Weighting positive versus negative: the fundamental nature of valence asymmetry

The relation between weighting of valence information in attitude generalization and evaluation of novel/hypothetical situations was explored.

Undergraduate participants played a computer game requiring them to learn which stimuli (beans) would increase/decrease their points. Later, participants classified the valence of game beans and novel ones varying in resemblance to game beans. The weighting bias in attitude generalization was estimated as the average response to novel beans, controlling for game bean learning. We examined whether this bias related to judgments of hypothetical situations concerning interpersonal relationships (Study 1), threat assessment (Study 2), risk propensities (Study 3), and risk behavior (Study 6). We also assessed whether valence weighting is specifically predictive of novel situations (Studies 4 and 5). Finally, we explored participants' ability to self-report their weighting bias (Study 7).

Valence weighting in attitude generalization was related to judgments of novel situations and risk behavior. The performance-based measure did not correlate with responses to various questionnaires related to valence weighting.

There is a fundamental individual difference associated with valence weighting, resulting in the relation between two processes unrelated in content, but sharing the essential characteristic of valence weighting-attitude generalization and evaluation of novel situations.