This paper presents the findings of a study looking into the valuation of travel time savings (VTTS) in Switzerland, across modes as well as across purpose groups. The study makes several departures from the usual practice in VTTS studies, with the main one being a direct representation of the income and distance elasticity of the VTTS measures. Here, important gains in model performance and significantly different results are obtained through this approach. Additionally, the analysis shows that the estimation of robust coefficients for congested car travel time is hampered by the low share of congested time in the overall travel time, and the use of an additional rate-of-congestion coefficient, in addition to a generic car travel time coefficient, is preferable. Finally, the analysis demonstrates that the population mean of the indicators calculated is quite different from the sample means and presents methods to calculate those, along with the associated variances. These variances are of great interest as they allow the generation of confidence intervals, which can be extremely useful in cost-benefit analyses.