thm

asgo

nd behaviors with respect to children with SEBD.ns was a key inuence.essionsless wi

Keywords:less willing to work with children with SEBD. Findings suggest that school principals have a central role

(e.g., Fuchs & Fuchs, 1995; Kavale & Forness, 2000; Lindsay, 2007;Walther-Thomas, Korinek, & Willliams, 2000; Woolfson, 2011).

Mainstream teacher attitudes may be a contributory barrier tosuccessful inclusive practices (Avramadis, Bayliss, & Burden, 2000;Bender, Vail, & Scott,1995; De Boer, Pijl, &Minnaert, 2010). Teacherstend to be broadly positive about the principle of inclusion while at

Kalyva, 2007) to inuence teacher attitudes, while others havenot (Brady & Woolfson, 2008; Woolfson & Brady, 2009).

The inclusion of children with social, emotional and behavioraldifculties (SEBD1) has consistently been reported as a particularlyproblematic for teachers, and is accompanied by negative teaching

* Corresponding author. Tel.: 44 141 548 2580; fax: 44 141 548 4001.1 Different countries have adopted differing terminology to describe the same

group of children. As the present study was located in Scotland, the Scottish term,social, emotional and behavioral difculties (SEBD) was adopted in this paper.

Contents lists available at SciVerse ScienceDirect

Teaching and Teacher Education

.e

Teaching and Teacher Education 29 (2013) 46e52E-mail address: lisa.woolfson@strath.ac.uk (L.M. Woolfson).Segregating children into special needs and mainstreamschools prevents equal access to social and curricular opportunitiesand labels children (United Nations Convention on the Rights of theChild, 2009; UNESCO, 1994). Thus inclusion of children withdisabilities and special educational needs in mainstream educationhas been encouraged through legislation internationally for over 20years. It is unclear though whether inclusive education is effectivein terms of promoting positive educational and social outcomes,due to mixed results and a lack of well-designed studies in this area

the same time viewing its practical implementation as problematic(e.g., Avramidis & Norwich, 2002; Scruggs & Mastropieri, 1996).However it has been argued that neutral, even negative, attitudestoward inclusion may better characterize teacher viewpoints (DeBoer et al., 2010; Soodak, Podell & Lehman, 1998). Indeedteachers in mainstream schools were less positive about thepotential of children with learning disabilities than special schoolteachers. Some studies have reported in-service training (INSET),support (Cagran & Schmidt, 2011), and experience (Avramidis &Behavioral difculties in the classroomInclusive educationTheory of planned behaviorTeacher attitudesINSET

in promoting an inclusive ethos within their schools. INSET could focus more on challenging beliefs. 2012 Elsevier Ltd. All rights reserved.< The study examined teacher beliefs a< Perception of head teacher expectatio< Teachers who attended more INSET s< But more experienced teachers were

a r t i c l e i n f o

Article history:Received 30 April 2012Received in revised form20 July 2012Accepted 21 August 20120742-051X/$ e see front matter 2012 Elsevier Ltd.http://dx.doi.org/10.1016/j.tate.2012.08.006were more positive.lling to work with this group.

a b s t r a c t

The Theory of Planned Behavior (TPB) was used to examine relationships between teacher attitudes andbehavior toward children with social, emotional and behavioral difculties (SEBD). One hundred andeleven elementary school teachers completed questionnaires. Teacher perception of their schoolprincipals expectations (subjective norm) predicted teacher behaviors. Teachers who had attended morein-service training (INSET) sessions held more positive feelings, but teachers with more experience wereh i g h l i g h t sTeacher attitudes and behavior towardemotional and behavioral difculties inof the theory of planned behavior

Kate MacFarlane, Lisa Marks Woolfson*

School of Psychological Sciences and Health, University of Strathclyde, 40 George St., Gl

journal homepage: wwwAll rights reserved.e inclusion of children with social,ainstream schools: An application

w G1 1QE, Scotland, UK

lsevier .com/locate/ tate

g anattitudes (Cook, 2001; Cook, Cameron & Tankersley, 2007; Hastings& Oakford, 2003; Shapiro, Miller, Sawka, Gardill, & Handler, 1999).These are childrenwhose learning in the classroom is compromisedby complex and long-term difculties in managing their behavior,emotions and relationships (Simpson, Bloom, Cohen, Blumberg, &Bourdon, 2005). Unlike other groups of students with specialneeds, they are still as likely to be placed in specialist provision nowas 30 years ago (Cooper, 2004). This group is mainly male, witha majority from low socio-economic status backgrounds, and withlower educational attainment than their peers (Farrell &Tsakalidou, 1999; Simpson et al., 2005).

1. Applying the theory of planned behavior (TPB) toinvestigate teacher attitudes and behavior

The theory of planned behavior (Ajzen, 1991) provides a usefulframework for addressing the relationship between attitude andbehavior. It is an extension of the theory of reasoned action (Ajzen &Fishbein, 1980) and proposes that in order to predict a specicbehavior, attitudes, subjective norm, perceived behavioral controland behavioral intention in relation to that behavior need to betaken into consideration. Subjective norm is an individualsperception of how signicant others will approve of their behavior,perceived behavioral control is the perception of how easy/difcultit is to perform the behavior, and behavioral intention is willingnessto carry out the given behavior. TPB predicts that the behavior ismore likely to be performed if each component of these compo-nents is favorable. Specically the model posits that behavioralintention is determined byattitudes, subjective norm and perceivedbehavioral control. In turn, behavior is determined by attitudes,subjective norm and perceived behavioral control, mediated bybehavioral intention. Attitudes furthermore may be considered asconsisting of both cognitive (beliefs), and affective (feelings)dimensions (Ajzen, 1991; Eagly & Chaiken, 1993). Widely applied inhealth settings, TPB has been used to a lesser extent in educationalsettings, to predict, for example, teachers use of cooperativelearning in science classes (Lumpe, Czerniak & Haney, 1998),students learning behavior (Norwich & Rovoli, 1993), and studentbehavior toward peers with disabilities (Roberts & Smith, 1999).

There is a gap in the literature however with respect to theapplication of TPB to teacher attitudes and behavior toward chil-dren with special needs in inclusive settings. Using the searchengines Science Direct, Web of Knowledge and Google Scholar, thesearch term, theory of planned behavior combined with inclusion,teacher, education, mainstream and school, produced only sixstudies. Of these only two, Stanovich and Jordans (1998) Canadianstudy and Kuyini and Desais (2007) Ghanaian study, measuredactual behavior as an outcome variable, the others measuringbehavioral intention. Of these two studies, one (Stanovich & Jordan,1998) reported relationships that differed from the typical TPBmodel outlined above, in that they found subjective norm to be thestrongest predictor of effective teaching behavior with studentswith disabilities. This suggests the importance of school ethos asa predictor of teacher behavior. The other study measuringbehavior as an outcome (Kuyini & Desai, 2007) did not nd thisrelationship with subjective norm but its nding may have beenspecic to the educational context in a developing country.

Problems in operationalising TPB measures in a classroomcontext may also have contributed to these apparently conictingndings. Because it has been argued that perceived behavioralcontrol can be conceptualized as teacher efcacy (Ajzen, 1991;Eagly & Chaiken, 1993; Fishbein & Cappella, 2006), Stanovich andJordan employed the Teacher Efcacy Scale (TES; Gibson & Dembo,1984) as ameasure of perceived behavioral control. Teacher efcacy

K. MacFarlane, L.M. Woolfson / Teachinmay be more accurately measured using the Teachers Sense ofEfcacy scale (TSES) which has a more unied and stable factorstructure than the TES (Tschannen-Moran & Woolfolk-Hoy, 2001).

Measurement of subjective norm too may have been problem-atic. Stanovich and Jordan (1998) measured it by assessing schoolprincipals attitudes, whereas Kuyini and Desai (2007) measuredschool principals expectations. Subjective norm though is denedas an individuals perceptions of how signicant others will ratetheir behavior (Ajzen, 1991; Eagly & Chaiken, 1993). As it is widelyaccepted that the principal is the key gure in a school environ-ment (Goddard, Neumerski, Goddard, Salloum, & Berebitsky, 2010;Praisner, 2003), s/he is often referred to as the signicant other inthis area of research. Thus, rather than directly measuring princi-pals attitudes or expectations, subjective norm might be moreeffectively assessed by measuring teachers views about how theirschool principal would react to their inclusive behaviors.

The remaining four studies retrieved in the literature search(Batsiou, Bebetsos, Panteli, & Antoniou, 2008; Elik, Weiner, &Corkum, 2010; Oh et al., 2010; Palou & Norwich, 2002), did notactually measure behavior but focused on behavioral intention asthe main outcome variable. Batsiou et al.s study suggested thatexperience and training inuenced teachers beliefs and intentionsin relation to teaching pupils with special needs, while Palou andNorwich found that teachers who had more positive feelingstoward children with behavioral difculties, accompanied byhigher levels of perceived behavioral control, were more likely toreport the intention to behave inclusively. Subjective norm did notpredict behavioral intention in either Batsiou et al.s study or Palouand Norwichs study. Both Elik et al.s. study and Oh et al.s studywere with student teachers, whom their authors argued, had notyet had sufcient experience of teaching to be aware of subjectivenorms.

TPB studies of teacher inclusive behaviors are limited in number,may not have measured the intended variables, and have compo-nents of the model omitted. All aspects of the model need to beconsidered within a study in order to study the link between atti-tudes and behavior (Ajzen, 1991). Indeed omitting the measure-ment of behavior is perhaps the most important limitation ofstudies if they wish to make inferences regarding teachers inclu-sive practices.

1.1. The present study

There is then a lack of research investigating the link betweenteachers attitudes and behavior toward childrenwith special needsin mainstream schools, using TPB. The studys main aim was toinvestigate mainstream teachers attitudes and behavior withrespect to students with SEBD in particular, assessing all TPB vari-ables, measuring both cognitive and affective components of atti-tudes, and considering subjective norm and perceived behavioralcontrol as closely as possible to Ajzens (1991) conceptualizationof these. A second aim was to examine the teacher-relatedvariables which predict attitudes. In doing so it was hypothesizedthat:

1. Teachers beliefs, feelings, subjective norm and perceivedbehavioral control will each predict behavioral intention.

2. Teachers beliefs, feelings, subjective norm and perceivedbehavioral control will predict behavior, mediated by behav-ioral intention.

3. INSET and teaching experience will predict teachers beliefstoward the inclusion of children with SEBD.

4. INSET and teaching experience will predict teachers feelingstoward pupils with SEBD individually.

5. INSET and teaching experience will predict teachers willing-

d Teacher Education 29 (2013) 46e52 47ness to work with pupils with SEBD.

2. Method

2.1. Participants

General classroom teachers in mainstream schools froma convenience sample of 24 Scottish school districts were invited toparticipate in the study, out of 32 school districts in Scotland.Twelve districts granted permission and all 483 schools wereinvited to participate. Following up those 61 schools who respon-ded positively, two hundred and eighty-three questionnaire packswere distributed, of which 92 were completed and returned. Anemail with the link to the online survey was forwarded to 32schools instead of paper questionnaires, and seven schools askedfor both paper copies and online survey links (see Procedure). Theonline survey resulted in 32 further responses. Of the 124 ques-tionnaires that were returned, 13 were omitted from the analysisbecause they were not completed by general classroom teachers.This resulted in a sample of 111 participants (105 female and 6maleteachers) from a possible population of 22,851 elementary school

K. MacFarlane, L.M. Woolfson / Teaching an48teachers in Scotland as recorded in 2011 (Summary Statistics forSchools in Scotland, 2011). Mean class size of the sample was21.70 (SD 5.28).

As our regression analyses were designed to employ 2e5independent variables, a sample size of 111 was appropriate formultiple regression analysis as it was greater than 104 thenumber of independent variables, assuming a medium effect size(Tabachnick & Fidell, 2007). Post-hoc power analysis using theG*power program (Faul, Erdfelder, Lang, & Buchner, 2007) showedstatistical power of .97 for a xed-predictor model linear multipleregression predicting behavioral intention from beliefs, feelings,subjective norm and perceived behavioral control (four predictors)with a large effect size (f2 .35), a .001, andwith 110 participants.Similarly for the linear multiple regression: xed-predictor modelpredicting behavior from beliefs, feelings, subjective norm,perceived behavioral control and behavioral intention with a largeeffect size (f2 .35), a .05, 110 participants and ve predictors,post-hoc power analyses showed a statistical power of .99. For bothabove models, power is above Cohens (1988) recommended levelof .80.

Demographic teacher information is presented in Table 1 butbecause responses were anonymous at both teacher and schoollevel, we do not know which schools within a district completedquestionnaires and so cannot report school level demographicinformation.

Table 1Participants characteristics.

Characteristic Frequency (%) Mean (SD) Range

Highest level of teachingqualicationBachelor in Education 56 (50.5%)Postgraduate Certicate 21 (18.9%)Postgraduate Diploma 18 (16.2%)Masters 3 (2.7%)Other 13 (11.7%)

Age level taught4e6 year olds 30 (27.03%)7e10 year olds 38 (34.23%)11e12 year olds 43 (38.74%)

Teacher age (years) 40.20 (11.49) 22e63General teaching experience

(years)13.78 (10.25) 1e42

Teaching experience (SEBD)(years)

11.47 (9.43) 1e42

No. of SEBD INSET sessions 4.49 (3.63) 0e20

attended2.2. Materials

2.2.1. Multidimensional Attitudes Toward Inclusive Education Scale(MATIES: Mahat, 2008)

Teacher attitudes, both beliefs and feelings, were measuredusing the cognitive and affective subscales of the MultidimensionalAttitudes Toward Inclusive Education Scale (MATIES: Mahat, 2008).Each subscale contains six items modied in the present study bydescribing the child in each item as a child with SEBD. Participantswere asked to indicate the degree to which they agreed or dis-agreed with each statement on a nine point Likert rating scaleanchored at ve points with: strongly disagree, disagree, neitheragree nor disagree, agree, strongly agree. As all scales which reectthe determinants of behavior should have an equal number of scalepoints for equal comparison (Ajzen, 1991), it was altered in thepresent study to a nine point scale to correspond with the otherquestionnaires. Two of the items on the cognitive subscale and eachitem on the affective subscale were reverse coded to ensure thathigher scores indicated more positive attitudes toward the inclu-sion of children with SEBD. With the present sample, Cronbachsalphas were adequate with both scales at .75, similar to reliabilitiesreported by Mahat (2008).

2.2.2. Teachers Sense of Efcacy Scale (TSES: Tschannen-Moran &Woolfolk Hoy, 2001)

The TSES was used to measure teachers perceived behavioralcontrol toward the inclusion of children with SEBD, as percep-tion of self-efcacy, in this case self-efcacy in the classroom,can be conceptualized as a form of perceived behavioral control(Eagly & Chaiken, 1993; Fishbein & Cappella, 2006). The shorter12 item version was employed in order to minimize the timerequired to complete the questionnaire pack. It contained threeefcacy subscales in relation to: instructional strategies, class-room management, and student engagement, with high internalconsistency reported (a .9: Tschannen-Moran & Woolfolk Hoy,2001). In the present study the scale was adapted to measureteaching self-efcacy toward working inclusively with childrenwith SEBD in particular. Participants were asked to indicate theextent to which they felt they could do something to supportchildren with SEBD on a nine point Likert scale anchored at vepoints with the statements: nothing, very little, some inuence,quite a bit and a great deal. Higher scores reect higherperceptions of teaching efcacy. Reliability of the revised scalewas .93.

2.2.3. Teachers Subjective Norm Scale (modied version of theTSES: Tschannen-Moran & Woolfolk Hoy, 2001)

As was argued earlier, because the principal is the key gure ina school, the measurement of teachers views about how theirschool principal would react to their inclusive behaviors wouldseem to be an appropriate measure of subjective norm in a schoolenvironment. Thus was measured using a modied version of theTSES was used to measure subjective norm (teachers perceptionsof their school principals views) about their working inclusivelywith pupils with SEBD.

It should be noted that in Scotland the school principal isreferred to as the head teacher, with the same school managementrole and responsibilities that accompany the title of school prin-cipal in other countries. Each question therefore now began Rateyour head teachers approval of how you., followed by the sameitems as in the modied TSES, referring to children with SEBD.Reliability of the scale was high with Cronbachs alpha .95.Participants were asked to rate their principals approval of howthey performed each of the twelve strategies on a nine point Likert

d Teacher Education 29 (2013) 46e52scale anchored at ve points with: strongly disapprove, disapprove,

tion will change given a one standard deviation change in eachpredictor. They therefore provide a better indication of each

Table 2Participants mean scores.

Mean (SD) Standard error Range

AttitudeCognitive 6.44 (1.15) .11 3.83e9.00Affective 6.20 (1.39) .14 2.50e9.00

Subjective norm 7.76 (.83) .08 5 .00e9.00Perceived behavioral control 6.58 (1.08) .10 4.17e9.00Behavioral intention 7.74 (.01) .10 1.00e9.00Behavior 4.03 (.60) .06 2.83e5.00

Table 3Correlations between each of the theory of planned behavior measures.

Variable 2 3 4 5 6

1. Beliefs .51** .17* .44** .52** .112. Feelings .30* .44** .43** .28*3. Subjective norm .50** .24* .49**4. Perceived behavioral control .64** .37**5. Behavioral intention .31**6. Behavior

g and Teacher Education 29 (2013) 46e52 49neither approve nor disapprove, approve, strongly approve. Higherscores represent higher perceptions of approval.

2.2.4. Teachers Willingness to Work with Severe Disabilities Scale(TWSD: Rakap & Kaczmarek, 2010)

The behavioral subscale of the TWSD was employed to measurebehavioral intention to promote the inclusion of children withSEBD. It consisted of a vignette describing a child with SEBD fol-lowed by eight items which asked teachers how committed theywould be to having the child in their classroom and to learningskills to support this. Teachers indicated the extent to which theyagreed/disagreed with each item on a nine point Likert scaleanchored at ve points with: strongly disagree, disagree, undecided,agree, strongly agree. Higher scores indicated a stronger intention toinclude children with SEBD. Rakap and Kaczmarek (2010) reporteda high internal consistency of a .94, which was replicated withthe present sample.

2.2.5. Adaption Evaluation Instrument (AEI: Schumm & Vaughan,1991)

Behavior was measured using the AEI, which evaluates views onthe feasibility and desirability of 30 planning, social, instructionaland curricular adaptions that promote the inclusion of childrenwith additional support needs (Schumm & Vaughan, 1991). Thescale was revised in this study to measure teachers inclusivebehavior specically toward children with SEBD. The term main-streamed student (MS) with replaced with SEBD and item 18, usecomputers was removed as using computer-aided technology isnow routine. Following initial piloting of the questionnaire packwith nine teachers, in order to reduce possible social desirabilityeffects, scale points were reworded into a ve point Likert scalewhere teachers were asked to indicate the frequency with whichthey felt they engaged in each adaption: occasionally, sometimes,quite often, frequently, all the time. Unlike responses on the otherscales, the pilot groups responses had been mainly distributedacross the two points, frequently and nearly all the time and it wasfelt that teachers were unlikely to disclose that they never or hardlyever engaged in inclusive behaviors. Higher scores reect higherengagement in inclusive behavior. Schumm and Vaughan (1991)reported high internal consistency values of .97 when they askedteachers to rate the feasibility of the adaptions and .97 whenteachers rated their desirability. With the present sample, Cron-bachs alpha was .95.

2.3. Procedure

Schools who agreed to participate were given the option ofcompleting the online questionnaire pack via Qualtrics SurveySoftware or the paper version by post with stamped addressedenvelopes for return. Questionnaires were presented in randomorder to reduce the possibility of order effects.

3. Results

Log transformations were applied to deal with outliers. Thevariable, teaching experience, had one outlier (reporting 42 yearsexperience), subjective norm had two outliers (both with a lowscore of 5), behavioral intention had two outliers (low scores of 5.05and 5.15), and INSET had three outliers (a high number of 15,16 and20 sessions). Log transformation was applied, as suggested byTabachnick and Fidell (2007) which pulled outliers from the vari-ables teaching experience (SEBD), INSET sessions attended (SEBD)and subjective norm closer to the center of the distribution.However one outlier remained within the variable behavioral

K. MacFarlane, L.M. Woolfson / Teachinintention: case 103 (low score of 5.05). When square roottransformation and reciprocal transformation also had no impactthe second option was to change the extreme score to one unitlarger or smaller than the next highest score (Field, 2009;Tabachnick & Fidell, 2007). As this procedure was not successful inbringing case 103 nearer to the center of the distribution, it wasremoved. Deletion of outliers is only recommended after trying thetwo options rst (Tabachnick & Fidell, 2007). After the log trans-formation and outlier deletion, assumptions of homoscedasticityand normality were now met for all regression analyses.

Participants mean scores are displayed in Table 2. On average,participants held moderately positive attitudes and perceivedbehavioral control, and high levels of subjective norm and behav-ioral intention concerning the inclusion of pupils with SEBD in theirclassroom. They also reported on average that they frequentlyengaged in inclusive practices. Bivariate correlations are presentedin Table 3.

3.1. Predicting behavioral intention to engage in inclusive practicesfrom beliefs, feelings, subjective norm and perceived behavioralcontrol

Standard multiple regression analysis demonstrated thatoverall there was a signicant relationship between predictorsand behavioral intention to include children with SEBD,F(4,105) 24.81, p < .001, R2 .49. R2 > .25 meet the criterion fora large effect size (Cohen, 1988). Thus a large proportion (49%) ofthe variability in teachers behavioral intention toward the inclu-sion of children with SEBD was predicted by their beliefs, feelings,subjective norm and perceived behavioral control in relation toworking inclusively with pupils with SEBD. Unstandardizedregression coefcients (B), standardized regression coefcients (b),semipartial correlations e the relationship between (sr2), R2 andadjusted R2 are displayed in Table 4. Unstandardized regressioncoefcients (B) highlight the nature of the relationship betweenbehavioral intention and each predictor, both in terms of directionand how much behavioral intention scores would change givena one unit increase in each of the predictor scores. However,whereas unstandardized regression coefcients are not directlycomparable, standardized regression coefcients (b) are, as theyindicate the number of standard deviations that behavioral inten-**p < .001; *p < .05.

g anpredictors relative importance in the model. The R2 value suggestshow much of the variance in behavioral intention is accounted forby the regression model in the sample, whereas the value ofadjusted R2 indicates the amount of variance in behavioral inten-tion would be accounted for if the model is from the populationfrom which the sample was drawn. Given that R2 and adjusted R2

are similar at .49 and .47, the model may be considered to gener-alize to the wider population of teachers. The semipartial correla-tions (sr2) when combined indicate the amount of R2 that can beattributed to the signicant predictors: beliefs and perceivedbehavioral control.

Table 4 therefore highlights that when considered individually,only teachers beliefs and perceived behavioral control predictedteachers intention to include children with SEBD in a positivedirection. Teachers feelings or subjective norm did not signicantlypredict behavioral intention. Teachers who held more positivebeliefs and felt they had a higher level of perceived behavioralcontrol reported a higher level of intention to promote the inclu-sion of children with SEBD. Of these two predictor variables,however, teachers perceived behavioral control was slightly moreimportant as indicated by the squared semipartial correlations.Although bivariate correlations between feelings and behavioralintention and between subjective norm and behavioral intentionwere signicant at, r (110) .43, p < .001 (one-tailed) and r(110) .24, p< .02 (one-tailed), respectively, these variables did notpredict behavioral intention individually. This may indicate that therelationship between variables was mediated by, or redundant to,the relationship between behavioral intention and the otherindependent variables (Tabachnick & Fidell, 2007). Post-hoccorrection indicated that this was not the case for subjectivenorm, F(4,105) 1.12, p > .05, but for feelings, the test was signif-icant, F(4,105) 6.09, p < .05. Thus the relationship betweenteachers feelings toward children with SEBD and behavioralintention was mediated by the relationships between beliefs andintention and between perceived behavioral control and intention.

Table 4Predicting teachers behavioral intention from their beliefs, feelings, subjectivenorm and perceived behavioral control.

B SE b sr2(unique)

Beliefs .04** .02 .25 .04Feelings .02 .01 .08Subjective norm .12 .10 .12Perceived behavioral control .10* .02 .37 .17

R2 .49aAdjusted R2 .47R .70**

*p < .05, **p < .001.a Unique variability .21; shared variability .28.

K. MacFarlane, L.M. Woolfson / Teachin503.2. Predicting inclusive behavior from beliefs, feelings, subjectivenorm, perceived behavioral control and behavioral intention

There was a signicant relationship between predictors andteachers inclusive behavior toward children with SEBD,F(4,105) 7.59, p < .001, R2 .31. Thus a large proportion (31%) ofthe variability in teachers behavior toward childrenwith SEBDwaspredicted by their beliefs, feelings, subjective norm, perceivedbehavioral control and behavioral intention, in relation to workinginclusively with pupils with SEBD. However individually, onlyteachers subjective norm predicted their inclusive behavior, asdemonstrated by the largest normalized beta weights and signi-cance level (b .35, p < .001). Thus, teachers with high levels ofsubjective norm reported that they acted more inclusively withchildren with SEBD than teachers with lower levels. Bivariatecorrelations between feelings, perceived behavioral control,behavioral intention and behavior were signicant (see Table 3) butdid not predict behavior on an individual basis. This may indicatethat these relationships were mediated by, or redundant to, therelationship between behavior and subjective norm. Again usingpost-hoc correction for each variable, the relationships betweenbehavior and feelings and behavior and behavioral intention werefound to be non-signicant at F(5,105) 1.76, p > .05 andF(5,105) 2.20, p > .05, respectively, indicating that these rela-tionships were not mediated by the relationship between subjec-tive norm and behavior. However, for the relationship betweenperceived behavioral control and behavior the test was signicant;F(5,105) 3.28, p < .05. Thus the relationship between perceivedbehavioral control and behavior was mediated by the relationshipbetween subjective norm and behavior.

3.3. Predicting teacher beliefs toward the inclusion of pupils withSEBD from INSET and experience

The relationship between the predictors and teachers beliefstoward the inclusion of children with SEBD was non-signicant,F(2,107) 1.00, p .37, R2 .02. In addition neither INSET(b .13, p .191) nor experience (b .02, p .995), individuallypredicted beliefs. Teachers with less experience and training heldsimilar beliefs toward the inclusion of children with SEBD asteachers with more experience and training.

3.4. Predicting teachers feelings toward the inclusion of childrenwith SEBD from INSET and experience

There was a signicant relationship between predictors andteachers feelings toward including children with SEBD,F(2,107) 3.28, p < .05, R2 .06, indicating a small proportion ofthe variability (6%) in teachers feelings can be explained by INSETcourses attended and teaching experience of SEBD.

INSET signicantly predicted teachers feelings in a positivedirection (b .20, p < .05) but teaching experience signicantlypredicted teachers feelings in a negative direction (b .21,p < .05), such that teachers who attended more INSET courses heldmore positive feelings toward children with SEBD than teacherswho attended less. On the other hand, those with less experience inteaching children with SEBD held more positive feelings thanteachers with more experience.

3.5. Predicting teachers willingness to work with pupils with SEBDfrom INSET and experience

There was a signicant moderate relationship between predic-tors and teachers willingness to include children with SEBD,F(2,107) 3.49, p< .05, R2 .06. Therefore only 6% of the variabilityin teachers willingness to include pupils with SEBD was explainedby INSET and teaching experience with pupils with SEBD. Teachingexperience (SEBD) predicted teachers willingness to work withchildren with SEBD in a negative direction, (b .25, p < .02)whereas INSET did not have any predictive power, (b .15, p .115).As teachers experiences of teaching children with SEBD increased,their willingness to work with this particular group of childrendecreased.

4. Discussion

The present study employed TPB to examine teachers attitudesand behavior toward childrenwith SEBD. It was found that teacherswho held more positive beliefs and higher levels of perceived

d Teacher Education 29 (2013) 46e52behavioral control (teaching self-efcacy) had a higher level of

g anbehavioral intention to engage in inclusive practices in workingwith children with SEBD. These ndings support those of Oh et al.(2010) and Palou and Norwich (2002). Furthermore, subjectivenorm, as measured by teachers views of their school principalsexpectations, predicted teacher behavior (social, instructional,curricular adaptations) but not behavioral intention (willingness toworkwith childrenwith SEBD). Teachers with high subjective normlevels had similar levels of intention to engage in inclusive behav-iors compared to teachers with lower levels of subjective norm. Inthe wider theory of planned behavior literature too, subjectivenorm is often either found to be a weak predictor of behavioralintention or has no predictive power (Armitage & Conner, 2001).

We found subjective norm though to be the only predictor ofteacher behavior suggesting it may therefore be the most impor-tant determinant of teaching adaptive behaviors in relation toworking inclusively with children with SEBD. If this is the case,school principals have an important role in communicating clearexpectations of an inclusive ethos to staff, providing them withappropriate support and training, and promoting a collective senseof efcacy. While this nding of the importance of subjective normin predicting teacher inclusive behavior supports that of Stanovichand Jordans (1998) study, it corresponds less well with TPB nd-ings in health settings, where behavioral intention is typically thestrongest predictor of behavior (Ajzen, 1991). Also, with regards topredicting behavior intention rather than behavior, teachers beliefsand level of perceived behavioral control were found to predictbehavioral intention. The relationship between teachers feelingsand behavioral intention was mediated by the relationshipbetween beliefs and intention, and perceived behavioral controland intention. This suggests that holding positive feelings towardchildren with SEBD may lead to positive beliefs and high perceivedbehavioral control levels, which in turnmay lead to a higher level ofbehavioral intention.

In TPB though, attitudes, subjective norm and perceivedbehavioral control predict behavior, mediated by behavioralintention, while in this study subjective norm alone predictedbehavior but not behavioral intention. Perhaps TPB does nottransfer directly from health settings to education, or to thesespecic teacher behaviors regarding inclusive education? In orderto be perceived by their principal as behaving professionally,teachers may require to suppress personal attitudes and theirperceived behavioral control, in favor of the subjective norm. It hasbeen argued that by adding additional variables to the TPB modelthe prediction of a specic behavior can be substantially improved(e.g., Heath & Gifford, 2002), so it may be that additional teacher-,environment- and child-related variables need to be added tobetter predict teacher intention and behavior relationships inschools. For example, teacher-related variables such as self-identity, openness to experience, and enthusiasm have beenidentied as important for effective teaching (Avramidis &Norwich, 2002; Kunter, Frenzel, Nagy, Baumert, & Pekrun, 2011).

Attendance at INSET sessions positively predicted teachersfeelings towardpupilswith SEBD, althoughdidnotpredict beliefs, orwillingness to work with children with SEBD. INSET was measuredhere as the number of training courses attended. It should be notedhowever that data were not gathered in the present study oncontent, duration and intensity of courses, which may be usefulvariables for future studies to examine (Snyder et al., 2011). Thisnding on INSET, together with our nding that teachers withmorepositive beliefs and higher levels of self-efcacy reported greaterintention and commitment to having children with SEBD in theirclassrooms and to learn new skills to facilitate this, suggests thataiming to challenge positive beliefs and develop high levels ofteaching efcacy in relation to working with children with SEBD

K. MacFarlane, L.M. Woolfson / Teachinshould be a key focus of initial teacher education, INSET, and post-graduate professional development courses. Effective teachereducation on exceptional groups of children should not only includespecialized knowledge and information, but also should addressteacher skill development for inclusion, as well as being contextu-alized to a greaterextent to challenge teacher beliefs about problemsin learning being located within the child (e.g., Brady & Woolfson,2008; Symeonidou & Phtiaka, 2009; Woolfson & Brady, 2009).

In addition, teaching experience predicted teachers feelings andwillingness to work with children with SEBD but in a negativedirection. Thus teachers had greater experience of children withSEBD held less positive feelings and were less willing to work withthis group than teachers with less experience, a nding thatsupports Forlin, Douglas and Hatties (1996) study. Two Greekstudies though reported apparently contradictory ndings: Batsiouet al. (2008) reported experience and attitudes to be positivelycorrelated, while Avramadis et al. (2000) found no relationship.However it should be noted that these studies were not focused onchildren with SEBD but on the wider group of children with specialeducational needs. It may be that it is the nature of the experiencethat is important, with positive experiences associatedwith positiveteacher attitudes (Praisner, 2003). As children with SEBD oftenpresent a range of challenging behaviors it is possible that the moreteachers experience these behaviors the more negative theybecome. Another plausible explanation for this nding may be thatteacherswithmore experience are thosewho completed their initialteacher training before inclusion was a routine part of mainstreamteaching and who have retained these attitudes formed duringteacher training (Brady & Woolfson, 2008; Hastings & Oakford,2003;Woolfson & Brady, 2009;Woolfson, Grant, & Campbell, 2007).

A study limitationwas its use of self-report measures. Where allquestionnaires in a study utilize self-report methods, there isa possibility of measurement bias because all variables have beenmeasured by the same respondent using a common method(Podsakoff, MacKenzie, Lee, Podsakoff, 2003). Self-report measurescan increase the likelihood of social desirability bias (Ganster,Hennessey & Luthans, 1983; King & Bruner, 2000; Thomas &Kilmann, 1975) in teacher responses about inclusive attitudes andprofessional practices. In the present study as condentiality wasassured and indeed as there was a range of responses for eachmeasure in both positive and negative directions, social desirabilitywas unlikely to have had a major effect.

4.1. Conclusions and implications

Investigating the determinants of teachers attitudes andbehavior and their relative importance is crucial for improvingteaching practices, initial teacher education and professionaldevelopment opportunities for effective inclusion of children withspecial needs and in particular with SEBD, the focus of the study.Teacher perception of their principals expectations was identiedas the only signicant predictor of teaching behavior in relation tochildren with SEBD. This implied that school principals havea crucial rolewithin their school to communicate their expectationsregarding inclusive practices clearly to their teaching staff. Inaddition, teachers beliefs and perceived behavioral control pre-dicted behavioral intention. This suggests that initial teachereducation, INSET and post-graduate courses should focus on chal-lenging teacher beliefs and developing teaching self-efcacy con-cerning children with SEBD, to a greater extent. As well asimplications for teacher education and day-to-day teaching prac-tices, there are also research implications. Future research shouldcontinue to take into consideration the multidimensional nature ofattitudes, the role of other teaching staff in the promotion ofinclusion and the level of specicity in relation to variables which

d Teacher Education 29 (2013) 46e52 51may inuence attitudes and behavior.

Kunter, M., Frenzel, A., Nagy, G., Baumert, J., & Pekrun, R. (2011). Teacher enthu-

K. MacFarlane, L.M. Woolfson / Teaching and Teacher Education 29 (2013) 46e5252References

Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and HumanDecision Processes, 50, 179e211.

Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and predicting socialbehavior. Prentice-Hall: Englewood Cliffs NJ.

Armitage, C. J., & Conner, M. (2001). Efcacy of the theory of planned behavior:a meta-analytic review. British Journal of Social Psychology, 40, 471e499.

Avramadis, E., Bayliss, P., & Burden, R. (2000). Student teachers attitudes towardsthe inclusion of children with special educational needs in the ordinary school.Teaching and Teacher Education, 16(2), 277e293.

Avramidis, E., & Kalyva, E. (2007). The inuence of teaching experience andprofessional development on Greek teachers attitudes towards inclusion.European Journal of Special Needs Education, 22(4), 367e389.

Avramidis, E., & Norwich, B. (2002). Teachers attitudes towards integration/inclu-sion: a review of the literature. European Journal of Special Needs Education,17(2), 129e147.

Batsiou, S., Bebetsos, E., Panteli, P., & Antoniou, P. (2008). Attitudes and intentions ofGreek and Cypriot primary education teachers towards teaching pupils withspecial education needs in mainstream schools. International Journal of InclusiveEducation, 12(2), 201e219.

Bender, W., Vail, C., & Scott, K. (1995). Teachers attitudes towards increasedmainstreaming: implementing effective instruction for students with learningdisabilities. Journal of Learning Disabilities, 28(2), 87e94.

Brady, K., & Woolfson, L. (2008). What teacher factors inuence their attributionsfor childrens difculties in learning? British Journal of Educational Psychology,78, 527e544.

Cagran, B., & Schmidt, M. (2011). Attitudes of Slovene teachers towards the inclu-sion of pupils with different types of special needs in primary school. Educa-tional Studies, 37(2), 171e195.

Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.).Mahwah, NJ: Lawrence Erlbaum Associates.

Cook, B. G. (2001). A comparison of teachers attitudes toward their includedstudents with mild and severe disabilities. Journal of Special Education, 34(4),203e213.

Cook, B. G., Cameron, D. L., & Tankersley, M. (2007). Inclusive teachers attitudinalratings of their students with disabilities. Journal of Special Education, 40(4),230e238.

Cooper, P. (2004). Is inclusion just a buzz word? Emotional and Behavioral Dif-culties, 9(4), 219e222.

De Boer, A., Pijl, S. J., & Minnaert, A. (2010). Regular primary schoolteachers atti-tudes towards inclusive education: a review of the literature. InternationalJournal of Inclusive Education, 1(1), 1e24.

Eagly, A. H., & Chaiken, S. (1993). The psychology of attitudes. Fort Worth TX: Har-court Brace Jovanovich College Publishers.

Elik, N., Weiner, J., & Corkum, P. (2010). Pre-service teachers open-mindedthinking dispositions, readiness to learn, and attitudes about learning andbehavioral difculties in students. European Journal of Teacher Education, 33(2),127e146.

Farrell, P., & Tsakalidou, K. (1999). Recent trends in the re-integration of pupils withemotional and behavioral difculties in the United Kingdom. School PsychologyInternational, 20(4), 323e337.

Faul, F., Erdfelder, E., Lang, A.-G., & Buchner, A. (2007). G*Power 3: a exiblestatistical power analysis program for the social, behavioral, and biomedicalsciences. Behavior Research Methods, 39, 175e191.

Field, A. (2009). Discovering Statistics Using SPSS (3rd ed.). London: SagePublications.

Fishbein, M., & Cappella, J. N. (2006). The role of theory in developing effectivehealth communications. Journal of Communication, 56, 1e17.

Forlin, C., Douglas, G., & Hattie, J. (1996). Inclusive practices e how accepting areteachers? International Journal of Disability, Development and Education, 43(2),119e133.

Fuchs, D., & Fuchs, L. (1995). Whats so special about special education? Phi DeltaKappan, 76(7), 552e530.

Ganster, D., Hennessey, H., & Luthans, F. (1983). Social desirability response effects:three alternative models. Academy of Management Journal, 26, 321e331.

Gibson, S., & Dembo, M. (1984). Teacher efcacy: a construct validation. Journal ofEducational Psychology, 76, 569e582.

Goddard, Y. L., Neumerski, C. M., Goddard, R. D., Salloum, S. J., & Berebitsky, D.(2010). A multilevel exploratory study of the relationship between teachersperceptions of principals instructional support and group norms for instructionin elementary schools. The Elementary School Journal, 111(2), 336e357.

Hastings, R. P., & Oakford, S. (2003). Student teachers attitudes towardsthe inclusion of children with special needs. Educational Psychology, 23(1),87e94.

Heath, Y., & Gifford, R. (2002). Extending the theory of planned behaviour: pre-dicting the use of public transportation. Journal of Applied Social Psychology,32(10), 2154e2189.

Kavale, K., & Forness, S. (2000). History, rhetoric and reality: analysis of the inclu-sion debate. Remedial and Special Education, 21(5), 279e296.

King, M. F., & Bruner, G. C. (2000). Social desirability bias: a neglected aspect ofvalidity testing. Psychology and Marketing, 17(2), 79e103.siasm: dimensionality and context specicity. Contemporary EducationalPsychology, 36(4), 289e301.

Kuyini, A. B., & Desai, I. (2007). Principals and teachers attitudes and knowledge ofinclusive education as predictors of effective teaching practices in Ghana.Journal of Research in Special Educational Needs, 7(2), 104e113.

Lindsay, G. (2007). Educational psychology and the effectiveness of inclusiveeducation/mainstreaming. British Journal of Educational Psychology, 77, 1e24.

Lumpe, A. T., Czerniak, C. M., & Haney, J. J. (1998). Science teacher beliefs andintentions regarding the use of cooperative learning. School Science and Math-ematics, 98(3), 123e132.

Mahat, M. (2008). The development of a psychometrically-sound instrument tomeasure teachers multidimensional attitudes toward inclusive education.International Journal of Special Education, 23(1), 82e92.

Norwich, B., & Rovoli, I. (1993). Affective factors and learning behavior in secondaryschool mathematics and English lessons for average and low attainers. BritishJournal of Educational Psychology, 63(2), 308e321.

Oh, H. K., Rizzo, T. L., So, H., Chung, D. H., Park, S. J., & Lei, Q. (2010). Preservicephysical education teachers attributes related to teaching a student labelledADHD. Teaching and Teacher Education, 26(4), 885e890.

Palou, M., & Norwich, B. (2002). Cognitive, emotional and behavioral responses tostudents with emotional and behavioral difculties: a model of decision-making. British Journal of Educational Psychology, 28(1), 111e138.

Podsakoff, P., MacKenzie, S., Lee, J.-Y., & Podsakoff, N. (2003). Common methodbiases in behavioral research: a critical review of the literature and recom-mended remedies. Journal of Applied Psychology, 88(5), 879e903.

Praisner, C. L. (2003). Attitudes of elementary school principals towards theinclusion of students with disabilities. Exceptional Children, 69(2), 135e145.

Rakap, S., & Kaczmarek, L. (2010). Teachers attitudes towards inclusion in Turkey.European Journal of Special Needs Education, 25(1), 59e75.

Roberts, C. M., & Smith, P. R. (1999). Attitudes and behavior of children toward peerswith disabilities. International Journal of Disability, Development and Education,46(1), 35e50.

Schumm, J. S., & Vaughan, S. (1991). Making adaptions for mainstreamed students:general classroom teachers perspectives. Remedial and Special Education, 12(4),18e27.

Scruggs, T. E., & Mastropieri, M. A. (1996). Teacher perceptions of main-streaming/inclusion, 1985e1995: a research synthesis. Exceptional Children,63(1), 59e74.

Shapiro, E., Miller, D., Sawka, K., Gardill, C., & Handler, M. (1999). Facilitating theinclusion of children with EBD into general education classrooms. Journal ofEmotional and Behavioral Disorders, 7, 83e127.

Simpson, G., Bloom, B., Cohen, R., Blumberg, S., & Bourdon, K. (2005). US childrenwith emotional and behavioural difculties: data from the 2001, 2002, and2003 National Health Interview Surveys. Advance Data: From Vital and HealthStatistics, 360.

Snyder, J., Low, S., Schulz, T., Barner, S., Moreno, D., Garst, M., et al. (2011). Theimpact of brief teacher training on classroom management and child behaviorin at-risk preschool settings: mediators and treatment utility. Journal of AppliedDevelopmental Psychology, 32, 336e345.

Soodak, L., Podell, D., & Lehman, L. (1998). Teacher, school, and student attributes aspredictors of teachers responses to inclusion. Journal of Special Education, 31(4),480e497.

Stanovich, P. J., & Jordan, A. (1998). Canadian teachers and principals beliefs aboutinclusive education as predictors of effective teaching in heterogeneous class-rooms. The Elementary School Journal, 98(3), 221e238.

Summary Statistics for Schools in Scotland, 2011, Scottish Government: Edinburgh.Symeonidou, S., & Phtiaka, H. (2009). Using teachers prior knowledge, attitudes

and beliefs to develop in-service teacher education courses for inclusion.Teaching and Teacher Education, 25(4), 543e550.

Tabachnick, B. G., & Fidell, L. S. (2007). Using multivariate statistics (5th ed).. Boston:Allyn & Bacon.

Thomas, K., & Kilmann, R. (1975). The social desirability variable in organizationalresearch: an alternative explanation for reported ndings. Academy ofManagement Journal, 18, 741e752.

Tschannen-Moran, M., & Woolfolk Hoy, A. (2001). Teacher efcacy: capturing anelusive construct. Teaching and Teacher Education, 17, 783e805.

United Nations Educational, Scientic and Cultural Organisation (UNESCO). (1994).The Salamanca statement and framework for action on special needs education.New York: UNESCO.

UN General Assembly, Convention on the Rights of the Child 2009, United Nations,Treaty Series, vol. 1577, p. 3.

Walther-Thomas, C., Korinek, L., & Willliams, B. (2000). Collaboration for inclusiveeducation. Boston: Allyn & Bacon.

Woolfson, L. (2011). Educational psychology: The impact of psychological research oneducation. London: Pearson Education.

Woolfson, L., & Brady, K. (2009). An investigation of factors impacting on main-stream teachers beliefs about teaching students with learning difculties.Educational Psychology, 29(2), 221e238.

Woolfson, L., Grant, E., & Campbell, L. (2007). comparison of special, general andsupport teachers controllability and stability attributions for childrens dif-culties in learning. Educational Psychology, 27(2), 293e304.

Teacher attitudes and behavior toward the inclusion of children with social, emotional and behavioral difficulties in mains ...1. Applying the theory of planned behavior (TPB) to investigate teacher attitudes and behavior1.1. The present study

2. Method2.1. Participants2.2. Materials2.2.1. Multidimensional Attitudes Toward Inclusive Education Scale (MATIES: Mahat, 2008)2.2.2. Teachers' Sense of Efficacy Scale (TSES: Tschannen-Moran & Woolfolk Hoy, 2001)2.2.3. Teachers' Subjective Norm Scale (modified version of the TSES: Tschannen-Moran & Woolfolk Hoy, 2001)2.2.4. Teachers' Willingness to Work with Severe Disabilities Scale (TWSD: Rakap & Kaczmarek, 2010)2.2.5. Adaption Evaluation Instrument (AEI: Schumm & Vaughan, 1991)

2.3. Procedure

3. Results3.1. Predicting behavioral intention to engage in inclusive practices from beliefs, feelings, subjective norm and perceived beha ...3.2. Predicting inclusive behavior from beliefs, feelings, subjective norm, perceived behavioral control and behavioral intention3.3. Predicting teacher beliefs toward the inclusion of pupils with SEBD from INSET and experience3.4. Predicting teachers' feelings toward the inclusion of children with SEBD from INSET and experience3.5. Predicting teachers' willingness to work with pupils with SEBD from INSET and experience

4. Discussion4.1. Conclusions and implications

References