「中文部件組字遠距聯想測驗」之發展與其信效度研究

The Development of Chinese Radical Remote Associates Test

本研究擬發展「中文部件組字遠距聯想測驗」（Chinese Radical Remote Associates Test，以下簡稱CRRAT），期能形成華人遠距聯想創造力的完整測量系統。 測驗編製者首先挑選出60 組具有同樣部件之三個衍生字組，編成CRRAT 預試題 本，計60題。接著以196 名大學生進行預試，依項目分析結果挑選出30 題做為正 式試題，並區分為難度、鑑別度相近的甲、乙兩式題本（各15 題）。其次，甲式 以173名、乙式以187 名大學生進行正式施測及信效度研究。結果發現，在信度方 面，甲、乙兩版本之內部一致性係數分別為.706 與.685；在效度方面，兩版本皆與 「中文遠距聯想測驗」（Chinese Remote Associates Test， CRAT）（任純慧、陳學 志、練竑初、卓淑玲，2004）有顯著相關（r 值分別為.317與.375），與「中文詞 彙遠距聯想測驗」（Chinese Word Remote Associates Test， CWRAT）（黃博聖、陳 學志、劉政宏，2012）有顯著相關（r值分別為.585與.350），與「頓悟性問題測 驗」（邱發忠、陳學志、徐芝君、吳相儀、卓淑玲，2008）亦有顯著相關（r 值分 別為.417與.448）。另外，本測驗與「新編創造思考測驗」（吳靜吉，1998）各指 標皆無相關存在（r 介於-.103～.151 之間），顯示本測驗具有良好的聚斂效度與區 辨效度。此外，不同性別大學生在甲、乙式均無顯著差異，中文相關系所與非中文 相關系所大學生在乙式亦無顯著差異。本研究最終針對測驗的信效度與未來可能發 展提出討論與建議。

The purpose of this study was to develop “Chinese Radical Remote Associates Test” (CRRAT)， which is a comprehensive system to measure the cognitive ability of Chinese remote associates test for Chinese people. To develop CRRAT， firstly， we choose 60 sets of derivative characters; every set contained three characters that derived from a same radical. These 60 sets was the pretest of CRRAT， and 196 undergraduates participated. Then 30 items of them were selected into CRRAT according to the result of item analysis; in particular， two alternative forms， version A and version B， were edited based on their item difficulty and discrimination， each version was composed of 15 items. With respect to the reliability and validity of CRRAT， we recruited 173 undergraduates for version A and 187 ones for version B to join the formal test. As for the reliability， the coefficients of internal consistency for version A and version B of CRRAT were .706 and .685 respectively. With regard to the validity of CRRAT， it is significantly correlated (r = .317， r = .375) to the “Chinese Remote Associates Test” (CRAT) (Chun-Hui Jen， Hsueh-Chih Chen， Hung-Chu Lien， & Shu-Ling Cho， 2004)， and to the “Chinese Word Remote Associates Test” (CWRAT) (Po-Sheng Huang， Hsueh-Chih Chen， & Cheng-Hong Liu， 2012) (r = .585， r = .350)， and to the “Insight Problems” (r = .417， r = .448). However， it wasn’t correlated to all the index of “New Creativity Test” (r = -.103 to .151)， it shows that CRRAT has the convergent validity and discrimination validity. Besides， there was no gender difference between the responses of two versions of CRRAT， and the response of version B was not different between the Chinese literature major and other majors undergraduates. We have the discussion and suggestion for the reliability， the validity， and the future potential development of the CRRAT.