SEM何以優於回歸分析(以組織研究為例)

antipower

SEM can simultaneously examine different paths/relationships in one model, while regression can only test one specific relationship (i.e. direct effects, mediation, moderation). However, SEM is not always expected to use. Reviewers usually like a simpler method technique. So, if you are only testing a simple relationship such as direct effects, mediation, or moderation. A regression is preferred to used than SEM. However, if your research model is complicated, in order to simultaneously consider all the relationships in one model, SEM is preferred.