Escaping unidimensional analysis limits and linear regression irrelevancy, the duration model incorporates impacts of covariates on the duration variable and permits to test the dependence of daily travel times on elapsed time. In the perspective of a discussion of Zahavi's hypothesis, the duration model approach is applied to the daily travel times of Lyon (France). The relationships between daily travel times and socio-economic attributes and activity duration only support the “weak version of TTB stability hypothesis”. Furthermore the non-monotonic estimated hazard questions the minimisation of daily travel times.