Modeling Individual Decisions from Information Search

Background

In order to study people’s search and choice behaviors in the laboratory, a “sampling paradigm” has been proposed in the DFE research (Hertwig & Erev, 2009). In the sampling paradigm, people are presented with two or more options to choose between. These options are represented as blank buttons on a computer screen. People are first asked to sample as many outcomes as they wish from different button options (information search). Once people are satisfied with their sampling of the options, they decide from which option to make a single final choice for real.

In the sampling paradigm, two classes of models have been proposed and these classes include the associative learning models and cognitive heuristics (Hertwig, 2011). Among the associative learning class, human choice is conceptualized as a learning process (for example, Busemeyer & Myung, 1992; Bush & Mosteller, 1955). Learning consists in changing the propensity to select a gamble based on the experienced outcomes. Good experiences boost the propensity of choosing the gamble associated with them, and bad experiences diminish it (e.g., Barron & Erev, 2003; Denrell, 2007; Erev & Barron, 2005; March, 1996). Some of the models in this class include the Instance-Based Learning (IBL) model (Gonzalez & Dutt, 2011; Lejarraga, Dutt, & Gonzalez, 2012), value-updating model (Hertwig et al., 2006), and fractional adjustment model (March, 1996). Among all the models in the associative class, the IBL model has been shown as the best performing model at the aggregate level (Gonzalez & Dutt, 2011, 2012). Thus, we choose the IBL model as a first model for evaluation in this article.