Learning Analytics for Entrepreneurship Education A Reproducible Moderated\u2013Mediation Study of Innovative Behavior and Motivation

Innovation Behavior (IB) and Entrepreneurial Motivation (EM) are said to be improved by entrepreneurship education, however there is little reproducible evidence of this relationship. In this paper, an end-to-end analytic workflow was defined and applied in a longitudinal single group pre\/post design (n=177) that tested the effects of structured entrepreneurship education on Innovation Behavior (IB) and Entrepreneurial Motivation (EM). Direct effects of education on IB and EM were estimated, entrepreneurial self-efficacy (ESE) was tested as a mediator, and the risk-taking propensity (RTP) of students was tested as a moderator of the direct education outcome relationships. An end-to-end workflow was defined to include data capture, preprocessing, analysis and public release of anonymized or synthetic data, executable notebooks, and an environment lock file. Multi-item scales used to measure IB and EM were validated and effect sizes were reported along with their confidence intervals, robustness checks, and assumptions diagnostics. The results indicated that statistically significant increases in both IB and EM occurred                         post-education course, and that entrepreneurial self-efficacy (ESE) was a partial mediator of both IB and EM post-course gains, and that risk-taking propensity (RTP) moderated the direct effects of education on both IB and EM. The study provides a transparent, replicable, and reusable framework for assessing curriculum-based entrepreneurship programs and identifying specific types of learners to provide targeted support.