(together with Werner Hemsing and Lorenz Gräf, Globalpark AG)
In order to study panel attrition and optimal gratification strategy for survey participation, a study has been carried out at Globalpark AG. A commercial online access panel has been analyzed with regard to drop-out and incentives given to participants. An online access panel is a group of internet users that are invited to participate in surveys on a regular basis. Participants might be recruited both in the world wide web, offline (e.g. by means of CATI), or based on pre-existing databases. The representativity of an online access panel largely depends on the recruiting of the panel members. For surveys, panel members are drawn randomly or by selecting them according to certain characteristics. Consequently, different panel members participate in different surveys.
Within the scope of a study undertaken by the research department of Globalpark AG, the consequences of different gratification strategies and frequency of invitations for surveys on panel attrition has been analyzed. 699 panel members were recruited in August and September 2000 and filled out the initial questionnaire. These panel members habe been surveyed over the next two and a half years up to 12 times. The participants of three of these surveys took part in raffles. Three surveys did not come with any incentive. Six out of the twelve surveys came with bonus points that could be exchanged into shopping vouchers. On average, the 699 panel members have been invited to 6.3 surveys. The average participation rate (answering at least one question) was 61%. In April and May 2002, four surveys had been conducted that again invited the 699 panel members. These four surevys serve as a bench mark whether or not the panel members are still active. 406 panel members reacted to the invitations and answered at least one question. This is equivalent to a participation rate of 58% (see figure 1).
Frequent invitations to surveys foster the readiness to remain in the panel. Those panel members that left the panel had been invited to 5.5 surveys on average. Members that remained in the panel, in contrast, had been invited to 7.3 surveys on average (see figure 3). The relation between frequency of invitation and panel attrition is shown in more detail in figure 4. Almost 70% of the panel members that received nine ore more invitations were still active in April and Mai 2002. Conversely, panel members that received only few invitations had a high probability of leaving the panel.
Hemsing, W., Müller, J. & Gräf, L. (2002). Panel attrition in relation to different kinds of incentives. Presentation at the 5th German Online Research Conference, Hohenheim.