Abstract: In many organizations, ideal workers are conceived of as male, which disadvantages female employees. To investigate this phenomenon, we integrate cultural theory, structural linguistics, and natural-language processing to capture shades of gender in cultural conceptions of workers and organizations. We analyze large-scale data on employee discourse about organizations and use word embeddings to extract a gender axis in semantic space. We study tech firms, which have highly masculine cultures, and analyze associations in tech-worker discourse between the gender axis and cultural constructs related to gender. We find that discourse about tech firms is sometimes “degendered”: the stereotypically male traits independence and leadership competence appear gender-neutral, while instrumental competence appears female-shaded. Although discourse about tech firms is generally male-shaded, there is considerable variation across employees and firms. Male employees, less-satisfied employees, and those in privately held and smaller firms use more male-shaded language; language-use differences between men and women are wider among less-satisfied employees and those in publicly traded and larger firms. Our approach to quantifying shades of gender in organizational cultures moves us closer to determining how those cultures promote or reduce inequality and exclusion. It also points the way to quantifying other dimensions of organizational culture content at scale.
