串流音樂平台使用者心理特質與推薦工具使用關係之研究

Abstract

隨著串流音樂平台的普及，演算法推薦已成為使用者探索音樂的重要途徑，然而個體對於新音樂的接受程度存在明顯差異，此一差異背後的心理機制仍待系統性釐清。本研究旨在探討音樂涉入程度、音樂開放傾向與串流平台推薦工具使用行為三者之間的關聯，並發展適用於音樂聆聽情境之問卷工具。 本研究採兩階段研究設計。質化前導階段以焦點團體訪談蒐集音樂涉入與探索行為之語料，作為問卷題項發展之基礎；量化階段以線上問卷檢驗問卷之因素結構與構面間關聯。透過探索性因素分析，本研究發展出音樂涉入問卷（四構面：吸引力／中心性、自我探索、社會連結、身份認同）與音樂開放傾向問卷（四構面：選擇性開放、一般性開放、背景陪伴、偏好熟悉），兩份問卷信度均達可接受標準。 路徑分析結果顯示，音樂情緒反應與涉入各構面透過開放傾向，對使用者的主動使用、程式推薦與半主動推薦三類功能使用行為產生差異化影響。其中，選擇性開放是唯一對三種功能使用均有顯著正向效果的類型；背景陪伴型使用者高度依賴演算法推薦，偏好熟悉型使用者則傾向主動迴避演算法介入。開放傾向在多數路徑中扮演重要的中介角色，尤其對演算法驅動之功能中介效果最為完整。 本研究提出的四象限開放傾向模型，為推薦情境下的音樂探索行為提供了更精細的理論架構，並可作為串流平台設計差異化推薦策略之實證依據。研究亦討論樣本代表性、問卷發展屬探索性階段，並提出未來研究方向。

As music streaming platforms have become ubiquitous, algorithmic recommendation has emerged as a central pathway for music discovery. Yet individuals vary considerably in their openness to new music, and the psychological mechanisms underlying this variation remain insufficiently understood. This study examines the relationships among music involvement, music openness tendencies, and the use of recommendation tools on streaming platforms, and develops questionnaire instruments suited to the music-listening context. A two-phase design was employed. In the qualitative pilot phase, two focus group interviews were conducted to generate corpus data on music involvement and exploration behavior, informing the development of questionnaire items. In the quantitative phase, an online survey was used to examine the factor structures of the questionnaires and test relationships among constructs. Exploratory factor analysis yielded a four-factor music involvement questionnaire (attraction/centrality, self-exploration, social connection, and identity) and a four-factor music openness questionnaire (selective openness, general openness, background companionship, and familiarity preference), both demonstrating acceptable reliability. Path analysis revealed that emotional responsiveness and the involvement dimensions exert differentiated effects on three types of feature use—active use, programmed recommendations, and semi-active recommendations—largely through openness tendencies. Selective openness was the only type positively associated with all three feature-use behaviors, whereas users high in background companionship relied heavily on algorithmic recommendations, and those high in familiarity preference tended to actively avoid algorithmic intervention. Openness tendencies functioned as a substantial mediator across most paths, particularly for algorithm-driven features. The proposed four-quadrant model of music openness offers a more nuanced theoretical framework for understanding music exploration behavior in recommendation contexts, and provides an empirical basis for platforms to design differentiated recommendation strategies. Limitations regarding sample representativeness and the exploratory stage of questionnaire development are discussed, along with directions for future research.