A social comparison theory meta-analysis 60+ years on.

These meta-analyses of 60+ years of social comparison research focused on 2 issues: the choice of a comparison target (selection) and the effects of comparisons on self-evaluations, affect, and so forth (reaction). Selection studies offering 2 options (up or down) showed a strong preference (and no evidence of publication bias) for upward choices when there was no threat; there was no evidence for downward comparison as a dominant choice even when threatened. Selections became less differentiable when a lateral choice was also provided. For reaction studies, contrast was, by far, the dominant response to social comparison, with ability estimates most strongly affected. Moderator analyses, tests and adjustments for publication bias showed that contrast is stronger when the comparison involves varying participants' standing for ability (effect estimates, -0.75 to -0.65) and affect (-0.83 to -0.65). Novel personal attributes were subject to strong contrast for ability (-0.5 to -0.6) and affect (-0.6 to -0.7). Dissimilarity priming was associated with contrast (-0.44 to -0.27; no publication bias), consistent with Mussweiler (2003). Similarity priming provided modest support for Collins (1996) and Mussweiler (2003), with very weak assimilation effects, depending on the publication bias estimator. Studies including control groups indicated effects in response to upward and downward targets were comparable in size and contrastive. Limitations of the literature (e.g., small number of studies including no-comparison control conditions), unresolved issues, and why people choose to compare upward when the most likely result is self-deflating contrast are discussed. (PsycINFO Database Record