Sustainable Intra-Campus Micromobility at an Ecuadorian University: Multinomial Logit and Mixed Logit Models for Bicycle and E-Scooter Choice

Expanding university campuses face a dual challenge: meeting higher internal travel demand while reducing congestion and emissions. Using stated-preference data (412 respondents; 9 choice tasks per person), this study quantifies the expected adoption of two low-emission micromobility options—bicycle and electric scooter—relative to walking and car for intra-campus trips at an Ecuadorian university. Under Random Utility Theory, we estimate a multinomial logit (MNL) and a panel mixed logit (MIXL) model, treating MIXL as the preferred specification. Simulated maximum likelihood with 12,000 draws shows MIXL substantially improves fit and reveals marked heterogeneity in time sensitivity. In out-of-sample prediction (TEST), the two prospective modes achieve comparable average choice probabilities (bicycle = 0.289; electric scooter = 0.286), with a joint acceptance of 0.575, supporting the feasibility of a micromobility program under the evaluated scenarios. Environmental impact (gCO2/km) exhibits limited aggregate influence compared with operational factors (time and cost). A complementary WTP-space MIXL provides VOT estimates and confirms that 10–20% time reductions increase expected acceptance. Overall, results indicate that adoption is most likely when sustainability objectives are translated into tangible service improvements in effective speed, reliability, and affordability.