We have established new values of time (VOTs) and values of travel time reliability (VORs) for use in cost-benefit analysis (CBA) of transport projects in The Netherlands. This was the first national study in The Netherlands (and one of the first world-wide) to investigate these topics empirically in a joint framework. Stated preference (SP) questionnaires were designed for interviewing travellers, where the hypothetical alternatives were described in terms of travel time, travel costs and travel time reliability, the latter being presented to the respondents in the form of five possible travel times having equal probability.For passenger transport, we first collected interviews using an existing internet panel. Additional data collection recruitment was done by asking travellers at petrol stations/service areas, parking garages, stations, bus stops, airports and ports to participate in the survey. One important conclusion is that the SP survey using members of this internet panel leads to substantially lower VOTs than the SP survey with en-route recruitment, probably because of self-selection bias in the internet panel.We estimated discrete choice models in which the values of time differ between trips with different time and costs levels, different time and costs changes offered in the SP, and different observed characteristics of the respondents (e.g. education, income, age, household composition). By using a panel latent class model, we also account for unobserved differences between respondents in the value of time and for repeated measurements/panel effects. The reference values of time and the reference reliability ratios were estimated on the 2011 sample only, but the effect of time and cost level, time and cost changes offered and socio-economic attributes was estimated on both the 2009 and 2011 samples.