Skills mismatch between the output of Technical and Vocational Education and Training (TVET) institutions and universities and the actual demands of the labor market remains a persistent challenge to human capital utilization and youth employment outcomes in Uganda and across sub-Saharan Africa. This study examined the institutional, curricular, and individual-level determinants of the severity of skills mismatch among 450 TVET and university graduates who had transitioned into the labor market, treating mismatch severity as an ordered categorical outcome comprising no/low, moderate, and severe mismatch. Data were analyzed in three sequential stages: univariate description of sociodemographic and institutional characteristics; bivariate analysis using Pearson chi-square tests for categorical predictors and Kruskal-Wallis H tests for continuous predictors; and multivariable analysis using a proportional-odds ordinal logistic regression model, with robustness assessed through 1,000-replicate bootstrap resampling and a trimmed-sample re-estimation excluding high-leverage observations. The results showed that 57.6% of graduates experienced moderate or severe mismatch, with TVET graduates disproportionately represented among those reporting severe mismatch. In the multivariable model, TVET status significantly increased the odds of higher mismatch severity (adjusted OR = 1.958, 95% CI: 1.342-2.856, p < 0.001), while internship completion (adjusted OR = 0.660, p = 0.033), curriculum-industry linkage score (adjusted OR = 0.404 per SD, p < 0.001), digital skills score (adjusted OR = 0.620 per SD, p < 0.001), soft skills score (adjusted OR = 0.734 per SD, p = 0.001), and work experience (adjusted OR = 0.736 per SD, p = 0.002) were each significantly protective, whereas sex was not a significant predictor (p = 0.605). Sensitivity analyses confirmed the stability of these estimates across bootstrap resampling and after exclusion of high-leverage cases. The study concluded that weak curriculum-industry linkage, limited internship exposure, and underdeveloped digital and soft skills were the principal drivers of skills mismatch severity in Uganda, and recommended prioritized curriculum-industry alignment, expanded work-integrated learning,