Are Occupations Associated With Women Devalued? A Causal Test

Empirical evidence from the labor market suggests that when an occupation is associated with women, it is devalued; but without tight experimental control these studies cannot establish that the relationship is causal. In a series of preregistered survey experiments, I test whether U.S. residents devalue three distinct occupations when they are feminine-typed rather than masculine-typed. I independently vary two gendered dimensions of each occupation: whether it is described as majority-women (versus majority-men) and as necessitating feminine-typed communal (versus masculine-typed agentic) traits. I find modest evidence that the gender ratio or gendered traits of an occupation independently shape its perceived value net of the other, and robust evidence that a doubly feminine-typed occupation (majority-women and typified by communal traits) is devalued relative to a doubly masculine-typed occupation (majority-men and typified by agentic traits). This study provides causal evidence that an occupation’s association with women shapes its perceived monetary value, and shows that the effect is most pronounced when an occupation is feminine-typed on both dimensions. The devaluation of domains associated with women illustrates women’s limited options for evading gender bias: Although women are least likely to face gender bias in feminine-typed domains, those very domains are subject to gendered devaluation.