A Meta-Analysis of the Effects of Sexuality Education Programs for StudentsFacilities
關鍵詞 Key words : 性教育;教學效果;後設分析;sexuality education;teaching effect;meta-analysis
研究目的：本研究旨在經由後設分析，整合個別研究所發現的性教育教學效果與調節因素，作為學校推行性教育之參考。研究方法：本研究搜尋1990－2012年性教育教學文獻，經篩選後納入97篇論文，共獲得432項個別變項效果量，其包含實驗組與對照組國小至大學之研究對象，共69,206位學生；本研究運用Study DIAD評估研究的品質；採用隨機效果模式，計算性教育教學效果整體及分層面的加權平均效果量，並以Mann-Whitney U test和Kruskal-Wallis test考驗不同文件與介入特徵其性教育教學效果的差異。研究結果：性教育教學的整體效果為0.37；分層面認知、情意及行為之加權平均效果為0.63、0.32及0.21，其中以認知層面效果最大；受試者類別、發表語言、出版類別、教學方案類型以及教學者的不同而有顯著的差異。研究結論：整體而言，性教育教學效果屬於中等程度，性教育教學對認知、情意以及行為層面效果不一，受到許多調節因素所影響。其中，以國小階段、貞潔加上避孕的教育方案以及健康專家或護士為教學者的教學效果較高。
Purpose: The purpose of this meta-analysis was to examine the effects of sexuality education programs for students and the impact of potential moderator factors. Methods: Relevant electronic databases, journals, documents, and reference lists were searched for eligible studies. We identified 97 studies on sexuality education programs with 432 independent effect sizes that were published in either Chinese or English from 1990 to 2012, comprising a total of 69,206 students in elementary school to college level. The Study Design and Implementation Assessment Device (Study DIAD) was used to address the quality of the studies. Weighted mean effect sizes were calculated based on the random-effects model. Additionally, Mann-Whitney U test and Kruskal-Wallis test were performed to examine the moderating effects. Results: The results of this study revealed that the overall weighted mean effect size for the sexuality programs was 0.37 with a 95% CI 0.31~0.43. The weighted mean effect sizes for the cognitive, affective, and behavioral outcomes were 0.63, 0.32, and 0.21, respectively. Moreover, moderator analyses showed that the grade level, the publication language, the article type, the program type, and the instructor type significantly influenced the weighted mean effect sizes. Conclusions: Overall, the sexuality education programs for students were moderately effective. The effects of sexuality education programs differed in students' cognitive, affective, and behavioral outcomes and were influenced by several moderator factors.