感恩介入之後設分析

本研究以後設分析的技術來探討感恩介入的成效。無論是從人性的角度或以社會生活的觀點來說，感恩的議題向來備受重視，尤其近期興起的正向心理學更將感恩視為六大美德下的二十四個強項之一，並且許多感恩介入的研究更顯示，感恩可以激發許多正向的情緒，促使幸福感的提升；本研究透過「臺灣博碩士論文知識加值系統」、「中華民國期刊論文索引影像系統」、「PsycINFO」、「PsycARTICLES」、「PubMed」等電子資料庫收集國內外2006年至2015年2月之前，最終收錄38篇符合納入標準之實驗組控制組前後測設計與單組前後測設計之樣本文獻，研究對象共3973人。本研究目的除了計算感恩介入的效果量之外，同時也探討可能影響介入成效的執行層面、個案特徵、方法學特徵、介入過程特徵等變項。　　結果發現：(1)國內外感恩介入具有中度效果( =0.574)；整體維持效果是低度效果( =0.298)，且受出版偏誤影響不大。(2)成效指標中，對感恩的介入成效最高（高度效果），其次是幸福感（中度效果），對正向情緒、負向情緒和憂鬱則具有低度效果。(3)可能影響介入成效的變項包括：「執行層面」、「招募地區」、「出版型態」、「流失率」和「介入持續時間」。上述調節變項雖與整體研究文獻的效果量變化明顯有關，但尚有其他潛在調節變項的影響。本研究根據上述研究結論，提供建議給未來研究者與實務工作者，作為未來研究方向的調整與參考之依據。

關鍵字

感恩；介入；幸福感；憂鬱症狀；後設分析

並列摘要

The purpose of this study is to adopt meta-analysis methods to investigate the effects of gratitude interventions.Whether it is from the human point of view or perspective of social life, the gratitude issue has always been taken seriously, especially the more recent emergence of positive psychology gratitude regarded as one of the six virtues under twenty-four strengths, and many studies involved more show gratitude, gratitude can inspire a lot of positive affects, promote well-being promotion. 　　This study collects nearly nine years of “gratitude interventions” from 2006 to February 2015 to compile 38 studies with single group or between-groups design are included, and the total subject numbers are 3973. This meta- analysis study computes effect sizes of gratitude interventions and explore the effects of moderators which include executive level, subject variables, outcome variables, and methodological variables in an attempt to search for possible moderator variables. 　　The results of this study reveal that: (a)the overall immediate effect( =0.574) of gratitude Interventions is medium. (b)the immediate effect of gratitude is medium to high, the immediate effect of well-being is close to medium , the immediate effect of positive affect, negative affect, and depression are low. (c)examination of potential moderators indicated that executive level, recruitment, publication patterns, attrition rate, and duration. Although the aforementioned moderators are correlated to effectiveness change of the overall research literature, the explaning power is not strong enough. There are still other effects of latent moderators.