The perception of COVID-19 infection risk in Long-distance travel

The dataset contains sensitive personal information and could therefore not be included here. A sample dataset with 10 rows of data is included as an example. Note that the values shown are random and do not correspond to any particular respondent.

The original dataset includes 705 valid responses obtained through the Dutch Railways (NS) panel. The survey is an Hierarchical Information Integration experiment, meaning it is made up of a rating experiment and a bridging experiment.

Rating experiment:
Each respondent firstly evaluated five rating experiment scenarios, where they are presented with a scenario with varying attribute levels. They need to indicate on a 5-point Likert scale their perceived level of infection risk with COVID-19, where 1 is very low risk and 5 is very high risk.

Bridging experiment:
Secondly, respondents answered eight briding experiment choice situations, where they had to choose their preferred mode of travel, based on price, travel time, comfort and perceived infection risk. Four situations are for trips of approximately 500km, with the other four being a trip of roughly 1000km.

Socio-demographic information:
In addition to perceived risk and choice information, respondents were asked socio-demographic and long-distance travel related questions, to better understand their choice behaviour and attitudes on this topic.

Here, you will find:
example_dataset.csv     --> Example dataset (10 rows)
final_choice_model.py   --> Biogeme Pandas code used to estimate the model
model_outcome.html      --> Biogeme output file of the final model
survey_transcript.docx  --> Transcript of the survey in both Dutch and English, with example graphics

