For some time now, the users of EUSTAT have been raising new information demands for our Tourist Occupancy Survey. The survey design does not allow certain temporary and geographical disaggregation to be provided immediately. The initial sample design produces monthly estimates for geographical strata, based on a time sampling schedule. The methodological problem raised is how to obtain a total time and geographical disaggregation to answer any requests for estimates. The solution has been

more »

... essive: first, the design of the sample has been changed and the survey has been set as daily for the highest hotel category strata (introducing the electronic data collection); and now, we are studying the way to compute a macrotable containing the main information (available places, travelers and overnight stays) for al the units in the population (hotels) and for each day of the month. We have chosen the method of the imputation to obtain the information relating to the missing data (due both to the sample design or to the total nonresponse), as it guarantees the consistency of the imputed values. In the paper we propose a hot-deck imputation procedure, more precisely, the nearest neighbor method. This choice ensures the effect of the geographical location on the occupancy.