Development of a Social Acceptance Scale for Inclusive Education

In this research, the development stages of the "Social Acceptance Scale for Inclusive Education" developed to measure the social acceptance levels of inclusive students were presented in detail. The research was conducted in 4 primary and 2 secondary schools located in Kayseri city center. A total of 488 students receiving education in 4th, 5th, 6th, 7th and 8th grades participated in the study. During the development phase of the test form, 48 students in 2 classes in an elementary school were asked 3 open-ended questions about inclusive education and asked to explain their feelings and thoughts. At the same time, the opinions of 5 teachers in these schools were also evaluated. A scale consisting of 56 items was prepared, and the number of items in the scale was reduced to 44 as a result of the expert opinions in the field. Students, who needed special education in their classrooms, who were affected by migration and terrorism, and who were under temporary protection, were reached by using criterion sampling method in the selection of sample. In order to provide evidence for the validity of the scale, exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) were applied to the measuring instrument within the scope of structural validity. The statistical calculations of the obtained data were performed in SPSS-21 and Lisrel 8.80 software. As a result of factor analysis, Cronbach's alpha reliability coefficient of the scale consisting of a single factor and 30 items was found as 0.94. As a result of the findings, it was determined that the scale measured the social acceptance level of the inclusive students in a valid and reliable way.


Introduction
Inclusive education is a teaching service ensuring that all disadvantaged students co-exist with their non-disadvantaged peers and receive education on equal terms.
Inclusive education, as recognized by UNESCO (2005), is the process of responding to the different needs of learners by increasing their participation in education, culture and society, and by reducing discrimination within the education system.
When literature is reviewed, it is seen that there are various definitions of inclusive education. In general, inclusive education is defined by focusing on the concepts of disability and special education, within the framework that children with disabilities should participate in educational processes together with other children. Some views focusing on the concept of diversity in the context of students' individual characteristics and differences define inclusive education within the framework of an approach that can respond to the different needs of students (Westwood, 2013).
Although inclusive education "is closely related to the right to education of children with disabilities, it is not limited to the right to education of children in this group; it is an approach that is necessary to ensure the right to education of all children who are disadvantaged for different reasons in a broader sense" (UNICEF, 2014). Unlike inclusive education, inclusiveness not only provides education for individuals with special needs, but also ensures that all disadvantaged individuals coexist with their peers.
To give examples of disadvantageous situations, teaching Turkish as a second language, working with children with disabilities, working with children exposed to violence, working with children under temporary protection, working with children affected by migration and terrorism, and working with children affected by natural disasters are the fields of inclusive education.
Inclusive education should be based on three different rationales: educational, social and economic. According to the educational ground, inclusive education offers differentiated and diversified instruction considering the student's ability and needs. Thus, it is aimed that all students participate in the learning process. According to the social rationale, this education model provides the basis for a fairer and more inclusive society. The economic ground is that schools that provide education to different students together are less costly, instead of establishing separate schools for different groups of students. The main rationale for inclusive education is to prevent discrimination. (Aktekin, 2017: p.16). With the inclusive education model, it is aimed that students are given education considering their differences.
When the studies in the literature are taken into consideration especially in the context of inclusive students whose mother tongue and ethnic origin are different, it can be stated that the group of students in question face problems such as lack of harmony, discrimination, exclusion and mother tongue problems in the classroom environment.
"The vast majority of Syrian children, who theoretically could have attended public schools in Turkey, do not attend these schools in practice" (Human Rights Watch, 2015: 13). Syrian students were found to have psychological problems such as lack of harmony with their classmates, loneliness and loss of self-confidence in public schools due to the fact that they lagged behind because of the forced migration (Human Rights Watch, 2015: 25) In the report of Fırat (2010) published as a result of the Role of Education for Social Reconciliation Project by the History Foundation, it is pointed out that children whose mother tongue is different from Turkish have difficulties in reading and writing when they start school as well as learning a new language.
Interviews with individuals from various ethnic backgrounds in the report of the Gündem Children's Association ("Gündem Çocuk Derneği") on child discrimination in Turkey, prepared by interviewing individuals between the ages of 18 and 25, reveal their experiences of ethnic-based exclusion in childhood (Gündem Children's Association, 2014).
In a study conducted on the Romani population in Kocaeli, it is stated that one of the reasons for Romani children being distanced from education is the perception of exclusion from schools. Again in this research, it is mentioned that the Romani people give importance to education and want to educate their children (Demirel, 2012: 92). In another study, it is stated that biased approach to Romani children attending school by non-Romani children, parents and teachers play a role in the dropout of these children (Alp and Taştan, 2011).
The literature review shows that students with special needs have problems of adaptation in the classroom besides the students whose mother tongue, race and ethnic origin are different, and that their social acceptance levels are low. Students with special needs usually have fewer friends and have lower social status than their peers who do not have special needs in inclusive settings (Ciechalski, 1995, cited by Sucuoğlu andÖzokçu, 2005, 43) It was determined that students with special needs were not preferred by their friends for studying together, and they were perceived as problematic in terms of adaptation and communication (Kabasakal, Girli, Okun, Çelik and Vardarlı, 2008). Individuals with special needs may exhibit more problem behaviors than their peers for various reasons or sometimes due to their inadequacies. Thinking that problem behaviors are caused by inadequacy by the classroom teacher may cause prejudices, refusal to accept and staying away. Individuals with special needs may lag behind in terms of social skills, which are a prerequisite for academic skills, compared to their peers, and these inadequacies lead to low social acceptance within the classroom, with the addition of problem behaviors. As a result of low social acceptance, peer interaction decreases and it becomes difficult to acquire social skills (Sucuoğlu and Kargın, 2014).
When these problems are examined, it is seen that they are mainly caused by the attitude of accepting students with different characteristics in the classroom. Thurstone (1928cited by Muller, 1986: 3) defines attitude as "the intensity of positive or negative feelings towards a psychological object". In order to understand the source of the problems experienced in inclusive education, it is necessary to look at the tendencies of other students to react positively or negatively towards inclusive students in the classroom.
The social acceptance levels of the students, who consider themselves part of the class, are independent of the adults, do not engage in harmful behaviors such as aggression, are welcome by the class, and have knowledge about their inadequacy, increase (Boutot, 2007). Therefore, in inclusive education, it is necessary to provide learning environments in which all disadvantaged students are accepted and feel good about themselves.
The inadequate social skills of the inclusive students and their inability to use these skills when the appropriate educational environment is not provided are the causes of the academic and social failures they face. Because, social skills are necessary for the individual to establish positive social relationships with his/her peers, teachers, family and others, to maintain these relationships, and these skills facilitate social cohesion through peer acceptance, and help to cope with social expectations of the environment (Gresham, 1986, Kolb, Hanley and Maxwell, 2003, Gresham, Sugai and Horner, 2001, cited by Sucuoğlu, 2005.
In this context, the concept of social acceptance in classes in which Inclusive education is applied can be expressed as the acceptance of all disadvantaged students by their peers and other individuals, both within and outside the classroom, communicating with them, the willingness of peers and other individuals to engage in social activities with them, and ensuring positive continuity in social relations.
These results indicate that it is important to measure social acceptance levels of inclusive students in schools. In the literature, there is the Social Acceptance Scale of Siperstein, which was adapted to Turkish by Civelek (1990). The scale consisting of 22 items has a 5-point rating structure. Since the items are expressed with positive structure, the high score received from the scale is interpreted as the high social acceptance of the inclusive student among their peers who have normal development. In the scale developed by the researchers, it is aimed to measure the social acceptance level of not only inclusive students but also all disadvantaged students who are under temporary protection, and affected by migration and terrorism. Therefore, it is important to develop such a scale in order to measure the social acceptance level of inclusive students.

Method
In this section, information about the model used in the research, study group, development process of the measurement tool, and data analysis are given.

Research Model
This research was designed in descriptive survey model. Since the aim of the study was to develop the scale, the steps of determining the characteristics to be tested, writing the items to be included in the scale, obtaining expert opinion and re-arranging the items, and performing the validity and reliability analyzes were followed (Cronbach, 1984;Altun and Büyüköztürk, 2011).

Study Group
The study group consisted of the students in the classes with inclusive students. Criterion sampling, which is one of the most frequently used purposeful sampling methods (Patton, 2014), was used in the research. In criterion sampling, the researcher can create a criterion or criteria meeting a predetermined set of criteria (Yıldırım and Şimşek, 2011). For the students who participated in the study, the criterion for having at least 1 of the students who need special education in their class, and those who are affected by migration and terrorism and who are under temporary protection, and the criterion for having studied together with this student or students for at least 1 year were taken as the basis. In this context, the study was conducted with data collected from a total of 488 students studying in 4th, 5th, 6th, 7th and 8th grades of 4 primary schools and 2 secondary schools determined by purposeful sampling from among different socio-economic regions of Melikgazi, Kocasinan and Talas districts of Kayseri province. The data of students who filled out the scale carelessly, did not respond sincerely and made erroneous markings were excluded from the scope of the study.
Information on the gender and grades of the students in the sample is given in Table 1. When Table 1 is analyzed, it is seen that 47.3% of the sample are females and 52.7% are males. Büyüköztürk (2002) stated that it is a general rule that the sample size should be at least five times the number of variables. Accordingly, considering that the 30-item scale form was applied to the study group, it can be said that the number of participants were sufficient for the statistical procedures. When Table 2 is examined, it is observed that 93 (19.1%) of the inclusive students in the classes of the students in the sample group were selected from the classes of the students who need special education, 213 (43.6%) were selected from classes with students affected by migration and terrorism, and 182 (37.3%) were selected from both classes with students needing special education and those with students affected by migration and terrorism.

Creation of the Pool of Items
At this stage, firstly literature was reviewed for the scale. In the literature review, measurement tools used in different previous studies were examined. 48 students in two classes in an elementary school affiliated to Kayseri Provincial Directorate of National Education were asked 3 open-ended questions regarding all subjects and parties covered by the process, considering the principles of inclusive education, and they were asked to express their feelings and thoughts. Opinions of 5 teachers working in this school were also taken. During the creation of the item pool, opinions obtained from students and teachers were used. In addition, relevant research on inclusive education was also utilized in the item writing process. In this way, a pool of 56 items was created.

Determination of Content Validity and Preparation of the Draft Scale
In this study, it was paid attention that the measurement tool which was prepared by using expert opinion had content validity. Therefore, first of all, the expressions that could be used as attitude statements from the item pool created after the implementation of the first stage were chosen meticulously. Afterwards, these expressions were arranged in accordance with the rules of item writing in order to express attitude, and the items in the item pool were subjected to a preliminary screening in accordance with the views of 3 faculty members specialized in the field. The rearranged attitude phrases were evaluated by considering the behavioral, cognitive and affective elements under the headings of three different factors. During the process, care was taken to ensure that attitude expressions were positive and negative, and that every item is understood by the students in the same way. The number of items, which was 56 initially, was reduced to 44 as a result of feedback received from experts. A 5-point Likert-type scale consisting of 33 positive and 11 negative items was made ready for the trial stage.

Pre-application of the Scale
In accordance with the expert opinion, the measurement tool was applied directly to 52 students with inclusive students in their classes, and the clarity and comprehensibility of the items were tested. The 44-point draft scale was finalized, with corrections being made on some words in line with expert opinions and pre-application stage.

Determination of Construct Validity and Calculation of Reliability
The trial application of the measurement tool developed within the scope of the research was applied to 280 students. The factor structure of the scale was determined by using the measurement results obtained as a result of the trial application. For this purpose, Principal Component Analysis, which is one of the Exploratory Factor Analysis (EFA) methods, was used. Kaiser-Meyer-Olkin (KMO) test and Bartlett's test of sphericity were used to determine the suitability of the data for factor analysis. When deciding the factor structure, the eigenvalues of the factors, the total explained variance and the factor loads of the items were taken into account. Analyzes were performed using SPSS-21 software.
Confirmatory Factor Analysis (CFA) was performed by applying the final version of the scale to a new sample (N=208) in order to determine the model data fit of the factor structure obtained by the trial application. In the CFA analysis to provide evidence of construct validity, some fit indices were taken into consideration to determine whether the model fitted the data. For this purpose, goodness of fit tests such as Chi-Square Goodness, Comparative Fit Index (CFI), Normed Fit Index (NFI), Root Mean Square Error of Approximation (RMSEA), and Standardized Root Mean Square Residual (SRMR) were used. In addition, Cronbach Alpha (α) reliability coefficient, Spearman-Brown Split Half Reliability Coefficient and corrected item-total correlation coefficient were calculated to provide evidence for the reliability of the measurement results obtained from the scale. The analyzes were performed using LISREL 8.80 software.

Finalizing the Scale
As a result of the findings, the items eliminated in the above analyzes were excluded from the study, and the scale was finalized (Table 5).

Data Collection Tool
As a result of the factor analysis performed, items that do not fit the structure of the scale or load more than one factor were removed from the scale consisting of 44 items. In this process, 14 items that did not meet the criteria were excluded from the analysis. The "Acceptance of Students with Different Characteristics Scale", designed to measure social acceptance levels of students with different characteristics in the classroom, consists of 30 items and one factor. The scale, which was developed in order to reveal the social acceptance levels of inclusive students in the classroom, was prepared as a 5-point Likert type scale. The answer options of the items included in the 5-point Likert scale were created as "1 = I strongly disagree", "2 = I disagree", "3 = I am undecided", "4 = I agree", and "5 = I strongly agree". Since the scores in the scale were between 1 and 5, it was assumed that the students' social acceptance levels of the inclusive students were high in the classroom as the propositions in the items approached 5, and low as they approached 1. The values of negative items were reversed at the scoring stage (1=5, 2=4, 3=3, 4=2, 5=1).

Data Analysis
The 44-item scale created was applied to 280 students with inclusive students in their classes. The data obtained were analyzed with SPSS-21 and Lisrel 8.80 software in order to provide evidence for validity and reliability, and item analyses were evaluated.

Findings
In this section, the explanations of the validity and reliability study of the Acceptance of Students with Different Characteristics Scale are included.

Construct Validity
The construct validity, defined as the degree to which the measuring instrument is able to measure the theoretical structure it is trying to measure, reveals the relationship of each item in the scale to each other (Seçer, 2015). The path used in the construct validity of the measuring instrument is factor analysis (Büyüköztürk, 2003). In this study, exploratory factor analysis and confirmatory factor analysis were applied to test the construct validity of the Acceptance of Students with Different Characteristics Scale.

Exploratory Factor Analysis (EFA)
After the expert opinion, 44 items remained in the final version of the scale. This final version of the scale was applied to 280 students in trial application. Firstly, in order to determine the extreme values in the data collected, Z scores of the total scores each student received from the scale were calculated. Seven participants whose Z score was outside the ±3 range were deleted from the data file. Exploratory Factor Analysis (EFA) was performed for the construct validity of the measurement results obtained from the scale. EFA was conducted on 273 individuals. Kaiser-Meyer-Olkin (KMO) test and Bartlett's test for sphericity were performed to determine the suitability of the data to be analyzed for factor analysis. The findings are given in Table 3. As shown in Table 3, KMO value was 0.89, and the Bartlett's test was seen to be significant (χ²=4276.27; p<0.05) In order to reveal the factor structure, the KMO value must be at least 0.50. The findings indicate that the KMO value is ideal and the data are suitable for factor analysis. First, EFA was performed without any limitation. According to the findings, 10 factors with eigenvalues above 1 were created in the first case. 7 items with a factor load below 0.40 were excluded from the scale (M10, M13, M19, M24, M30, M32 and M37). Analyzes were repeated without any limitation over the remaining 37 items. The eigenvalue and explained variance ratios for the factors obtained from the second factor analysis are given in Table  4. When Table 4 is examined, the eigenvalue of the first factor is seen to be 4 times greater than the eigenvalue of the second factor (10.093/2.257=4.47), and the variance ratio it explains by itself is approximately 27%. The fact that the explained variance ratio of the first factor is greater than 20%, and that the eigenvalue of the first factor is three times higher than the eigenvalue of the second factor suggests that the scale has a dominant factor (Hattie, 1985;Lord, 1980;Reckase, 1979). Thus, it was decided that the scale had a single-factor structure. After deciding on the factor structure of the scale, EFA was performed again by limiting the number of factors to 1. As a result of this last factor analysis, the factor analysis was repeated by removing 7 items with factor load values below 0.40 starting from the lowest factor load value item (M2, M5, M9, M11, M20, M25, M39). In the final case, 30 items remained on the scale. In the 30-item scale, 3 items, 19, 23 and 30, were negative. The items for the final version of the scale consisting of single factor and 30 items, factor load values and total explained variance ratio are given in Table  5.
As can be seen in Table 5, factor load values of all 30 items in the scale are higher than 0.40 taken as a criterion, and factor load values of these items vary between 0.42 and 0.74. The total variance ratio explained by the single factor structure is approximately 32%. It is sufficient to obtain at least 30% of the total explained variance in measurement instruments consisting of a single factor and in which the affective characteristic is measured (Büyüköztürk, 2012). According to the EFA results of the study, a single factor scale consisting of 30 items was re-applied to a similar sample (N=208). CFA was performed to determine whether the single-factor model fitted the data. The path diagram of the established model is given in Figure 1.  Figure 1, there are 30 items in the scale. As a result of the CFA analysis, it is observed that the standardized regression coefficients of the items range between 0.33 and 0.75. All these coefficients are significant at 0.05 level. Error variances are between 0.44 and 0.89. The fact that the standard regression coefficients (factor load) for each item are above 0.32 indicates that the model data fit is acceptable (Tabachnick and Fidel, 2013).
In order to evaluate the fit of the model to the data, many model fit statistics are used. Within the scope of this research, χ 2 (chi-square), χ 2 / sd (chi-square / degree of freedom), RMSEA, SRMR, CFI, NFI and NNFI values were interpreted. The values of the model data fit of the measurement results obtained in the actual application of the scale are given in Table 6 below. Examining Table 6, the RMSEA value is obtained as 0.073 and the SRMR value as 0.068. The CFI value was obtained as 0.96, NFI=0.93, NNFI=0.96 and χ 2 / sd as 2.11. Since the chi-square value is generally affected by the sample size, χ 2 / sd is taken into account. The fact that the "χ 2 / sd" ratio is less than 5, the value of RMSEA and SRMR is less than 0.08, whereas the value of CFI, NFI and NNFI is higher than 0.90 indicates that the model fitted to data (Hu and Bentler, 1999;Tabachnick and Fidel, 2001). The goodness of fit values obtained as a result of the CFA analysis generally shows that the model fits well to the data.

Reliability
In order to determine the reliability of the measurement results obtained in addition to the evidence given above for the construct validity of the developed measurement tool, Cronbach Alpha (α) reliability coefficient, Spearman-Brown Split Half Reliability and corrected item-total correlations were calculated and given in Table 7.

Item Analyses
As shown in Table 7, the average scores for 30 items on the scale range from 2.75 to 3.96. Item-total correlation is the correlation of an item with the total score in the relevant factor. For an item, this value is generally expected to be greater than 0.30 (Nurosis, 1994). The total score of the items included in the scale and corrected correlation values range between 0.32 and 0.72. These values indicate that item-total correlations are sufficient. While the Cronbach's Alpha value for the reliability of the measurement results obtained from the main application of the scale was 0.94, the reliability coefficient for the Spearman-Brown methods was 0.91.
According to these results, considering that the reliability coefficient must be 0.70 and higher in order for the scale to be considered reliable (Tavşancıl, 2005;Büyüköztürk, 2006;Büyüköztürket al., 2008), it is seen that the whole scale (0.94) is highly reliable.

Discussion
In this study, a measurement tool was developed to measure the acceptance level of students with different characteristics in the classroom in a valid and reliable way. In order to develop the measurement tool, firstly a literature review was conducted, expert opinion was taken and a pool of 56 items was created. A 44-item draft scale was created in accordance with the pool of items, pilot implementation and expert opinion. Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA) were performed to ensure the construct validity of the Acceptance of Students with Different Characteristics Scale. For exploratory factor analysis, Kaiser-Meyer Olkin test and Bartlett's Test of Sphericity were performed. KMO value was found as 0.89, and Bartlett's Test also was significant (χ²=4276.27; p<0.05). According to these results, it was demonstrated that the data was suitable for factor analysis. As a result of EFA, a 30-item scale was obtained, in which the single-factor structure explained 32.018% of the total variance.
In the confirmatory factor analysis results of the Social Acceptance Scale for Inclusive Education, fit index values of the Social Acceptance Scale for Inclusive Education, χ 2 / sd, were found as 2.11, the RMSEA value as 0.073 and the SRMR value as 0.068. The CFI value was obtained as 0.96, NFI=0.93, and NNFI=0.96. The fact that the "χ 2 / sd" ratio was less than 5, the value of RMSEA and SRMR was less than 0.08, whereas the value of CFI, NFI and NNFI was higher than 0.90, indicated that the model had sufficient fit indices.
In order to test the reliability of the scale, the internal consistency coefficient of the Cronbach's Alpha was calculated and the internal consistency coefficient of the overall scale was found as 0.94, and the scale was proven to be highly reliable.

Conclusions
These results indicate that the Social Acceptance Scale for Inclusive Education is a valid and reliable measurement tool for determining the level of social acceptance of inclusive students in the classroom. It is thought that this form of the scale can contribute to the literature as a valid and reliable measurement tool that can be used to determine the social acceptance level of inclusive students.
It is thought that the results obtained from the actual applications of the scale can determine the level of social acceptance of the inclusive students in the classroom, that the measurement tool developed can be examined in applications in different samples, and that it can be used in experimental and descriptive studies to determine the level of social acceptance of primary and secondary school students for inclusive students.
Since the study included students in a limited sample, it may be suggested to study in larger samples. A next study could not be so locally oriented ('Kayseri city center'). It is thought that the validity and reliability of the scale can be re-tested and used in research including teachers as well as in studies including students. 26. I can feel what he/she is going through by putting myself into his/her shoes. 27. I chat with my friend, who can't hear, in body language. 28. I give him/her my pocket money so that he/she can meet his/her needs. 29. I teach/explain to him/her what he/she doesn't understand because of his/her absence from school. 30. I'd rather have another friend than play with him/her in games.