科技融入於教育的時代下，數位遊戲式學習已成為發展趨勢，而第二語言學習過程中，自我效能與學習焦慮對外語習得具有影響力，其中，又以口語學習所產生之焦慮較大。為了解決口語教學現場上，時間與人力的限制，本研究建置數位遊戲式英語學習系統，並加入語音辨識功能，提供學習者更多口語練習機會，分析數位遊戲式學習與傳統學習對於學習成效、自我效能與口語焦慮之差異，並探討自我效能與口語焦慮是否能預測學習成效表現，除此之外，也進一步調查學習者學習成效與遊戲成效之相關性與影響。本研究研究對象為國小四年級學童（N=67），按原班級隨機分為使用數位遊戲式英語學習系統之實驗組（N=47），及進行傳統學習之控制組（N=20），兩組實驗時間皆為420分鐘，資料分析方式包含描述性統計、t檢定、無母數檢定、Pearson相關及多元迴歸，並於實驗結束後進行訪談，以了解學習者的學習感受。 研究結果顯示數位遊戲式學習與傳統學習能有效幫助學習者在學習成效上的表現，但實驗組與控制組於學習成效後測未達顯著差異，可能因口語任務的困難度較高，學習者於實驗期間無法完成所有的學習任務。學習者在自我效能與口語焦慮後測皆未達顯著差異，且兩組口語焦慮於後測上皆略提升，透過訪談發現，可能因遊戲系統所提供的辨識標準較不易達到，學習者對於口語表現不佳會感到緊張。不同學習環境下，自我效能與口語焦慮存有顯著負相關。透過數位遊戲式英語學習系統之學習者，其自我效能與口語焦慮表現可預測學習成效，也證實自我效能與口語成效之間存有顯著正相關，此外，學習者在遊戲中所獲得的星等總數，也可預測學習成效表現。;In an era with technology highly integrated in the field of education, digital game-based learning (DGBL) has become the mainstream. In second language learning, self-efficacy and learning anxiety hold the keys to language acquisition, among which English speaking leads to greater anxiety. With an attempt to overcome the limit of time and labor in traditional speaking courses, the current study developed a digital game-based English learning system with embedded speech recognition to provide learners with ample opportunities for practicing English speaking. Therefore, the current study aimed to examine the differences in learning performance, self-efficacy, and speaking anxiety between DGBL and traditional language learning. Furthermore, this study not only probed into whether self-efficacy and speaking anxiety predicted learning performance, but also investigated the effects as well as relationship between game performance and learning performance. The participants were 67 fourth graders in an elementary school in northern Taiwan, with random grouping employed by class to assign 47 students who used the DGBL system as the experiment group while 20 students who learned through traditional way as the control group. The experiment group and control group had 420 mines to learn English by different way. Data analysis included descriptive statistics, non-parametric test, t-test, Pearson correlation and multiple regression. Interview was conducted to collect students’ responses to the given learning experience. The results revealed that both DGBL and traditional learning effectively enhanced the students’ learning performance. Significant differences in learning performance was not found in the pre-test and post-test respectively for the two groups. The degree of difficulty speaking task were higher for learners probably be a cause that learner couldn′t finish all learning tasks during the experiment. Significant differences in self-efficacy and speaking anxiety also were not found in the pre-test and post-test, and learners’ speaking anxiety in post-test was higher than pre-test respectively for the two groups. The result of interview for the experiment group found that learners were nerves about their worse speaking performance and the standard of speech recognition was not easy to achieve probably. In different learning contexts, self-efficacy and speaking anxiety were negatively correlated at a significant level. Findings from the digital game-based English learning context showed that self-efficacy and speaking anxiety could predict learning performance, and confirmed a significant positive relationship between self-efficacy and speaking performance. The amount of stars earned in the game could predict learning performance also in the experiment group.
