1995 Consumer Panel Data

Panel Participant Maintenance

In recruiting panel participants, particular emphasis is placed on achieving a good regional distribution, which is accomplished in part by using random selection procedures. Recruitment occurs partly through the use of interviewers (who are to meet quotas in the areas of German federal state, age, and household size), and partly through direct mailings and telephone contact (that is based on a random selection of addresses stratified by administrative districts and community size). In a panel of this type, where longer-term cooperation is highly valued, successful recruitment is considerably more difficult than for a survey where only a single face-to-face interview is conducted. Methodology research has indicated that though one can achieve higher initial recruitment rates by relying heavily on interviewers, the drop-out rate in the first weeks of panel participation is also correspondingly higher.

A variety of methods are used to motivate panel participants: regular information about the study and communication from and with the organizers of the study, presents, bonuses, and raffles. Thus, several times a year participants receive letters and brochures that broadly describe the use to which the data they have provided will be put. They receive congratulatory messages on their birthdays, and are reminded either by letter or telephone if they have forgotten to mail in information about their purchases. A telephone hotline exists to answer questions participants may have, and if problems appear in the data submitted, participants are contacted by telephone. Each quarter, a VW Golf car is raffled off, as well as a trip that includes a generous cash allowance. There are also drawings which have cash prizes, prize competitions, raffles for trips to Nürnberg to visit the GfK (which organizes the panel study), and various small presents. After a year of regular cooperation, participants can select a bonus worth approximately DM 70 from a catalogue.

Panel quality control occurs at two levels. One is the continuous monitoring of the quality of panel cooperation, and in cases where participants report purchases poorly or irregularly, they may be removed from the panel. The other level consists of spot-checking reported purchases for plausibility as well as for minor, correctable errors in the entries.

In addition, a system has been developed to check cooperation against a base reference that is derived from a monthly compilation sample. At the end of a month, the reporting of all participants is checked first to see whether data has been submitted to the institute (vacations or illnesses are permissible notifications), and then whether the scope of the data provided about purchases lies within a particular range of tolerance. This range is determined by the mean value for the household size category to which a specific household belongs. Only data from the participants who meet these quality guidelines are utilized.

At regular intervals, investigations are also conducted about longer-term reporting behaviour, with the occasional result that GfK terminates collaboration with a particular panel participant.

In checking reported purchases, the key issue is to recognize errors that have been made in the entries by the households submitting the data. Most such errors occur out of carelessness in noting prices (thus, entering 1.08 € rather than 10.80 € in the household budget book). Appropriate checks, such as through comparison with the average price for that particular purchased good, make it possible to recognize the nature of the error and to correct it.



© GESIS Georg Papastefanou 02. Januar 2008