| 英文摘要 |
The present research assesses the causal effects of being assigned to high-ability groups in junior high according to 9th graders’ academic achievement in Taiwan. Using data gathered by the Taiwan Education Panel Survey in 2001 and 2003, the sample is divided into three school types: schools with no ability grouping, schools grouping students only at certain grades, and schools grouping students in all grades. Within each school type, this research focuses on estimating the average treatment effect (ATE) by comparing separately the achievement of 9th graders who have self-identified to high-ability groups in both 8th and 9th grade, or in 9th grade only, to those having no tracking experience and enrolled in schools without ability grouping. The research uses multilevel linear modeling, single-level propensity score matching (PSM), and the combination of single-level PSM and cross-classified random effects model (CCREM) to estimate the ATE of being assigned to high-ability groups. According to the analysis of the model combining PSM and CCREM, the research finds that only students assigned to high-ability groups at schools which group students at certain grades have positive impacts on their 9th grade achievement. |