New values of travel time, reliability and comfort in the Netherlands

Objective
The objective of this study was to generate a new and consistent set of values for changes in travel time, reliability and comfort for the various modes of transport in The Netherlands. These values serve as key input for Social Cost Benefit Analyses (SCBA) for mobility projects.

Scope
Compared with the previous executed Dutch VTT study date back from 2013, national values of travel time have been laid down for walking and cycling in this study for the first time. The value of travel time for transport to and from airports (in The Netherlands) for travel by airplane have also been determined for the first time. Also new are the values of travel time of comfort. The comfort multipliers indicate the level of comfort in relation to passenger volume for train and BTM, and the comfort of the walking and cycling infrastructure.

Methodology
The national averaged value of travel time (VTT) and value of travel time reliability (VTTR) for passenger transport are derived from Stated Preference (SP) choice experiments. A questionnaire is distributed among a target audience, containing questions regarding trip characteristics, personal characteristics and several choice experiments. In the experiment, respondents are asked to choose between hypothetical alternatives, each describing a trip in terms of travel time, travel cost and possibly other characteristics. Respondents are asked each time which alternative is preferred. From these choices the value of travel time and travel time reliability can be inferred.
The VTT and VTTR values are subsequently estimated by employing discrete choice models: models aiming to explain the observed choice made by the respondent in the experiment. Coefficients and interaction factors are added to these models to improve the model fit: boost the model's ability to explain the data. Once values have been obtained from these models, these number are applied to the sample of respondents to calculate their respective VTT/VTTR. Finally, these values are weighted to match statistics from the recent Dutch national travel survey. As a result, a VTT and VTTR is derived for each mode - purpose (business, commute and other) combination.

Data collection
About 9,700 respondents participated in the passenger travel survey designed for the national VTT study. Roughly 80% of respondents were recruited from an internet panel and about 20% were recruited by an interviewer intercepting them during their travel. The survey contained 10 unique SP experiments, dedicated to various modes of transport (car, train, bus tram metro (BTM), airplane, recreational navigation, walking and cycling) and valuations (value of travel time, reliability, comfort and more). Each respondent participated in two of them, based on their current travel activities. The choice experiments were dynamically constructed by a pivotal experimental design, so that each respondent received realistic choice tasks relatable to their current travel pattern.

The main data collection phase occurred in June and September 2022. Response rates for both the internet panel and intercept recruitment were high compared to previous studies, especially compared to the previous Dutch VTT survey from 2009/2011. The use of a high-quality internet panel and the increased rewards for participation in this new study are the likely cause for this improvement. A filtering procedure based on a list of predetermined conditions was applied to the respondent data. Conditions for exclusion were entries for which it was suspected that something had gone wrong during the survey, either because of a problem in the survey, a misunderstanding, or a mistake made by the respondent. About 80% of the observations remained for further analysis after this procedure. This procedure has resulted in a final dataset of more than 7.500 high-quality reliable responses.

Significance was responsible for data gathering and analysis on behalf of the Netherlands Institute for Transport Policy Analysis (KiM).

Dataset
Note that the uploaded datasets contains all (anonymised) data, which includes data from the various pilot studies and other data which are filtered out based on the filtering procedure to obtain the final dataset as above-mentioned.