The demand for data analytical skills by gender: Evidence from a field experiment

This paper examines the return to advanced data analysis skills among job applicants from economics undergraduate programs employing a resume audit experiment. We randomly assigned fictitious resumes with three levels of data analysis skills (basic, medium, and strong) and submitted them to online job postings. Resumes with basic data analysis skills indicated proficiency in Excel. Resumes with medium data analysis skills demonstrated proficiency in Stata and SPSS, while resumes with strong data analysis skills indicated proficiency in Python and SQL, in addition to Stata and SPSS. Compared to resumes with basic skills, those with medium and strong skills received callback rates that were 2.5 and 2.8 percentage points higher, representing increases of 19.2 % and 21.5 %, respectively. For female applicants, resumes with medium and strong skills received callback rates that were 3.4 and 5.1 percentage points higher, corresponding to increases of 29.8 % and 44.7 %, respectively. These differences in callback rates were statistically significantly different from zero for both the overall sample and female applicants. On the other hand, no statistically significant effect was observed for male applicants. Interview evidence suggests that employers demand data analysis skills as tangible skills, rather than merely considering them as signals of ability. This finding is consistent with human capital theory, as opposed to signaling theory. Moreover, we find evidence of gender discrimination among applicants with basic data analysis skills, where women received statistically significantly lower callback rate than men. However, for resumes indicating advanced data analysis skills, no significant gender differences emerged, suggesting statistical discrimination.