Gender Representation and Stereoyping in TV Advertising

Advertising is crucial for gender representation, yet systematic evidence of gender representation in TV ads remains scarce. This project aims to provides the first large-scale analysis of gender representation in mass media advertising. The project uses historical data on 2.1M TV ads aired between 2010 and 2020, and conducts large-scale surveys of representative samples of the consumer populations. The project develops a novel, flexible method to quantify gender representation in advertising by combining presence and stereotyping indicators, using unstructured text widely available to researchers and practitioners. The preliminary results show that advertisers systematically rely on gender stereotyping to represent both men and women in their ads, and men are twice as likely as women to be portrayed stereotypically. Industry-level patterns echo stereotypical gender norms, and men are overwhelmingly confined to stereotypical portrayals. The actual gender gaps in product consumption explain little of the variation in stereotyping. Instead, perceived usage gaps (often inaccurate and especially biased among marketing professionals) strongly predict representation choices. At the brand level, large and male-consumed brands depict men more and more stereotypically, whereas brands with women founders depict more women and represent men less stereotypically. The project highlights the role of advertising in perpetuating or challenging gender stereotypes, often driven by beliefs more than consumer behaviors.

Goals of the project: 1) Creating the first large-scale quantitative analysis of gender representation in TV advertising spanning over 10 years of data and over 1900 product categories. 2) Developing a novel method to measure representation of social categories (gender, race, ethnicity, etc) from unstructured text data, via text analysis methods, machine learning, and survey methods. 3) Developing interactive dashboards for policy-makers and institutions, to visualize the state of social representation in mass media, identify product industries/brands with sustained inequitable practices, and propose new industry policies to promote equitable mass media communications.

Workplan: 1) Phase 1 - early 2025: process the quantitative dataset of TV advertising, analyze the text to extract gender representation information. 2) Phase 2: mid-2025: run the large-scale representative surveys to measure stereotypical beliefs and product usage from the general population. Connect the survey with the historical data, and draft the analysis plan. 3) Phase 3: late-2025: conduct the analyses, process sales data, work on causal modeling, finalize writing the descriptive working paper, present the paper to conferences/seminars, and submit the paper to FT50 academic journals.