Estimation of Discontinuities from the Introduction of Tablet Based Data Collection on the International Passenger Survey

The International Passenger Survey (IPS) is undertaken by the Office for National Statistics to measure tourism flows and tourist expenditure, and international migration. It is interviewer-administered, and the questionnaire instrument was changed in 2017 to 2018 from a paper questionnaire (completed by the interviewer) to an electronic questionnaire administered with a tablet. For operational reasons no parallel run was possible, but the new questionnaire was rolled out progressively to sampling locations. This phased introduction supported the estimation of the effects of the new questionnaire on the main outputs from the survey. We describe initial simulations designed to estimate the power of the phased introduction approach to detect important difference in the IPS outputs, and analyses of the survey estimates at different stages up to the end of 2019 using state-space models, to estimate the discontinuities in the survey outputs. We make an assessment of the effectiveness of the overall approach.