This dissertation investigates the psychological processes underlying consumers use of online product reviews in their purchase decision-making and implications for online review web site design. Specifically, the dissertation examines (1) the information selectivity question, i.e., how do people select which reviews to read, (2) the preference construction question, i.e., how do the reviews they read influence their criteria for evaluating a product, and (3) the design question, i.e., how can we design an online review web site to help people make better informed decisions. The dissertation proposes that exploratory search and goal-directed search are two mechanisms underlying information selectivity. Exploratory search is driven by consumers curiosity and cues for validity including review star rating, helpful vote, and review age. Goal-directed search is driven by information scent, defined as the perceived relevance of a review to a consumers criteria for evaluating a product. Consumers preference construction is affected by learning and forgetting, and biased information processing based on the intention to reduce cognitive costs and dissonance. Based on the conceptual model, a new review presentation design named Attribute Overview is proposed. The dissertation tests and refines the conceptual model and compares the Attribute Overview design with the traditional online review web site using both variance and process approaches. Data are collected in an experiment using process tracing methods (i.e. monitoring information acquisition and verbal protocol analysis). The variance model confirms (1) the impact of cues for validity including review star rating (especially one star review) and helpful vote, consumers curiosity, and information scent on information selectivity, and (2) the impact of memory processes and biased information processing on preference construction. The process approach shows how the psychological processes unfold over time. Although the variance model shows that the same set of factors affect peoples information selectivity and preference construction when using the traditional web site and Attribute Overview web site, the process approach reveals processual differences. The experimental data show that the Attribute Overview web site is better in terms of mitigating inappropriate selectivity and in helping reduce evaluation bias.