Eurasian Journal of Educational Research, Issue 61, 2015, 23-44 Psychometric Characteristics of the Social Justice Scale’s Turkish Form and A Structural Equation Modeling1 İlker CIRIK* Suggested Citation: Cirik, I. (2015). Psychometric characteristics of the social justice scale’s Turkish form and a structural equation modeling. Eurasian Journal of Educational Research, 61, 23-44. http://dx.doi.org/10.14689/ejer.2015.61.2 Abstract Problem Statement: In order to provide equal educational opportunities for students, teachers should encourage their students to have an effective voice concerning social justice. Studies reveal that teachers face trouble when transferring from the concept of social justice as theory to social justice as practice. A scale which will be developed on social justice may enable teachers to better comprehend the process of movement from attitude to action. Moreover, examining which factors affect teachers’ behaviors towards social justice will contribute to the arrangement of social justice studies in learning environments. Purpose of the Study: The primary aim of the study is to analyze the psychometric characteristics of the Social Justice Scale’s Turkish form. The secondary aim is to determine the effects of social justice attitudes, subjective norms and perceived behavioral control on behavioral intentions. Method: The scale adaptation process was conducted with 515 participants (376 female, 139 male), and structural equation modeling was conducted with 410 (313 female, 97 male) participants. The data were collected through the Social Justice Scale. In analyzing the data, Pearson Correlation Coefficients, confirmatory factor analysis and Cronbach Alpha coefficients were utilized. The effects of social justice attitudes, subjective norms and perceived behavioral control on behavioral intentions were analyzed through structural equation modeling. Findings and Results: Linguistic equivalence of the scale was obtained p=.00, p<.01. After confirmatory factor analysis, the fit indices χ2=671.15, 1 This study was presented at the 2nd Eurasian Educational Research Congress in Ankara, 8-10 June, 2015 * Dr. Mimar Sinan Fine Arts University, Faculty of Arts and Sciences, E-mail: [email protected] 24 İlker Cırık df=246, χ2/df=2.72, GFI=.90, AGFI=.88 were found to be at an acceptable level; RMSEA=.05, NFI=.96, NNFI=.97, CFI=.97, SRMR=.04 indices demonstrated good fit. The internal and external consistency coefficients and corrected item-total correlations of the scale were found to be high. After the model test, the values of χ2=570.93, df=246, χ2/df=2.32, GFI=.90, AGFI=.87 showed that model data consistency was at an acceptable level; the values of RMSEA=.05, NFI=.95, NNFI=.97, CFI=.97, SRMR=.05 demonstrated a good fit. Conclusion and Recommendations: It was determined that the Turkish form of the scale has 24 items that fall under four factors. The reliability coefficients of the scale were found to be high. It was revealed that the effects of attitudes towards social justice, perceived behavioral control and subjective norms on behavioral intentions were significant. It is concluded that the Turkish form of the scale will help measure prospective teachers’ views about social justice and thus will contribute to filling a gap in the field. Further researches are recommended to provide evidence for the concurrent validity of the scale. Keywords: Equality, educational opportunity, confirmatory factor analysis, teacher education Introduction It can be stated that educators who are supposed to keep a balance of equality and justice among their students cannot sometimes hold the balance and give adequate support to all their students. As Bradley, Werth, and Hastings (2012) specified, the means of providing such a balance of equality and justice for all the classes in a society have been argued by different philosophers from Aristotle and Socrates to contemporary activists such as Freire and Beauvoir on a large scale. In the very essence of these arguments, there lies the interaction between education and society. When this interaction is considered from the point of education, it can be said that “education is not the ultimate lever for social transformation, but without it transformation cannot occur” (Freire, 1998, p. 37). When assessed from the society’s perspective, “it is vital that the school and community recognize the importance of community involvement in education and the involvement of students in the community” (Goulet & Goulet, 2014, p. 210). Altogether, these two outlooks disclose the fact that the interaction between education and society has an undeniable significance. This perspective has enhanced studies by educators intending to provide equal educational opportunities for students who have come from different social structures (Capper & Young, 2014; Constantine, Hage, Kindaichi, & Bryant, 2007; Goodman et al., 2004; Lemley, 2014; Leonard & Moore, 2014). In the essence of these studies, there exist social justice objectives which are intended to support cultural pluralism. According to Prilleltensky (2001), social justice should “promote fair and equitable allocation of bargaining powers, resources, and obligations in society in consideration of people’s differential power, needs, and abilities to express their Eurasian Journal of Educational Research 25 wishes” (p. 754). Correspondingly, Fouad, Gerstein, and Toporek (2006) also argue that social justice requires being fair and equitable in distributing social resources for every member of the society, regardless of their race, gender, ability status, sexual orientation, physical makeup, or religious creed. To realize all of these, the society itself has to work in a collaborative, democratic, participatory and inclusive way (Bell, 2007). Therefore, it can be argued that allowing social justice applications in learning environments which display themselves in a micro-dimensional size in social structure will lead to positive contributions to the structure of the society. In social justice applications, the interaction among teacher, student and learning environment is crucial. Teachers and students do not only bring their own cultural backgrounds into the classroom, but also interpret the classroom culture and social life they find there (Lalas, 2007). In light of these interpretations, the notions of equality and justice gain profound importance for the stakeholders in the learning environments as many studies point out. For instance, Aydin and Tonbuloglu (2014) stated that educators primarily emphasize equality, democracy and justice in learning environments. Similarly, Fuentes, Chanthongthip, and Rios (2010) concluded that university students who had social justice education have more attention to equality and justice. In another study, Torres-Harding, Diaz, Schamberger, and Carollo (2015) mentioned that taking service-learning courses was related to university students’ social justice attitudes, intentions and actions. Also Littenberg-Tobias (2014) stated that participating in some service programs may influence university students’ attitudes and understandings of social justice. Thus, the students’ sensitivity to these concepts in human life is closely associated with social justice education (Banks, 2009; Goodman & Burton, 2012; Leonard & Moore, 2014). It is because social justice education, which has a student-centered, cooperative, empirical, intellectual, analytical and multicultural structure (Wade, 2004), enables students to see not only their differences, but also their common ground (Capper & Young, 2014). The key role in this context is played by teachers. Teachers who aim to create differences in their students’ lives should be sensitive to the notion of equal opportunities in education (Villegas, 2007), should help students develop strategies to solve social problems (Brown & Brown, 2011; Wade, 2004) and should support their students in having an effective voice in terms of social justice (Garii & Rule, 2009). In order to realize this necessity, it is thought that teacher training education should include social justice subjects. Related studies carried out in this context also support this view. For example, Leonard and Moore (2014) and Page (2009) found that prospective teachers taking social justice education gave more attention to equality and justice. Tinkler, Hannah, Tinkler, and Miller (2015) mentioned that social justice service-learning experiences support the social justice goals of prospective teachers. Rios and Montecinos (1999), in a similar study, found that prospective teachers pointed out the importance of social justice in their education and that the notion of social justice should be a part of their instructional curriculum. On the other hand, a study conducted by Tomul, Celik, and Tas (2012) revealed that according to prospective teachers, in-service teachers behave in a discriminative way in terms of students’ socio-economic features, political views and 26 İlker Cırık beliefs. In a similar study by Ersoy (2014), it was detected that, since teachers do not have adequate knowledge and experience concerning effective and democratic citizenship education, primary students belonging to lower socio-economic levels cannot exercise their rights adequately compared to those who come from higher socio-economic levels. Thomas (2007) in his study argues that teachers who intend to maintain social justice implementations are not supported enough. The inadequacy of social justice implementations of in-service teachers reveals the fact that problems emerge in the transition of social justice from theory to practice. Even though the notions of equality and justice are frequently emphasized in literature, there have been arguments over how to implement social justice in learning environments (Speight & Vera, 2009). Vera and Speight (2003) argue that solely focusing on theoretical knowledge could be important for researchers; however, allotting efforts to implementation will enable social justice to exist as a dynamic structure. Correspondingly, Torres-Harding, Siers, and Olson (2012) underline the fact that the relationship between attitudes and actions requires more applied research. Therefore, though not sufficient for now, it can be stated that in the literature there is an increasing tendency concerning implementation of the notion of social justice in studies (Bradley et al., 2012). At this point, there appears to be a question over how the teachers’ behaviors towards social justice implementations will be evaluated and directed. Torres- Harding et al. (2012) hold that a scale that can be developed in the field of social justice will help researchers and educators in terms of understanding the process of moving from attitudes to action. Analyzing the literature, it is clear that there are scales which evaluate individuals’ behaviors concerning social justice in general; however, the number of scales used to evaluate which factors at what levels effectively influence behaviors is limited (Colquitt, 2001; Rasinski, 1987). In order to remove the stated limitation, Torres-Harding et al. (2012), developed the Social Justice Scale (SJS) based on Ajzen’s (1991) social cognitive model. Ajzen (1991) in his model states that individuals’ attitudes towards the action, subjective norms around the action and their perceived behavioral control of the action predict the behavioral intention. The component of behavioral intention in the model points out that an individual’s intention to act affects their behavior in the future; the component of attitudes towards the action is related with the individual’s evaluation of the behavior’s convenience level; the component of subjective norms relates to the individual’s perception of social support or pressure from his/her environment in terms of performing the action; the component of perceived behavioral control of action refers to the individual’s perception of self-efficacy about himself/herself regarding the difficulty of behavior. The three elements in the model could show differences in accordance with the significance level of behavioral intention and situation, as well as the feature of the behavior. In other words, these three predictors may affect behavioral intentions independently or all together. Eurasian Journal of Educational Research 27 All in all, SJS was developed in order to measure attitudes concerning social justice, values, perceived behavioral control and behavioral intentions. It is thought that searching the psychometric characteristics of SJS’s Turkish form, and thus bringing it into the literature, will contribute to the researches regarding social justice and social justice implementations to become integrated into learning environments. Therefore, the primary aim of this study is to analyze linguistic equivalence, validity and reliability of SJS. The secondary aim is to test Ajzen’s (1991) model upon whose basis the scale was constructed. Within the framework of the second aim, the effects of “social justice attitudes,” “subjective norms” and “perceived behavioral control” (independent variables) on “behavioral intentions” (dependent variable) were investigated. Method Research Design This research is comprised of two different studies: scale adaptation and structural equation modeling (SEM). In scale adaptation, since the existing situation remained to be determined, the data gained through SJS were analyzed by applying a survey method, as Fraenkel and Wallen (2009) suggested. In the SEM study, as it was thought that, based on Ajzen’s (1991) model, there would be a cause and effect relationship between “social justice attitudes,” “perceived behavioral control,” “subjective norms,” and “behavioral intentions,” causal research design was preferred. Research Sample This research was carried out in two different study groups; the first group being in the scale adaptation study, the other being in the SEM. The first study group was comprised of 515 prospective teachers attending pedagogical formation programs at Marmara University, Mimar Sinan Fine Arts University and Yildiz Technical University in the 2013-2014 Academic Year. The second study group included 410 prospective teachers attending pedagogical formation programs at Mimar Sinan Fine Arts University in the 2014-2015 Academic Year. The demographic characteristics of the participants are displayed in Table 1. Table 1. Frequency and Percentage Distribution of Participants’ Demographic Characteristics First Group Second Group f % f % Gender Female 376 73.0 313 76.3 Male 139 27.0 97 23.7 Education Level Undergraduate 411 79.8 346 84.4 Graduate 104 20.2 64 15.6 28 İlker Cırık Table 1 continue… First Group Second Group f % f % Graduation Turkish 95 18.4 113 27.6 History 24 4.7 62 15.1 Music 90 17.5 78 19.0 Painting 101 19.6 49 12.0 Mathematics 29 5.6 57 13.9 Physics 36 7.0 - - Theology 62 12.0 - - Art and Design 78 15.1 - - Sociology - - 51 12.4 Total 515 100 410 100 Research Instrument and Procedure The data of the study were collected through SJS developed by Torres-Harding et al. (2012). The scale has four factors (social justice attitudes, perceived behavioral control, subjective norms and behavioral intentions) and 24 items in total. The first factor consists of 11 items concerning social justice attitudes. An example of these items includes, “Allow others to have meaningful input into decisions affecting their lives.” In the second factor, there are five items intended to measure the individuals’ perceived behavioral controls about social justice. An example of these items includes, “I am capable of influencing others to promote fairness and equality.” The third factor contains four items to measure subjective norms about social justice. An example of these items includes, “Other people around me are supportive of efforts that promote social justice.” The fourth factor is comprised of four items to measure behavioral intentions toward social justice in the future. An example of these items includes, “In the future, I will do my best to ensure that all individuals and groups in my community have a chance to speak and be heard.” Responses were scored on a 7- point scale ranging from strongly disagree (1) to strongly agree (7). In order to adapt SJS into Turkish, permission was obtained from S.R. Torres- Harding via e-mail. Three lecturers who are experts in psychological counseling and guidance, curriculum development and instruction, measurement and evaluation were asked for their opinions on the scale’s cultural convenience, clarity etc. During the process of translation, two English Language experts translated the scale into Turkish and then two other experts made the reverse translation. In accordance with the common opinions of the experts and the researcher, the problematic items were corrected on the Turkish form of the scale. Then, the Turkish form was examined by Turkish Language experts in the context of language and expression. English and Turkish forms of the scale, respectively, were applied to English Language Teaching Eurasian Journal of Educational Research 29 students attending their 4th class in a four-week interval, and the consistency between the forms was examined. Applying the scale to the participants was realized by four lecturers including the researcher. The participants were informed of the aim of the study, features of the scale and privacy of the data in order to get them to involve themselves in the study voluntarily. Validity and Reliability The factor structure of the original scale was analyzed through confirmatory factor analysis (CFA). After the analysis, it was discovered that the fit indices, χ2=789.14, df=246, p=.00, RMSEA=.09, NFI=.95, CFI=.97, GFI=.80, AGFI=.75, PGFI=.65, were at an acceptable level. The factor loading values of the items were calculated between .55 and .91, and after reliability analysis, the Cronbach Alpha coefficients in factors were found to range from .82 to .95. Finally, the inter-factor correlation coefficients were determined to be between .34 and .58. Data Analysis In the process of linguistic equivalence, the consistency between English and Turkish applications was analyzed through Pearson Correlation Coefficient. As the scale’s factor numbers and consistency between factors and indicators (Kline, 2011) were specified in the original form, factor consistency with the data obtained from the first study group was tested through CFA (Bentler & Bonett, 1980; Brown, 2015). In order to specify the internal consistency of the scale, the Cronbach Alpha Coefficients of the factors and the total of the scale were examined. For the purpose of specifying on what level the items differentiate the participants, corrected item- total correlation was calculated and independent samples t test was realized between the lower 27% and the upper 27% groups. The external consistency of the scale was calculated through test-retest applications performed in four-week intervals. The effects of social justice attitudes, subjective norms and perceived behavioral control on behavioral intentions were analyzed by means of SEM. SEM is a comprehensive and flexible approach in modeling the relationships between observed and latent variables (Hoyle & Smith, 1994; Hu & Bentler, 1998). Additionally, its capability to provide the chance of evaluation and correction about a theoretical model (Anderson & Gerbing, 1988), its offering the chance of controlling measurement errors and its utility to provide information about the consistency level of the model make SEM an influential method (Frazier, Tix, & Barron, 2004). LISREL 8.80 and SPSS 17.0 were used for the analysis of the data. Results Linguistic Equivalence Between four-week interval applications of the SJS’s English and Turkish forms (N=33), a positive and significant correlation was found in items .63 and .84; in factors .76 and .84; in the total of the scale .86, p=.00, p<.01 (Table 2). Findings show 30 İlker Cırık that the consistency between applications is at an acceptable level. Thus the scale’s linguistic equivalence was acquired. Table 2. Correlation Coefficients of the SJS’ Linguistic Equivalence Item r Item r Item r Item r 1 .76* 7 .76* 13 .73* 19 .75* 2 .75* 8 .79* 14 .74* 20 .68* 3 .75* 9 .76* 15 .82* 21 .63* 4 .68* 10 .63* 16 .77* 22 .73* 5 .80* 11 .63* 17 .72* 23 .83* 6 .70* 12 .81* 18 .82* 24 .84* Factor Attitude .81* Behavior .84* Norm .76* Intention .84* Total .86* (N=33), *p<.01 First-Order Confirmatory Factor Analysis In the factor analysis, as Brown (2015) suggested, to arrange the relationships between indicators, maximum likelihood technique, which enables a statistical evaluation over how to perform a better factor analysis, was applied. After CFA, the fit indices χ2=671.15, df=246, χ2/df=2.72, GFI=.90, AGFI=.88 were found to be at an acceptable level; RMSEA=.05, NFI=.96, NNFI=.97, CFI=.97, SRMR=.04 indices showed good fit (Hu & Bentler, 1998; Hoyle & Smith, 1994; Schermelleh-Engel, Moosbrugger, & Muller, 2003). The explained variances ranged from .29 to .51 in social justice attitude factor; ranged from .48 to .54 in perceived behavioral control factor; ranged from .56 to .83 in subjective norms factor; and, ranged from .55 to .76 in behavioral intention factor. As for factor loadings (Figure 1), they ranged from .54 to .71 in the first factor; from .70 to .73 in the second factor; from .75 to .91 in the third factor; and, from .74 to .87 in the fourth factor. Findings showed that the scale, as in the original form, consisted of 24 items falling under four factors. Eurasian Journal of Educational Research 31 Figure 1. SJS first-order confirmatory factor analysis. Second-Order Confirmatory Factor Analysis As Kline (2011) suggested, in order to test for whether the four factors are components of a higher latent structure (social justice behaviors), something which cannot be measured directly, the second order CFA was applied. The fit indices of the second order CFA were found to be similar to that of the first order CFA fit indices: χ2=671.64, df=248, χ2/df=2.70, GFI=.90, AGFI=.88, RMSEA=.05, NFI=.96, NNFI=.97, CFI=.97, SRMR=.04. When explained variances were examined, it was found that 45% of the variability of social justice behavior was expressed by social 32 İlker Cırık justice attitude factor; 49% by perceived behavioral control factor; 24% by subjective norms factor and 88% by behavioral intention factor. The factor loadings of the second order CFA were found to be the same as the first order CFA factor loadings (Figure 1, Figure 2). When the levels of factor effect on social justice behaviors are taken into consideration, it was specified that standardized direct effect size of the first factor was .67; of the second factor .70; of the third factor .49 and of the fourth factor .94. Kline (2011) determined standardized direct effect sizes as <.10 small; about .30 medium and >.50 large. Values found showed that factors significantly explained the latent variable of social justice behavior.