A Data-Driven Approach to Understanding the Attrition of Women in Software Engineering
| Autor: | Golison, Madeleine A. |
|---|---|
| Rok vydání: | 2024 |
| Druh dokumentu: | Diplomová práce |
| Popis: | Data from large tech companies shows that 15% or fewer software engineers are women. While Tech companies blame the university pipeline, studies from McKinsey and Accenture found that Tech company “bro culture” was influencing the pipeline of women out of Tech. However, in the MIT Women in Software Engineering survey, of the 183 respondents, most women reported planning on staying in Tech when leaving SWE roles. This formed the hypothesis that female software engineers were leaving SWE roles for reasons other than “bro culture.” Understanding and improving the attrition of women in the software engineering career path is important because the representation of women in the field is already so small, so any attrition is consequential. Overall, many factors were found to have influenced the retention of women in software engineering roles. Notably, culture was not the single most important reason for women leaving the software engineering career path. The primary reason directly stated in the open-ended survey responses was “burnout,” but this was closely followed by reasons such as finding other opportunities outside of Tech, a desire for better work-life balance, and the lack of diversity. While these explicitly stated reasons were easily noted, predictive models (using logistic regression and tree-based methods) were needed to illuminate factors that were not explicitly identified by respondents. The predictive models identified the primary reasons women leave SWE roles by comparing women who planned to remain in the SWE career path and those who did not. The top reasons identified were not enjoying programming, believing that better opportunities existed outside of software engineering, and being co-located with their team. The last reason, team co-location, was identified as being related to various other environmental factors related to imposter syndrome and was likely a proxy for these other factors. Women in the age range of 25 – 44 seemed to be particularly at risk of leaving the career path, and between the general population and the specific 25 – 34 and 35 – 44 age groups, each had different factors that were most important. Given these results, several recommendations exist for improving attrition for women in the software engineering career path. The key recommendations include improving manager feedback processes, diversity, work-life balance, and opportunities to work on high-visibility initiatives. S.M. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|