Adaptation of Self-Perception Profile for Children and Testing Measurement Invariance According to Different Variables

Abstract

The present study aims to adapt Self-Perception Profile for Children (SPPC) and to test its factor structure equation regarding several variables. To attain this goal, the SPPC developed by Harter (1985), being originally in English, is adapted into Turkish for 5th to 8th grade students in primary schools and equation of the five-factor-structure of the adapted tool is tested with regard to gender, socio-economic status, and grades. Survey method is conducted in this study and the data proceeded in this study are collected from two study groups within the framework of the pilot (n=320) and the empirical studies (n=1798). During the adaptation process of the SPPC, experts are asked for review after translation has been completed, the pilot form is translated back and the adapted form is implemented to a group of individuals who are high proficient both languages in order to examine whether linguistic equivalence of the tool is ensured. Upon gathering evidence on the linguistic equivalence of the translation, the pilot study is conducted. The collected data are analysed and field experts are asked for review for the items with psychometric problems. Considering the feedbacks of
the field experts and the participants during the pilot study, some items have been revised, thus the final form is prepared. By using the collected data, exploratory and confirmatory factor analyses are computed, correlations between the subscales and item-test correlations ar e calculated for reliability studies; Cronbach’s Alpha values are calculated to find out internal consistency coefficients, and finally, test-retest method is performed to determine form stability coefficients. Multi-group confirmatory factor analysis is computed to determine the factor structure equation of the SPPC for the groups with respect to gender, socio-economic status and grades. The analysis results attest that the five–factor–structure of the tool is equal for the relevant variables; in other words, measurement invariance is supported.