本研究為透過整合分析了解青少年中輟之危險因子，及影響中輟之路徑分析。蒐集 中文22 篇，英文共17 篇，共39 篇，透過整合分析的應用程式分析423 個效應量。並 透過AMOS 軟體進行路徑分析。研究結果發現個人因素中人口學變項屬高效果量，認 知信念、心理疾患及人格特質屬中效果量。家庭因素中的家庭事件、家庭結構及教養態 度變項屬高效果量，學校因素中有10 項屬高效果量，其中學習興趣、出缺席率及學習 態度分居前三位，社會因素偏中低效果量。透過驗證性因素分析得知本研究所提出的理 論模式尚佳。在路徑分析上，十條路徑係數皆達顯著，在標準化總效果值中學校的影響 力最高，其次為家庭，兩者皆呈正向效果，但個人及社會因素呈負向效果。

The study understands the risk factors of the dropout teenagers through metaanalysis, and conducts the path analysis for dropout by using AMOS software. A total of 39 literatures were collected, including 22 Chinese and 17 English ones, to analyze 423 effect sizes by using the application program of meta-analysis. According to the study results, regarding the individual factor, demography variable shows high effect size, while perceived faith, mental illness and trait show medium effect size. In terms of the family factor, family event, family structure, character and attitude variables show high effect size. Moreover, 10 items under school factor show high effect size, among which the learn interest, absence rate and learning attitude are the top three. However, the social factor shows medium and low effect size. It is learned from the confirmatory factor analysis (CFA) that the theoretical model proposed by the study achieves good effect. In the part of path analysis, the coefficients of 10 paths all achieve significance. In terms of the overall effect value of standardization, school shows the highest influence, followed by family, both of which show positive effect. Nevertheless, the individual and social factors show negative effect.