虛擬寵物對線上學習注意力和表現的影響

Abstract

隨著線上學習的迅速普及，學生在遠距教學環境中經常面臨多工處理的挑戰，進而降低專注力並影響學習成效。為提升學習者的注意力維持，過去研究提出了「Fidget Widgets」概念，指出適度的動作行為可有助於減輕分心。然而，在虛擬學習情境下，如何設計虛擬工具（如虛擬寵物）以促進學習者的專注與學習表現，仍缺乏系統性的實證探討。

本研究旨在檢驗虛擬寵物的「互動性」與「課程相關性」兩項設計特徵，對於學生在線上學習過程中的專注程度、學習表現及使用者滿意度之影響。研究採用二因子混合實驗設計（interactivity 為組內變項，course relevance 為組間變項），結合自陳量表、眼動追蹤與半結構式訪談進行資料收集與分析。

研究結果顯示，虛擬寵物的高度互動性顯著提升使用者的滿意度、趣味性與視覺美感，但並未顯著促進學習成效（知識保留與知識遷移）。相較之下，僅具視覺課程相關性的設計並未對學習表現或專注力產生明顯影響，顯示僅靠外觀上的內容連結無法有效促進學習者的認知參與。此外，僅有視覺型學習者在課程相關設計條件下，其有意識的走神行為明顯減少，指出學習風格可能在特定設計情境中扮演調節角色。

綜上所述，本研究強調虛擬工具設計應超越表層的視覺連結，融入功能性與情境感知，才能有效支持學生專注與學習。研究成果可為未來虛擬學習平台與教學輔具設計提供實證參考，進一步提升線上教學的學習體驗與教育成效。

With the rapid expansion of online learning, students frequently encounter multitasking challenges that compromise attention and hinder academic performance. To address this issue, prior studies have explored ``fidget widgets'' as tools to sustain learner focus through moderate physical engagement. However, little is known about how such tools—particularly virtual agents like digital pets—can be designed to support attention and engagement in asynchronous and self-paced educational environments.

This study investigates the effects of two design features of virtual pet-based fidget widgets—\textit{interactivity} and \textit{course relevance}—on learners' attention, academic performance, and user satisfaction. A mixed-method experimental design was employed, incorporating self-report surveys, eye-tracking metrics, and semi-structured interviews.

Findings reveal that while high interactivity significantly enhances user satisfaction, playfulness, and visual appeal, it does not lead to measurable gains in knowledge retention or transfer. Visual course relevance, when implemented through static design elements, showed no significant effects on attention or learning, suggesting that surface-level contextual cues are insufficient for meaningful engagement. Notably, visual learners exhibited reduced deliberate mind-wandering when interacting with course-relevant pets, pointing to the nuanced role of individual learning styles in attention regulation.

These results offer empirical insights into how non-instructional virtual agents can be designed to support focused engagement in online learning. We discuss design implications for incorporating adaptive, context-aware behaviors in virtual agents, with the goal of supporting sustained attention in digitally mediated educational settings.