Understanding the third-person perception: Evidence from a meta-analysis

The third-person effect hypothesis has generated a vibrant research area. The expansion of this literature poses need for a systematic synthesis of the empirical evidence and assessment of the viability of various theoretical accounts. For this purpose, a meta-analysis is conducted on all published empirical studies of the perceptual component. This analysis covers 60 papers, 106 studies, and 372 effect sizes. Results from a series of multilevel models show that the third-person perception is robust and not influenced by variations in research procedures. Desirability of presumed message influence, vulnerability of referent others, referent others depicted as being similar to self, and others being likely audience of the media content in question are significant moderators. A vote counting analysis is conducted on 124 self-other comparisons in 29 studies from 13 additional papers that do not have the necessary statistical information. Further analyses are conducted to detect and estimate potential publication bias. Based on the empirical synthesis, the paper evaluates several explanatory factors and offers suggestions for future research.