The full set of survey data was edited to remove directly identifiable information in free text such as participant names, parkrun names, uniquely identifiable illnesses etc.

The dataset is restricted to fields relevant only to the analysis required for the paper this is attached to in order to minimise the risk of participant identification.