A meta-analytic review of accuracy and bias in romantic partner perceptions

Abstract

People's perceptions of their romantic partners are somewhat accurate but also contain biases. In the current meta-analytic review, we sought to examine overall levels of accuracy and bias in romantic partner perceptions and moderators of accuracy and bias. We examined tracking accuracy (k = 157), projection (k = 157), indirect accuracy (k = 69), and mean-level bias (k = 153) in perceptions of a romantic partner across samples that estimated tracking accuracy and projection simultaneously. Gender, relationship length, study design, geographic region, publication year, and judgment type were examined as potential moderators. Results revealed significant mean effect sizes for tracking accuracy (β = .24), projection (β = .42), similarity (β = .22), and indirect accuracy (β = .11). Overall, people tended to have mean-level accuracy (d = -0.006) even when constructs were coded as conceptually positive or negative (d = -0.03). Men seemed to view their partners more negatively. Publication year and relationship length did not matter for any perceptual process. Tracking accuracy was higher in overall judgments, whereas projection was higher in daily judgments. Positive interaction traits and mind-reading judgments showed a negative bias. Greater projection and positive bias in noninteraction judgments were associated with higher relationship quality. Our results are the first to provide effect sizes for projection, tracking accuracy while controlling for projection, and indirect accuracy, and build on prior research concerning the complex ways in which partners perceive each other. (PsycInfo Database Record (c) 2023 APA, all rights reserved).