THE MEDIATING ROLE OF STUDENT’S ACADEMIC SELF-BELIEFS IN STUDYING EDUCATIONAL EQUITY A comparison between Sweden and Ukraine in TIMSS 2011

Sammanfattning: Aim: The aim of this study is to enrich the understanding of student-level mechanisms involved in shaping (in)equity in educational outcomes among eight-graders in Sweden and Ukraine by examining the role of academic self-beliefs in science, represented by motivational constructs of self-concept, interest and utility value, as mediators of the association between student socioeconomic status (SES) and achievement in science. Theory: Expectancy-value theory (EVT) of achievement motivation (Eccles (Parsons) et al., 1983) laid out the fundament for the author’s empirical model. Their perspective was used to identify within TIMSS 2011 two sets of hierarchical self-beliefs constructing academic identity. To further establish them as the mediators between SES and academic achievement and thus build a conceptual framework, the reference to Eccles’ EVT model, Bandura’s sociocognitive perspective (2012) and Krapohl & Plomin’s (2015) finding on genetical mediation along with a literature study were used. Method: Structural Equation Modeling (SEM) is applied in this study to TIMSS 2011 data for Sweden and Ukraine using Mplus version 7.4 software. Only single-level model with complex option is fit into data for reporting results, as the author’s focus is on studentlevel mediating constructs. Intraclass correlation coefficients are represented to support this choice. It was initially planned that five latent indices would be used within the specified model, however, one latent mediator representing student’s interest in science was excluded due to multicollinearity issues. Measurement invariance was performed to check for parameter estimates' comparability across Sweden and Ukraine. Results: Only academic self-concept or self-confidence in science was found to be a significant non-cognitive mediating factor in the interplay between student socioeconomic status and science achievement in both Sweden and Ukraine. In Sweden, the total indirect effect of two academic self-beliefs makes up for ¼ of SES – science achievement relationship, whereas academic self-concept alone mediates one third of this relationship. In Ukraine, half of SES – science achievement association is mediated by the effect of academic self-concept. Students’ value of science (utility) had a nonsignificant indirect effect, although it added up to self-concept indirect effect in Ukraine and reduced the total indirect effect in Sweden by .01 points. Interest had to be excluded from the structural model due to its high correlation with both measures discussed above. The finding confirms that in order to get a fair image of educational equity in the country it is important to go from the individual level up to family, classroom, school and society, and no other way around, as they are additional layers built up around the individual’s non-cognitive and cognitive characteristics.