合作有助於規則發現嗎？訊息交換方式與主導性對新想法產生的影響

Abstract

在如何促進新想法的產生這個議題上，合作的功效是常被提起。過去研究發現新角度假設的產生是受試者能否在規則發現作業（2 4 6問題）成功解題的關鍵因素，本論文旨在承繼上述的觀點，探討合作（兩人小組）是否有助於新角度假設的產生，並瞭解何種互動方式最能促進新角度假設的產生與解題表現，以找出合作有效的機制。實驗一比較兩種常見的合作方式（訊息交換方式不同，為面對面自由討論與非面對面假設交換）對於解題的影響，並與個別解題組比較，以初步瞭解合作是否有利於解題表現以及新角度假設的產生。結果發現兩者均可有效提升新角度假設的產生以及解題正確率。實驗二則在資訊來源與數量相同的狀況下，操弄交換訊息的差異（是否進行假設交換或自由討論）與成員對於測試的貢獻（是否能夠主動進行測試），企圖瞭解主導性與訊息交換量（交換訊息的多寡）如何影響新角度假設的產生。結果發現在一人主導進行解題的情況，即使有訊息的交換也無法提升解題表現，若兩人能夠對等進行解題互動，則交換的訊息量與解題表現呈線性關係。

Past research showed that generating new-perspective hypotheses is crucial for solving a rule discovery task, such as ‘2 4 6 problem’. Collaboration is usually used for facilitating the generation of new ideas. In the current research the author intended to find out whether in which way that collaboration can improve the performance of rule discovery task, particularly the generation of new-perspective hypotheses. In experiment 1, the team members (two in each team) were either allowed to freely interact with each other while they were solving the problem together or to exchange hypotheses alternatively while they were solving problem on they own. Compared to the control group (solving problem individually), both types of collaboration could increase the correct rate and facilitate the generation of new-perspective hypothesis. In experiment 2, three types of information exchange were applied to each of 18 teams. Team members must exchange hypotheses with each other in all conditions and the number of testing trials was kept the same (twelve times) across conditions as well. However, the second member of each team might (1) had no control of what to test in all testing trials (dominant condition), or (2) took charge of what to test for half of the times (six times. non-dominant condition), or (3) allow to freely discuss with the other member before they tested any instance (freely interactive condition). Compared to the control condition (the fourth group in which an individual solved the same problem, with a bystander watching), exchanging hypothesis was found to be in vein in the dominant condition but had facilitating effect in the non-dominant condition. In addition, freely discussion after exchanging hypothesis and before testing further enhances the performance. How information exchange in collaborative problem solving influenced performance is discussed in the thesis.