網路科技的普及與電子商務的興起雖意謂著銷售通路的改變,但是由於電子商務中的買賣雙方往往因空間上的分隔而必需進行非即時的交易,加上網站間轉換成本極低的特性,信任與忠誠度的建立相較於實體商店更加的困難。 在本研究中,我們以 Tan 與 Sutherland(2004)的概念模型為基礎,探討包含機構型、人際型、與人格特質型信任的整體信任模型,以及其與網站忠誠度的關係,除了在機構型信任構面上加入網路安全知覺及網際網路口碑等新的信任來源以符合網站信任研究的需求之外,為了進一步探討過去研究結果的不一致性,我們以個人信任傾向作為干擾變數來探究高低信任傾向對信任建立與忠誠度的影響。我們以調查研究的形式取得實證的資料,以結構方程式驗證模型的正確性,並進一步分析不同的信任來源對整體信任與忠誠度建立之影響與其優先順序。 雖然調查資料並未高度支持本研究所提出的概念模型且個人信任傾向並未具有顯著的干擾效果,指示模式對現實情境的解釋能力或有不足,但透過分析我們仍獲得許多值得注意的結果。其中,統計結果顯示網站能力和整體信任的形成並未具有顯著的相關性。此外,若直接研究各信任來源和網站忠誠度的關係,則可發現個人信任傾向的干擾效果,在高信任傾向的消費者方面,其忠誠度主要可由正直此一變數解釋,低信任傾向的消費者方面,其忠誠度主要則由網路安全知覺來解釋,研究結果對於電子商務整體發展及企業策略發展上都具有管理上的意義。
While trust is considered one of the most important factors that contribute to the success of an e-business, the characteristics of online transactions such as lacking of face-to-face interactions between sellers and buyers, unsynchronized exchanges of money and goods, and frequently requested sensitive personal information from customers, can all make trust-building an even more difficult task. In this thesis, we develop a framework based on the multi-dimensional trust concept suggested by Tan and Sutherland (2004). We consider both interpersonal-based trust and institutional-based trust, and we add perceived internet security and word-of-mouth as antecedents of trust in order to build a new model that is suitable for e-business settings. Also, to clarify the noted inconsistency and to further verify the effect that an individual’s disposition to trust has on trust-building, we use an individual’s disposition to trust as a moderating variable. In other words, we divide the consumers into sub-groups based on their trust propensity, and we see whether the difference in propensity can significantly affect the relative importance of antecedents of trust. A survey design was chosen for testing the model of customers’ trust, and several statistical procedures (including both traditional approaches such as regression, correlation, and factor analysis for verifying the hypotheses and structural equation modeling for assessing the overall model fit) were employed in this study. While the results indicate that the proposed conceptual model cannot be well supported by our data, and the moderating effect of trust propensity does not exist, some managerial implications can still be derived from our analyses. In particular, we found that when exploring the relationships between loyalty and various sources of trust, a significant moderating effect of trust propensity might exist. The main determinant of customer loyalty for customers with high trust propensity is integrity. On the other hand, for customers with low trust propensity, online loyalty is mostly explained by the perceived internet security.