Model-based analyses
The accuracy-based analyses suggest that participants could achieve good performance in the two rule-based conditions, but not in the information-integration condition.
Yet, it is important to know whether each participant eventually adopted a decision strategy of the optimal type.For instance, the inability to perform in the information integration condition can stem from an incorrect categorization
strategy or a difficulty in consistently applying the appropriate strategy. To answer this question, we fit four different types of decision-bound models (e.g., Maddox &
Ashby, 1993) to the data from each individual participant:rule-based, information-integration, similarity-based and guessing models. The rule-based models assumed either a single vertical or a horizontal bound, or that participants used a conjunction rule. The information-integration model assumed that the decision bound was a single line of
arbitrary slope and intercept. For all the above models, it was assumed that the participants individually categorized the stimuli and then compared the outcomes.