Coder gender and potential for hostility ratings

Abstract

This study examined the effects of coder gender on Potential for Hostility ratings. Six trained coders (3 men and 3 women) who were unaware of the coder gender effect hypothesis coded 30 male and 30 female undergraduates for Potential for Hostility. Although reliability estimates as calculated by Cronbach's alpha suggested that all coders were consistent, an analysis of variance revealed a significant main effect for coder gender, wherein female coders rated participants as displaying significantly less Potential for Hostility than did male coders. This significant difference was also meaningful, as coder gender accounted for 32% of the variance in Potential for Hostility scores. Thus, future Potential for Hostility investigations need to consider the gender of those coding, as this factor both significantly and substantially influences reported Potential for Hostility ratings.