基於自我及他人感知的單車行為分析：以國立臺灣大學為例

Abstract

自行車是一種可持續的城市交通方式，但基礎設施不足迫使騎士與車輛和行人共享道路，增加了衝突和安全隱憂。本研究基於擴展的計劃行為理論（The Theory of Planned Behavior）框架，運用結構方程模型（Structural Equation Modeling）探討心理因素如何影響騎士的認知和行為。關鍵因素包括態度、道德規範、風險認知、環境認知、他人影響、行為意圖、自我行為評估以及對他人行為的評估。 本研究在國立臺灣大學進行問卷調查，該校園內超過25,000輛私人自行車服務將近4,0000人，橫跨110公頃的校園面積，導致騎士、行人和其他車輛之間產生相當大的衝突，校園內經常可見違規停放的自行車以及單車騎士的事故，由此可見單車管理為校園重要課題之一，因此本研究分別建立騎乘行為和停車行為模型，以捕捉各行為類型的不同心理機制，希望藉此對於校園單車管理提供建議。 SEM結果顯示模型具有優異的適配度（騎乘：CFI = 0.921；停車：CFI = 0.926），並揭示正向態度矛盾地增加危險行為意圖(β=0.245)，而社會意識則起強烈的嚇阻作用(β=-0.332)。道德規範在兩個模型中都是影響風險認知和社會敏感性的核心構念。研究提供實證證據顯示自利偏誤的存在，即騎士在意圖從事不當行為的同時，仍保持對其不當性的認知。研究結果強調需要同儕教育計畫、道德規範提升策略、平衡的基礎設施設計，以及整合性干預框架，以促進更安全的騎乘環境，讓自我覺察和規範的騎乘文化能夠解決潛在的行為問題。

Cycling is a sustainable urban transport mode, but inadequate infrastructure forces cyclists to share roads with vehicles and pedestrians, increasing conflicts and safety concerns. This study explores how psychological factors shape cyclists' perceptions and behaviors using Structural Equation Modeling (SEM) based on an extended Theory of Planned Behavior (TPB) framework. Key factors include attitudes, moral norms, risk perception, environmental perception, effect of others, behavioral intention, self-behavior assessment, and assessment of others' behavior. A questionnaire survey was conducted at National Taiwan University (NTU), where over 25,000 private bicycles serve around 40,000 people across 110 hectares, leading to considerable conflicts between cyclists, pedestrians, and other vehicles, with disorderly parked bicycles frequently observed on campus. The study developed separate models for riding and parking behaviors to capture distinct psychological mechanisms governing each behavior type. The SEM results demonstrate excellent model fit (riding: CFI = 0.921; parking: CFI = 0.926) and reveal that positive attitudes paradoxically increase risky behavioral intentions (β=0.245), while social awareness serves as a strong deterrent (β=-0.332). Moral norms emerge as central constructs influencing risk perception and social sensitivity across both models. The study provides empirical evidence for self-serving biases where cyclists retain awareness of inappropriate behaviors while intending to perform them. The findings highlight the need for peer education programs, moral norm activation strategies, balanced infrastructure design, and integrated intervention frameworks to promote safer cycling environments where self-aware and regulated cycling culture addresses underlying behavioral issues.