Inclusion.pdf

9
Comparative outcomes of two instructional models for students with learning disabilities: inclusion with co-teaching and solo-taught special educationPhilippe Tremblay Université Laval Key words: Inclusive practices, collaboration/consultation, special education, programme, learning disabilities. We compared two instructional models (co-teaching inclusion and solo-taught special education) for stu- dents with learning disabilities (LD) with regard to their effect on academic achievement and class attendance. Twelve inclusive classes (experimental group) and 13 special education classes (control group) participated in the study. In grade 1, there were eight inclusive classes and nine special edu- cation classes with a total of 353 students (195 without disabilities, 58 with LD in inclusion and 100 with LD in special education classes). The data were collected from academic tests. Although our results revealed no significant difference between the two models in terms of target population, objectives and assigned resources, significant differences were observed in the effects on student outcomes in reading/writing and on attendance, as the inclusion model was shown to be globally more effective com- pared with the special education setting. The role of the special educator in the inclusive classroom has gradually evolved towards a greater collaboration with the general education teacher. Formerly provided outside of school or the regular classroom, special education services are now taking place within the general classroom (pull-in) in a co-teaching approach with the general educator. For example, in 1995, the National Center on Education Restructuring and Inclusion reported that this co-teaching collaboration involving general and special education teachers was the most used service organisation model in the inclusion setting. Co-teaching is defined as ‘two or more professionals deliv- ering substantive instruction to a diverse, or blended group of students in a single physical space’ (Cook and Friend, 1995, p. 2) and thus comprises four basic characteristics: two qualified teachers (i.e., a general education teacher and a special education teacher), teaching that is dispensed by both teachers, a heterogeneous group of students (i.e., both general education and special needs students) and a shared setting (i.e., classroom) (Friend and Cook, 2007). This col- laboration consists of either temporary (a few hours per day, week, etc.) or permanent (full time, yearlong) co-teaching activities. Five major configurations compose the co-teaching operations: (1) support teaching, (2) parallel teaching, (3) station teaching, (4) alternative teaching, and (5) team teaching (Friend and Cook, 2007; Walther-Thomas et al., 2000). In a meta-analysis of qualitative research on co-teaching, Scruggs, Mastropieri and McDuffie (2007) found that in the traditional classroom setting, the dominant configuration was support teaching, where one taught and the other observed or assisted, and where the special education professional assigned to the class often held a subordinate role. Despite the enormous popularity of co-teaching, there is surprisingly little literature on the effectiveness of this approach. In examining meta-analyses on the effectiveness of a co-teaching model in an inclusive setting, Murawski and Swanson (2001) concluded that insufficient data pre- vented a clear determination, as only 6 out of the 99 studies reviewed met the meta-analysis selection criteria. The latter demonstrated a moderately significant size effect (+0.40). The authors thus called for more research with experimental and control groups with a more defined characterisation of the populations involved to better determine how co-teaching differs from other practices or when no special educational services are provided. In a literature review, Magiera and Zigmond (2005) observed that between 1986 and 2003, only 13 studies on co-teaching addressed student achievement. Seven of these studies showed significant positive academic gains for students with disabilities. Rea, McLaughlin and Walther-Thomas (2001) compared two integrative models for students with learning disabili- ties (LD): a pull-in model with co-teaching and a pull-out model in a resource class. Compared with the other groups, the outcomes of the pull-in students were superior in first language, mathematics and science. These authors also examined the social integration of these students and reported less negative behaviours and greater attendance. Fontana (2005) noticed that a co-taught model had a highly significant impact on the self-esteem of secondary level LD students in math but not in writing. In comparing the out- comes of students with LD over 2 years following a solo- teaching/co-teaching experiment, Hang and Rabren (2009) Journal of Research in Special Educational Needs · Volume 13 · Number 4 · 2013 251–258 doi: 10.1111/j.1471-3802.2012.01270.x 251 © 2012 The Author. Journal of Research in Special Educational Needs © 2012 NASEN. Published by John Wiley & Sons Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA

Transcript of Inclusion.pdf

Page 1: Inclusion.pdf

Comparative outcomes of two instructional modelsfor students with learning disabilities: inclusion withco-teaching and solo-taught special educationjrs3_1270 251..258

Philippe TremblayUniversité Laval

Key words: Inclusive practices, collaboration/consultation, special education, programme, learning disabilities.

We compared two instructional models (co-teachinginclusion and solo-taught special education) for stu-dents with learning disabilities (LD) with regard totheir effect on academic achievement and classattendance. Twelve inclusive classes (experimentalgroup) and 13 special education classes (controlgroup) participated in the study. In grade 1, therewere eight inclusive classes and nine special edu-cation classes with a total of 353 students (195without disabilities, 58 with LD in inclusion and 100with LD in special education classes). The data werecollected from academic tests. Although our resultsrevealed no significant difference between the twomodels in terms of target population, objectives andassigned resources, significant differences wereobserved in the effects on student outcomes inreading/writing and on attendance, as the inclusionmodel was shown to be globally more effective com-pared with the special education setting.

The role of the special educator in the inclusive classroomhas gradually evolved towards a greater collaboration withthe general education teacher. Formerly provided outside ofschool or the regular classroom, special education servicesare now taking place within the general classroom (pull-in)in a co-teaching approach with the general educator. Forexample, in 1995, the National Center on EducationRestructuring and Inclusion reported that this co-teachingcollaboration involving general and special educationteachers was the most used service organisation model inthe inclusion setting.

Co-teaching is defined as ‘two or more professionals deliv-ering substantive instruction to a diverse, or blended groupof students in a single physical space’ (Cook and Friend,1995, p. 2) and thus comprises four basic characteristics:two qualified teachers (i.e., a general education teacher anda special education teacher), teaching that is dispensed byboth teachers, a heterogeneous group of students (i.e., bothgeneral education and special needs students) and a sharedsetting (i.e., classroom) (Friend and Cook, 2007). This col-laboration consists of either temporary (a few hours per day,

week, etc.) or permanent (full time, yearlong) co-teachingactivities. Five major configurations compose theco-teaching operations: (1) support teaching, (2) parallelteaching, (3) station teaching, (4) alternative teaching, and(5) team teaching (Friend and Cook, 2007; Walther-Thomaset al., 2000). In a meta-analysis of qualitative research onco-teaching, Scruggs, Mastropieri and McDuffie (2007)found that in the traditional classroom setting, the dominantconfiguration was support teaching, where one taught andthe other observed or assisted, and where the specialeducation professional assigned to the class often held asubordinate role.

Despite the enormous popularity of co-teaching, there issurprisingly little literature on the effectiveness of thisapproach. In examining meta-analyses on the effectivenessof a co-teaching model in an inclusive setting, Murawskiand Swanson (2001) concluded that insufficient data pre-vented a clear determination, as only 6 out of the 99 studiesreviewed met the meta-analysis selection criteria. The latterdemonstrated a moderately significant size effect (+0.40).The authors thus called for more research with experimentaland control groups with a more defined characterisation ofthe populations involved to better determine howco-teaching differs from other practices or when no specialeducational services are provided. In a literature review,Magiera and Zigmond (2005) observed that between 1986and 2003, only 13 studies on co-teaching addressed studentachievement. Seven of these studies showed significantpositive academic gains for students with disabilities.

Rea, McLaughlin and Walther-Thomas (2001) comparedtwo integrative models for students with learning disabili-ties (LD): a pull-in model with co-teaching and a pull-outmodel in a resource class. Compared with the other groups,the outcomes of the pull-in students were superior in firstlanguage, mathematics and science. These authors alsoexamined the social integration of these students andreported less negative behaviours and greater attendance.Fontana (2005) noticed that a co-taught model had a highlysignificant impact on the self-esteem of secondary level LDstudents in math but not in writing. In comparing the out-comes of students with LD over 2 years following a solo-teaching/co-teaching experiment, Hang and Rabren (2009)

bs_bs_banner

Journal of Research in Special Educational Needs · Volume 13 · Number 4 · 2013 251–258doi: 10.1111/j.1471-3802.2012.01270.x

251© 2012 The Author. Journal of Research in Special Educational Needs © 2012 NASEN. Published by John Wiley & Sons Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK and350 Main Street, Malden, MA 02148, USA

Page 2: Inclusion.pdf

found that these students scored higher in reading and inmath under co-teaching than they did the previous year inthe presence of only one teacher. However, absenteeism washigher for these students during the second year (co-teaching) compared with the previous year. Moreover, theseauthors observed no differences between the studentswithout disabilities and their peers with LD on standardisedtests.

On the other hand, Murawski (2006) noted that studentswith LD in co-taught classrooms did not achieve betterstandard test scores than did those in resource, self-contained special education classrooms. Idol (2006) alsoreported that scores on high-stakes tests were little affectedby co-teaching (students with and without disabilities).Research by Walsh and Jones (2004) corroborated thatco-teaching was ‘a moderately effective procedure forinfluencing student outcomes’.

Finally, we found that very few studies addressedco-teaching in the inclusive model in the primary schoolsetting, particularly in grades 1 and 2. Indeed, most of theresearch on the subject concern secondary education.

This study consisted of a comparative analysis of twoinstructional models for students LD in the Communautéfrançaise de Belgique (French Community of Belgium):co-teaching in an inclusive setting and solo teaching in aself-contained special education classroom with pull-out(speech therapy, remedial teacher, etc.). We focused on theindicators that enabled us to compare the two models and tomeasure their effect on the students’ outcomes in reading/writing, mathematics and attendance. We performed a char-acterisation and a comparison of the populations in the twosubsamples (inclusion and special education), followed byan analysis of the resources assigned to each educationmodel, with a final comparison of the students’ outcomes toexternal testing as well their attendance levels.

In the French Community of Belgium, special educationdiffers from regular instruction in that it is proposed forstudents with specific needs in all three levels (kindergarten,primary and secondary) and consists of eight teachingmodels corresponding to different needs (physical, sensory,intellectual, etc.). Type 8 special education, reserved forchildren with learning difficulties, is defined as ‘specialeducation provided for the educational needs of childrenwith instrumental disabilities and designed for students forwhom the pluridisciplinary evaluation (. . . ) concludes thatdespite having normal levels of intelligence, hearing, andsight, they present difficulties in language or speech devel-opment and/or the acquisition of reading, writing, or calcu-lation, with a level of gravity requiring specific interventionwhich regular instruction alone cannot provide’ (Commu-nauté française de Belgique, 2004; author’s translation).

Over the last 20 years, the population benefiting from type8 instruction has considerably increased. For example, from1996–1997 to 2007–2008, the clientele went from 5138 to6086 students in type 8 for the entire French Community of

Belgium (Communauté française de Belgique, 2010), andduring the year preceding our study, only 15 students withLD had been integrated into regular classrooms.

Orientation of students towards type 8 special education(entry) is led by a ‘neutral’ organisation [PsychologicalMedical and Social Center (CPMS)] that provides non-binding recommendations supported by standard protocols(social, pedagogical, psychological, and medical evalua-tions and conclusions) and is mandatorily established toorient students towards special education.

MethodParticipantsMatched comparison groups were chosen for this study. In2007–2008, four grade 1 regular classes participated in ajoint inclusion/co-teaching experiment with students withLD. These four classes continued this experiment the fol-lowing year in grade 2, at which time four new grade 1classes were added to the initial sampling to form theexperimental group, chosen on a voluntary basis. Thespecial education classes thus formed the control group andwere subsequently selected by the author accordingto sociogeographic proximity (urban/rural, privileged/underprivileged) and academic level. For analysis purposes,the classes were grouped according to level. There were4.83 students with LD per class in inclusion compared with8.15 in special education. In grade 1 (eight inclusive classesand nine special education classes), a total of 228 studentsparticipated in the study, including 133 students withoutdisabilities, 37 students with LD in inclusion and 58 withLD in special education. In grade 2, there were only fourinclusive classes and four special education classes for atotal of 125 students: 62 students without disabilities, 21students with LD in inclusion and 42 students with LD inspecial education.

Following a multidisciplinary evaluation by the neutralorganisation (CPMS), the students with LD were referred toone or the other instructional model. Informed writtenconsent was obtained by the students’ parents for eachorientation, and the parents were informed of both theresearch objectives and their child’s participation in in-classassessments at the beginning and end of the school year.Each school relayed this information to the parents andcollected any eventual refusal to participate in the study.Only one parent refused consent.

The population under study was characterised by analysingthe independent CPMS evaluation reports and studentrecords. The analysis centred on intelligence quotient (IQ)(mandatory for this type of referral), gender, socioprofes-sional status, nationality, language spoken in the home andage. Results were compared at that time with those ofTremblay (2007), who examined the characteristics of thispopulation.

The average IQ was 80.95 [standard deviation (SD) = 8.21]in the inclusion model and 80.26 (SD = 11.24) in the specialeducation model. The comparative t-test for independent

Journal of Research in Special Educational Needs, 13 251–258

252 © 2012 The Author. Journal of Research in Special Educational Needs © 2012 NASEN

Page 3: Inclusion.pdf

samples confirmed the absence of significant differencesbetween the two groups (P = 0.728). Tremblay (2007)arrived at an average of 82 IQ points with a sample in theFrench Community of Belgium using the identical method-ology (n = 440).

Our analysis of the average age at the beginning of theschool year in each subsample revealed that this averagewas higher by approximately 1 year in the special educationclasses (M = 7.82 years; SD = 307 days) compared with theinclusion classes (M = 6.88 years; SD = 190 days) in grade1. The t-test showed significant differences between the twosamples (P = 0.010). The following year, the differencebetween the two groups was exactly 1 year: 6.99 years ininclusion (SD = 210 days) and 7.99 years in special educa-tion (SD = 317 days).

Concerning gender, the two samples were composed iden-tically of 2/3 boys. An over-representation of boys wastherefore present in the two models. In comparing this pro-portion with that in the total number of schools for studentswith LD within the French Community of Belgium (Com-munauté française de Belgique, 2010) and with Tremblay(2007), we found this same over-representation (2/3 boys)to be equally present for this type of special education.

The socioprofessional level of the parents (occupation) wasdivided into two categories: more favourable (managementand employee) and less favourable (manual worker andunemployed). Of interest was that 82.86% of the students ininclusion were in the less favourable category comparedwith 66.66% of students in special education. The chi-square independence test showed no significant difference(P = 0.089) between the two groups, although we didobserve significant inter-school variations because of thesize of the recruitment pool.

Regarding nationality, an average of 7.50% foreign studentswas recorded in the inclusion setting against 10.11% in thespecial education model, with no significant differenceobserved (P = 0.356). As for origin (parents’ place of birth),42% of the students in the inclusion model were of foreignorigin compared with 39% in the special education groups,which were very similar and thus showed no significantdifference.

Finally, the first language of 17.39% of the students inspecial education was other than French compared with

35% of the students in inclusion (French being the languageof instruction). The chi-square independence test revealedthat the two groups differed significantly on this point (P =0.027); however, this factor, among others, was found to belargely dependent upon the sociogeographical backgroundof each school.

Treatment and comparison conditionsThe first instructional approach consisted of a full-time1

co-teaching context involving a general education teacherand a special education teacher and centred on the inclusionof a group of students with LD within a larger group ofstudents without disabilities (Table 1). The participatingeducators were free to choose any teaching method. Thesecond instructional approach consisted of special educa-tion classes at the same academic level and was reserved forstudents with LD. Here, a special education teacherinstructed a smaller group of students. Speech therapy andremedial interventions were also provided during the day.On the one hand, this type of operation involved separatingthe students with LD within the classes and specialisedschools and, on the other, using pull-out for speech therapyand other remedial sessions (Table 2).

Regarding the available human resources (teacher, speechtherapist), on average, the students in the inclusion modelreceived 0.25 full-time equivalent (FTE) (SD = 0.071),although those in special education received 0.23 FTE(SD = 0.091). The t-test showed no significant difference onthis point between the two samples (P = 0.475). Theseresources varied from 0.14 to 0.44 FTE, depending on theclass. We also found that on average, the students with LDreceived more collective, small-group or individual in-classinterventions provided by a remedial teacher, psychomotortherapist, kinesiologist or speech therapist. Lastly, thespecial education classes distinguished themselves by anappreciable use of speech therapy (M = 1.34; SD = 1.03compared with M = 0.68; SD = 0.28). Thus, the amount ofintervention in speech therapy differed significantly (P =0.042) between the two groups. The included students ben-efited primarily from resources related to the presence inclass of a special education professional.

As for the teachers’ qualifications, each one possessed adegree in primary education, with the exception of twospecial education teachers in the inclusive setting who did

1 Grade 1 classes in one school shared the same special educator on a part-time basis.

Table 1: Characteristics of inclusion classesClasses iA iB iC iD iE iF iG iH iI iJ iK iL

Grade 1 1 1 1 1 1 1 1 2 2 2 2

Total number of students 17 18 23 24 24 29 17 18 17 19 23 24

Students with LD 5 7 4 3 6 6 3 3 6 6 6 3

Average speech therapy/remedial teacher/week 5 4 2 0 6 3 2 2 6 4 4 3

Average ratio Full-Time Equivalent 0.26 0.18 0.30 0.36 0.23 0.21 0.21 0.21 0.20 0.21 0.21 0.41

LD, learning disabilities.

Journal of Research in Special Educational Needs, 13 251–258

253© 2012 The Author. Journal of Research in Special Educational Needs © 2012 NASEN

Page 4: Inclusion.pdf

not have the necessary credentials (kindergarten teachers)and who worked together in the same class. The specialeducation teachers also differed in terms of their comple-mentary specialised training and were in fact almost half(3 out of 7) to have this degree in the inclusive co-teachingsetting compared with those in special education (3 out of13). Moreover, one teacher had an additional bachelor’sdegree in speech therapy and another, a masters in educa-tion sciences. In fact, in the inclusion model, five out ofseven teachers possessed an extra job-related degree. Incontrast, only one general educator had an additionaldiploma (masters in education sciences). The remedialteachers in the inclusive setting were shown to have lessexperience (M = 4.43 years, SD = 3.74) than were those inthe special education classes (M = 9.5 years, SD = 2.49). Incontrast, the general educators working in inclusion had anaverage of 7.75 years of experience (SD = 6.13). The pre-vious integrative experiment was very limited, as only oneteacher had experience in an inclusion setting.

The participating teachers initially attended an informationsession and one meeting with their future colleague duringthe school year prior to implementing the co-teachingmodel. The first part consisted of an individual meetingwith the volunteer teacher, the principal and their immedi-ate supervisor (school inspector, school counsellor, etc.)during which the teachers were informed of the projectgoals and limitations and were able to address any concerns.During the second interview, this time with their co-teacher,we went over the various possible configurations andexplored several themes of the co-teaching model and itsimplementation in terms of planning, schedules, academicprogramme, intervention plan, etc. One half-day was thenset aside to enable both co-teachers to meet alone in one oftheir classes (observation) so as to plan their next meeting.

Thereafter, two training days were organised during theyear with all of the participating co-teaching teams. Thefirst year, the first day was divided into three segments: atraditional training session during which the teachers weregiven content relative to co-teaching and individualisedmethods; the presentation of a video filmed in one of theclasses, which enabled the group to analyse the practicesthey saw and to apply the acquired training content; andfinally, focus groups where the teachers were asked to reactto various situations proposed by the moderator that

recalled various aspects that were potentially problematic tothe implementation of the co-teaching model in a classroom(advantages and disadvantages, major challenges, rolesharing, etc.). The second training day proceeded along thesame lines, with the exception of a discussion period in themorning and training content that addressed educationaldifferentiation.

MeasuresTo assess the effectiveness of the two instruction models,student achievement was investigated. In terms of studentoutcomes, the students were tested in class in reading/writing and mathematics in October and June of eachschool year. The first evaluation at the beginning of grade 1was done on an individual basis. The reading/writing testswere taken from the Observation Survey (Clay, 2003),although those in math were based on TEDI-Math (VanNieuwenhoven, Grégoire, & Noel, 2001). The validity andreliability of each task in the Observation Survey have beendocumented (Clay, 2003; Denton, Ciancio and Fletcher,2006), and the Observation Survey highly correlates withthe Iowa Test of Basic Skills (Rodgers et al., 2005; Tang andGómez-Bellengé, 2007). The French version was testedwith 360 students (Clay, 2003), although the TEDI-Mathwas experimented with 583 students (Belgian Francophoneand French) in levels kindergarten through grade 3. Thisbattery of tests constitutes a complete evaluation of thedifferent skills composing the basic competencies in arith-metic (Lion, 2002).

The June evaluation (PEDA-1Ca) consisted of both an indi-vidual assessment as well as a test for the entire class(Simonart, 1998) and was experimented with 290 BelgianFrancophone students. In grade 2, the October evaluationconsisted again of a group assessment (PEDA-1Cb),although that in June (PEDA-2C) was an individual test(Simonart, 1998). This evaluation involved 232 Belgianfrancophone students and the items pertained to the educa-tion programme in the French Community of Belgium. Thistest concerned three subjects: reading (oral reading andcomprehension), spelling (sentence dictation) and math-ematics (calculation, problems).

The data were analysed by means of SPSS. The level ofstatistical significance was determined as 0.05. A pairedsamples t-test was used to determine the presence ofany significant differences between the two subsamples

Table 2: Characteristics of special education classesClasses sA sB sC sD sE sF sG sH sI sJ sK sL sM

Grade 1 1 1 1 1 1 1 1 1 2 2 2 2

Total number of students 10 4 10 13 9 9 8 10 9 12 10 8 14

Students with LD 10 4 10 11 9 3 8 3 5 12 10 8 12

Average speech therapy/remedial teacher/week 6 11 7 13 30 4 24 7 8 7 15 18 21

Average ratio Full-Time Equivalent 0.15 0.44 0.30 0.20 0.29 0.15 0.30 0.14 0.19 0.13 0.19 0.29 0.16

LD, learning disabilities.

Journal of Research in Special Educational Needs, 13 251–258

254 © 2012 The Author. Journal of Research in Special Educational Needs © 2012 NASEN

Page 5: Inclusion.pdf

(co-teaching inclusion and solo-taught special education).A ranks analysis was also performed of all of the students’outcomes (LD and without disabilities).

ResultsAt the beginning of grade 1 (October), our student out-comes analysis (pre- and post-test) showed no significantdifferences between the two groups under study (studentswith LD in co-teaching inclusion and students with LD inspecial education) in reading/writing and math (t = 0.110,P = 0.913; t = -1.67, P = 0.097, respectively). In June, thet-test showed that these differences were significant (t =3.271, P = 0.002) in reading/writing for the students in theinclusion setting, but not in math (t = -0.363, P = 0.718)(Figure 1).

For the students with LD in inclusion, compared with theirpeers, the Wilcoxon test showed a decrease in rank betweenthe beginning and the end of grade 1, with the exception ofthe students with the lowest scores in reading/writing. Thiswould suggest that the students with lower outcomes pro-gressed better than the others did and explains the differ-ence in average between the two groups in this subject ingrade 1. However, the Wilcoxon test results were not sig-nificant for the two subjects evaluated (P = 0.147 and P =0.326, respectively). For the students with LD in the specialeducation model, we observed a significant drop in rankbetween the beginning and the end of grade 1 in bothsubjects evaluated (P = 0.000 and P = 0.000, respectively).

In grade 2, the students with LD in special education faredbetter in reading/writing and math compared with the stu-dents in inclusion, although the differences were not sig-nificant (t = -0.883, P = 0.382; t = -1.132, P = 0.263,

respectively). In June, we observed that the average devia-tion became positive in favour of the included students, asthe difference went from -0.67 to 1.35 in reading/writingand from -0.82 to 1.95 in math. The t-test did reveal,however, that these differences were not significant forthese two subjects (t = 1.802, P = 0.091; t = 1.726, P =0.079, respectively). The ranks analysis with the Wilcoxontest was significant in reading/writing (P = 0.035), but not inmath (P = 0.086). For the students in special education, weagain observed a significant decrease in achievement inmath (P = 0.021), but not in reading/writing (P = 0.070)(Figure 2).

For the grade 1 co-taught students with LD, compared withthe students without disabilities, the latter were shown toscore higher in reading/writing (t = -0.186, P = 0.853) thanin math (t = -2.588, P = 0.011), although these differenceswere only significant in math. At the end of grade 1, themeans difference between the two groups grew was signifi-cant (t = 2.155, P = 0.036; t = -5.680, P = 0.000). Early ingrade 2, we observed significant differences between thetwo groups to the advantage of the students without dis-abilities (t = -3.546, P = 0.001; t = -3.661, P = 0.001). InJune, however, despite the fact that the two means contin-ued to differ significantly in both subjects, the gap stabilisedand even decreased in math (t = -2.401, P = 0.022) and inreading/writing (t = -2.664, P = 0.009).

Although we observed a progression of the rank differencesbetween the students without disabilities and those with LDin the special education setting at the beginning of grade 1,the difference in the means between the two groups werelow and thus not significant in reading/writing and math(t = -0.399, P = 0,30; t = -0.424, P = 0.672). However, atthe end of this year, greater differences were apparentbetween the two groups and were statistically significant(t = -7.317, P = 0.000; t = -5.338, P = 0.000). In grade 2,marked differences between the two groups were observed

Figure 1: Performance gap between grade 1 studentswith learning disabilities (LD) in inclusion and in specialeducation

Figure 2: Performance gap between grade 2 studentswith learning disabilities (LD) in inclusion and in specialeducation

Journal of Research in Special Educational Needs, 13 251–258

255© 2012 The Author. Journal of Research in Special Educational Needs © 2012 NASEN

Page 6: Inclusion.pdf

in reading/writing (t = -3.429, P = 0.001) and in math(t = -2.502, P = 0.015), and in June, these differences wereshown to be even greater (t = -5.066, P = 0.000; t = -5.169,P = 0.000).

DiscussionWe examined and compared two instructional approachesfor students with LD (full-time co-teaching in an inclusivesetting and solo teaching in a self-contained special educa-tion setting). More specifically, we sought to measure theeffect of these two education models on student outcomesin reading/writing and mathematics.

The two groups (students with LD in co-teaching inclusionand in special education) presented generally similar char-acteristics. In addition, the t-test early in grade 1 alsorevealed no palpable differences between the two groups inboth reading/writing and mathematics. On the whole, bothsettings were provided with the same level of humanresources. Only in average age and language spoken in thehome was there any distinction. The students in these twogroups appeared to be representative of the reality observedelsewhere in special education with regard to previousstudies on the subject (Tremblay, 2007).

The impact of the two instructional models on studentachievement demonstrated that compared with students inspecial education, the students in the inclusive settingnoticeably progressed on the external evaluations inreading/writing between the beginning and the end of grade1 and grade 2, but the differences were only statisticallysignificant for grade 1. In mathematics, no significant dif-ference was observed in both years, yet in grade 2, theaverage deviation between the two cohorts was quite differ-ent. The student rank analysis shows that in grade 1, thenoticeable progress made by the included students inreading/writing pertained to those students with the lowestscores. In grade 2, the gains in reading/writing were bothgeneralised and significant. As for the special educationstudents, their scores decreased significantly between thebeginning and the end of grade 1, regardless of rank, and ingrade 2, the same tendency was observed, with the gapbetween groups increasing over time. These results alsoappear to correlate with those of Rea, McLaughlin andWalther-Thomas (2001), who reported superior outcomesby students in a pull-in setting over those in a pull-outsetting.

We sought to determine whether this difference in averageage between the two samples was likely to affect the testresults. To control this possible association between age andacademic outcome, we proceeded with a rank analysis.According to this variable, the results show no significantdifference in performance in math, whereas the older stu-dents in both models were more likely to have better out-comes in reading/writing early in the year and to maintaintheir rank at year’s end.

Our findings reveal relatively unequal outcomes by thethree groups under study, namely students with LD in inclu-

sion, students with LD in special education and studentswithout disabilities. The first two groups scored lower thandid the students without disabilities in the general educationsetting. However, the gap between the included studentswith LD and the students without disabilities appeared tostabilise or decrease by the second year. As for the studentsin special education, the gap significantly and systemati-cally increased compared with that of the two other groups.Hang and Rabren (2009) demonstrated that the rate of pro-gression of students with disabilities was not significantlydifferent from that of students without disabilities. Thesefindings appear to show that inclusion with co-teachingprovided students with LD with the necessary support foracademic achievement on standardised tests. Finally, con-trary to Fontana (2005), who observed an effect in math butnot in writing, our results tend to show the opposite, with asignificant effect in reading/writing rather than in math.Although not significant, the outcomes in math for the stu-dents in the inclusive model are positive. Furthermore, Fon-tana’s study regarded secondary level students, althoughour study focuses on the early primary school years forwhich little or no research of this type has been done in thispopulation of students with LD.

LimitationsAlthough a relatively apparent stabilisation of the perfor-mance gaps in grade 2 was observed between the includedstudents with LD and their pairs without disabilities, themore rapid pace of the general education model neverthe-less tends to ‘naturally’ produce these differences and toaccentuate them over time. As a result, these increasinglyobvious gaps in achievement levels make it difficult tomaintain certain students in these classes. Indeed, weobserved that 25% of the students in inclusion were redi-rected into special education at the end of grades 1 and 2. Inthe French Community of Belgium, a hierarchy between theinclusion model and special education model transpires inthe decisions to redirect these students towards largely seg-regative special education models. Aside from this, logi-cally speaking, these losses have the collateral effect ofincreasing its effectiveness by ‘lightening’ the load for thosestudents having the most difficulty keeping up. The mereexistence of these ‘losses’ thus constitutes a definite causeto question the effectiveness of the inclusion model.

Because of the relatively limited sample, it is difficult toextrapolate further. Indeed, on the basis of the availabledata, we can argue that the two subsamples were globallycomparable. Thus, despite being representative of certainaspects of the population in this type of special education inthe French Community of Belgium, the sampling was rela-tively limited in both groups, as only 12 classes in inclusionand 13 in special education were studied. As a result, it isdifficult to assert that these classes are representative ofwhat is being done (or can be done) in each setting.

Our assessment of the first 2 years of implementation of theinclusive co-teaching model in the French Community ofBelgium was basically a study of their ‘start-up’ period,as the classes were prototypes of the inclusive model.

Journal of Research in Special Educational Needs, 13 251–258

256 © 2012 The Author. Journal of Research in Special Educational Needs © 2012 NASEN

Page 7: Inclusion.pdf

Moreover, having only considered the first 2 years ofprimary school, this study lacks sufficient longitudinal dataon the compared achievement of the students within thecontext of these instructional models.

Implications for future researchThis study on the relative effectiveness of co-teaching withdata collected from experimental groups (inclusion) andcontrols (special education) enabled us to determine howco-teaching differs from other instructional models withregard to student achievement. Future research should belongitudinal so as to evaluate the effects of these twomodels on students over a long period and to verify whetherthese effects can in fact be maintained over time underroutine conditions.

In this study, the co-teaching involving a general educationteacher and a special education teacher was either full-timeor part-time, but in actual fact, this collaboration takes placemost often only a few hours a week. It would thus be ofinterest to compare various co-teaching configurations byconsidering this temporal dimension on the actual durationand intensity of this collaboration between these two edu-cators. This would be particularly relevant for smallerschools with limited resources. For example, if a specialeducation teacher co-teaches with several general teachers,certain factors related to effectiveness and efficiency maybe compromised (i.e., time allotted for planning, type ofco-teaching involved, etc.). Further studies should deter-mine the impact of the number of general educators withwhom a special education teacher can effectively co-teach.

Though most often present in the inclusive setting, thesestrategies are, however, not exclusive and may also be foundin or transferred to the special education setting; therein liesthe importance of extending and reinvesting the knowledgeproduced by research on the flexibility of these two instruc-tional models. In fact, the connection between educationmodels and practices merit further exploration.

Address for correspondencePhilippe Tremblay,Département d’études sur l’enseignement etl’apprentissage,Faculté des sciences de l’éducation,Université Laval,2320, rue des Bibliothèques,Bureau 1034,Québec, QC, G1V 0A6Canada.Email: [email protected].

ReferencesClay, M. (2003) Le Le sondage d’observation en

lecture-écriture [An observation survey of earlyliteracy achievement]. (G. Bourque, Trans.) Montréal,Canada: Éditions Chenelière/McGraw-Hill.

Communauté française de Belgique (2004) ‘Décret surl’enseignement spécialisé. Communauté française deBelgique.’ Moniteur belge, 03 June 2004.

Communauté française de Belgique (2010) ‘Lesindicateurs de l’enseignement.’ Brussels: ETNIC –Services des Statistiques de la Communauté françaisede Belgique.

Cook, L. & Friend, M. (1995) ‘Co-teaching: guidelinesfor creating effective practices.’ Focus on ExceptionalChildren, 28, pp. 1–16.

Denton, C. A., Ciancio, D. & Fletcher, J. (2006) ‘Validity,reliability, and utility of the observation survey ofearly literacy achievement.’ Reading ResearchQuarterly, 41, pp. 8–34.

Fontana, K. C. (2005) ‘The effects of co-teaching on theachievement of eighth-grade students with learningdisabilities.’ The Journal of At-Risk Issues, 11 (2), pp.17–23.

Friend, M. & Cook, L. (2007) Interactions: CollaborationSkills for School Professionals. (5th edn). New York:Pearson Education.

Hang, Q. & Rabren, K. (2009) ‘An examination ofco-teaching: perspectives and efficacy indicators.’Remedial and Special Education, 30 (5),pp. 259–68.

Idol, L. (2006) ‘Toward inclusion of special educationstudents in general education: a program evaluation ofeight schools.’ Remedial and Special Education, 27(2), pp. 77–94.

Lion, P. (2002) ‘Test Diagnostique des Compétences deBase en Mathématiques (TEDI-MATH).’ Cahiers de laSBLU, 11, pp. 29–33.

Magiera, K. & Zigmond, N. (2005) ‘Co-teaching inmiddle school classrooms under routine conditions:does the instructional experience differ for studentswith disabilities in co-taught and solo-taught classes?’Learning Disabilities Research and Practice, 20 (2),pp. 79–85.

Murawski, W. W. (2006) ‘Student outcomes inco-taught secondary English classes: how canwe improve?’ Reading/Writing Quarterly, 22, pp.258–68.

Murawski, W. W. & Swanson, H. L. (2001) ‘Ameta-analysis of co-teaching research: where are thedata?’ Remedial and Special Education, 22 (5), pp.258–67.

Rea, P., McLaughlin, V. L. & Walther-Thomas, C. (2001)‘Outcomes for students with learning disabilities ininclusive and pullout programs.’ Exceptional Children,68 (2), pp. 203–23.

Rodgers, E. M., Gómez-Bellengé, F. X., Wang, C. &Schultz, M. M. (2005, April) ‘Examination of thevalidity of the observation survey with a comparisonto ITBS.’ Paper presented at the annual meeting of theAmerican Educational Research Association inMontreal, Quebec.

Scruggs, T. E., Mastropieri, M. A. & McDuffie, K. A.(2007) ‘Co-teaching in inclusive classrooms: ametasynthesis of qualitative research.’ ExceptionalChildren, 73 (4), pp. 392–416.

Journal of Research in Special Educational Needs, 13 251–258

257© 2012 The Author. Journal of Research in Special Educational Needs © 2012 NASEN

Page 8: Inclusion.pdf

Simonart, G. (1998) Tests pédagogiques de premier cycleprimaire. Braine-le-Château, Belgium: ATM.

Tang, M. & Gómez-Bellengé, F. X. (2007, April)‘Dimensionality and concurrent validity of theobservation survey of early literacy achievement.’Paper presented at the annual meeting of the AmericanEducational Research Association in Chicago, Illinois.

Tremblay, P. (2007) ‘Évaluation de la validité et del’efficacité interne de l’enseignement spécialiséprimaire de type 8 en Wallonie.’ Éducation –Formation, pp. 286, 9–21.

Van Nieuwenhoven, C., Grégoire, J. & Noel, M. P. (2001)TEDI-MATH, Test Diagnostique des Compétences deBase en Mathématiques. Louvain-la-Neuve, Belgium:C. Edition ECPA.

Walsh, J. M. & Jones, B. (2004) ‘New models ofcooperatively teaching.’ Teaching ExceptionalChildren, 36, pp. 14–20.

Walther-Thomas, C. S., Korinek, L., McLaughlin, V. L. &Williams, B. T. (2000) Collaboration for InclusiveEducation: Developing Successful Programs. Boston,MA: Allyn and Bacon.

Journal of Research in Special Educational Needs, 13 251–258

258 © 2012 The Author. Journal of Research in Special Educational Needs © 2012 NASEN

Page 9: Inclusion.pdf

Copyright of Journal of Research in Special Educational Needs is the property of Wiley-Blackwell 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.