Gender stereotyping in the labor market: A descriptive analysis of almost one million job ads across 710 occupations and occupational positions

This study presents patterns of gender stereotyping in job ads in the German labor market and examines its association with the unequal distribution of men and women across occupations. Using a large dataset of job ads from the "BA-Jobbörse", one of the largest online job portals in Germany, we apply a machine learning algorithm to identify the explicitly verbalized job descriptions. We then use a dictionary of agentic (male-associated) and communal (femaleassociated) signal words to measure gender stereotyping in the job descriptions. We collect information for 710 different occupations. Our first result shows that more jobs are femalestereotyped than male-stereotyped. We then take the example of two occupational groups that reveal clear differences in tasks contents and are highly relevant regarding important megatrends like digitalization and the demographic change: On the one hand, Science, Technology, Engineering, and Mathematics (STEM) and, on the other hand, Health and Social Services occupations. Additionally, we investigate the hierarchical aspect of occupational gender segregation. We distinguish jobs according to their required skill level and whether or not they are supervisory and leadership positions. In contrast to our first result, we find within STEM occupations as well as in supervisory and leadership positions that the majority of jobs is malestereotyped. Our findings indicate a positive association between gender stereotyping and occupational gender segregation, suggesting that gender stereotyping in job ads might contribute to the underrepresentation of women in certain occupations and occupational positions.