Teaching Students with Moderate Intellectual Disabilities ...daddcec.org/Portals/0/CEC/Autism_Disabilities/Research/... · Teaching Students with Moderate Intellectual Disabilities

Teaching Students with Moderate Intellectual Disabilities toRead: An Experimental Examination of a Comprehensive

Reading Intervention

Jill H. Allor, Patricia G. Mathes, J. Kyle Roberts, Francesca G. Jones, andTammi M. Champlin

Southern Methodist University

Abstract: The primary purpose of our research was to determine if a comprehensive, phonics-based, directinstruction reading program would be effective in teaching early reading and language skills to students withmoderate intellectual disabilities (ID). Participants were 28 elementary students from 10 public schools in anurban school district and one urban private school who were randomly placed into treatment and contrastgroups. Students in the treatment condition received daily, comprehensive reading instruction in small groupsof 1–4 students for approximately 40 minutes per session. A broad array of measures was studied, includingphonemic awareness, phonics, word recognition, comprehension, and oral language. Means favored theintervention group on all measures, with moderate to strong effect sizes. Statistically significant differences werefound on most measures, including phonemic awareness, phonics, vocabulary, and comprehension. Thesefindings demonstrated that students with moderate ID can learn basic reading skills given consistent, explicit,and comprehensive reading instruction across an extended period of time.

In recent years, there has been growing na-tional recognition that literacy is a civil right.The national rhetoric suggests that all chil-dren have the right to scientifically-basedreading instruction and that it is not accept-able for any child to leave school with lowliteracy skills (No Child Left Behind Act,2002). However, within the rhetoric about allchildren, references to students with intellec-tual disabilities (ID), or mental retardation,are typically vague or absent. One might ques-tion how one set of children could be over-looked in discussions about all children. Inour experience, the answer appears to be that

all has really referred to all children who arebelieved to be capable of learning to read(Katims, 2000). We define reading as the abil-ity to process individual words in connectedtext resulting in understanding the author’sintended meaning. With this definition inmind, many educators assume that childrenwith ID are not capable of learning to read(Katims). The expectation has been that, atbest, students with ID, particularly those withmoderate ID, can learn to identify a specificlist of words memorized by sight. The result isthat typically little effort is made to teach thesestudents to become fully literate and only 1 in5 children with mild or moderate ID achieveseven minimal literacy skills (Katims, 2001).

Research on Reading and Intellectual Disabilities

Although much progress has been made inrecent years regarding the education of stu-dents with ID, to date, very little reading re-search has been conducted with these stu-dents. What research has been done hasfocused primarily on students with mild ID(see Browder, Wakeman, Spooner, Ahlgrim-Delzell, & Algozzine, 2006) and has focused

The work presented in this article was supportedby Grant No. H324K040011-05 from the Institute ofEducation Sciences. This article does not necessarilyreflect the positions or policies of this fundingagency and no official endorsement should be in-ferred. We thank the teachers and students whoparticipated in this research. Correspondence con-cerning this article should be addressed to Jill H.Allor, Department of Teaching and Learning, An-nette Caldwell Simmons School of Education andHuman Development, Southern Methodist Univer-sity, P.O. Box 750381, Dallas, TX 75275-0381.

Education and Training in Autism and Developmental Disabilities, 2010, 45(1), 3–22© Division on Autism and Developmental Disabilities

Comprehensive Reading Intervention / 3

only on isolated subskills of reading, ratherthan on comprehensive reading interventionsthat integrate all essential components ofreading (defined below). Currently, no re-search has been conducted to determinewhether students with ID can learn to read byfully processing the print and meaning of con-nected text, as is consistent with current the-ories of reading development (see reviewsBrowder & Xin, 1998; Browder et al., 2006;Conners, 2003; Joseph & Seery, 2004).

In spite of the paucity of research, the re-search that does exist is promising, suggestingthat students with ID are capable of learningvarious aspects of reading. Sight word recog-nition has received the greatest attention fromresearchers and the preponderance of evi-dence demonstrates that students with evenmoderate and severe levels of ID can learn toautomatically recognize a fairly large corpusof words with systematic instruction (Browderet al., 2006). Even so, these students have littleability to generalize their learning beyond thespecific words included in instruction, andthus, are far from achieving even basic literacy(Browder et al.). Research on the effectivenessof phonics instruction is primarily limited tostudents with mild ID, but that research is alsopromising. Conners (1992) and Joseph andSeery (2004) found fourteen studies that ex-amined phonics instruction for students withID and these studies lend preliminary supportto the effectiveness of phonics interventions.Unfortunately, these studies were all relativelybrief, providing at most a few months of in-struction, and they focused on isolated sub-skills of phonics, rather than a comprehen-sive, systematic approach that might result inskilled decoding. Further, none of these stud-ies focused specifically on phonemic aware-ness (PA). In 1996, O’Connor, Notari-Syver-son, and Vadasy, described the progress ofseveral students with mild ID who participatedin a PA intervention study. Of the nine stu-dents with ID who participated in the six-month PA intervention, three made substan-tial progress. In a recent study, students withID receiving instruction for approximately 10weeks made significantly more progress onsounding out activities than a similar controlgroup (Conners, Rosenquist, Sligh, Atwell, &Kiser, 2006). Studies on vocabulary and com-prehension are even more limited, only in-

cluding demonstrations of very basic skills,such as using a sight word in the context of afunctional activity or matching a word to apicture (Browder et al.).

Taken in its totality, the research base onteaching students with ID to read is sparse andinadequate. At the present time, there are nostudies that have examined the effectivenessof a comprehensive reading intervention de-livered over a sustained period of time. With-out this type of research, we cannot determinewhether “all” as described in No Child LeftBehind should or should not include studentswith ID. In short, we simply do not know whatis possible for students with ID. The mission ofthe research reported here is to take impor-tant steps toward addressing this question.Specifically, we seek to determine what is pos-sible for teaching children with moderate lev-els of ID to read.

Conceptual Framework

While there is little research on reading toguide decision making for children with mod-erate ID, much research has been conductedwith other populations who also find learningto read very difficult, and thus, should informresearch on teaching students with ID to read.This research provides the conceptual frame-work for our study. As is consistent with cur-rent research, we see reading as an integratedprocess, rather than a set of isolated skills. Ina relatively simple view of reading, good read-ers effortlessly recognize words and buildmental representations of the message of thetext (Ehri, 2005; Perfetti, Landi, & Oakhill,2005). Studies examining the underlying pro-cesses of word recognition are clear. Goodreaders fully process print, attending to theinner structure of each word that is read (Ad-ams, 1990; Ehri, 2005; Ehri & McCormick,1998; Torgesen, 2002). They do this quicklyand effortlessly. Researchers believe goodreaders are able to focus attention on themeaning of print because word recognitionprocesses are automatized. The underlyingprocesses of comprehension are arguablymore complex, depending on a variety of fac-tors including listening comprehension, lin-guistic abilities, relevant knowledge, and gen-eral intelligence (Perfetti et al.). Specific towritten language understanding are factors in-

4 / Education and Training in Autism and Developmental Disabilities-March 2010

cluding sensitivity to story structure, inferencemaking, and comprehension monitoring(Perfetti et al.). We know that studentsprogress through predictable stages as theirword recognition and comprehension skillsdevelop (Chall, 1996; Ehri; Ehri & McCor-mick, 1998). In early stages students developphonological awareness and print awareness,along with expressive and receptive oral lan-guage skills. In later stages decoding and mor-phographic knowledge increases, eventuallyleading to the quick and effortless retrieval ofwords from long-term memory, enabling stu-dents to read fluently and, most importantly,focus on making sense of the message of text.Good readers make inferences and monitortheir own comprehension, ensuring that sto-ries and information are cohesive (Perfetti etal.).

Research on Early Reading Interventions

Over the past 30 years numerous studies fo-cused on the prevention and correction ofreading problems with students who struggleto learn to read who do not have ID. A pri-mary finding from this research is that inter-vention provided to small groups of childrenin the primary grades can be highly effectivein preventing reading problems for most chil-dren and greatly reducing the depth of read-ing problems for those who continue to expe-rience difficulty, (e.g., Foorman & Torgesen,2001; Mathes et al., 2005; Mathes & Denton,2002; Denton & Mathes, 2003; National Read-ing Panel, 2000; Snow, Burns, & Griffin,1998). Likewise, we now understand the criti-cal content students must acquire if they areto become competent readers. Effective inter-ventions in early reading target multiple com-ponents of the reading process in an inte-grated and comprehensive manner, includingconcepts of print, oral language, phonologicaland phonemic awareness (PA), letter knowl-edge, word recognition, fluency, and compre-hension (see Foorman & Torgesen, 2001; Na-tional Reading Panel, 2000; Pressley, 1998;Rayner, Foorman, Perfetti, Pesetsky, & Seiden-berg, 2001; Snow et al., 1998). Many experi-mental studies demonstrate that teaching PAresults in improved reading and spelling out-comes (see Ehri et al., 2001). Letter knowl-edge, including letter naming and letter-

sound recognition, is also an importantpredictor of reading achievement (Share,Jorm, Maclean, & Matthews, 1984; Adams,1990), and these skills influence other keyearly literacy skills, such as PA and phonemicdecoding (Blaiklock, 2004; Evans, Bell, Shaw,Moretti, & Page, 2006; Foy & Mann, 2006;Roberts, 2003; Treiman, Tincoff, & Rich-mond-Welty, 1996; Treiman, Tincoff, Rodri-guez, Mouzaki, & Francis, 1998). Many chil-dren who have difficulty learning to read alsostruggle with the development of good orallanguage skills (Perfetti et al., 2005). Onemethod demonstrated to be effective for stu-dents with language delays is interactive story-book reading (Arnold & Whitehurst, 1994;Dickinson & Smith, 1994; Karweit & Wasik,1996; Valdez-Menchaca & Whitehurst, 1992).Explicit instruction in basic comprehensionstrategies is also a critical component of suc-cessful early reading interventions (Mathes etal., 2005).

As we approach the task of intervening withchildren with moderate ID, it is important thatwe provide instruction that not only teachesthe critical content of reading, but also syn-thesizes what is known to be effective in teach-ing students with ID. For these students, abehavioral approach appears to be most ap-propriate (Adams & Engelmann, 1996; Brow-der et al., 2006; Joseph & Seery, 2004). Therole of the teacher in a behaviorist model is toexplicitly teach content and model skills, pro-viding systematic review of skills and reinforce-ment for mastery.

Purpose of the Study

The purpose of this study was to analyze theeffectiveness of a carefully crafted, compre-hensive reading intervention built on behav-ioral principles in teaching primary-grade stu-dents with moderate ID to read. Students inthis study participated in our intervention forone to one and a half years. Specifically, weimplemented and expanded an explicit, sys-tematic reading intervention that had beenempirically validated with students at-risk forlearning disabilities (Mathes et al., 2005) andwith students who are both struggling readersand English Language Learners (Vaughn,Mathes et al., 2006). This intervention, nowpublished as Early Interventions in Reading


(Mathes & Torgesen, 2005a), was (a) rootedin behavioral theory, (b) comprised of all ofthe content demonstrated to be critical forstruggling readers without ID, and (c) supple-mented with additional language develop-ment support.

This study adds to the literature in severalways. First, the reading intervention is com-prehensive in nature, with instruction target-ing oral language, phonemic awareness, al-phabetic knowledge, phonemic decoding,and basic comprehension strategies. Second,the intervention in this study includes manycomponents that have been previously vali-dated with students at-risk for reading failurewho have IQs in the average range, thus ex-tending that research to the population ofstudents with moderate ID. Third, the studyextends these techniques with the addition oforal language activities and modified teachingtechniques. Fourth, the study employs a lon-gitudinal, randomized trial design. In this ar-ticle, we report data collected after the stu-dents had been in the study for at least onefull academic year. Future reports will followthe students for approximately four academicyears. Finally, phonemic awareness and pho-nemic decoding were measured repeatedly al-lowing for the use of advanced statistical tech-niques. We addressed the following specificresearch question: Does a comprehensivereading program taught to primary-grade stu-dents who have moderate ID (IQs rangingfrom 40–55) result in better reading out-comes than typical special education instruc-tion on measures of (a) phonemic awareness,(b) alphabetic knowledge, (c) word recogni-tion/phonemic decoding, and (d) oral lan-guage/comprehension?

Previously Validated Intervention Components

The intervention included components previ-ously validated for students without ID. Thefirst, and most comprehensive, is Early Inter-ventions in Reading (Mathes & Torgesen,2005a; Mathes et al., 2005; Vaughn, Mathes etal., 2006). We also built upon oral languagestorybook techniques successfully used withEnglish Language Learners (Vaughn, Cirinoet al., 2006; Vaughn, Linan-Thompson,Mathes, Duradola, & Cardenas-Hagan, 2007).Finally, we used a simple game to provide

students with extensive modeling, practice,and feedback in phonemic awareness segmen-tation and blending, as well as the applicationof those skills to print (Allor, Gansle, &Denny, 2006). (See Method section for fur-ther details about the intervention.)

Method

Research Design

This study focused on students with moderateintellectual disabilities (i.e. IQs ranging from40-55) who were participants in a larger, lon-gitudinal study examining the effectiveness ofa comprehensive reading program for stu-dents with low IQs (ranging from 40–79: Al-lor, Roberts, Mathes, Roid, & Cheatham,2010). Students were randomly assignedwithin each school to either (a) an interven-tion group that participated in daily, smallgroup reading instruction delivered by re-search teachers or (b) a contrast group receiv-ing typical special education.

Participants

Schools. The study took place in 10 ele-mentary schools in a large, southwestern ur-ban school district and one private school forstudents with special needs. District personnelworked with the researchers to select schoolswith a relatively large number of students withID and that would provide a balanced sample,racially and economically. An urban, privateschool that served students with special needswas added to increase the size of our sample ofstudents with moderate ID.

Teachers. Six certified special educationteachers were hired to provide instruction tostudents in the research study. The highestdegree held by five of the teachers was a bach-elor’s degree, while one teacher also held amaster’s degree. Five were female and one wasmale. Five were Caucasian and one was Afri-can American. Five were jointly hired and su-pervised by district personnel and researchers,teaching at two or three different schools eachday. One taught exclusively at the privateschool. At the outset of the study, two werenew to teaching and the others had 5, 9, 12,and 35 years of teaching experience, respec-tively. Three of the teachers had prior experi-


ence working with students with reading diffi-culties, one had prior experience withstudents with behavioral disorders, one wasbilingual, and one had prior experienceteaching students with ID.

Students. At the outset of the study, re-searchers and school district personnel iden-tified all students in each of the schools withmoderate ID (IQ scores between 40 and 55)and who were in grades 1 to 4. All students inthis IQ range were included regardless of thecause or comorbid conditions (i.e., Down Syn-drome, autism, William’s Syndrome, physicaldisability, etc.). Students were randomly as-signed within each school into either the in-tervention group or the contrast group. Dueto the small number of students within eachschool, students were not matched on othervariables. Twenty-four students began thestudy in the first year and another seven stu-dents joined the study at the beginning of thesecond year (these seven were also randomlyassigned to the treatment or contrast group).Of these 31 students, two moved during thestudy and one was removed from the sampledue to misidentification, resulting in a sampleof 28 students (treatment, n � 16; contrast,n � 12). The mean age of the participants was9.46 (SD � 1.19) for the treatment group and

9.25 (SD � 1.76) for the contrast group. Thisdifference was not significant (t � �.106).Other demographic information is presentedin Table 1. Chi-square analyses revealed nosignificant differences on any demographicvariables, including race, gender, socioeco-nomic status, and educational placement.

Measures

We employed two types of measurementschemes. First, we assessed at pretest and post-test. Second, we collected continuous progressmonitoring data every four weeks during thefirst year of the intervention and every sixweeks during the second year.

Pre-post

All students were assessed prior to the inter-vention and at the end of the Spring semesterof the second year. Pretesting during the firstyear occurred between October and Februaryon a staggered schedule with students in thetreatment and contrast groups tested at ap-proximately the same time. The 7 studentswho entered the study in the second year werepretested in August or September of that year.

TABLE 1

Student Demographic Data by Group

VariableTreatment (n � 16)

n (%)Contrast (n � 12)

n (%) �2 (df)a

Gender 1.77 (1)M 14 (88%) 8 (67%)F 2 (12%) 4 (33%)

Race .58 (3)Caucasian 4 (25%) 4 (33%)African American 9 (56%) 5 (42%)Hispanic 2 (13%) 2 (17%)Other 1 (6%) 1 (8%)

Free Lunch Program ParticipationFree 5 (31%) 3 (25%) .15 (2)None 10 (63%) 8 (67%)Unknown 1 (6%) 1 (8%)Special Education Placement 1.39 (2)

Self contained class for students with ID 12 (75%) 11 (92%)Self contained class for students with autism 1 (6%) 0 (0%)General education w/ resource 3 (19%) 1 (8%)

a No differences were statistically significant.


The following measures comprised the com-prehensive battery:

Peabody Picture Vocabulary Test-III (PPVT-III;Dunn & Dunn, 1997). The PPVT-III mea-sures receptive vocabulary. The technicalmanual reports reliability coefficients rangingfrom .91 to .98 and adequate content, crite-rion, and construct validity.

The Expressive Vocabulary Test (EVT; Williams,1997). The EVT measures expressive vocab-ulary. The technical manual reports internalreliability alphas ranging from .90 to .98 witha median of .95 and test-retest reliability coef-ficients range from .77 to .90. Data on con-tent, criterion, clinical and construct validityare reported in the technical manual.

The Woodcock Language Proficiency Battery-Revised. (WLPB-R; Woodcock, 1991). We in-cluded memory for sentences and listeningcomprehension from the language composite.We included letter-word identification (realword reading), word attack (nonsense wordreading), and passage comprehension fromthe reading composite. The WLPB-R has goodreliability (internal consistency ranged from.81–.92; test-retest ranged from .75 to .95).Adequate content, concurrent, predictive,and construct validity data are also reported inits technical manual.

The Comprehensive Test of Phonological Process-ing (CTOPP; Wagner, Torgesen, & Rashotte,1999). Five subtests of the CTOPP wereused: Blending Words, Blending Nonwords,Segmenting Words, Sound Matching (firstsound and last sound), and Rapid Letter Nam-ing. The CTOPP has good reliability (internalconsistency ranged from .83 to .95; test-retestranged from .70 to .92). Adequate content,concurrent, predictive, and construct validitydata are also reported in its technical manual.

Test of Word Reading Efficiency (TOWRE:Torgesen, Wagner, & Rashotte, 1999). Bothsubtests were administered: phonemic decod-ing efficiency and sight word efficiency.Reliability coefficients are .95 and .96, respec-tively. Data on content-description, concur-rent, construct identification, and item valid-ity are reported in its technical manual.

Continuous Progress Monitoring

In order to assess progress continuously acrossa school year, we used Dynamic Indicators of

Basic Early Literacy Skills (DIBELS; Good &Kaminski, 2002). DIBELS measures are com-monly used for collecting continuous progressmonitoring data. We administered 4 subtests:Initial Sound Fluency (ISF), Phoneme Seg-mentation Fluency (PSF), and NonsenseWord Fluency (NWF). ISF, PSF, and NWF re-liability coefficients range from .72 to .92 onsingle probes and .91 to .98 on the means ofmultiple probes (3 to 5 probes). Concurrentand predictive validity with a variety of readingtests ranges from .36 to .82. In addition, theLetter Naming Fluency (LNF) test was given atpre and posttest. The alternative form reliabil-ity coefficient for LNF was .88. Validity coeffi-cients for this measure ranged from .65 to .71.

Intervention Overview

Students in the intervention condition re-ceived approximately 40 to 50 minutes of in-struction daily in small groups of one to fourfrom one of our six highly trained and sup-ported intervention teachers across the dura-tion of the study. The intervention was com-prehensive, including systematic and explicitinstruction in multiple content strands (i.e.concepts of print, phonological and phone-mic awareness, oral language, letter knowl-edge, word recognition, vocabulary, fluency,and comprehension) woven together so skillsand strategies were integrated and applied incontext.

The intervention built on a curriculum pre-viously validated with students without ID,Early Interventions in Reading (Mathes &Torgesen, 2005a, 2005b), which was com-prised of 240 lessons split evenly into two lev-els. However, students in the current researchdid not possess the prerequisite skills neces-sary to profit from this curriculum. Thus, wecreated an additional 60 lessons we called theFoundation Level (Allor, Mathes, & Jones,2010). An oral language component was alsodeveloped and included in both the Founda-tion Level and Level One. In total, 300 lessonshave been designed to take students from be-ing nonreaders with very little or no letterknowledge or phonological awareness to read-ing at approximately an ending 3rd gradereading level. Since no students had yet begunLevel Two at the time of the article, only details


about the Foundation Level and Level One areprovided here.

Based on pre-test DIBELS scores, 13 stu-dents began the intervention in the FoundationLevel and three began in Level One. The threestudents beginning in Level One were amongthe older students in the study (2 third grad-ers and 1 fourth grader). Groups were deter-mined by DIBELS pretest scores as well asother practical considerations. Thus, two ofthe 16 students were taught individually, whileothers were taught in groups of two to four.Grouping arrangements changed as needed,based on rate of progress.

Instructional Design and Features

All of the lessons in the program were fully-specified and employed the principles of Di-rect Instruction (Carnine, Silbert, Kame’enui,& Tarver, 2004; Coyne, Kame’enui, & Sim-mons, 2001; Englemann, 1997; Englemann &Carnine, 1982; Kame’enui & Simmons, 1990).We chose this model of instruction because ofits long standing record of success with variouspopulations at-risk for school failure (Adams& Engelmann, 1996; Borman, Hewes, Over-man, & Brown, 2003; Carlson & Francis, 2002;Ligas, 2002). Instructional content was care-fully analyzed and organized into a systematicscope and sequence intended to reduce stu-dent confusion and target big ideas and keystrategies. Errors were reduced through inte-gration of new learning with previous learn-ing, ongoing review, and opportunities forgroup and individual responding. The goalwas to integrate skills and strategies over time,resulting in a set of daily lesson plans withoverlapping content strands and extensive cu-mulative review and application (i.e., conceptsof print, phonological and phonemic aware-ness, letter knowledge, word recognition,connected text fluency, comprehension strat-egies, vocabulary, and oral language develop-ment). Following a behavioral approach, les-sons provided for (a) frequent reinforcementon both an interval and intermittent schedule,(b) carefully orchestrated time delay tech-niques between stimuli presentation and stu-dent responses, and (c) multiple opportuni-ties to practice each item of content.

Each lesson plan was highly detailed, pro-viding exact wording to ensure teacher lan-

guage was clear and kept to a minimum. Byfollowing these plans, teachers delivered ex-plicit instruction in integrated instructionalstrands, responding to individual studentlearning needs by scaffolding instructionwhen necessary. Thus, while lesson plans wereprescribed, the way in which lessons were ac-tually delivered required teachers to make onthe spot decisions and minor adjustments inthe plans in order to focus on specific targetareas needed by students within a group. Ac-companying these lesson plans, teachers wereprovided storybooks for read-alouds, picturesfor vocabulary support, student activity books,magnetic pictures (Foundation Level only),daily reading books using decodable stories(Level One only), a puppet with a fully articu-lated mouth, letter-sound picture cards, “au-tomatic” word cards, and lesson mastery track-ing forms. Additionally, the Foundation Levelincluded a game designed to provide studentswith opportunities to practice the PA skills ofblending, segmenting, and letter-sound corre-spondence (See Allor et al., 2006 for details).

Instructional Strands

Concepts of print. During the FoundationLevel, students developed various concepts ofprint. These included pointing to the title andauthor of a book, tracking from left to right,and pointing to individual words while repeat-ing a sentence.

Phonological and phonemic awareness. Activ-ities in this strand spanned the FoundationLevel and Level One and addressed skills alongthe continuum of phonological and phone-mic awareness, including clapping words insentences, clapping syllables within a multi-syllabic word, initial sound isolation, pho-neme segmentation, phoneme blending, andphoneme discrimination. Over time the com-plexity of words included in segmentation andblending activities increased.

Letter knowledge. In this strand, studentslearned letter names and the sounds of indi-vidual letters and letter combinations, as wellas worked on speeded retrieval (i.e., rapidautomatic naming tasks). Starting in the 21st

lesson of the Foundation Level, students weretaught to map phonemes to letters, with newletter-sound correspondences introduced ev-


ery few days and followed by daily cumulativereview.

Word recognition. This strand includedboth phonetically regular and irregular words.Toward the end of the Foundation Level, stu-dents were taught a small number of sightwords; these words were high-frequency, pho-netically irregular words presented as trickywords to be recognized automatically. Stu-dents were also taught to decode very simplephonetically regular words (i.e. closed sylla-ble, consonant-vowel-consonant: CVC) byblending the sounds represented by the let-ters. As students progressed through LevelOne, additional sight words were taught andthe time allowed to sound out the words wasreduced, while the complexity of the wordswas increased (i.e. variant spelling patterns,blends, additional syllable types, and multisyl-labic words). Students were also taught to beflexible decoders.

Fluency with connected text. Beginning veryearly in Level One, word recognition strategieswere applied as students read decodable sto-ries. As students acquired greater mastery ofmore elements, as well as the ability to decodemore difficult words, this text became morechallenging. To promote fluency, repeatedreading of these stories was built into dailylessons. Typically students read a story in uni-son on the first reading, followed by reading apage or two individually on the second read-ing. The third reading was typically read inpairs, with the teacher timing the reading rateof one student.

Comprehension strategies. A major objectivewas for children to read strategically to in-crease understanding. Thus, prior to readinga story, students “browsed the story” looking atthe pictures and predicting story content. Stu-dents then read to find out if their predictionswere true. With expository text, teachers acti-vated prior knowledge by asking students totell what they already knew about the topicand to read to learn more. After reading thestory, students then engaged in a number ofactivities depending on the students’ compe-tence and text structure. Initially, studentswere only asked to tell about what they read.Information in any order was accepted. Overtime, students sequenced information untilthey were able to sequence only the most im-portant information. In later lessons, students

identified story grammar elements for narra-tive text and new information learned in ex-pository text.

Vocabulary and oral language development.Language goals were addressed through sto-rybook read-alouds, with direct teaching ofspoken vocabulary and key backgroundknowledge, as well as extensive discussion. Inthe Foundation Level, teachers explicitly taughtvocabulary and engaged students in conversa-tion using open-ended questions and buildingon student language (Arnold & Whitehurst,1994). When students began Level One, theStorybook routine became more complex,with books organized into themes to facilitatevocabulary and concept review. One book wasread from and discussed for 3 to 5 days, withtwo to three new vocabulary words taughteach day. Students listened for and discussedthe “target words” during the reading of thestory. After the passage was read aloud, stu-dents provided an oral retell and discussedthe story with the teacher who employed feed-back and scaffolding to encourage the use ofcomplete sentences and new vocabularyterms.

Staff Development

During the first year of the intervention, theteachers attended a total of six days of trainingon the intervention, four at the beginning ofthe school year and two later in the schoolyear. Teachers were visited by two experi-enced reading coaches every other month toaddress their individual needs and the needsof their students. The coaches were formerteachers who had previously taught the EarlyInterventions in Reading (Mathes & Torgesen,2005a) curriculum under similar researchconditions. Teachers also attended threemeetings with the entire research team, in-cluding the coaches and lead research inves-tigators who had created the curriculum.

During the second year, teachers partici-pated in three days of training, two days at thebeginning of the school year and one day inthe middle of the school year. The number ofcoaching visits was reduced to two per semes-ter. Research team meetings with the teacherswere increased in frequency to once permonth and focused on using student data to


make instructional decisions, including bothacademic and behavioral modifications.

Implementation Fidelity and Intensity

Three fidelity observations were conductedeach year to measure the degree to which theintervention was implemented. After each ob-servation, the research assistants shared feed-back with teachers. A 3-point rating scale wasused to evaluate the fidelity of implementa-tion across several categories including teach-ing to mastery, maintaining a good pace,maintaining student attention, and providingerror correction and scaffolding. A score of 3indicated that the teacher implemented thecategory exactly as intended. A score of 2 in-dicated that the category was implementedacceptably but with some error. A score of 1indicated that the category was poorly repre-sented. A score of 0 indicated that the behav-ior was expected but not observed. The mea-sure included a global checklist for readinessof materials, appropriate seating arrange-ment, and instructor warmth and enthusiasm.Interrater agreement was calculated and ex-ceeded 85%. Averaged across six fidelity ob-

servations, teachers’ scores ranged from 2.29to 2.96 out of 3 with a mean of 2.75 (SD �0.25). The mean, calculated as a percentagescore, was 90.9% (SD � 8.63).

Total instructional time for each studentvaried depending on when they began theintervention and attendance. As a result, in-struction for the students varied from 30 to 53weeks, with a mean of 42.8 weeks (SD �10.34). The average length of an instructionalsession was 40 minutes (SD � 6). Studentsparticipated in an average of 119 (SD � 11)instructional sessions during the study.

Results

Pretest Equivalence

Pretest data were analyzed using independentt-tests. These indicated no statistically signifi-cant differences between the treatment andcontrast groups on any pretest measure. Pre-test equivalency data are presented in Table 2.

Growth from Pretest to Posttest

Independent t-tests on difference scores of thepretest and posttest measures were conducted

TABLE 2

Pretest Equivalencies

Measure

Treatment (n � 16) Contrast (n � 12)

t (1, 27)nsM SD M SD

CTOPPBlending Words 2.00 3.50 .92 1.98 �.96Blending Nonwords .88 1.78 .83 1.64 �.86Segmenting Words .37 1.50 .00 .00 �.86Sound Matching 2.00 2.45 1.50 4.30 �.36

EVT 34.94 13.28 30.08 13.85 �.94PPVT 40.81 21.70 33.92 19.52 �.87TOWRE

Sight Word Efficiency 2.69 5.65 6.08 16.26 .78Phonemic Decoding Efficiency .38 1.50 2.42 8.37 .84

WLPBMemory for Sentences 23.44 10.02 21.17 8.27 �.64Listening Comprehension 2.81 3.25 1.25 1.55 �1.54Letter-Word Identification 12.25 7.61 11.25 7.42 �.35Passage Comprehension 2.62 3.30 1.58 2.28 �.94Word Attack 1.00 2.85 .75 2.30 �.25

ns No significant differences found at .05 level on any measure


TA

BL

E3

Gro

wth

onP

rete

stto

Pos

ttes

tM

easu

res

Mea

sure

Tre

atm

ent

(n�

16)

Con

tras

t(n

�12

)

tEf

fect

Size

Pret

est

Post

test

Diff

eren

cePr

etes

tPo

stte

stD

iffer

ence

MSD

MSD

MSD

MSD

MSD

MSD

CT

OPP

Ble

ndi

ng

Wor

ds2.

003.

504.

444.

942.

444.

66.9

21.

981.

251.

87.3

31.

61�

1.68

.57

Ble

ndi

ng

Non

wor

ds.8

81.

782.

312.

941.

442.

75.8

31.

64.3

3.8

9�

.50

1.24

�2.

49.8

7*Se

gmen

tin

gW

ords

.37

1.50

3.06

3.87

2.69

3.75

.00

.00

.17

.58

.17

.58

�2.

64.8

8*So

und

Mat

chin

g2.

002.

454.

693.

932.

692.

911.

504.

302.

332.

81.8

32.

44�

1.78

.68

EV

T34

.94

13.2

842

.38

11.5

17.

449.

3130

.08

13.8

534

.42

11.8

44.

337.

70�

.94

.36

PPV

T40

.81

21.7

050

.81

25.3

810

.00

10.2

233

.92

19.5

233

.25

14.1

8�

.67

19.6

4�

1.87

.71*

TO

WR

ESi

ght

Wor

dE

ffici

ency

2.69

5.65

11.3

811

.12

8.69

9.45

6.08

16.2

68.

3316

.38

2.25

8.35

�1.

87.7

2*Ph

onem

icD

ecod

ing

Effi

cien

cy.3

81.

505.

006.

534.

635.

382.

428.

371.

754.

07�

.67

5.23

�2.

611.

00*

WL

PB-R

Mem

ory

for

Sen

ten

ces

23.4

410

.02

27.8

87.

224.

447.

6621

.17

3.27

22.0

07.

99.8

35.

87�

1.36

.52

Lis

ten

ing

Com

preh

ensi

on2.

813.

255.

446.

112.

634.

241.

251.

552.

172.

89.9

22.

35�

1.25

.48

Let

ter-

Wor

dId

enti

fica

tion

12.2

57.

6118

.75

7.72

6.50

4.05

11.2

57.

4214

.00

8.11

2.75

3.38

�2.

60.9

9*Pa

ssag

eC

ompr

ehen

sion

2.62

3.30

5.31

3.93

2.69

2.63

1.58

2.28

2.67

2.93

1.08

1.83

�1.

81.6

9*W

ord

Att

ack

1.00

2.85

2.94

2.82

1.94

2.49

.75

2.30

1.08

2.94

.33

.78

�2.

15.8

2*

*p

�.0

5


to determine whether students in the treat-ment condition made greater gains than stu-dents in the contrast condition. Because ofpositive results in previous studies with EarlyInterventions in Reading, we anticipated the di-rectionality of any differences (Mathes et al.,2005) and, therefore, we analyzed the datausing a one-tailed test of the null hypothesis(Gall, Gall, & Borg, 2007). T-test and effectsize results are presented in Table 3. Statisti-cally significant results were found on the fol-lowing measures: CTOPP Blending Non-words, CTOPP Segmenting Words, CTOPPSound Matching, PPVT, TOWRE Sight WordEfficiency, TOWRE Phonemic DecodingEfficiency, WLPB-R Letter-Word Identifica-tion, WLPB-R Passage Comprehension, andWLPB-R Word Attack. No statistically signifi-cant differences were found on CTOPPBlending Words, EVT, WLPB-R memory forsentences, and WLPB-R listening comprehen-sion, although all means favored the treat-ment group and effect sizes were moderate tostrong.

We also applied the Bonferroni correctionprocedure because we employed multiple, re-lated measures of various reading constructs.This adjustment was made to help control forType I error (Dunn, 1961). We adjusted ourcritical p value by dividing .05 by the numberof measures in a given construct, i.e. phone-mic awareness, phonemic decoding, real wordrecognition, reading comprehension, andoral language measures. After making this cor-rection, differences on PPVT, TOWRE SightWord Efficiency, and WLPB-R Passage Com-prehension were no longer statistically signif-icant. Other findings remained the same. Ad-ditionally, Analysis of Covariance tests wereconducted on the gain scores using pretestmeasures as covariates. However, results werevery similar to the t-test analyses, includingsignificant findings on all of the same mea-sures, as well as significant findings on CTOPPBlending Words, CTOPP Sound Matching,and WLPB-R Memory for Sentences. There-fore, these results were not included.

Growth on Continuous Progress MonitoringMeasures

We used a hierarchical linear modeling(HLM) approach to examine student gains on

the three DIBELS measures: initial sound flu-ency (ISF), phoneme segmentation fluency(PSF), and nonsense word fluency (NWF).The advantage of HLM over simple regres-sion, ANOVA, or repeated measures ANOVAis that it allows the researcher to look at hier-archically structured data and interpret resultswithout ignoring these structures. This is ac-complished by including a complex randompart that can appropriately account for com-plex covariance structure in the data (Roberts,2004). In the present analysis, a two-levelmodel was examined with measurement occa-sions at level-1 and students at level-2. Previ-ously, only t-tests were conducted on gainscores to note differences between the treat-ment and contrast groups. These initial anal-yses were not performed in a HLM environ-ment because of a lack of power in the HLMdesign. Therefore, independent samplest-tests were used because they were more ap-propriate and more parsimonious. With thecontinuous progress monitoring data, HLMgrowth curve analysis is more appropriate as itcan model complex covariance structures andeffectively model explanatory variables thatare able to mediate the changes in growthfrom student to student.

The HLM model investigated differencesamong the success of the intervention for stu-dents with moderate ID (IQ scores between 40and 55). HLM combines the strength of sim-ple ANOVA (mean difference analysis) andregression (correlational analysis) to build amodel that both considers differences acrossstudents (the second-level or student-level)and incorporates a correlational componentfor each of these students (the time covari-ate). This assumption is fundamental to ouranalysis since we are hypothesizing that be-longing to the intervention group and thecross-level interaction between time and inter-vention (a student-level variable) will have animpact on growth.

For each of the three dependent variables,PSF, ISF, and NWF, two models were tested.Because these models included interaction ef-fects, it was important to code time with ameaningful zero (Hox, 2002). Therefore,time was centered with zero being the day thata student began the program and increasingnumbers representing the number of weeksthe student was involved in the intervention.


The first model included only the time covari-ate with random effects for the intercept andfor the time variable and was presented as:

yti � �00 � �10 * time � u0i � u1i * time � eti,

(1)

where yti is the dependent variable for thatmodel, �00 is the average fluency at time � 0,�10 is the average student increase in fluencyfor each week since the program began, u0i isthe random effect of �00, u1i is the randomeffect for �10, and eti is the random effect ofthe measurement occasions within individualstudents.

The second model included a student-level(level-2) effect to identify whether or not thestudent was in the intervention or controlgroup. In this model, the intercept �00 has aslightly different interpretation than in thefirst model as it now represents the averagefluency for a student in the control group attime�0 (the intervention group receives a “1”for the level-2 grouping variable). This newmodel is:

yti � �00 � �10 � time � �01 � group

� �11 � time � group � u0i � u1i�time � eti,

(2)

where �01 is the effect of a student belong-

ing to the intervention group at time � 0 and�11 represents the cross-level interaction be-tween time and the intervention effect.

Hox (2002) has noted that it is typical toinclude both of the main effects in a model inthe presence of a statistically significant inter-action effect. As can be seen from the resultsof our three models in Tables 4–6, the modelstructures are the same across all four analy-ses. The only change between each analysiswas the dependent variable.

The analysis in Table 4 represents the effectof the intervention on ISF across time forstudents. There was no statistical differencebetween the intervention and contrast groupsat the initial time-point (�1.026, p � 0.711).The interaction effect tested to see whether ornot the amount of difference between theintervention group and contrast groupchanged over time. For example, a large pos-itive value for �11 would mean that studentsinvolved in the intervention tended to havelarger gains in ISF over the contrast groupstudents the longer they were involved in theintervention. In this analysis, however, thevalue for this interaction (0.167) was not sta-tistically significant over time (p � 0.058) in-dicating that students in the intervention andcontrol groups tended to have the same rateof change over time.

TABLE 4

Model Fit Estimates for Initial Sound Fluency with Students with IQs in the Moderate Range

M0: Null model M1 : � group & interaction

estimate s.e. p-value estimate s. e. p-value

Fixed Effects:Intercept �00 5.000 1.326 � 0.001 5.579 2.060 0.007Time �10 0.238 0.046 � 0.001 0.142 0.066 0.033Group �01 �1.026 2.735 0.711Time*Group �11 0.167 0.088 0.058

Random Effects:�2

e 30.706 30.746�2

u0 38.957 41.130�2

u1 0.040 0.034COV (u0.u1) 0.017 0.041

Fit:AIC 1678.826 1678.642BIC 1699.882 1706.652�2 1666.826 1662.642


Table 5 shows the effect of the interventionon PSF across time for students. As can beseen from this analysis, there was no statisticaldifference between the intervention and con-trast groups at the initial time-point (�0.199,p � 0.927), thus indicating that they werestatistically equivalent in terms of PSF whenthe program began. Also in model M1, the

value for the interaction effect (0.417) wasstatistically significant over time (p � .001)thus indicating that students in the interven-tion group tended to have a larger rate ofgrowth in PSF over time than did the studentsin the contrast group.

Table 6 shows the effect of the interventionon students’ NWF across time. Again, there

TABLE 5

Model Fit Estimates for Phoneme Segmentation Fluency with Students with IQs in the Moderate Range



Fixed Effects:Intercept �00 0.681 1.032 0.509 0.883 1.608 0.584Time �10 0.369 0.066 � .001 0.124 0.081 0.130Group �01 �0.199 2.139 0.927Time*Group �11 0.417 0.108 � .001

Random Effects:�2

e 27.819 27.832�2

u0 20.332 22.102�2

u1 0.107 0.062COV (u0.u1) 0.231 0.247

Fit:AIC 1666.922 1657.986BIC 1687.979 1685.996�2 1654.922 1641.986

TABLE 6

Model Fit Estimates for Nonsense Word Fluency with Students with IQs in the Moderate Range



Fixed Effects:Intercept �00 2.786 1.794 0.122 4.845 2.668 0.071Time �10 0.327 0.065 �0.001 0.140 0.085 0.102Group �01 �3.725 3.586 0.309Time*Group �11 0.337 0.114 0.003

Random Effects:�2

e 37.252 37.266�2

u0 73.907 73.004�2

u1 0.088 0.064COV (u0.u1) 0.455 0.828

Fit:AIC 1673.274 1666.205BIC 1694.057 1693.847�2 1661.274 1650.205


was no statistical difference between the inter-vention and contrast groups at the initial time-point (�3.725, p � 0.309), thus indicatingthat they were statistically equivalent in termsof NWF when the program began. Also inmodel M1, the value for the interaction effect(0.337) was statistically significant over time(p � .003) thus indicating that students in theintervention group tended to have a largerrate of growth in NWF over time than did thestudents in the contrast group. It should benoted that one student in the treatment groupwas excluded from this analysis because thestudent began the study above benchmarkand maintained scores above benchmark.

Graphs of scores for individual students onPSF and NWF are presented in Figures 1 and2. The graphs on the left show the scores forthe 12 students in the contrast group; thegraphs on the right show the scores for the 16students in the treatment group.

Discussion

In this article, we report the results from alongitudinal study examining the effectivenessof a comprehensive early literacy interventionfor students with moderate intellectual disabil-ities (ID). This article reports on studentprogress after participating in the interven-tion, or typical special education instruction,for one to one and a half years. The purposeof the study is to determine if students partic-ipating in the intervention make significantlymore progress on a variety of reading andlanguage measures than similar students par-ticipating in typical special education. Ouroutcomes strongly support the effectiveness ofthe intervention with students with moderateID. These findings are discussed in detail be-low.

Research Question: Does a comprehensive readingprogram taught to primary-grade students who havemoderate ID (IQs ranging from 40-55) result inbetter early reading outcomes than typical specialeducation instruction on measures of (a) phonemicawareness (PA), (b) phonics, (c) word recognition,and (d) oral language/comprehension? The an-swer to this question is clearly yes. On alloutcome measures, means favored the inter-vention group, with moderate to strong ESs onall measures. Despite low statistical power dueto the small sample size (16 in the interven-

tion and 12 in the contrast group), statisticallysignificant differences were found on multiplemeasures, including measures of phonemicawareness, phonics, word recognition, vocab-ulary, and comprehension.

The clearest, and arguably one of the mostimportant findings in the study were on mea-sures of phonemic awareness (PA). Studentsparticipating in the intervention consistentlyoutperformed students in the contrast groupon measures of PA. Effect sizes on the fourCTOPP subtests (gains from pretest to post-test) ranged from a moderate effect of .57 to astrong effect of .88. The differences on bothBlending Nonwords and Segmenting Wordswere statistically significant (See Table 3). Ad-ditionally, these differences remained signifi-cant after the Bonferroni correction. Al-though differences on CTOPP BlendingWords and Sound Matching were not statisti-cally significant, effect sizes were moderate tostrong (.57 and .68, respectively). Results fromthe HLM analysis also revealed that the stu-dents in the intervention group tended tohave a higher rate of growth on DIBELS-PSFover time, with this interaction statistically sig-nificant (p � .001; see Table 5 and Figure 1).This finding is particularly compelling be-cause it indicates that students in the interven-tion group consistently outperformed stu-dents in the contrast group over a long periodof time. Unexpectedly, the same pattern ofresults was not evident on DIBELS-ISF, butthis was likely because students with ID foundthe language and cognitive demands of thetask challenging, preventing them from dem-onstrating their ability to isolate phonemes.

Consistent differences in favor of the treat-ment group were also evident on multiplemeasures of alphabetic decoding. Effect sizeson nonsense word reading measures, TOWREphonemic decoding and WLPB word attack,were 1.0 and .66, respectively, with statisticallysignificant differences on the former measure.These differences remained significant afterthe Bonferroni correction procedure. Addi-tionally, HLM analyses of NWF measuresacross time revealed a statistically significantinteraction in favor of the treatment group (p�.003; See Table 6 and Figure 2).

Data also indicate that students in the treat-ment group consistently made more growthon word recognition (i.e., real word reading)


than students in the contrast group. Two mea-sures directly assessed this skill, TOWRE SightWord Efficiency and WLPB-R Letter-WordIdentification. Differences on these measureswere statistically significant and ESs werestrong (.72 on TOWRE-word reading effi-ciency and .99 on WLPB-R Letter-Word Iden-tification).

Outcomes for oral language and compre-hension were also positive. Effect sizes were

moderate to strong on language measures,ranging from .36 to .71, with significant differ-ences on receptive vocabulary (PPVT). Differ-ences on passage comprehension were statis-tically significant. (See Table 3.)

Conclusions

This study provides clear support for raisingexpectations related to reading for students

Figure 1. Individual Graphs on Phoneme Segmentation Fluency.

Figure 2. Individual Graphs on Nonsense Word Fluency.


with moderate ID. Students with moderate IDshould not be left behind; they should beprovided with scientifically-based reading in-struction. The findings of this study stronglysupport the conclusion that students withmoderate IDs can make important gains inreading and language skills when providedwith intensive and comprehensive instructionover an extended period of time. A broadarray of measures was studied, including PA,phonics, word recognition, comprehension,and oral language. ESs on all measures weremoderate to strong, with means consistentlyfavoring the intervention group. Statisticallysignificant differences were found on multiplemeasures, including phonemic awareness,phonics, word recognition, and comprehen-sion. These findings are consistent with exist-ing research and extend that research in sev-eral ways.

First, explicit, systematic instruction in PAand phonics that has proven to be effective forstudents with IQs in the average range (Ehri etal., 2001; Mathes et al., 2005) is also effectivefor students with moderate ID. Prior researchon teaching PA and phonics to students withID focused on those with mild ID and waslimited to relatively brief instructional periodstargeting isolated skills (Joseph & Seery, 2004;O’Connor et al., 1996). The current studydemonstrates that with an integrated and sys-tematic approach, students with moderate IDcan successfully combine isolated skills in PAand phonics to decode unfamiliar words.

Second, this study is consistent with previ-ous research demonstrating the effectivenessof systematic approaches in improving sightword recognition (Browder et al., 2006). Inthis study sight word instruction was one com-ponent of the comprehensive reading pro-gram implemented. Effect sizes on measuresof sight word recognition were high and dif-ferences between the treatment and contrastgroups were statistically significant.

Third, we found that a comprehensive read-ing intervention can positively impact oral lan-guage and comprehension. With moderateESs on oral language measures and strong,statistically significant differences on readingcomprehension and receptive vocabulary, thecurrent study extends previous research thathad demonstrated only very basic, isolatedcomprehension skills (Browder, 2006). As is

similar in research with students without ID, itis likely that gains in comprehension arestrongly influenced by gains in word recogni-tion. It appeared that the students in the treat-ment group were able to identify more wordsthan the students in the contrast group, en-abling them to answer a few basic comprehen-sion items on the standardized measure.

Fourth, the longitudinal design of this studyprovides information about the level of read-ing performance that can be expected afterone to one and a half years of consistent in-struction in a comprehensive reading pro-gram. Eight of the 16 students in the treat-ment group were approximately halfwaythrough Level One or further. At this level,students were able to identify the most com-mon sound for all individual letters and readwords made up of those letters. For example,students were able to successfully say thesounds in words such as last, mom, slip, andstep, as well as blend those sounds together toform the word. Further, students at this levelwere working on basic comprehension strate-gies, such as retelling stories, sequencing mainevents, and story grammar. Generally, stu-dents in this study took approximately twicethe amount of time to successfully completelessons than struggling readers in previousstudies. Further, a closer look at the graphs inFigures 1 and 2 reveals that gains on DIBELSmeasures of PA and phonics (PSF and NWF)were typically not evident until students hadbeen participating in the intervention for ap-proximately 15 to 20 weeks of instruction. Thetime needed to evidence gain was muchlonger in duration than is typical of strugglingreaders without ID. Thus, while the content ofinstruction for both groups is the same, whatdifferentiates them is the persistence neededon the part of schools to provide this instruc-tion.

Practical Implications

The findings of this study have importantpractical implications for educators in thefield of intellectual disabilities. First, and mostimportantly, our findings support educatorswho choose to provide reading instructionthat is comprehensive and not limited to sightword memorization, even with students withIQs in the moderate range. Second, reading


programs should be selected that are consis-tent with the techniques of the interventiondescribed in this study, including (a) system-atic, explicit instruction in all components ofreading; (b) repetitive, routine activities im-plemented with consistent instructional lan-guage; and (c) fast-paced, short activities thatare highly motivating. Third, to be effectivewith students with ID, programs must be im-plemented with extremely high degrees of fi-delity. This requires initial and ongoing pro-fessional development. Fourth, practitionersneed to make data-based decisions about howto modify instruction and provide positive be-havioral support. In addition to using exist-ing progress monitoring measures, such asDIBELS, observation of student performanceduring lessons and other informal measures iskey to making appropriate decisions.

Limitations and Future Research

One limitation of the current study is thevariability of student performance on out-come measures, as is common among stu-dents with IDs. We met the challenge of elic-iting optimal performance from our studentson study measures by ensuring familiarity ofexaminers and discontinuing testing whennecessary. We also addressed this issue by in-cluding repeated measures across time, whenpossible. This enabled us to employ data ana-lytic techniques (i.e., HLM) that analyzedtrends across time and minimized the impactof variability of the data. Due to this limita-tion, findings related to measures only admin-istered at pretest and posttest should be inter-preted cautiously. Further research is neededto develop reading and language tests that userepeated measures of progress, especially un-timed measures as existing repeated measuresare usually timed.

Another limitation of the study is the smallsample size. This is a common problem whenstudying low-incidence populations because itis logistically challenging and resource-inten-sive to increase sample size with a low-inci-dence population. Even after carefully select-ing schools with as many students withmoderate IDs as possible and with the addi-tion of a school that focuses on students withIDs, our sample remained quite small for agroup design study. This is problematic be-

cause it increases the probability of Type IIerror and it is possible that significant differ-ences between the groups on some measureswere not detected simply because of the smallsample size. A competing limitation is that byconducting multiple t-tests on related mea-sures we increased the possibility of Type Ierror. We addressed this limitation by apply-ing the Bonferroni correction procedure. Ourfindings held up under the scrutiny of thisconservative procedure.

Further research is needed to address mul-tiple questions related to teaching studentswith ID to read. One need is further explora-tion of the relationship between IQ and re-sponse to reading instruction. Currently, weare examining this issue with our larger studyin which we are following the progress of stu-dents with IQs ranging from 40 to 79 over fouracademic years. In that study, we are also ad-dressing the question of the level of readingcompetence that can be achieved by studentswith low IQs. In this article, language mea-sures were administered only at pretest andposttest. Further analyses of language mea-sures, especially measures across time, are alsoneeded. Given the variability of student per-formance, language measures that can be ad-ministered frequently would be useful for re-search and for teachers to use in theirclassrooms for ongoing progress monitoring.Finally, further research is needed to deter-mine progress over a longer period of time,especially on measures of advanced reading,including fluency and comprehension.

Summary

In summary, students with moderate IDs canlearn basic reading skills given consistent, ex-plicit, and comprehensive reading instructionacross a long period of time. Success requiresthat we apply key instructional features thathave been demonstrated to be effective withstruggling readers with average IQs, as well astechniques known to be effective for studentswith IDs. Teachers must be provided with up-to-date materials and extensive professionaldevelopment and continued support in orderto implement research-based instruction withhigh degrees of fidelity. Additionally, teachersmust monitor student progress in order tomake academic and behavioral modifications


needed to ensure success. Teachers also needaccess to coaches with expertise in reading.Although we hope this study raises expecta-tions for students with IDs, particularly mod-erate IDs, we also wish to emphasize that pro-viding effective reading instruction tostudents with IDs is extremely challenging.Finally, we need to continue to explore what ispossible for students with ID if they are pro-vided consistent, comprehensive reading in-struction for an extended period of time.

References

Adams, G., & Engelmann, S. (1996). Research ondirect instruction: 25 years beyond DISTAR. Seattle,WA: Educational Achievement Systems

Adams, M. J. (1990). Beginning to read: Thinking andlearning about print. Cambridge, MA: Massachu-setts Institute of Technology.

Allor, J. H., Gansle, K. A., & Denny, R. K. (2006).The stop and go phonemic awareness game: Pro-viding modeling, practice, and feedback. Prevent-ing School Failure, 50(4), 23–30.

Allor, J. H., Mathes, P. G., & Jones, F. G. (2010).Early Interventions in Reading: Foundation Level. Un-published manuscript.

Allor, J. H., Roberts, J. K., Mathes, P. G., Roid, G., &Cheatham, J. P. (2010). The relationship betweenIQ and response to intervention: An experimen-tal examination. Unpublished manuscript.

Arnold, D. S., & Whitehurst, G. J. (1994). Acceler-ating language development through picturereading: A summary of dialogic reading and itseffects. In D. K. Dickinson (Ed.), Bridges to literacy:Children, families, and school (pp. 102–128).Cam-bridge, MA: Blackwell.

Blaiklock, K. E. (2004). The importance of letterknowledge in the relationship between phonolog-ical awareness and reading. Journal of Research inReading, 27, 36–57.

Borman, G., Hewes, G. M., Overman, L. T., &Brown, S. (2003). Comprehensive school reformand achievement: A meta-analysis. Review of Edu-cational Research, 73, 125–230.

Browder, D. M., Wakeman, S. Y., Spooner, F., Ahl-grim-Delzell, L., & Algozzine, B. (2006). Researchon reading instruction for individuals with signif-icant cognitive disabilities. Exceptional Children, 72,392–408.

Browder, D. M., & Xin, Y. P. (1998). A meta-analysisand review of sight word research and its impli-cations for teaching functional reading to individ-uals with moderate and severe disabilities. TheJournal of Special Education, 32, 130–153.

Carlson, C. D., & Francis, D. (2002). Increasing the

reading achievement of as-risk children throughdirect instruction: Evaluation of the Rodeo Insti-tute for Teacher Excellence (RITE). Journal ofEducation for Students Placed at Risk, 7, 141–166.

Carnine, D. W., Silbert, J., Kame’enui, E. J., &Tarver, S. G. (2004). Direct instruction reading (4thed.). Upper Saddle River, NJ: Pearson MerrillPrentice Hall.

Chall, J. S. (1996). Stages of reading development (2nded.). Fort Worth, TX: Harcourt Brace.

Conners, F. A. (2003). Reading skills and cognitiveabilities of individuals with mental retardation. InL. Abbeduto (Ed.), International review of researchin mental retardation (Vol. 27, pp. 191–229). SanDiego: San Diego Academic Press.

Conners, F. A., Rosenquist, C. J., Sligh, A. C., Atwell,J. A., & Kiser, T. (2006). Phonological readingskills acquisition by children with mental retarda-tion. Research in Developmental Disabilities, 27, 121–137.

Conners, F. A. (1992). Reading instruction for stu-dents with moderate mental retardation: Reviewand analysis of research. American Journal of MentalRetardation, 96, 577–597.

Coyne, M. D., Kame’enui, E. J., & Simmons, D. C.(2001). Prevention and intervention in beginningreading: Two complex systems. Learning Disabili-ties Research and Practice, 16, 62–73.

Denton, C. A., & Mathes, P. G. (2003). Interventionfor struggling readers: Possibilities and chal-lenges. In B.R. Foorman (Ed.), Preventing andremediating reading difficulties: Bringing science toscale (pp. 229–251). Timonium, MD: York Press.

Dickinson, D. K., & Smith, M. W. (1994). Long-termeffects of preschool teachers’ book readings onlow-income children’s vocabulary and story com-prehension. Reading Research Quarterly, 29, 105–122.

Dunn, O. J. (1961). Multiple comparisons amongmeans, Journal of the American Statistical Association,56, 52–64.

Dunn, L. M. & Dunn, L. M. (1997). Peabody PictureVocabulary Test (3rd ed). Circle

Pines, MN: American Guidance Service.Ehri, L. C. (2005). Learning to read words: Theory,

findings, and issues. Scientific Studies of Reading, 9,167–188.

Ehri, L. C., & McCormick, S. (1998). Phases of wordlearning: Implications for instruction with de-layed and disabled readers. Reading & WritingQuarterly, 14, 135.

Ehri, L. C., Nunes, S. R., Willows, D. M., Schuster,B. V., Yaghoub-Zadeh, Z., & Shanahan, T. (2001).Phonemic awareness instruction that helps chil-dren learn to read: Evidence from the NationalReading Panel’s meta-analysis. Reading ResearchQuarterly, 36, 250–287.

Engelmann, S. (1997). Preventing failure in the pri-


mary grades. Eugene, OR: Association for DirectInstruction.

Engelmann, S., & Carnine, D. W. (1982). Theory ofinstruction: Principles and applications. New York:Irvington.

Evans, M. A., Bell, M., Shaw, D., Moretti, S., & Page,J. (2006). Letter names, letter sounds, and pho-nological awareness: An examination of kinder-garten children across letters and of letters acrosschildren. Reading and Writing 19, 959–989.

Foorman, B. R., & Torgesen, J. (2001). Critical ele-ments of classroom and small-group instructionpromote reading success in all children. LearningDisabilities Research & Practice, 16, 203–212.

Foy, J. G., & Mann, V. (2006). Changes in lettersound knowledge are associated with develop-ment of phonological awareness in pre-schoolchildren. Journal of Research in Reading, 29, 143–161.

Gall, J. P., Gall, M. D., & Borg, W. R. (2007). Apply-ing educational research: A practical guide (8th ed.).Boston: Allyn and Bacon. Good, R. H., & Kamin-ski, R. A. (2002). Dynamic indicators of basic earlyliteracy skills, 6th edition. http://dibels.uoregon.edu/

Hox, J. J. (2002). Multilevel analysis: Techniques andapplications. Mahway, NJ: Erlbaum.

Joseph, L. M., & Seery, M. E. (2004). Where is thephonics? A review of the literature on the use ofphonetic analysis with students with mental retar-dation. Remedial and Special Education, 25(2), 88–94.

Kame’enui, E. J., & Simmons, D. C. (1990). Design-ing instructional strategies: Prevention of academiclearning problems. Columbus, OH: Merrill.

Karweit, N., & Wasik, B. A. (1996). The effects ofstory reading programs on literacy and languagedevelopment of disadvantaged preschoolers. Jour-nal of Education for Students Placed at Risk, 1, 319–348.

Katims, D. S. (2000). Literacy instruction for peoplewith mental retardation: Historical highlights andcontemporary analysis. Education and Training inMental Retardation and Developmental Disabilities,35, 3–15

Katims, D. S. (2001). Literacy assessment of studentswith mental retardation: An exploratory investiga-tion. Education and Training in Mental Retardationand Developmental Disabilities 36, 363–371.

Ligas, M. (2002). Evaluation of Broward Countyalliance of quality schools project. Journal of Edu-cation for Students Placed at Risk, 7, 117–139.

Mathes, P. G., & Denton, C. A. (2002). The preven-tion and identification of reading disability. Sem-inars in Pediatric Neurology, 9, 185–191.

Mathes, P. G., Denton, C. A., Fletcher, J. M., An-thony, J. L., Francis, D. J., & Schatschneider, C.(2005). The effects of theoretically different in-

struction and student characteristics on the skillsof struggling readers. Reading Research Quarterly,40, 148–182.

Mathes, P. G., & Torgesen, J. K. (2005a) Early Inter-ventions in Reading, Level 1. Columbus, OH: SRA/McGraw-Hill.

Mathes, P. G., & Torgesen, J. K. (2005b) Early Inter-ventions in Reading, Level 2. Columbus, OH: SRA/McGraw-Hill.

National Reading Panel. (2000). Teaching children toread: An evidence-based assessment of the scientific re-search literature on reading and its implications forreading instruction. Washington, D.C.: National In-stitute of Child Health and Human Development.

No Child Left Behind Act of 2001, Pub. L. No.107–110, 115 Stat. 1425 (2002).

O’Connor, R. E., Notari-Syverson, A., & Vadasy, P. F.(1996). Ladders to literacy: The effects of teacher-led phonological activities for kindergarten chil-dren with and without disabilities. Exceptional Chil-dren, 63, 117–130.

Perfetti, C. A., Landi, N., & Oakhill, J. (2005). Theacquisition of reading comprehension skill. InM. J. Snowling & C. Hulme (Eds.), The science ofreading: A handbook (pp. 248–266). Malden, MA:Blackwell.

Pressley, M. P. (1998). Elementary reading instructionthat works. New York: Guilford Press.

Rayner, K., Foorman, B. R., Perfetti, A., Pesetsky, D.,& Seidenberg, M. S. (2001). How psychologicalscience informs the teaching of reading [Mono-graph]. Psychological Science, 2, 31–74.

Roberts, T. A. (2003). Effects of alphabet-letter in-struction on young children’s word recognition.Journal of Educational Psychology, 95, 41–51.

Roberts, J. K. (2004). An introductory primer onmultilevel and hierarchical linear modeling.Learning Disabilities: A Contemporary Journal, 2(1),30–38.

Share, D. L., Jorm, A. F., Maclean, R., & Matthews,R. (1984). Sources of individual differences inreading acquisition. Journal of Educational Psychol-ogy, 76, 1309–1324.

Snow, C. E., Burns, M. S., & Griffin, P. (Eds.).(1998). Preventing reading difficulties in young chil-dren. Washington, DC: National Academy Press.

Torgesen, J. K. (2002). The prevention of readingdifficulties. Journal of School Psychology, 40, 7.

Torgesen, J. K., Wagner, R. K., & Rashotte, C. A.(1999). The Test of Word Reading Efficiency. Austin,TX: PRO-Ed. Treiman, R., Tincoff, R., Richmond-Welty, E. D. (1996). Letter names help childrento connect print and speech. Developmental Psychol-ogy, 32, 505–514.

Treiman, R., Tincoff, K. R., Rodriguez, K., Mouzaki,A., & Francis, D. J. (1998). The foundations ofliteracy: learning the sounds of letters. Child De-velopment, 69, 1524–1540.


Valdez-Menchaca, M. C., & Whitehurst, G. J. (1992).Accelerating language development through pic-ture book reading: A systematic extension to Mex-ican day care. Developmental Psychology, 28, 1106–1114.

Vaughn, S., Mathes, P., Linan-Thompson, S., Cirino,P., Carlson, C., Pollard-Durodola, S., et al. (2006).Effectiveness of an English intervention for first-grade English language learners at risk for read-ing problems. The Elementary School Journal,107(2), 153–180.

Vaughn, S., Cirino, P. T., Linan-Thompson, S.,Mathes, P. G., Carlson, C. D., Cardenas-Hagan, etal. (2006). Effectiveness of a Spanish interventionand an English intervention for English-LanguageLearners at risk for reading problems. American Ed-ucational Research Journal, 43, 449–479.

Vaughn, S., Linan-Thompson, Mathes, P. G. Dur-adola, S., & Cardenas-Hagan, E. (2007). First-grade English language learners at-risk for read-ing problems: Effectiveness of an Englishintervention. Elementary School Journal, 107, 153–180.

Wagner, R., Torgesen, J. & Rashotte, C. (1999).Comprehensive Test of Phonological Processing. Austin,TX: PRO-ED.

Williams, K. T. (1997). Expressive Vocabulary Test.Circle Pines, MN: American Guidance Service.

Woodcock, R. W. (1991). Woodcock Language Profi-ciency Battery-Revised. Itasca, IL: Riverside.

Received:I 4 August 2008Initial Acceptance: 12 October 2008Final Acceptance: 30 March 2009


Documents

Teaching Students with Moderate Intellectual Disabilities ...daddcec.org/Portals/0/CEC/Autism_Disabilities/Research/... · Teaching Students with Moderate Intellectual Disabilities