Li Weiling, McDavid Martha Lindley, San Miguel Sandra F, Parker Loran Carleton
Evaluation and Learning Research Center, Purdue University.
College of Veterinary Medicine, Purdue University.
J Youth Dev. 2021;16(4):166-182. doi: 10.5195/jyd.2021.1055.
This paper presents the application of a meta-analysis approach to the evaluation of youth-learning data from the nationally distributed This is How We "Role" program. The application of meta-analysis for examining the impact of other multisite youth programs encountering similar data analysis challenges is discussed. At each This is How We "Role" program site, university partners collected data to examine youth-participant learning. Data analysis from these unique sites was challenging as the approach had to accommodate the innate heterogeneity across sites due to differences in implementation, sample size, and learning context. The meta-analysis method revealed details of the underlying variation between sites that could be masked by typical regression approaches, estimated overall program effects, examined subgroups and identified heterogeneity across project sites. The results showed the This is How We "Role" program generally increased learning at each site and as a whole, even though the program effects varied across sites. This example demonstrates the utility of using the meta-analysis approach to similar multi-site youth development programs.
本文介绍了一种元分析方法在评估全国范围内开展的“这就是我们如何‘扮演角色’”项目的青年学习数据中的应用。文中还讨论了运用元分析来检验其他面临类似数据分析挑战的多地点青年项目的影响。在每个“这就是我们如何‘扮演角色’”项目地点，大学合作伙伴收集数据以考察青年参与者的学习情况。对这些独特地点的数据分析颇具挑战性，因为该方法必须适应因实施方式、样本规模和学习环境差异而导致的各地点内在异质性。元分析方法揭示了各地点之间潜在差异的细节，而这些细节可能会被典型的回归方法所掩盖，它估计了项目的总体效果，检验了亚组情况，并识别了各项目地点之间的异质性。结果表明，尽管各地点的项目效果有所不同，但“这就是我们如何‘扮演角色’”项目总体上在每个地点以及整体上都提高了学习效果。这个例子展示了将元分析方法应用于类似的多地点青年发展项目的效用。
