| This dissertation develops a framework for understanding how the information environment and time preferences affect search behavior and choices over time. The information environment is conceptualized in terms of two temporally distinct search-cost categories that are traded off: initial set-up costs and on-going evaluation costs. The tradeoff of these two search-cost categories is hypothesized to affect consumers' initial selections and their propensity to search for, discover, and adopt new options, resulting in a two-stage lock-in process. First, time preferences and failure to anticipate the impact of switching costs, cause selection of a lower set-up and higher evaluation cost option. Second, after incurring set-up cost, intertemporal preferences result in less likelihood of considering and switching to a new alternative. An individual might initially select an option for a single period, but failing to anticipate the impact of their investment, continues using it in subsequent periods. Six experiments support this framework. Experiments 1, 2, and 3 address the first stage of lock-in (initial selection of retailer), in which participants selected between two Internet vendors differing only in set-up and evaluation costs---a low set-up, high evaluation cost option, and vise versa. The main treatment included anticipated usage occasions and time horizon manipulations Collectively, the experiments demonstrated that participants tend to select a low set up option. Experiments 1 and 2 demonstrated that when time is not explicitly mentioned, participants behave as if it's a one period decision. Experiments 2 and 3 show that the choice pattern is not sensitive to whether decisions occur over a 3- or 12-month period, suggesting that a myopic time horizon accounts for effects, rather than future time discounting per se. Experiments 4, 5 and 6 addressed the second lock-in stage (propensity to switch after initial investment). Experiment 4 demonstrated that people anticipate a greater likelihood of future switching than they exhibit at the decision time. Experiments 5 and 6 demonstrated that the best future decision predictor is prior selection, and once set-up cost is incurred, there is less likelihood of switching. Both experiments show that lock-in results in real monetary loss, but in higher satisfaction. |
