Activating stereotypes through sports media: effects of racialized athlete portrayals on subsequent social judgments

ABSTRACT

This research investigates whether stereotypes activated by common media portrayals of Black athletes that emphasize physicality influence social judgments outside the sporting context. Results of a 2 (athlete race) X 2 (athlete frame) between-subjects experiment indicated that when an athlete was framed in terms of physical attributes, participants evaluated a suspect in a subsequent news story based on overlapping qualities related to criminality, regardless of athlete race. However, when an athlete was described in terms of their mental skills, physical qualities were more likely to be ascribed to a Black athlete and to impact evaluations of the crime suspect.

KEYWORDS:

• Media priming
• race
• stereotype
• sport communication

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. A pooled confirmatory factor analysis (CFA) was employed to examine the efficacy of the primary measures: perceived physicality of the athlete, perceived criminality of the accused in the subsequent story, and perceived guilt of the accused in the subsequent story. The normal theory-weighted least squares chi-square, Bentler’s (1990) comparative fit index (CFI), Steiger and Lind (1980) root-mean-square error of approximation (RMSEA), and the standard root-mean-square residual (SRMR) of the 17-item three-factor CFA model were assessed.

Results showed the model fit the data poorly, χ2(116) = 711.41, p < .001, CFI =.85, RMSEA =.11 [.10–.12], SRMR =.06. Examination of the test statistics revealed low factor loadings for two items, one item from the criminality traits scale (nice, [reverse-coded], b = .56), and another from the physicality traits scale (fast, b = .61). These items were dropped from the model. The new model showed improved but inadequate fit, χ2(101) = 554.34, p < .001, CFI =.87, RMSEA =.10 [.09–.11], SRMR =.06. Modification indices were consulted and suggested estimating (rather than setting to zero) the covariance between errors for items in the criminality traits scale to covary. Making these adjustments seemed appropriate; one set of items were the reverse-coded items in the scale (peaceful and nice), and the other set of items are conceptually similar indicators of criminality (threatening and dangerous). The resulting model showed acceptable fit, χ2(85) = 274.93, p < .001, CFI =.94, RMSEA =.07 [.06–.08], SRMR =.05. Next, reliability and validity were assessed by computing Cronbach’s α reliability coefficients, composite reliability scores (σ), and average variance extracted (AVE) values. Observed values of α and σ exceeded .80 and observed AVE scores exceeded .50, indicating each factor captured more variance than error (Hair et al., 2010), and therefore providing evidence of reliability and convergent validity. Evidence for discriminant validity was mixed; whereas correlations between latent factors did not exceed .85 (Zainudin, 2010), the factor correlation between criminality traits and guilt (.75) was greater than the √AVE of criminality traits (.72). Given one of two criteria for discriminant validity were met and the conceptual distinction between traits and behavior represented by these measures, we proceeded to data analysis.

Additional information

Notes on contributors

Erin Ash

Erin Ash (Ph.D., Pennsylvania State University, 2013) is an Associate Professor of Communication at Clemson University. Her research focuses on media representations of social groups and issues in media and their effects.

Gregory A. Cranmer

Gregory A. Cranmer (Ph.D., West Virginia University, 2015) is an Associate Professor of Sport Communication at Clemson University, Fellow of the Robert H. Brooks Sports Science Institute, and Research Fellow of the US Center for Mental Health & Sport.

Rebecca N. Pool

Rebecca N. Pool is a graduate student in the Department of Psychology at Clemson University. Her research focuses on diversity initiatives and quantitative analyses.