《Coleman報告書》指出學校因素對學生學業成績的影響很小，引發後續不少探討學校有 無效能的研究。在這類眾多的研究中，並無一致性的定論。而這些研究往往忽略學校的階層 特性，沒有採用符合此特性的階層線性模式（Hierarchical Linear Modeling, HLM）來分析。本 研究以三階層HLM分析「臺東縣教育長期資料庫」的國中資料，發現學校之間的變異雖然很 小，但是學校內班級之間的變異並不小，這是《Coleman報告書》中沒有發現的。本研究發現 學校的班級因素有：班級平均家長社經地位、導師族別和人力資本等，對學生學業成績具有 不小的影響力。這些學校因素當中，導師的人力資本是教育作為可以致力提升的著力點，這 也是學校發揮效能的重要關鍵。

Because Coleman et al. argued that the influence of school factors on the academic achievement of students was low, humerous subsequent studies have focused on examining school effectiveness. The findings of these studies were not coherent because of the limitation of disregarding the hierarchical data structure of schools. To reexamine the conclusion presented by Coleman et al., this study applied Taitung Educational Panel Survey data in three-level hierarchical linear modeling. The results reveal that, although the variance in the academic achievement of students among schools was low according to Coleman et al., the variance among homeroom classes was high, which contradicts their conclusion. This current study indicated that class factors, such as average socioeconomic status in class, and the ethnic and human capital of homeroom teachers, substantially influence the academic achievement of students. Regarding school factors, the human capital of homeroom teachers affects the academic achievement of students and is key to school effectiveness.