2026 ASEE Annual Conference & Exposition

University-affiliated accelerators are now common features of campus entrepreneurship ecosystems, yet administrators have limited comparative guidance on which program design choices deserve attention. This study develops a public-evidence measurement framework for evaluating accelerator architecture using records that administrators, researchers, and external stakeholders can inspect: cohort pages, program descriptions, and publicly documented venture outcomes. Across 1,288 ventures from 10 U.S. university-affiliated accelerators, the study codes program design features, links admitted ventures to source-level evidence, and produces diagnostics for funding and survival/activity outcomes. Mentorship and advising intensity shows the strongest directional association with documented funding rate across programs (Spearman rho = 0.563, N = 10), while resource and infrastructure intensity is also positively associated, though more modestly (rho = 0.347). Missingness is substantial: 14.4% for funding-event evidence and 56.8% for survival/activity evidence. These estimates should therefore be interpreted as public-documentation measures rather than full venture-performance measures. The main contribution is a reproducible accelerator design scorecard that makes mentorship cadence, in-kind support depth, and evidence observability comparable across university entrepreneurship programs.