Academic entrepreneurship, universities and fields of study: a multilevel analysis of scientists’ entrepreneurial intention

Purpose This paper aims to examine entrepreneurial intention (EI) among FONDECYT-funded scientists in Chile, analyzing how theory of planned behavior (TPB) profiles and university-level perceptual climates relate to EI. Attention is paid to gender and STEM differences to assess how inequality and disciplinary contexts shape academic entrepreneurship. Design/methodology/approach Survey data were collected from 1,027 academics across 39 Chilean universities and 20 fields. A two-level multilevel structural equation modeling (SEM) approach with Bayesian estimation was applied, including random intercepts and a latent EI factor at both levels. University-level variables were derived from aggregated perceptions, capturing perceptual climate. Findings The multilevel structure of EI is modest: 6%–7% of variance lies between universities, with most at individual level. Personal attitudes toward entrepreneurship are the strongest predictor, followed by perceived behavioral control. Subjective norms show no direct effect once attitudes and control are included, suggesting indirect influence. Between universities, differences reflect average attitudinal climate. Gender differences in EI diminish after accounting for attitudes, perceived control and prior entrepreneurial experience. STEM differences are better understood as disciplinary – TPB configurations rather than a simple STEM/non-STEM contrast. Social implications Findings highlight the need for policies that strengthen attitudinal and control-related resources, expand entrepreneurial experiences – especially for women – and address STEM-specific tensions. Universities can support regional development by aligning entrepreneurship initiatives with territorial needs and socially relevant knowledge transfer. Originality/value This study provides multilevel evidence from Chile, showing university-level effects are modest and primarily attitudinal, whereas gender and STEM differences are explained mainly by TPB configurations and experiences.