A new integrated model of noncompensatory and compensatory decision strategies

Abstract

We describe and test a model that captures conjunctive, disjunctive, and compensatory judgment and choice strategies, as well as selected hybrid combinations of these. This model: (a) can be estimated solely from nonexperimental outcome data, (b) remains true to the conceptualization of noncompensatory heuristics as cognitively less demanding for decision makers, (c) is truly noncompensatory and not just approximately, (d) tests for a “pervasive” influence of cutoffs, (e) allows for the possibility that decision makers use different strategies across attributes, and (f) provides a more plausible account of behavior than competing models. We show empirically that decision makers may sometimes devalue objects for almost failing a conjunctive criterion or value objects more favorably for almost passing a disjunctive criterion—what we term a pervasive influence of a cutoff. The superiority of the proposed model relative to two other state-of-the-art models is demonstrated using both actual admit/reject decisions of an MBA admissions office as well as 10 simulations of various decision tasks.