abilities, 12, 172–175.

Snow, C. E., Burns, S. M., & Griffin, P. (Eds.).(1998). Preventing reading difficulties inyoung children. Washington, DC: NationalAcademy Press.

*Stout, T. W. (1997). An investigation of the ef-fects of a repeated reading intervention on thefluency and comprehension of students with

language-learning disabilities. Unpublisheddoctoral dissertation, Georgia State Uni-versity, Atlanta.

*Sutton, P. A. (1991). Strategies to increase oralreading fluency of primary resource students.Unpublished manuscript, Nova Univer-sity.

*Swain, K. D., & Allinder, R. M. (1996). Theeffects of repeated reading on two typesof CBM: Computer maze and oral read-ing with second-grade students withlearning disabilities. Diagnostique, 21, 51–66.

Tan, A., & Nicholson, T. (1997). Flashcardsrevisited: Training poor readers to readwords faster improves their comprehen-sion of text. Journal of Educational Psychol-ogy, 59, 276–288.

Topping, K., & Ehly, S. (1998). Peer-assistedlearning. Mahwah, NJ: Erlbaum.

*Vaughn, S., Chard, D. J., Bryant, D. P.,Coleman, M., Tyler, B. J., Thompson, S., et al. (2000). Fluency and comprehension

interventions for third-grade students.Remedial and Special Education, 21, 325–335.

*Weinstein, G., & Cooke, N. L. (1992). Theeffects of two repeated reading interven-tions on generalization of fluency. Learn-ing Disability Quarterly, 15, 21–28.

Wolf, M., & Bowers, P. (1999). The “double-deficit hypothesis” for the developmen-tal dyslexias. Journal of Educational Psy-chology, 91(3), 1–24.

Wolf, M., & Bowers, P. (2000). The questionof naming-speed deficits in developmen-tal reading disabilities: An introductionto the double-deficit hypothesis. Journalof Learning Disabilities, 33, 322–324.

Wolf, M., & Katzir-Cohen, T. (2001). Read-ing fluency and its intervention. ScientificStudies of Reading, 5, 211–238.

Young, A., & Bowers, P. G. (1995). Individ-ual differences and text difficulty deter-minants of reading fluency and expres-siveness. Journal of Experimental ChildPsychology, 60, 428–454.

jld-v35n5-p386 8/22/02 2:34 PM Page 406

a synthesis of research on effective interventions for building … · 2017-11-29 · effective...

Documents