中文部件組字遠距聯想測驗編製及其解題認知成分之研究

Abstract

本研究擬發展以中文部件組字為基礎的《中文部件組字遠距聯想測驗》（Radical Remote Associates Test, RRAT），期能形成針對華人遠距聯想之創造力進行測量的完整系統。在測驗編製上，首先挑選出60組具有同樣的部件之三個衍生字組，編成RRAT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，本測驗在信度方面，甲、乙兩版本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值別為.585與.350），與「頓悟性問題測驗」（邱發忠，2005）亦有顯著正相關存在（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與.151之間），顯示本測驗具有良好的聚斂效度與區辨效度。其次，本研究擬針對國內現存三種中文遠距聯想測驗（RRAT、CWRAT、CRAT）進行解題認知成分之探究。研究參與者為36位大專院校學生。研究參與者須完成由三種測驗挑選之各5道題目所組成的作業試題，並於答題後進行主觀頓悟感受評等，最後接受「頓悟性問題測驗」。結果發現參與者在三種中文遠距聯想作業型態之正確率與解題前的嘗試次數為無相關或負相關，顯示三種作業型態與解題前的嘗試次數無關聯。其中以CWRAT最為符合聯結理論之假設；其次在頓悟成分方面，RRAT與客觀的「頓悟性問題測驗」有最高的相關，且參與者的主觀頓悟感受亦最高。顯示RRAT的作業型態與頓悟思考的關聯較高。最後研究者依據研究結果針對遠距聯想及頓悟思考之測量與研究提出建議。The purpose of this study was to develop “Radical Remote Associates Test” (RRAT) with radical combinations, hoping to build the system of Chinese remote associates test for Chinese. To develop RRAT, the researcher first formed a 60-item pretest by choosing sixty radicals and their three derivative characters, and deleted the same component of derivative characters. This pretest was administrated to 196 college students, and 30 items of the pretest were selected as formal scale according to the item analysis of preparing research. The researcher then divided these items into two versions, version A and B, with the same difficulty and discrimination, and each of them included 15 questions. Next, the formal scale, including version A and B, was distributed to college students, and 173 (version A) and 187 (version B) data were collected respectively. As to reliability, the coefficient of internal consistency of the two RRAT versions was .706 and .685 each. As to validity, each of the two versions of RRAT was significantly positively correlated (r = .317, r = .375) with “Chinese Remote Associates Test” (CRAT) (Chun-Hui Jen, Hsueh-Chih Chen, Hung-Chu Lien, Shu-Ling Cho, 2004); each of the two versions of RRAT and “Chinese Word Remote Associates Test” (CWRAT) (Po-Sheng Huang, Hsueh-Chih Chen, Cheng-Hong Liu, 2012) had a significant positive correlation (r = .585, r = .350); still, each of the two versions of RRAT and “Insight Problems” were also significantly positively correlated (r are .417 and .448). Besides, the two RRAT versions and all the index of “New Creativity Test” had no significant correlation, demonstrating the good Convergent validity and discriminant validity of RRAT. Another purpose of this study is to analyze the cognitive components during answering process of the three Chinese remote associates tasks (RRAT, CWRAT,& CRAT) in Taiwan. 36 college students were participated in this study. Each of them first had to complete a task consisting of 15 items, respectively selected 5 items from the three Chinese remote associates tests with the same average difficulty. After finishing the task, they had to rate their subjectively insight feeling, and then an “Insight Problems” assessment was administered to them in the end. The result showed that no significant correlation or significant negative correlation was found between the three Chinese remote associates task and the trial frequency before answering, indicating that there was no relationship between these two variables. Among the three Chinese remote associates tests, CWRAT was the one being most consistent with the hypothesis of associative theory. As to the part of insight, RRAT and objective "Insight Problems" held the highest correlation, and the rating of subjectively insight feeling was also the highest, showing that there was a stronger relationship between the type of RRAT and insight thinking. In the last, the researcher would make some suggestions for the measurement of remote association and insight thinking based on the research conclusion.