不同協作學習模式對運算思維學習成效之影響：以相互信任、溝通效能、人際喜好、自我效能為中介變項

Abstract

隨著資訊科技的蓬勃發展，生活中許多產品與服務都是需要藉由大量數據進行運算與分析，顯示出人們仰賴資訊科技來解決生活問題的頻率日益升高，於是越來越多研究紛紛闡述運算思維之重要性與教育上的應用，並詮釋其為數位時代下，每個人所需具備的基礎能力。本研究以電腦中介傳播（Computer-Mediated Communication, CMC）的觀點切入，預期協作學習（collaborative learning）可以提升運算思維（Computational Thinking）之學習成效，並探討在協作學習中，不同溝通情境與不同小組組成對運算思維之影響，及學習者之間的相互信任（mutual trust）、溝通效能（communication effectiveness）、人際喜好（interpersonal liking）與自我效能（self-efficacy），是否會具有中介效果。本研究採實驗研究法（experiment research），以2（溝通情境：匿名／非匿名）X 2（小組組成：異質／同質）設計，進行隨機抽樣（random sampling）班級、隨機分派（random assignment）溝通情境，再分層隨機抽樣（stratified random sampling）小組組員，將研究對象（N = 156）分為四組實驗組與一組控制組，實施為期四週之教學實驗。

本研究透過獨立樣本t檢定、雙因子共變數分析與調節式中介分析完成資料處理與分析，研究結果顯示：（一）協作學習對運算思維學習成效具正向影響、（二）協作學習中的匿名溝通情境對運算思維歷程具正向影響、（三）匿名協作學習中的運算思維歷程，同質分組高於異質分組、（四）同質分組會正向調節人際喜好對於匿名溝通情境與運算思維歷程之中介效果，以及（五）協作學習對自我效能具正向影響。最後根據研究結果，提出未來研究與教學實務之建議。

With the blooming of information technology, many products and services can be operated based on a large amount of data. People rely on information technologies to solve life problems more and more frequently. Thus, there's a growing body of research illustrating the importance and the application in education of computational thinking, and indicating that it is the basic ability everyone needs to have in the digital age.

This study investigates how computational thinking can be taught in computer-mediated learning environments through different collaborative learning modes. Using experiment as research method, we look into students’ learning performance under different mediated communication setups (anonymous/non-anonymous) and grouping compositions (homogenous/ heterogenous) as independent variables. Also, we include mutual trust, communication effectiveness, interpersonal liking and self-efficacy as mediator in this model.The two-by-two between experiment design used random sampling, random assignment and stratified random sampling to assign students (N = 156) into different conditions, including experimental groups and one control group. The result is analyzed with independent sample t-test, two-way ANCOVA and moderated mediation.

The main conclusions of this study show that: (1) collaborative learning condition positively contributed to students’ learning performance; (2) anonymity is positively related tothe skill of process building in computational thinking; (3) homogeneous groups performed better in computational thinking process than heterogeneous groups; (4) homogeneous groups moderate the indirect effect of anonymity in computational thinking process through interpersonal liking (moderated mediation); (5) self-efficacy is positively related to collaborative learning.