Understanding Belief-Behavior Correspondence: Beliefs and Belief-to-Behavior Inferences.

Psychologists and other scholars have long studied the influence of beliefs on behavior, proposing models that illuminate aspects of the belief-behavior relation. However, prior psychological theories have not explained how a particular belief influences behavior. This limitation coincides with the assumption, common among both psychologists and lay observers of human behavior, that beliefs are powerful drivers of behavior. To improve our understanding of the belief-behavior association, we first show evidence that the association between beliefs and behaviors is, generally speaking, small (r < .2). We then define different types of beliefs (i.e., existence beliefs like "I believe in God," descriptive beliefs like "Magnolias are white," and outcome beliefs like "Vaccines save lives") and propose that beliefs have a causal influence on behavior when people form a belief-to-behavior inference. Belief-to-behavior inferences are generally probabilistic and are made when individuals have a behavioral (vs. informational) goal. These inferences are also more likely when individuals rely on beliefs that are closer to behavior (e.g., descriptive versus existence beliefs) within the belief-to-behavior inference chain. The model clarifies inferential processes for different belief types and makes novel predictions about the effects of goals and cognitive capacity on the belief-behavior correspondence. In addition, it integrates memory-based and online decision processes, independent activation of behavioral attitudes and intentions, and proceduralization of belief-to-behavior inferences.