Stated and experienced availability as clues for call-scheduling in households longitudinal CATI surveys

In a telephone survey a large amount of time and money is spent in attempts to contact the sample units and obtain valid and complete interviews. Computer-assisted interviewing allows the implementation of call scheduling procedures aimed at maximising the number of successful attempts. In this contribution, we analyse data from a rotated panel survey on risk factors of social uneasiness at family level, conducted by the Inter-institutional Permanent Observatory on Padua Families, constituted by the Padua Local Health Service in co-operation with the Statistics Department of the University of Padua. Alternative estimation models for the probability of obtaining a valid interview are fitted on the basis of the three different levels of information collected during the first wave of the survey: household characteristics, call history and respondent’s stated availability. These models are then assessed and compared by means of the call histories recorded in the second wave.