Gender Bias in Consumer Evaluations and Organizational Reputation: Industry as Institutional Evaluative Field

• In Event: Organizations, Occupations, and Work Section Roundtables
  In Refereed Roundtable (90 minute): Table 23: Audiences and Evaluation

Abstract

Online consumer reviews have become a central site of organizational evaluation. While these evaluations are often perceived as value-neutral assessments, this perception obscures inherent biases rooted in societal status beliefs and industry stereotypes. Drawing on status characteristics theory, this study proposes that industries function as institutionalized evaluative fields that structure the activation and expression of gender bias. We analyze over six million Yelp reviews across 19 industry categories to train industry-specific word embedding models. We examine how linguistic associations related to competence and warmth, the two core dimensions of the Stereotype Content Model, distribute along the semantic gender axis. By constructing three structural indicators: bias direction, bias spread, and concept location, this study characterizes the heterogeneity of gendered evaluations across industry contexts. Findings reveal that: Gender bias in consumer evaluations is asymmetric, with male-favored evaluations linked to competence traits and female-favored evaluations linked to warmth traits. Gender bias is inherently ambivalent, that the direction of bias on the competence dimension can diverges from the direction of bias on the warmth dimension. Industry moderates these expressions, that strongly gender-typed industries exhibit a higher intensity of bias, and the vehicle for bias (competence vs. warmth) corresponds to the industry’s evaluative criterion. Methodologically, this research employs a two-stage approach: The first stage utilizes word embedding models to capture latent semantic biases, providing a structural foundation for mapping stereotype dimensions onto gendered evaluative language. The second stage outlines a proposed quantitative framework to link these semantic indicators to organizational reputation. By utilizing Concept Mover’s Distance (CMD) and moderated interaction modeling, we provide a path to evaluate how industry-level evaluative structures condition the translation of linguistic bias into consumer star ratings. This study contributes to the literature on social evaluation by providing a systematic framework for understanding the reproduction of gender bias in commercial contexts.

Authors

• Tongjian Li, Emlyon Business School

• Adam Storer, Emlyon Business School