透過fsqca(模糊集質性比較分析)重新詮釋SEM調查數據

Abstract

社會科學具有高複雜性的特徵。一般利用結構方程模式(SEM)來處理。然而，一種根據集理論(set theory)和布爾代數(Boolean algebra)來評估複雜性的較新技術(fsQCA)是可運用的。在這項研究中，我們應用模糊集定性比較分析(fsQCA)在SEM被用來尋找四個狀態的因果關係的實證實驗研究的調查數據。感知視覺新穎性、感知視覺複雜性為獨立變數，而喚起和認知投入作為中介變數來解釋方法並排除由大學提供的非正式在線視頻。研究中，88個個案回應動畫般視頻和95個演講般視頻。考慮VidType(動畫或演講視頻)的分組資料的三種比較方法是比較由fsQCA和早期SEM結果兩者的因果關係獲得的。這項研究顯示藉由必要性發現中介變數證明fsQCA的可能用途。例如:單一條件“認知投入”是必要條件，也恰好是一個中介變數。

Key words: fsQCA, SEM, 認知參與，複雜性

Social Science is characterized by high complexity. A common approach to deal with it is Structural Equation Modeling (SEM). However, a relatively new analytic technique (fsQCA) based on set theory and Boolean algebra aiming to assess complexity is available. In this study, Fuzzy-Set Qualitative Comparative Analysis was applied to online survey data borrowed from an empirical experimental research where SEM was used to find causal relations among four conditions; Perceived visual novelty, perceived visual complexity as independent variables, and arousal and cognitive engagement as mediator variables to explain approach and avoidance to Informal Online Videos offered by universities. From here, 88 cases correspond to people responding to animation-like videos and 95 cases to lecture-like videos. A comparison of three ways of grouping data considering VidType (animation or lecture video) was made before contrasting causal relations obtained with fsQCA against previous SEM results. This study shows evidence of a possible use of fsQCA for finding mediator variables through necessity. For example; single condition “cognitive engagement” is a necessary condition for approach and avoidance, and also happens to be a mediator variable.