在全球化與國際化快速發展的時代，英語能力已成為學習者進行跨文化溝通、國際交流與職場競爭的重要能力之一。其中，英語字彙是語言理解與應用的基礎，對多益英語學習而言，學習者不僅需要記憶單字意義，也需要理解字彙在職場、商務與日常溝通等情境中的實際用法。然而，字彙學習通常需要長時間累積與反覆練習，若僅依賴傳統背誦方式，學習者容易產生學習疲乏與動機不足等問題。 隨著數位科技與行動裝置的普及，數位遊戲式學習逐漸被應用於語言學習情境，其任務、回饋、獎勵與情境設計有助於提升學習者參與。寵物養成機制能透過角色成長與情感連結強化學習者的持續投入，而不確定性機制則可能透過未知結果、隨機回饋與故事探索提升學習者的期待感與參與意願。因此，本研究發展一套名為「星語之境」之寵物養成式數位遊戲多益英語字彙學習系統，探討有無不確定性機制對學習者學習表現與主觀感受之影響，並進一步分析不同多益英語字彙能力學習者在上述變項中的表現差異。 本研究採準實驗研究設計，研究對象為中南部某公立科技大學學生，共計 43 名有效樣本，其中有不確定性組 20 人，無不確定性組 23 人。兩組學習者皆使用「星語之境」進行多益英語字彙學習活動，系統包含學習區、練習區與複習區；其中有不確定性組使用包含抽盲盒、隨機合成增益與故事碎片蒐集之版本，無不確定性組則使用固定獎勵與固定故事呈現之版本。研究工具包含多益英語字彙能力前後測、學習動機量表、遊戲心流量表、學習觀感量表與系統紀錄檔，並以信度分析、魏克森符號等級檢定、曼－惠特尼 U 檢定與斯皮爾曼等級相關分析進行資料處理。 本研究結果主要包含以下幾點：第一，有不確定性組與無不確定性組學習者在使用系統後，多益英語字彙學習成效皆顯著提升，顯示本研究所設計之寵物養成式數位遊戲學習系統能有效支持多益英語字彙學習。然而，兩組在學習成效進步幅度上未達顯著差異，表示不確定性機制未直接造成兩組學習成效之明顯差距。第二，在遊戲成效方面，兩組多數指標未達顯著差異，但無不確定性組在複習區題目數、答對題數與總時間上顯著高於有不確定性組，顯示無不確定性組學習者可能較集中於核心複習任務。第三，兩組在學習動機、遊戲心流與學習觀感上多未達顯著差異，代表不確定性機制在短期學習活動中未必能直接提升學習者的主觀感受。第四，相關分析結果顯示，兩組學習者之學習動機、遊戲心流與學習觀感彼此之間皆呈顯著正相關，表示學習者若具有較高的學習動機，通常也會具有較高的遊戲心流與學習觀感。第五，依據多益英語字彙前測成績區分高、低能力學習者後發現，有無不確定性機制對不同能力學習者之影響並不完全相同；高能力學習者在部分練習區參與表現與心流向度上呈現有不確定性組較佳之趨勢，而低能力學習者多數指標則未呈現顯著差異。 綜合而言，「星語之境」能有效提升學習者之多益英語字彙能力，而不確定性機制在本研究短期實驗中雖未直接提升整體學習成效、學習動機、遊戲心流與學習觀感，但仍可作為寵物養成式數位遊戲學習環境中的延伸回饋設計。本研究結果可作為未來多益英語字彙數位遊戲式學習系統設計、不確定性機制應用，以及不同先備能力學習者適性化學習活動設計之參考。;In an era of globalization and internationalization， English proficiency has become an essential ability for cross-cultural communication， international exchange， and workplace competitiveness. Vocabulary is the foundation of English comprehension and application. For TOEIC English learning， learners need not only to memorize word meanings but also to understand how vocabulary is used in workplace， business， and daily communication contexts. However， vocabulary learning requires long-term accumulation and repeated practice， and traditional memorization may easily lead to learning fatigue and insufficient motivation. With the development of digital technology， digital game-based learning has increasingly been applied to language learning. Game elements such as tasks， feedback， rewards， and contextualized learning may enhance learner engagement. In addition， pet-nurturing mechanisms may promote sustained participation through character growth and emotional attachment， while uncertainty mechanisms may increase learners’ anticipation and willingness to participate through random feedback， unknown outcomes， and story exploration. Therefore， this study developed a pet-nurturing digital game-based TOEIC English vocabulary learning system named Starry Realm to investigate the effects of uncertainty mechanisms on learners’ TOEIC English vocabulary learning performance， in-game performance， learning motivation， flow， and learning perceptions. This study also further examined the differences among learners with different levels of TOEIC English vocabulary ability. This study adopted a quasi-experimental research design. The participants were students from a public university of science and technology in central-southern Taiwan， with a total of 43 valid samples. Among them， 20 participants were assigned to the group with uncertainty mechanisms， while 23 participants were assigned to the group without uncertainty mechanisms. Both groups used Starry Realm to conduct TOEIC English vocabulary learning activities， including a learning area， a practice area， and a review area. The group with uncertainty mechanisms used the version with blind-box drawing， random synthesis boosts， and story fragment collection， whereas the group without uncertainty mechanisms used the version with fixed rewards and fixed story presentation. The research instruments included a TOEIC English vocabulary ability pretest and posttest， a learning motivation questionnaire， a flow questionnaire， a learning perception questionnaire， and system log data. The collected data were analyzed using reliability analysis， the Wilcoxon signed-rank test， the Mann-Whitney U test， and Spearman’s rank correlation analysis. The results showed that both groups significantly improved their TOEIC English vocabulary learning performance after using the system， indicating that the pet-nurturing digital game-based learning environment developed in this study can effectively support TOEIC English vocabulary learning. However， there was no significant difference between the two groups in learning performance improvement， suggesting that the uncertainty mechanisms did not directly lead to greater vocabulary learning gains. In terms of in-game performance， most indicators showed no significant differences between the two groups， but the group without uncertainty mechanisms performed significantly better in the number of review-area questions attempted， the number of correct answers， and total time spent in the review area. This finding suggests that learners in the group without uncertainty mechanisms may have focused more on core review tasks. Regarding learning motivation， flow， and learning perceptions， most indicators did not show significant differences between the two groups， indicating that uncertainty mechanisms may not directly enhance learners’ subjective perceptions in a short-term learning activity. The correlation analysis showed that learning motivation， flow， and learning perceptions were significantly and positively correlated， suggesting that learners with higher learning motivation tended to experience higher flow and more positive learning perceptions. Furthermore， after learners were divided into high- and low-ability groups based on their TOEIC English vocabulary pretest scores， the results showed that the effects of uncertainty mechanisms were not entirely consistent across different ability levels. High-ability learners showed a tendency toward better performance in some practice participation indicators and flow dimensions when using the version with uncertainty mechanisms， whereas most indicators among low-ability learners showed no significant differences. In summary， Starry Realm effectively enhanced learners’ TOEIC English vocabulary ability. Although the uncertainty mechanisms did not directly improve overall learning performance， learning motivation， flow， or learning perceptions in this short-term experiment， they may still serve as an extended feedback design in pet-nurturing digital game-based learning environments. The findings of this study can provide references for future TOEIC English vocabulary digital game-based learning system design， the application of uncertainty mechanisms， and the development of adaptive learning activities for learners with different levels of prior ability.
