Online rating dispersion and consumer conformity: How prior variability moderates the anchoring effect

This study analyzes how the dispersion of prior online ratings influences individual rating behavior in digital review platforms. Drawing on social learning theory, social conformity, and prospect theory, the empirical application examines over 3 million restaurant reviews in New York City (2010–2024). Using fixed-effects regression models, the results show that greater variance in prior ratings weakens the anchoring effect of the average, leading individuals to post more divergent evaluations. The results reveal a valence-based asymmetry in social influence, where informational ambiguity reduces conformity, especially in the loss domain. This finding represents an important theoretical contribution by demonstrating that prior rating variance functions as a contextual moderator of social influence, thereby introducing asymmetry into the conformity process.