The live or digital interviewer - A comparison between CASI, CAPI and CATI with respect to differences in response behaviour

One of the core activities of market research is the collection of data by interviewing. Three developments have strongly influenced this activity: decreasing response rates, higher interviewing costs, and the growing awareness that the respondent needs to be treated like a 'real' customer. These developments have led to new ways of sampling (from ad hoc to access panels) and to using other methods of collecting data (from CATI/CAPI to CASI). The effects of the different methods or modes of collecting data are the focus of this paper. In this study the effects of the sampling procedure are explicitly distinguished from the effects of the modes of data collection. In particular, this split between sampling and modality effect provides new insights into the field of data collection. Research conducted by TNS NIPO and two universities shows that when using CASI, without in interviewer, socially unacceptable behaviour is more easily admitted. At the same time, using CASI, the respondent will be less sensitive to evaluation apprehension and show less socially desirable behaviour. CASI also leads to less extreme answers, more deliberation and more unaided recall of brand names, for example. Overall, the presence or absence of an interviewer appears to influence the interviewees. Therefore one has to be careful when data are collected through different modes.