Not Just Noise: A Goal Pursuit Interpretation of Stochastic Choice

This paper contributes to our understanding of individual decision making by testing the proposal that differential weighting of 2 (or more) goals can be an important factor leading to stochastic (probabilistic) choice. The tested models follow from the endogenous maximum entropy program (eMEP) paradigm (Swait & Marley, 2013), which proposes that stochastic choice is (partially or entirely) a consequence of balancing multiple goals. That framework leads to an interpretation of the scale factor in classic random utility models (such as the multinomial logit [MNL]) as an endogenous property of a decision maker-an interpretation that is in stark contrast to the standard interpretation of the scale as due to heterogeneity or other "noise." The new perspective is supported by data from a task that manipulates (by a prime) an individual's propensity to be either consistent or to seek variety, suggesting that balanced pursuit of exploitation and exploration goals is a reasonable interpretation of stochastic choice.