CAREER: Identification and integration of persistent behavioral biases into online search and purchase models

Georgia Tech Research Corporation, Atlanta GA

Investigators

Abstract

Discrete choice models are the backbone of empirical analysis in many fields. These models are grounded in random utility maximization theory and generally require that fundamental properties, including stable and transitive preferences, be observed. Recent work by behavioral economists suggests that these properties may not be valid in all decision making contexts, and that failure to properly account for these and other behavioral biases may lead to poor forecasting accuracy. The objectives of this CAREER proposal are to: 1) integrate behavioral biases into discrete choice models to enhance our fundamental knowledge of individual and firm decision making; 2) investigate whether behavioral biases found to be important in current market conditions persist in future market conditions; 3) develop and validate theories of individuals, search and purchase behaviors; and, 4) develop probability modules for Georgia high school teachers based on state-mandated problem-based learning techniques that use datasets representative of engineering problems encountered in practice. The research findings will be applied and tested in the airline industry, which affords ideal conditions for testing behavioral theories under a wide variety of market conditions. This is due to the ability to both track menus and decisions made by travelers in a developed industry that has flexible capacity. This proposal creatively integrates data from multiple online sources and complements this with online experiments conducted with the airline industry. The ability to analyze large volumes of data from naturally occurring online markets provides the opportunity to uncover new empirical insights that will drive the development of theoretical models. The developments from this CAREER proposal will explicitly integrate individuals, search and purchase behaviors with firm decision-making. This approach results in a theoretical foundation that is unique from that of other researchers and reveals that demand and behavioral bias assumptions are not trivial and can result in forecasts opposite those predicted by classic models. The most important broader impact is the potential to change the nature of online markets to make them more interactive with customers, which will result in both increased profits for firms as well as better product and service offerings for customers.