The Development of Learning to Learn Competence Scale for College Students

大學生學習素養量表編製及其相關因素之研究

廖凡毅(Liao, Fan Yi)

指導教授 ： 陳素燕

國立清華大學/其它/學習科學研究所/碩士(2015年)

摘要

近年來，儘管多數文獻已廣泛探討過學習素養之於現今社會的重要性，但相關的實徵研究仍相對匱乏，且欠缺客觀的量測工具。因此，本研究旨在探討學習素養的本質並編製一份適用於大學生的學習素養評量工具。以大學生為主要研究對象，樣本來自全台共計25所大專校院，有效樣本數合計為1744人。採取描述性統計、項目分析、信度分析、因素分析、相關分析、迴歸分析、差異分析及常模建置等統計方法，進行量表編製及其相關因素之考驗。茲將研究結果摘要如下：
一、學習素養量表係由三個分量表組成，在學習的覺察與認知分量表中，包括「自身為學習者」、「社會互動」及「學習潛力」等因素；在學習的態度與傾向分量表中，包括「接受挑戰」及「學習承諾」等因素；在學習的遷移與調節分量表中，包括「遷移能力」及「調節能力」等因素。共計38題，為六點李克特氏量表，得分愈高表示受試者的學習素養愈明顯。
二、在信度研究的測量中，學習的覺察與認知分量表信度係數介於.76至.89之間，學習的態度與傾向分量表信度係數.88至.90之間，學習的遷移與調節分量表信度係數介於.88至.90之間。各分量表的組合信度皆達.70以上，平均變異萃取量均在.30以上，題項間具有適度的共同變異，題項信度佳。分量表間隔一周的測量穩定係數介於.79至.90之間。結果顯示學習素養量表具備內部一致性與穩定性。
三、以結構方程式之測量模式進行驗證性因素分析，結果顯示學習素養量表評量的因素結構與實徵資料間的適配度佳。而學習素養與後設認知及動機間的關係，支持學習素養量表評量具備合理的效標關聯效度。此外，學習素養亦能有效預測大學生的獲益評估。
四、不同性別的大學生在學習素養的表現上並沒有顯著差異。然，不同年級、學院、班級排名、打工經驗及社團經驗的大學生，則表現出不同程度的學習素養。
最後，依據上述的研究結果，本研究針對大學生學習素養量表之實務應用及未來研究提出具體建議，俾供相關領域實務工作者與研究人員參考。

關鍵字

學習素養 ； 大學生 ； 量表編製

並列摘要

In recent years, increasing attention had been drawn to learning to learn competence politically, practically, and academically, but it still remained a fuzzy notion. No specific instrument was developed for higher education. Hence, the present study aimed at exploring the essence of learning to learn competence and developing the Learning to Learn Competence Scale (L2LCS) for college students. L2LCS was a 6-point-Likert scale with 38 Chinese items. Three subscales were included: knowledge, attitudes, and skills. In total, 1744 college students were recruited from 25 universities across Taiwan. The data collected were analyzed through descriptive statistics, item analysis, reliability analysis, factor analysis, correlation analysis, discriminant analysis, regression analysis, and norm development. Major findings were summarized as follows. First, L2LCS consisted of 7 factors. In the knowledge dimension, there were “Awareness of self as a learner”, “Awareness of social interaction”, and “Awareness of potential of learning”; in the attitudes dimension, there were “Readiness to challenges” and “Commitment to learning”; and in the skills dimension, there were “Capacity to transfer” and “Capacity to regulate”. Second, the internal consistency was estimated by Cronbach’s alpha, composite value, and average variance extracted. The coefficients of knowledge subscale lay between .76 and .89; the coefficients of attitudes subscale lay between .88 and .90; and the coefficients of skills subscale lay between .88 and .89. For the entire scale, the coefficient reached .965. All the CR values were higher above .70 and all the AVE values were higher above .30. These showed that the observed items could explain appropriate variance together with good reliability. The coefficient of test-retest reliability with one-week interval was between .79 and .90. Hence, L2LCS was proved to have good internal consistency and stability. Third, the results of confirmatory factor analysis indicated that the model was well supported by observed data. These results supported the scale structure. Learning to learn competence was significantly related to metacognition and motivation. It was also a predictor of students’ learning outcomes, estimate of gains in particular. Fourth, except gender, significant difference on learning to learn competence was found by grade, college, class ranking, job experience, and club experience. Eventually, the application of L2LCS and the suggestions for the future research were discussed.

並列關鍵字

learning to learn competence ； college students ； scale development

參考文獻

Cheng, Y. Y., Yeh, L. J., Liu, K. S., & Mok, M. M. C. (2011). Constructing the Indicators of Undergraduate Students’ Key Competences. Psychological Testing, 58(3), 531-558. (in Chinese)

Chiou, H. (2011). Principles and practices of structuring equation modeling with LISREL/ SIMPLIS. Taipei: Yeh Yeh. (in Chinese)

Tsai, C. T. (2011). Competence: DNA of curriculum reform. Taipei: Higher Education. (in Chinese)

Allan, J. (1996). Learning outcomes in higher education. Studies in Higher Education, 21(1), 93-108.

Bagozzi, R. P., & Yi, Y. (1988). On the evaluation of structural equation models. Journal of Academy of Marketing Science, 16(1), 74-94.