Understanding the Leaky Pipeline in Engineering

: Identifying Gender-Generic and Gender-Specific Predictors of Occupational Exit Decisions.

Abstract

Despite decades of workforce-diversity efforts, engineering continues to show disproportionately high rates of occupational exits (OCEs) among women, fuelling the sector’s ‘leaky pipeline’. To better understand this phenomenon, this thesis examined OCE dynamics in engineering in an explicit gender-comparative manner. By combining Rhodes and Doering’s (1983) Integrated Model of Career Change with an extensive literature review on potential exit predictors, an evidence-based framework of OCE dynamics was derived to guide the empirical investigation of three core research questions: Are there both gender-generic and gender-specific exit predictors in engineering? Do these predictors span different domains at work, including job-related, organisational, and/or occupational factors? Finally, which of these predictors vary across early-, mid-, and late-career engineers? To address these three interrelated questions, six empirical studies were conducted. Two qualitative studies examined whether the derived framework, though based largely on mid-career research, was sufficiently adaptable to inform exit investigations with early-career engineers (Study 1) as well as late-career engineers (Study 4). In addition, four quantitative studies with early-career engineers (Studies 2 and 3) and engineers across all career stages (Studies 5 and 6), assessed the predictive strength of key framework variables in a gender-conscious way. Across studies, several empirical patterns emerged. First, both gender-generic and gender-specific predictors of OCE cognitions were identified. Specifically, employment opportunities outside engineering were the strongest gender-generic predictor, whereas gender-specific effects indicated that men’s OCE cognitions were more closely tied to organisational factors (such as pay, peers, and leadership) than women’s. Second, though key OCE predictors spanned multiple domains of work-related experience, occupational, predictors such as occupational enjoyment and occupational fit emerged more consistently across studies than organisational and/or job-level predictors. Third, career stage altered the relative importance of key predictors, with occupational enjoyment most influential during early- and mid-career employment and life-course pressures (such as challenging personal and/or financial circumstances) during late-career employment. Together, these findings demonstrate that nuanced understanding of women engineers’ OCE decisions – necessary to guide more effective interventions to retain women in engineering – requires gender-comparative, multi-domain, and career-stage-sensitive research approaches.

Date of Award	9 Dec 2025
Original language	English
Awarding Institution	University of Bristol
Supervisor	Susanne Quadflieg (Supervisor) & Steve Eichhorn (Supervisor)