The Causes of Poor Eyesight among Primary School Students - a Longitudinal Study

This study examines the causes of poor eyesight among 1st to 5th grade primary school students. We consider whether gender, parental employment, and student achievement affect the severity of the poor eyesight. The SAS 9.1 PROC MIXED, a flexible program suitable for fitting individual growth models, was used. A number of statistical models were compared, and we concluded that the reduced model with random slope (assumed compound symmetry, different for each group covariance matrix) was the best fitting model (AIC = 404.6, BIC = 355.9). Using this selected model, we observed a statistically significant decreasing poor eyesight trends (¿² = 5.43, p = 0.0214). Female students and male students had initially different levels of eyesight but the same rate of decrease over time (¿² = 5.13, p = 0.0239). However, significant differences in eyesight trends were found between high and low achievement groups when considering gender as well (interaction effect, ¿² = 6.65, p = 0.0102). Finally, we observed no significant effect of parental employment on the response pattern to poor eyesight.