Browsing by Author "Bofah, Emmanuel Adu-tutu"

This study used the Trends in International Mathematics and Science Study (TIMSS 2011) dataset for Ghana, which measures the trends in Mathematics and Science Achievement at the fourth and eighth grades. The focus of the present study is on the eighth grade Ghanaian students’ with a sample size 7323 (47% girls) who participated in TIMSS2011. The mean age was 15.81 with standard deviation of 1.50. This paper first addresses the issue of measuring socio-economic background in the context of the TIMSS 2011 data set using students’ home resources. This is based on the fact that the literature indicates that data on home possessions/resources collected from young children are much more reliable compared to the information they provide about their parents’ education, jobs, and income as such this study uses data on home resources as a measure of students Socio-Economic Status (SES). Applying a two-level mixture modeling technique by accounting for the clustering in the data set, the study explored the profiles of students’ (SES). Latent class analysis was used for the profiling. The two-level latent class analysis takes into account the measurement error and the variation of the latent class indicators across different class/schools. The detail process involved in obtaining the index of students’ socio-economic profiles from home possessions/resources items using latent class analysis is described. Once the SES measure was obtained, a discriminant analysis was used to validate the students SES. The relationship between the demographic variables (e.g., parental education, language spoken at home, parental involvement and gender) and students’ SES were examined. The analysis identified three latent classes of students based on reported home resource namely: the high SES, the intermediate SES, and the low SES group. The discriminant analysis based on the eleven household items was able to correctly classify 92.2% of the individual students into their appropriate SES group. Furthermore, the variables that had the most significant association with students socio-economic profile were investigated. Multinomial logit latent-class regression models were posited. The final analysis used the hierarchical regression analysis to access the clusters of variables to evaluate the relative importance of the predictors for students’ mathematics achievement. The results indicated that gender, parental education, SES, students’ educational aspiration, language spoken at home, and parental involvement variables significantly predict students’ mathematics achievement. When the variables were entered as six blocks, students’ educational aspirations were found to have the greatest variance explained for mathematics achievement. Gender and parental education explained additional 2% and 2.9% respectively of the variance in mathematics achievement. Speaking English always at home and being in the low SES group did not have any significant effect on students’ mathematics achievement. The findings of this study provide information to educators, researchers, parents, teachers, and policy makers about the effect of home resources on students’ academic achievement. This thesis advocates that governments should provide financial support for students from low SES. In addition, financial incentives to schools in low income areas should be increase to help close the achievement gap between students’ from the low-SES and high-SES.