| 摘要: | 隨著環境議題備受重視，公共自行車系統(public bicycle system, PBS)在世界各國逐漸流行。本研究調查了影響使用公共自行車系統之因素及其相互關係，以計畫行為理論(theory of planned behavior, TPB)與科技接受模式(technology acceptance model, TAM)作為研究的基礎框架，利用 (1) 偏最小平方結構方程模式(partial least squares structural equation modeling, PLS-SEM)探究因素間之路徑關係； (2) 偏最小平方多群組分析(partial least squares multi-group analysis, PLS-MGA)闡述可觀測的異質性，如：城市與性別的調節效果，以及 (3) 有限混合偏最小平方法(finite mixture partial least squares, FIMIX-PLS)探索不可觀測的異質性。為了進行實證分析，本研究在台北及台中蒐集了520個有效樣本，結果顯示： (1) 在計畫行為理論與科技接受模式中，除了主觀規範外，其餘的因子均會直接或間接影響行為意向； (2) 其他納入的影響因素，如：習慣、環境關懷以及公共自行車系統的基礎設施滿意度皆顯著正影響行為意向； (3) 異質性分析表明，城市與性別對模型中部分路徑存有調節效果，並於使用者中辨別出兩潛在類別。最後，提出研究結論、意涵與未來研究方向。;As environmental concern increases, public bicycle system (PBS) is becoming more and more popular worldwide. This research investigates interrelationships of factors influencing PBS. The research framework is constructed mainly based on theory of planned behavior (TPB) and technology acceptance model (TAM). The research framework is then analyzed with (1) partial least squares structural equation modeling (PLS-SEM) for exploring path relationships of influential factors of PBS, (2) partial least squares multi-group analysis (PLS-MGA) for elaborating observed heterogeneity, like city and gender, and (3) finite mixture partial least squares (FIMIX-PLS) for unobserved heterogeneity. To conduct the empirical study, we collected a sample of 520 respondents from Taipei and Taichung cities in Taiwan. The result shows: (1) Except for subjective norm, all factors drawn from TPB and TAM can exert influence and explain, either directly or indirectly, the effect on the behavior intention of using PBS, (2) Additional factors, such as habit, environmental concern, and satisfaction with PBS infrastructure, indicate positive effect on behavior intention of using PBS, (3) Heterogeneous analysis reveals that city and gender exert partial moderation effect, and two latent classes can be identified among cyclists. In the end, discussion and implications for future research are given.
Highlights:
►Include TPB, TAM and some added relevant factors in the proposed research framework.
►Investigate the interrelationships among influential variables using PLS-SEM model.
►Elaborate observed heterogeneity caused by city and gender using PLS-MGA model.
►Explore unobserved heterogeneity of PBS cyclists, two latent classes, with FIMIX-PLS.
►Use FIMIX-PLS rather than traditional clustering that ignores path model relations. |
