2025/06/12 by Porzak, Robert, Cwynar, Andrzej, Weziak-Bialowolska, Dorota +1
Arts and Humanities · #Arts and Humanities #Business #Social and Behavioral Sciences
paper · doi:10.17605/osf.io/y3er5
https://onlinelibrary.wiley.com/doi/10.1111/ijcs.70083 Abstract We examine whether the phrasing of financial literacy test questions affects the test scores of women and men. To this end, we developed three language versions of a well-known set of financial literacy questions: feminized, masculinized, and neutral. We then surveyed a large, representative random sample of adult Poles (N=4,389), with respondents randomly assigned to one of these three language versions. Using item response theory (IRT), we examined how female and male respondents answered the questions and scored on the financial literacy test, depending on the language version. We found that both women and men performed differently depending on the phrasing of questions. Through our analysis, we identified a subset of questions free from language bias. Utilizing this de-biased set, we assessed financial literacy levels for both genders. Even in this language de-biased test, a significant gender gap favoring men persisted, with a difference of 0.775 on the standardized scale. In a series of alternative models, we found that the gender gap in financial literacy was noticeably smaller when feminized language is used compared to both masculinized language and language de-biased version of the test. This study confirms the existence of a gender gap in financial literacy, but suggests that its magnitude may be overestimated when masculinized language is utilized in testing instruments.