創造力相關變項效應量之後設分析

Abstract

本研究以後設分析方法評估創造力訓練之效果。資料收集是透過電腦資料庫之網 路查尋，如 ProQuest, ERIC, EBSCOhost, 及 Education Complete。所用關鍵字為\"Creativity training\"， 也 對 ”Journal of Creativity Behavior\"及\"Creativity Research Journal\"作地毯式搜尋，另外對每篇論文後面所列之參考書目加以追 蹤。總共找出 22 篇可計算出效應量 (effect size) 之論文。這 22 篇論文共有 202 個效應量，其平均值為 0.795，標準差為 0.76。但這 202 個效應量之殘差有自 我相關，不能用參數統計分析。故將每篇論文之效應量求平均值，得出 22 個殘 差無自我相關之效應量。此 22 個效應量之總平均值為 0.73，標準差為 0.45， 以單樣本之 t 考驗檢定此總平均值是否顯著不等於零，結果得出 t(21) = 7.72, P < .001，顯出就整體而言 ，創造力訓練有顯著果。然後對 202 個效應量作細部 分析。以非參數統計 Kruskal-Wallis Test 檢定結果發現不同的訓練方案有顯著的 不同效果。「Osborn-Parnes 的 Creative Problem-Solving 方案」及「根據 Osborn 的腦力激盪或 Gordon 的共辯 (Synectics) 原則所編成的自我訓喻方案」 效果最 大，而單獨針對創造力之個別元素所作的訓練，如「孵化期法」，「 SCAMPER 法」， 或單獨對創造力之態度作訓練，效果最小，其餘介乎其間。至於訓練實驗所使用 的評量創造力之工具、實驗設計的種類、受試者之年齡、及訓練期間之長短對創 造力訓練之效果無顯著影響。

The present study, by means of meta-analysis method, is to synthesize the effect of creativity training. The ProQuest, ERIC, EBSCOhost, and Education Complete on-line databases were scanned for researches evaluating the effectiveness of creativity training. The term used was \"creativity training\". \"Journal of Creativity Behavior\"and \"Creativity Research Journal\"were systematically, manually searched. Additionally some usable empirical articles were traced from the references of research papers. Altogether 202 effect sizes from 22 studies were converted from different statistics. The grand mean of the 202 effect sizes was 0.795 with a standard deviation 0.76. The grand mean of the 22 independent effect sizes was 0.73 with a standard deviation of 0.45. Result of one-sample t-test, t(21) = 7.72, p < .001, revealed a significant effect of the creativity training programs. By performing nonparametric statistics, Kruskal-Wallis Test indicated that effectiveness of different training programs were significantly different from each other. The training packages \"Osborn-Parnes Creative Problem-Solving Program\"and \"Self-instruction based on Osborn`s brainstorming or Gordon`s synectics\"had the greatest effect sizes while the single technique of ideation \"incubation\"and \"SCAMPER\", as well as the attitude training had the smallest effect. With the exception of training programs, no significance was found in the effect of moderators. It indicates that the kind of instruments measuring creativity, the experimental design, age of subjects and the duration of training would not significantly influence the evaluation of the effectiveness of creativity training programs.