Web Survey Bibliography
We used a tailored survey design to obtain a more representative response. Paradata from previous consumer sentiments surveys and register information were used to stratify the sample into groups that differed in contact and co-operation propensity. We approached an experimental sample of 3000 households with a Web–mail–computer-assisted telephone interviewing sequential mixed mode strategy. The choice of initial mode and the subsequent computer-assisted telephone interviewing approach were tailored to the expected contact and co-operation propensities of the sample units. In the computer-assisted telephone interviewing follow-up of non-respondents, co-operation was manipulated by assigning specific interviewers to specific sample units. Contact was manipulated by timing, spacing and prioritizing calls. The tailored fieldwork strategy was successful in significantly increasing representativeness, while maintaining the level of response and costs. Representativeness was determined by R-indicators.