Recent work on the survey's weighting system, particularly the non-response adjustment, improves the accuracy of the results.

The main possible sources of error stem from the fact that the sample is a client panel that is not random. Other possible sources of error include higher non-response for certain population categories such as youth and high-income earners, the memory effect and under-reporting of business trips. These are partly corrected by the weights (using auxiliary information or other statistical techniques).

Non-response can be high for expenditure data. In this case, the data are imputed by the provider.

The tourism demand survey sample is not a random sample. It is a client panel

Memory effect

Underestimation of business travel which is not always included as part of tourism.

Over-represented reports of the longest and most expensive trips.

Higher non-response for youth, especially