Review of experimental research on academic learning by students with moderate and severe intellectual disability in general education

Research & Practice for Persons with Severe Disabilities2013, Vol. 38, No. 1, 17-29

copyright 2013 byTASH

Review of Experimental Research onAcademic Learning by Stndents With

Moderate and Severe IntellectnalDisability in General Edncation

Melissa E. HudsonAmerican Institutes for Research

Diane M. Browder and Leah A. WoodUniversity of North Carolina at Charlotte

A review of the literature on academic learning in generaleducation settings for students with moderate and severeintellectual disability was conducted. A total of 17 experi-mental studies was identified and evaluated using qualityindicators for single-case design research. Studies that metor met with reservation the criteria established for qualityresearch were used to determine the evidence base of theinstructional strategies described in the literature. The reviewfound embedded instruction trials using constant time delayto be an evidence-based practice for teaciiing academic con-tent to students with moderate and severe intellectual dis-ability in general education. In addition, strategies that werenot yet evidence-based but showed promise in the literaturefor teaching academic content to students with moderateand severe intellectual disability in general education weredescribed. Last, implications for practice and directions forfuture research were discussed.

DESCRIPTORS: general education, special educa-tion, moderate and severe intellectual disability, academiclearning, embedded trial instruction, evidence-based prac-tice, constant time delay

To ensure students with moderate and severe intel-lectual disability have fuU access to general curriculum,effective procedures are needed to teach academic con-tent. Most states have now adopted the Common CoreState Standards (National Governors Association Cen-ter for Best Practices, Council of Chief State SchoolOfficers, 2010), which will be the foundation of schoolaccountability for students with and without disabilities,including those who participate in alternate assessmentbased on alternate achievement standards (AA-AAS).For example, Colorado has developed extensions for the

This work was completed when Melissa E. Hudson was a spe-cial education doctoral student at the University of North Garolinaat Gharlotte.

Address all correspondence and reprint requests to Melissa E.Hudson, American Institutes for Research, 1000 Thomas JeffersonStreet NW, Washington, DG 20007. E-maü: [email protected]

Common Core State Standards (http://www.cde.state.co/us/cdeassess/UAS/CoAcademicStandards) for studentswho participate in AA-AAS. To prepare students withmoderate and severe disabilities to meet these increas-ing academic expectations, educators need to know theevidence-based practices that promote student academiclearning (Copeland & Cosbey, 2008/2009). Identifyingsuch practices for general education settings may be espe-cially critical to promote inclusive practices concurrentwith high stakes accountability.

Evidence-based practices (Odom et al., 2005) are inter-ventions that have been determined to be effective basedon a body of high-quality research. Determining whetheror not a practice is evidence-based involves two steps.First, after conducting a comprehensive review of the lit-erature, the reviewer applies criteria for design quality toeach study identified. Next, the number of high-qualitystudies is considered to determine if there is a criticalmass to identify a practice as evidence-based. Qualityindicators and guidelines differ across methodologies.Homer et al. (2005) developed quality indicators andguidelines for use with single case research. Althoughother criteria have been developed. Homer et al.'s qualityindicators and guidelines are most frequently applied inreviews of single case research (e.g., Browder, Ahlgrim-Delzell, Spooner, Mims, & Baker, 2009; Browder,Spooner, Ahlgrim-Delzell, Harris, & Wakeman, 2008;Browder, Wakeman, Spooner, Ahlgrim-Delzell, &Algozzine, 2006; Chard, Ketterlin-Geller, Baker, Doabler,& Apichatabutra, 2009; Lane, Kalberg, & Shepcaro, 2009;Test, Richter, Knight, & Spooner, 2010).

There are two ways to approach a review of evidence-based practice. The first is to identify a practice (i.e., inde-pendent variable) and evaluate the quality and quantityof research to support it. For example, Browder et al.(2009) evaluated the efficacy of time delay for teachingliteracy to students with severe disabilities. The researchersfound 30 studies published between 1975 and 2007 thatmet the inclusion criteria, and the quality indicators rec-ommended by Homer et al. (2005) were modified to fit

17

18 Hudson et al.

the characteristics of time delay research. The researchersfound time delay to be an evidence-based practice forteaching symbols to students with moderate and severedisabilities.

A second approach is to identify the desired outcome(i.e., dependent variable) to identify practices that yieldthis outcome. Studies have used this approach to identifyevidence-based practices for teaching reading (Browderet al., 2006), mathematics (Browder et al., 2008), andscience (Spooner, Knight, Browder, Jimenez, & DiBiase,2011) to students with moderate and severe disabilities.Browder et al. (2006) conducted a comprehensive re-view of literature in which students with significant cog-nitive disabilities were taught reading. The researchersreviewed 128 studies and compared the studies to theNational Reading Panel's (National Institute of ChildHealth and Human Development, 2000) five critical com-ponents of reading instruction (i.e., phonemic awareness,phonics, fluency, vocabulary, and comprehension). Of the88 single case studies included in the review, 54 met theindicators and guidelines recommended by Horner et al.(2005) for high-quality research. Sight word instructionaccounted for the largest number of studies meeting cri-teria (n = 42). Based on these findings, the researchersidentified the tise of systematic prompting (e.g., timedelay) to teach sight words as an evidence-based practice.

Similarly, Browder et al. (2008) conducted a compre-hensive literature review and meta-analysis of studiesteaching mathematics to students with severe cognitivedisabilities. The authors identified 68 single case studiesthat were eligible for consideration of evidence-basedpractices. Most of the studies taught number and com-putation skills (e.g., counting, number matching) or mea-surement (e.g., money skills). Of the total number ofsingle case studies reviewed, 19 met all quality indica-tors. Specific prompt fading procedures with feedbackwere found to be an evidence-based practice for teachingmathematics.

Finally, Spooner et al. (2011) conducted a comprehen-sive review of studies published between 1985 and 2009in which science content was taught to students withsevere developmental disabilities. The researchers iden-tified 17 studies and applied a modified version of Horneret al.'s (2005) quality indicators to each study. Addition-ally, the researchers used decision rules for determiningresearch quality and evidence-based practices from theNational Secondary Transition Technical Assistance Cen-ter (NSTTAC, 2010; Test et al., 2009), which considersboth high-quality and acceptable quality research whendetermining evidence-based practices. The NSTTAC cri-teria weights indicators differently, placing the greatestemphasis on the results indicator. The researchers fotmdsystematic instruction to be an evidence-based practice forteaching science content to students with severe devel-opmental disabilities.

Each of these reviews contributed evidence that stu-dents with moderate and severe disabilities can leam

academic content and offered guidance for the use of spe-cific practices; however, most of the studies were focusedon self-contained settings. While it is feasible that many ofthese strategies would transfer to instruction in generaleducation, there is some risk in making this inferencewithout experimental evidence. For example, many of theinterventions in the prior reviews were implemented byspecial education teachers or researchers and the peoplemost often available to teach academic content in generaleducafion settings are general education teachers, peerswithout disabilities, and paraprofessionals.

Providing students with disabilities access to the sameleaming and social opportunities as students without dis-abilities is challenging (Jackson, Ryndak, & Wehmeyer,2008/2009). Moreover, it is particularly important to iden-tify effective instructional strategies for academic leam-ing in general education for these students because of thelack of inclusion in the past. Therefore, the purpose ofthis review was to evaluate experimental research on aca-demic leaming in general education classrooms for stu-dents with moderate and severe intellectual disabihfy.The research questions used to analyze the literature are:(a) What are the practices for teaching academic contentin general education for which there is a strong evidencebase? and (b) What practices show promise for use ingeneral education?

MethodLiterature Search and Selection Procedures

Studies for the literature review were identified in threeways. First, education databases (i.e.. Academic SearchPremier, Education Research Complete, Eric, PsycINFO,Primary Search, and MasterFILE Premier) on the EBSCOhost interface were searched electronically using full andtruncated keywords for participants (e.g., students withdisabilities, students with moderate and severe disabil-ities, intellectual disabilify, developmental disabilities);instructional methods (e.g., embedded instruction, sys-tematic instruction, constant time delay); types of peersupport (e.g., partner leaming, peer tutoring, classwidepeer tutoring); and settings (i.e., inclusive education, gen-eral education, inclusion). The abstracts of the studiesfound through the electronic search were read and eval-uated. When additional information was needed to makea determination about the study's inclusion in the review,the study was read entirely. Studies were included in thereview that (a) were published in a peer-reviewed journalbetween the years of 1975 and 2012, (b) were written inEnglish, (c) used a teaching intervention, (d) had at leastone dependent variable that measured academic leam-ing for participants with moderate or severe intellectualdisability, (e) included at least one participant with mod-erate or severe intellectual disability (i.e., IQ < 55), (f)were implemented in a general education classroom, and(g) used an experimental design. The electronic searchyielded 12 studies that met the inclusion criteria.

Academic Learning Evidence-Based Practice Review 19

Second, an ancestral search of these studies' referencesections was conducted to identify additional studiesthat might be appropriate for the literature review thatwere missed in the electronic search. Studies from theancestral search were located in tbe professional journalin which they were published and the article abstractswere read and evaluated using the same procedures asthe electronic search. Three additional studies were iden-tified from the ancestral search. Third, special educationexperts (i.e., first authors on the 15 identified studies)were contacted by email to identify studies that were inpress with a peer-reviewed joumal that might be appro-priate for the review. The email inquiry yielded twoadditional manuscripts and electronic copies of manu-scripts were obtained from the authors. Combined, theelectronic search of primary education databases, ances-tral search, and email inquiry netted 17 studies. Thesestudies are summarized in Table I.

Application of the Homer et al. Quality IndicatorsAll of the studies in the review used a single-case

experimental design to establish a functional relationshipbetween the intervention and the dependent variable.In a special issue of Exceptional Children, Homer et al.(2005) described seven quality indicators and 21 sub-indicators (i.e., the level under each quality indicator)useful in identifying evidence-based practices with single-case research. Based on the work of Homer et al., theseseven quality indicators and 21 subindicators were usedto review tbe studies included in this literature review.

The first quality indicator, participants and setting,included three subindicators. The subindicators evaluatedif the study described (a) the number, age, disability, andethnicity (optional) of participants; (b) the criteria usedto select participants (e.g., teacher nomination, neededskills); and (c) the type of classroom setting in which theintervention was delivered (e.g., fifth grade hteracy class).

The second quality indicator, dependent variable, in-cluded five subindicators. The first subindicator evaluatedwhether the dependent variable specifically described astudent academic behavior (not a teacher behavior) oracademic engagement such as contact with the generaleducation curriculum. The second subindicator evalu-ated whether there was a quantifiable measurement forthe dependent variable (e.g., number of correct com-prehension responses). The third subindicator evaluatedwhether the measurement was well described in thestudy, including who collected data and descriptions ofthe leaming targets (e.g., word lists, vocabulary, anddefinitions). The fourth subindicator evaluated whetherlinear graphed data were provided that showed ongo-ing measures. The fifth subindicator evaluated whetherinterobserver agreement was collected on at least 20%of the sessions and that interobserver agreement scoreswere >80%.

The third quality indicator, independent variable, in-cluded three subindicators. The subindicators required

that the independent variable be described with enoughdetail to replicate, the independent variable was sys-tematically manipulated, and overt measurement of thefidelity of implementation for the independent variablewas collected, which was defined as a minimum of 80%fidelity had to be reported.

The fourth quality indicator, baseline procedures, in-cluded two subindicators. The first subindicator evaluatedwhether baseline data were stable and the second eval-uated whether the baseline conditions were adequatelydescribed so that they could be replicated.

The fifth quality indicator described experimental con-trol and intemal validity and included three subindicatoraThe first subindicator evaluated the study's intemal valid-ity. Three demonstrations of experimental effect acrossparticipants, settings, or materials were required to meetthis subindicator. The second subindicator required thatthe type of design be stated and that, unless the designwas an altemating treatment design, a staggered baselinewas used. Tbe tbird subindicator evaluated whetherthere was a change in trend or level in the linear graphwhen visually analyzed.

The sixth quality indicator described the study's exter-nal validity and included one subindicator. To meet thesubindicator, the study was required to have at least threereplications across participants or materials.

The seventh quality indicator addressed social validityand included four subindicators. The first subindicatorevaluated the social importance of the dependent vari-able. This subindicator was met if the social importanceof the intervention was described or could be easily in-ferred or evaluated by others. The second subindicatorevaluated whether the magnitude or degree of changewas socially important. For this review, this subindicatorwas met if participants leamed a minimum of five skills(e.g., at least five different sight words, at least five dif-ferent vocabulary definitions). The third subindicatorevaluated whether the intervention was practical and costeftective. This subindicator was met if the authors statedthat the intervention was practical, the cost of implement-ing the study was reported, or if the intervention was im-plemented in a school. The fifth subindicator evaluatedwhether the research was conducted in typical contexts,defmed as a school.

Coding MethodologyA two-part coding instrument was created for review-

ing the studies. Tbe first part of the instrument includedseven quality indicators based on Homer et al.'s (2005)recommendations and 21 subindicators defined specifi-cally for this study. Each quality indicator was rated aspresent (i.e., -i-) or missing (i.e., - ) based on the ratingsof the subindicators. For example, if all or all but one ofthe subindicators for a quality indicator were rated -i-,the quality indicator was rated +. If two or more sub-indicators were missing, the quality indicator was ratedas missing.

20 Hudson et al.

iîra

m '. m

M •

< on

Í2

p yî ^

•S'S

1Í

Ëm

B(UT33t/î

i/lBO

ms•a3

(N

OD,0)

ipl

3

S

Bra.B-'•graa.

Ilg's.2

ïîllS a.g

£3

•a p M '-g 'P-n ho. ra

<

11^ra u ^

lilira t« ;z: o

liîi^ tfl =+- t-S u 4-" oa.T3 o ra

rt\ ' ^ #*4 ^ C T^ niä § & g S | g lE p "

g P

ü ra3 a2

ai0) ra

i/5

i3S

lel

"ra

par

i

cro

ra

obc

ant

o.

irtici

- : S r a | - 2 . ^ | o P - c a

S-IO "ra

E i>^-ra

i2

P OÔ^-,Q. B 00E - Bo " 'Su u O

•s •§ Í3 « C E .s :::: c w

.E-o .E 2 g-3 .E .S ^ . 2o. S D..E -0 -o a . - :§ S ^Itlailèl

VM

o •« 3 x:D-Í3 U « O •

IBJSU

O3

T3

XI

EW

_ra

B

1

E

tant

tin

c(J

ra+-J

te sp

ec:

2cdO-

i cla

ss>

raCJ

Où ' ) o

¡îî•g s 2^ ! |g SB

• «-a

„ i g3 -o rao Pra .ra a

Oíp

'>;:3 < I

w c- oj sr *- ajV- aj ;^ o ra -w

r«XÎ i y p £ : P , ^ Si s Hg ^ i ragË«jS0 H ^ g

I ^ i „ I 2U'^^PS^ p ^ . c p a

oí

.2

W

i-.3 S) g S - | p i g . S! ^3 ra Kfi •*—I ^ u CJ hr" .«_•

. 20) ra

3O

f Jtfl tfl O

^§;

e s «a « ra

"ra o 3 i ^ '—t •--

p

'U• ara00

XI•if

and

high

.a

B"Ôo

sch

istor

y

x :

o

sch

lish

and

p

«rab,

>th£

1 gr

ade

'Sira

lltl

ion

• |

a,O

gTí

•ërasb

XI• ^

ISS a

nde

ra o" S

ing

Seil

S«

9rau

II o

S

-17

y

CM

"O

(De

repo

not

IQs 2 T3

'Ils2

II oc2

l—i ï-H V l

« -a ^

SÍI

. . raQ g.

0) 2"ra "

•a onOo ra M

! " g g-2

XI

Cus

tu(HraO

• apra

lar

U

edy

pp

ô

Haï

p

oil

U

• aBra

son,

B2m

2

•3BKo o<•o.S3 ^ ^

Jd

Academic Learning Evidence-Based Practice Review 21

« s ao c o. c0 .= c -2

tilt2

o 0)

fiil

ilD (/3

. ^ cj3 CO

•o .2T3 T3

- .1W

O

ca o 3 rs ^ CO u

BC

c TJ COO .2 3I ^ B Ê

T3 -aV, "

II

•a '" |3.i.SS |

7 ¡ms^

Ö.2

nil

efi

• 0

.2CO0

cal

.g(D

U

dei

Scd

Lev

els o

f:io

i

cd

—

T3CO

resp

on

S3

S20

' >CO

s:0XIon.S

com

pel

c0

S

—

tfl

i

Wee

kly

p

30x:

Í2u 3x: ^

C "

•a

11 et

0

c . ^c ^

cDo

(20C

CD

20

9s

3

Bue

0x:

Mat

• accOC

i2

Tho

001

VH

tE

•-=! O

CA Mcd cdX) atu «

illPI

CO

C(U crt

'S c

SSO.

0ca

robe

mts

Ci .ai ' DO,'-£3¿1 CO3 Cu

SSO.

CO

robe

rs

a b(U '> •

•3.S

3 Xî

itie

rax>

DiS

£

_uD.w

3

o-a om

s c•?? eu

C L(A

en

m

WH (U

CD O

GCOto

'o .sc -a o css, 0.2c Ë " cO Ü 5 O

•a(U

IËI .Ë - o ™ -oW V

ils1§

1W

S X3 g

W

o rtJ tM*--— cd ^"'"' ^ '.^ ^ "5 oû

g £ g -S CO .;. au

co1 I §•o

2 S C A . ^ S £ S Í ; 2 S ^ Í £3 .-3 CO

« a 2

iS

•o

Ü

00 *

IÛ0 O

w ti -^u g uÇ2 o a>

D o > ,h- 3 CO

o

siascd o

I SË 3a) -o'S «

• a leu o Rdi fí O

(U> w bfl

•g §"2• - - S o

.Ë Sá-S

o 2

1 ICd íi P rsw . t í •<-• oJi o ^ «C cd (ü CO

u J3

liliI 551

o COC en

>>'u "o

-- CO g«

CJ0j3 Upi

Q ...

ú Í2COu | ;d g S"^ - - ^ ... ^• S "

.2o -

o

O)

S

1Q

o•SQ SË S2 3

II< § •

13

<

< 11.

22 Hudson et al.

Each of the seven quality indicators and 21 subindica-tors were rated in this manner, and one of three decisionsregarding the study's quality was made: (a) meets criteriafor quality research, (b) meets criteria with reservation, or(c) does not meet criteria. A study met criteria if all21 subindicators were rated -i-. A study met with reserva-tion if the study met the quality indicators as describedwith the following exceptions: did not meet criteria for onesubindicator for up to two quality indicators, but not twosubindicators from the same quality indicator. For exam-ple, a study met criteria with reservation if a subindicatorfor dependent variable (e.g., interobserver agreement) anda subindicator for social validity (e.g., magnitude of effect issocially valid) were not met, but the study did not meetcriteria if two subindicators for dependent variable (e.g.,interobserver agreement and repeated measurement ofdependent variable) were not met. This decision mle wasmade to avoid eliminating high-quality studies from theevidence base that would have been eliminated byincluding only studies that met every indicator, a problemnoted by other researchers (Lane et al., 2009; Test, Fowler,Brewer, & Wood, 2005). Excluded from these criteria wasthe subindicator "change in trend and level" from theExperimental Control/Internal Validity indicator. If thissubindicator was missing, the study did not meet criteriafor quality.

Part 2 of the review instrument was completed for allstudies that met criteria for quality research or met crite-ria with reservation. The second part of the review instru-ment contained descripfive data for each study, includingdetailed descriptions about the participants and set-ting, the academic content taught, the independent vari-able, the context of instruction, the research design, andinstructional materials.

Method to Establish Rater Consensus onQuality Indicators

Two doctoral students (i.e., first and third authors)reviewed the studies using the quality indicators. Beforedata were coded, a process of review and consensus wasundertaken to ensure that the quality indicators wereapplied consistently by reviewers. After reviewers in-dependently reviewed a study and applied the qualityindicators, their ratings were compared item by itemand discrepancies were discussed until a consensus wasreached about the rating. Descriptions for the subindica-tors were clarified using examples and nonexamples andthe quality review instrument was revised using the defi-nitions and examples. The minimum criterion was 90%agreement for the quality indicators. Once the criterionfor agreement was met, the remaining studies were ran-domly assigned to one of the two reviewers for review.

Method to Establish Interrater Reliability onQuality Indicators

To establish interrater reliability, six studies (i.e.,40%) were randomly selected for a second review by

either the first or third author from the 15 studies re-viewed independently by the raters (not the two studiesused for rater reliability). If the first author did thefirst review, the third author did the second review andvice versa. Ratings from the first and second reviewsof the seven quality indicators were compared. Anagreement (+) was found for a quality indicator ifboth reviewers rated a quality indicator the same anda disagreement ( - ) was found if the reviewers' rat-ings were not the same. Intcrrater reliability was deter-mined by taking the number of agreements dividedby the number of agreements plus disagreements andmultiplying the quotient by 100 (Billingsley, White, &Munson, 1980).

The reviewers also compared their overall determina-tion about the quality of each study in the review (i.e.,met criteria, met criteria with reservation, did not meetcriteria). An agreement was scored if both reviewersmade the same determination and a disagreement wasscored if the reviewers did not make the same deter-mination. Interrater rehabüity was calculated as previ-ously described.

Determination of Evidence-Based PracticesHomer et al. (2005) described a two-step process for

determining the evidence base of an instructional practiceusing single case research. The first step is identifyingquahty studies by applying quality indicators. The sec-ond step is evaluating if enough quality research existsto warrant the instructional practice as evidence based.Homer et al. recommended minimum standards for de-termining the evidence base of a practice including: fivequality studies, three different researchers across threedifferent geographical locations, and at least 20 par-ticipants. These standards were applied to the 17 studiesthat met or met with reservation the criteria for qualityresearch established for the review.

ResultsQuality of the Single-Case Experiments

A total of 17 experimental studies met the inclusioncriteria established for this review. Twelve studies metall 21 subindicators; five studies were missing one or twosubindicators and met the criteria for quality researchwith reservation. Of the five studies that met qualitycriteria with reservation, subindicators from dependentvariable, independent variable, and social validity qualityindicators were rated as missing. Four studies did notreport interobserver agreement on at least 20% of thesessions (i.e., a subindicator for the dependent variablequality indicator), two studies did not report overt mea-surement of the independent variable, and the partici-pants in two studies did not leam a minimum of five skills(i.e., a subindicator for the social validity quality indi-cator). Table 2 describes the results of the application ofthe quality indicators and subindicators for each studyin the review.

Academic Leaming Evidence-Based Practice Review 23

13

OO

+ + + + +

§8E'~"

r o

•t3q

O

24 Hudson et al.

laQ

-^-^-^--^-^

1 'GI C

s8 H- -t- + I -I- -H -I-

ca

Ot . OH ÖH

Academic Leaming Evidence-Based Practice Review 25

Interrater Reliability for Quality IndicatorsThe reviewers reached 100% agreement after a pro-

cess of review and consensus for 2 of the 17 studiesincluded in this review. Reviewers independently ratedthe remaitiing 15 studies and interrater reliability for theseven quality indicators was 100%. Interrater reliabilitywas also 100% for decisions made about study quality(i.e., met criteria, met criteria with reservation, or did notmeet criteria).

Characteristics of the StudiesTable 1 describes the 17 studies included in this lit-

erature review. Both primary and secondary generaleducafion classroom setfings were represented in thisliterature review. In the primary grades, five studies wereconducted in an elementary school, one in kindergarten,and one in both elementary and kindergarten classrooms.In the secondary grades, five studies were conducted inthe middle school, two in high school, and two in bothmiddle school and high school. One study was conductedacross grade levels and included elementary, middle, andhigh school students.

Participants in these studies included 44 students en-rolled in public schools who ranged in age from 5 to21 years. Most were classified as having moderate intel-lectual disability (n = 34; 10 range, 35-55); however,participants from other disability categories were in-cluded: severe intellectual disability {n = 1), Aufism Spec-trum Disorder (n = 2), Pervasive Developmental Disorder(n = 1), multiple disabilifies (n = 2), and developmentaldisabilities (n = 4). Participants received instruction in avariety of academic content: science (« = 7), composition/spelling {n = 3), mathematics (« = 4), social studies/geography/history (n = 5), ELA/literacy/reading (« = 8),art/health/PE (n = 4), and German (n = 1).

Determination of Evidence-Based PracticeOf the 17 studies reviewed, five instructional interven-

tions were implemented. Ten of the 17 studies used em-bedded trial instrucfion; two studies used system of leastprompts; two studies used a combinafion of peer deliv-ered cues, feedback, and error correction; one studyused peer supports; one study used teacher-deliveredstatements embedded during instruction (but not de-livered within instructional trials); and one study usedconstant time delay instruction. The instructional in-tervention used by most studies was embedded trialinstruction. Embedded trial instruction was explicit, un-ambiguous, and systematic instruction distributed acrossongoing roufines and acfivities within the general edu-cation classroom (McDonnell, Johnson, & McOuivey,2008). The 10 studies using embedded trial instrucfionwere implemented by researchers in five geographicallocations (i.e., universities) and included 26 participantswith moderate and severe intellectual disability. There-fore, enough research was fotmd to determine the useof embedded trial instrucfion for teaching academiccontent to students with moderate and severe intellec-

tual disability in general education as an evidence-basedpractice.

Furthermore, of the 10 embedded trial instructionstudies, nine used constant time delay as the instructionalprocedure. Consequently, when embedded trial instruc-tion using constant time delay is considered, enoughquality studies exist to determine the use of embeddedtrial instruction using constant time delay as an evidence-based practice: nine quality studies, four geographicallocations, and 22 participants.

While not enough research exists to establish theiruse as an evidenced-based practice for teaching academiccontent in general education, the system of least promptsand task analytic instruction hold promise. The system ofleast prompts intervenfion was used in two studies thatwere implemented by researchers in two geographicallocations and involved six participants. Two studies usedtask analj^ic instruction in conjunction with systemaficinstruction and were implemented by researchers in twogeographical locafions and involved four parficipants.

DiscussionOverall, embedded trial instruction, especially when

using constant time delay (embedded CTD), was foundto be an evidence-based practice for teaching academiccontent to students with moderate and severe disabilitiesin general education. Embedded trial instrucfion is ex-plicit, tinambiguous, systematic instruction that distributesinstrucfional trials across ongoing routines and activitieswithin the general education classroom (McDonnellet al., 2008). In this review, embedded trial instructionwith constant time delay provided flexibility for implemen-tation and was used effecfively across different interven-tionist. For example, in a study by Jameson, McDonnell,Johnson, Riesen, and Polychronis (2007), embedded con-stant fime delay instrucfion was implemented by a mid-dle school special education teacher and paraeducator toteach two students with moderate intellectual disability.One student leamed definitions for states of matter (i.e.,boil, melt, and freeze) and another leamed to identifyteen living themes (e.g.. Best Friend's Wedding—Preparefor marriage) on a symbol chart. In another study byJameson, McDonnell, Polychronis, and Riesen (2008),two general educafion middle school teachers and threepeer tutors without disabilifies used embedded constanttime delay instruction to teach three students with mod-erate intellectual disability key word definitions for health(e.g., lungs—get less air, can get cancer) and art (e.g.,firing—heating clay in a special oven).

Embedded CTD instrucfion was also used acrossa range of content. For example. Riesen, McDonnell,Johnson, Polychronis, and Jameson (2003) taught scienceand German vocabulary sight words (e.g., gram, mass,eins, zwie); Johnson, McDonnell, Holzwarth, and Hunter(2004) taught first-grade reading sight words (e.g., police,stop); and Jameson, Walker, Utley, and Maughan (2012)

26 Hudson et al.

taught phonemic awareness. In other studies, Polychronis,McDonnell, Johnson, Riesen, and Jameson (2004) taughtmath (i.e., identifying numbers zero to nine, telling timeat 15 and 30 min past the hour) and geography (i.e.,identifying state capitals); McDonnell et al. (2006) taughtdefinitions from science (e.g., atom—the smallest part ofan element) and U.S. history (e.g., values—principles welive by); and Jimenez, Browder, Spooner, and DiBiase(2012) taught science vocabulary (e.g., kinetic energy) andconcept statements (e.g., kinetic energy is the energy ofmotion). Embedded CTD instruction was also adaptableacross a range of grade levels, including kindergarten,elementary, and middle school.

In addition to the evidence base for embedded trialinstruction, promise was found for the use of the systemof least prompts and task analytic instruction for teach-ing academic content. For example, Collins, Branson,Hall, and Rankin (2001) used the system of least promptsand an 11-step task analysis to teach four letter-writingcomponents (i.e., date, greeting, body, closing) to threehigh school students in a twelfth grade compositionclass. In another study, Hudson, Browder, and Jimenez(in press) used the system of least prompts and read-alouds of adapted grade-level science content to teachfourth grade students with moderate intellectual dis-ability listening comprehension responses. Interestingly,when applied to academic learning, there were somemodifications in how the prompts were used. For ex-ample, in the Collins et al. study, students were givendirections for writing a component of the letter (e.g.,write "Dear John" on the left side of the paper) andgrammar rules (e.g., start each word with a capitalletter) in the prompts. Similarly, Hudson et al. incorpo-rated opportunities to hear text read again in the prompthierarchy.

These results are similar to the results found in pre-vious literature reviews conducted to identify evidence-based practices for reading (Browder et al., 2006),mathematics (Browder et al., 2008), and science (Spooneret al., 2011) for this population in that each supportedthe use of systematic prompting with feedback. Whilethese initial reviews each noted a restricted range of aca-demic content, more recent studies in self-contained set-tings have addressed a broader range of math/science(Browder et al., 2012) and language arts standards (Mims,Hudson, & Browder, 2012). What differs is that researchin inclusive settings more often has used constant timedelay than the range of prompting and feedback strate-gies found in research in self-contained settings. Also, theprompted target responses are more typically embeddedin the flow of a general education lesson. Because theselessons may be asking many other responses of the stu-dent that are not a focus of the intervention (e.g., listeningto the general education teacher, interacting with a peer,managing general education materials) but may be theresponsibility of whoever is providing instructional sup-port (e.g., peer, paraprofessional), it may be that the sim-

plicity of constant time delay is more amenable to thissetting. It also is feasible that researchers have not triedthe range of evidence-based practices available from re-search in self-contained settings to more inclusive con-texts. For example, in a dissertation study, Hudson (2013)found that a least intrusive prompting strategy could beused by peers to promote listening comprehension for anadapted chapter book.

Limitations and Implications for Future ResearchA limitation of this study is that reliability data for the

selection of articles to be included in the study werenot collected. Even though the search process was exten-sive, it is possible that relevant articles were missed. Fu-ture research may want to conduct two separate articlesearches to increase the likelihood that all relevant studiesare identified.

Expanding the intervention research in inclusive set-tings is an overall need for future research given thatonly 17 studies were identified by this review. As noted,future research is needed to see if some of the sys-tematic instruction strategies that have been well docu-mented in research in other settings (e.g., system of leastintrusive prompts; task analytic instruction) have ap-plicability for academic learning in inclusive settings.Research is also needed to expand the type of contenttargeted for instruction and to involve the student morein delivery of the intervention.

Future research needs to consider academic contentthat goes beyond simple recall of facts or identification ofsight word vocabulary. For example, elementary studentslearned to identify ftinctional (e.g., refrigerate, measure)and core content science words (e.g., electricity, pre-cipitation) in a general education class (Collins, Evans,Creech-Galloway, Karl, & Miller, 2007). Similarly, in astudy conducted by Jameson et al. (2007), middle schoolparticipants learned to define vocabulary terms from ageneral education science unit on states of matter and toidentify teen living themes used on a symbol chart bygeneral education students in Teen Living class. Bothdemonstrated an important starting point for inclusiveinstruction which is to be sure students have the vocab-ulary to communicate about the topic. In contrast, theCommon Core State Standards, which nearly all stateshave now adopted, require more complex engagementwith the content (National Governors Association Cen-ter for Best Practices, Council of Chief State SchoolOfficers, 2010). For example, many academic standardsrequire learning concepts. In the current review, Jimenezet al. (2012) taught middle school students with mode-rate intellectual disability not only to identify the vocab-ulary but also to use it to identify concepts (e.g., "What

word completes this concept statement: is theenergy of motion? Kinetic energy is the energy of mo-tion."). These students also learned the use of a KWHLchart (i.e., K = what do you Know?; W = What do youwant to know? H = How wül you ñnd out?; L = what did

Academic Leaming Evidence-Based Practice Review 27

you Leam?) so that they could participate in the inquiryprocess the general education teacher used. In future re-search, the academic target might be concepts, inquiry,synthesis, comparisons, or evaluation.

Besides expanding the academic content, another areafor future research is to involve students more in theirown intervention. One model for doing so is the Self-Determined Learning Model of Instruction (SDLMI;Mithaug, Wehmeyer, Agran, Marfin, & Palmer, 1998;Wehmeyer, Palmer, Agran, Mithaug, & Martin, 2000),in which educators teach students to set goals, take actionon those goals, and adjust their goals and plans as needed.Within the SDLMI, students pose questions (e.g.. Whatis my goal?; What is my plan?; What have I learned?) toset and assess progress towards their goals.

Studies have demonstrated how students with mildand moderate disabilities can experience improved ac-cess to the general curriculum with the SDLMI (Agran,Blanchard, Wehmeyer, & Hughes, 2002; Lee, Wehmeyer,Palmer, Soukup, & Little, 2008; Palmer, Wehmeyer,Gipson, & Agran, 2004). Future studies might build onthis methodology by using these strategies to demon-strate learning of specific academic content like com-prehension of informational text by setting a goal forreading. Another option to involve students in theirown intervention might be to have students direct theirown prompting like one study in this review did. In theHudson et al. (in press) study, elementary students withmoderate intellectual disability learned comprehensionresponses with a peer-delivered system of least promptsintervention. The prompts in the system of least promptswere self-paced. That is, the participants with disabilitiesdirected the amount of help they received from peertutors during instruction.

Recommendations for PracticePractitioners are encouraged to use embedded in-

structional trials with constant time delay for academicleaming given the strong research foundafion for theeffectiveness of this practice. To make this application,the planning team would identify the most salient contentof the instrucfional unit that students with moderate/severe disabilities should leam. For example, in a lan-guage arts class in which students are learning aboutthemes and character, a goal might be for the studentto identify the main characters (recall), determine howcharacters are alike/different (comparisons), and deter-mine if an event in the story does or does not relate tothe theme of friendship (evaluation). Next, the specifictarget responses are defined (e.g., use a friendship symbolto mark an event as being about friends) and a promptis identified that is applicable to the response and worksfor the student (e.g., model). In constant time delay, theprompt is given at 0 s delay for some number of trials(e.g., for the first three trials) and then provided at somespecific interval (e.g., 4 s) after giving the student theopportunity to anticipate the correct response. The team

will need to identify who wiU provide the prompt (e.g.,peer, teacher, paraeducator) and when during the lessonthe trials can be embedded without competing with theoverall fiow of the lesson (e.g., when the teacher asksthe whole class to write examples of events that advancedthe friendship theme). Practifioners may also want toconduct their own action research to try other systematicinstruction procedures that have strong evidence fromresearch in other settings (e.g., task analysis, system ofleast prompts) to academic content.

SummaryThis review identified 17 studies demonstrating stu-

dents with moderate and severe disabilities learned aca-demic content in general education settings. The targetcontent was derived directly from what students withoutdisabilities were learning in these settings (i.e., "gradeappropriate"). By applying experimental research de-signs, a strong inference can be made that leaming wasdue to the intervention and not to the student's matura-tion alone or other confounding variables. The strongestevidence-based practice for use in general educationis embedded trial instruction in the general educationactivity with constant time delay to fade systematicprompts. This practice has substantial flexibility for useacross content areas and interventionists (e.g., peers,paraprofessionals). There also was promise for the use ofsystem of least prompts, which can be modified to focuson the academic (vs. motoric) responses to be made. Forexample, the prompt system can include rereading keypassages of text to promote comprehension in a read-aloud. Researchers are encouraged to replicate and ex-tend these procedures to create an evidence base for theiruse. Given that many more practices have been describedfor use in inclusive settings than found in these 17 stud-ies (e.g.. Downing, 2008, 2010; Jorgensen, McSheehan,& Sonnenmeier, 2010), researchers are encouraged toconsider building an evidence base for these additionalstrategies.

ReferencesAgran. M., Blanchard, C , Wehmeyer, M., & Hughes, C. (2002).

Increasing the problem-solving skills of students with de-velopmental disabilities participating in general education.Remedial and Special Education, 23, 279-^288.

Billingsley, E, White, O. R., & Munson, R. (1980). Proceduralreliability: A rationale and example. Behavioral Assessment,2, 229-241.

Browder, D. M., Ahlgrim-Delzell, L., Spooner, E, Mims, P. J.,& Baker, J. N. (2009). Using time delay to teach pictureand word recognition to identify evidence-based practice forstudents with severe developmental disabilities. ExceptionalChildren, 75, 343-364.

Browder, D. M., Spooner, E, Ahlgrim-Delzell, L., Harris, A.,& Wakeman, S. (2008). A meta-analysis on teaching mathe-matics to students with significant cognitive disabilities. Excep-tional Children, 74, 407-432.

Browder. D. M., Trela, K., Courtade, G. R., Jimenez, B. A.,Knight, V, & Elowers, C. (2012). Teaching mathematicsand science standards to students with moderate and severe

28 Hudson et al.

developmental disabilities. The Journal of Special Education,46, 25-35.

Browder, D. M., Wakeman, S. Y., Spooner, F., Ahlgrim-Delzell,L., & Algozzine, B. (2006). Research on reading for studentswith significant cognitive disabilities. Exceptional Children,72, 392^08.

»Carter, E. W., Cushing, L. S., Clark, N. M., & Kennedy, C. H.(2005). Effects of peer support interventions on student' accessto the general curriculum and social interactions. Researchand Practice for Persons with Severe Disabilities, 30, 15-25.

Chard, D. J., Ketterlin-Geller, L. R., Baker, S. K., Doabler, C ,& Apichatabutra, C. (2009). Repeated reading interventionsfor students with learning disabilities: Status of the evidence.Exceptional Children, 75, 263-281.

*Collins, B. C , Branson, T. A., Hall, M., & Rankin, S. W. (2001).Teaching secondary students with moderate disabilities in aninclusive academic classroom setting. Journal of Developmentand Physical Disabilities, 13, 41-59.

*Collins, B. C , Evans, A., Creech-Galloway, C , Karl, J., &Miller, A. (2007). Comparison of the acquisition and main-tenance of teaching functional and core content sight words inspecial and general education settings. Focus on Autism andOther Developmental Disabilities, 22, 220-233.

*Collins, B. C, Hall, M., Branson, T. A., & Holder, M. (1999).Acquisition of related and unrelated factual information de-livered by a teacher within an inclusive setting. Journal ofBehavioral Education, 9, 223-237.

Copeland, S. R., & Cosbey, J. (2008/2009). Making progress inthe general curriculum: Rethinking effective instructional strat-egies. Research and Practice for Persons with Severe Disabilities,33-34, 214-227.

Downing, J. (2008). Including students with severe and multipledisabilities in typical classrooms (3rd ed.). Baltimore: PaulH. Brookes.

Downing, J. (2010). Academic instruction for students with mode-rate and severe intellectual disabiiities in inclusive ciassrooms.Thousand Oaks, CA: Corwin.

Horner, R. H., Carr, E. G., Halle, J., McGee, G., Odom, S., &Wolery, M. (2005). The use of single-subject research toidentify evidence-tsased practice in special education. Excep-tional Children, 71, 165-180.

Hudson, M. E. (2013). Effects of a peer-delivered system of leastprompts intervention package and academic read-aiouds onlistening comprehension for students with moderate intellectualdisability (Doctoral dissertation). Retrieved from ProQuest.(AAT3510205).

»Hudson, M. E., Browder, D. M., & Jimenez, B. (in press).Effects of a peer-delivered system of least prompts withadapted text read-alouds on listening comprehension for stu-dents with moderate intellectual disability. Education andTraining in Autism and Developmental Disabiiities.

Jackson, L. B., Ryndak, D. L., & Wehmeyer, M. L. (2008/2009).The dynamic relationship between context, curriculum, andstudent learning: A case for inclusive education as a researchbased practice. Research and Practice for Persons with SevereDisabiiities, 33-34, 175-195.

»Jameson, J. M., McDonnell, J., Johnson, J. W., Riesen, T., &Polychronis, S. (2007). A comparison of one-to-one embeddedinstruction in the general education classroom and one-to-onemassed practice instruction in the special education classroom.Education & Treatment of Children, 30, 23-44.

»Jameson, J. M., McDonnell, J., Polychronis, S., & Riesen, T(2008). Embedded, constant time delay instruction by peerswithout disabilities in general education classrooms. Inteiiec-tuai and Developmental Disabiiities, 46, 346-363.

»Indicates experiments included in the analysis.

»Jameson, J. M., Walker, R., Utley, K., & Maughan, R. (2012).A comparison of embedded total task instruction in teachingbehavior chains to massed one-to-one instruction for studentswith intellectual disabilities: Accessing general education set-tings and core academic content. Behavior Modification, 36,320-340. doi:10.1177/0145445512440574

»Jimenez, B. A., Browder, D. M., Spooner, E, & DiBiase, W.(2012). Inclusive inquiry science using peer-mediated em-bedded instruction for students with a moderate intellectualdisability. Exceptional Children, 78, 301-317.

»Johnson, J. W, McDonnell, J., Holzwarth, V. K, & Hunter, K.(2004). Tbe efficacy of embedded instruction for students withdevelopmental disabilities enrolled in general education classes.Journal of Positive Behavior Interventions, 6, 214—227.

Jorgensen, C M., McSheehan, M., & Sonnenmeier, R. M. (2010).The beyond access model. Baltimore: Paul H. Brookes.

Lane, K. L., Kalberg, J. R., & Shepcaro, J. C. (2009). An exami-nation of the evidence base for function-based interventionsfor students with emotional and/or behavioral disorders attend-ing middle and high schools. Exceptional Chiidren, 75,321-340.

Lee, S. H., Wehmeyer, M. L., Palmer, S. B., Soukup, J. H., &Little, T. D. (2008). Self-determination and access to the gen-eral education curriculum. The Journal of Special Education,42, 91-107.

McDonnell, J., Johnson, J. W., & McQuivey, C. (2008). Em-bedded instruction for students with developmental disabiiitiesin gênerai education classes (Vol. 6). Arlington, VA: Councilfor Exceptional Children.

»McDonnell, J., Johnson, J. W, Polychronis, S., Riesen, T.,Jameson, M., & Kercher, K. (2006). Comparison of one-to-one embedded instruction in general education classes withsmall group instruction in special education classes. Educa-tion and Training in Developmental Disabiiities, 41, 125-138.

»McDonnell, J., Mathot-Buckner, C , Thorson, R , & Fister, S.(2001). Supporting the inclusion of students with moderateand severe disabilities in junior high school general educationclasses: The effects of classwide peer tutoring, multi-elementcurriculum, and accommodations. Education & Treatment ofChiidren, 24, 141-160.

»McDonnell, J., Thorson, K, Allen, Q, & Mathot-Buckner, C.(2000). The effects of partner learning during spelling forstudents with severe disabilities and their peers. Journai ofBehavioral Education, 10, 107-121.

Mims, P., Hudson, M., & Browder, D. (2012). Using read aloudsof grade-level biographies and systematic prompting to pro-mote comprehension for students with moderate and severedevelopmental disabilities. Focus on Autism and Develop-mental Disabilities, 27, 65-78.

Mithaug, D. E., Wehmeyer, M. L., Agran, J., Martin, J. E., &Palmer, S. (1998). The self-determined leaming model of in-struction: Engaging students to solve their learning problems.In M. L. Wehmeyer & D. J. Sands (Eds.), Making it happen:Student involvement in educationai pianning, decision-makingand instruction (pp. 299-328). Baltimore: Paul H. Brooks.

National Governors Association Center for Best Practices,Council of Chief State School Officers. (2010). Common CoreState Standards. Retrieved from www.corestandards.org/.

National Institute of Child Health and Human Development.(2000). Report of the National Reading Panel. Teachingchildren to read: An evidence-based assessment of the scien-tific research Hterature on reading and its implications forreading instruction: Reports of the subgroups (NIH Pub-lication No. 00-4754). Washington, DC: US. GovernmentPrinting Office.

National Secondary Transition Technical Assistance Center.(2010). Decision ruies. Retrieved from http://www.nsttac.org/ebp/LiteratureReview.aspx.

Odom, S. L., Brantlinger, E., Gersten, R., Horner, R. H.,Thompson, B., & Harris, K. R. (2005). Research in special

Academic Learning Evidence-Based Practice Review 29

education: Scientific methods and evidence-based practices.Exceptional Children, 71, 137-148.

Palmer, S. B., Wehmeyer, M. L., Gipson, K., & Agran, M. (2004).Promoting access to the general curriculum by teachingself-determination skills. Exceptional Children, 70, 427^39.

*Polychronis, S. C , McDonnell, J., Johnson, J. W., Riesen, T.,& Jameson, M. (2004). A comparison of two trial distribu-tion schedules in embedded instruction. Focus on Autism andOther Developmental Disabilities, 19, 140-151.

»Riesen, T., McDonnell, J., Johnson, J. W, Polychronis, S., &Jameson, M. (2003). A comparison of constant time delayand simultaneous prompting within embedded instruction ingeneral education classes with students with moderate to se-vere disabilities. Journai of Behavioral Education, 12, 241-259.

Spooner, F., Knight, V., Browder, D., Jimenez, B., & DiBiase, W.(2011). Evaluating evidence-based practice in teaching sciencecontent to students with severe developmental disabilities. Re-search & Practice for Persons with Severe Disabilities, 36,62-75.

Test, D. W, Fowler, C. H., Brewer, D. M., & Wood, W. M. (2005).A content and methodological review of self-advocacy inter-vention studies. Exceptional Children, 72, 101-125.

Test, D. W., Fowler, C. H., Richter, S. M., White, J, Mazzotti, V,Walker, A. R., et al. (2009). Evidence-based practices in sec-

ondary transition. Career Development for Exceptional Indi-viduals, 32, 115-128. doi:10.1177/0885728809336859

Test, D. W., Richter, S., Knight, V., & Spooner, F (2010). A com-prehensive review and meta-analysis of the social stories liter-ature. Focus on Autism and Other Developmental Disabilities,26, 49-62.

Wehmeyer, M. L., Palmer, S. B., Agran, M., Mithaug, D. E.,& Martin, J E. (2000). Promoting causal agency: Theself-determined leaming model of instruction. ExceptionalChildren, 66, 439^53.

*Wolery, M., Anthony, L., Snyder, E. D., Werts, M. G, &Katzenmeyer, J (1997). Training elementary teachers toembed instruction during classroom activities. Education &Treatment of Children, 20, 40-58.

*Wolery, M., Werts, M. G., Snyder, E. D., & Caldwell, N. K.(1994). Efficacy of constant time delay implemented by peertutors in general education classrooms. Journai of BehavioralEducation, 4, 415^36.

Received: May 14, 2012Final Acceptance: October 10, 2012Editor in Charge: Susan Copeland

Copyright of Research & Practice for Persons with Severe Disabilities is the property ofTASH and its content may not be copied or emailed to multiple sites or posted to a listservwithout the copyright holder's express written permission. However, users may print,download, or email articles for individual use.