Facing panel non-response, consequences and solutions
In this paper additional support is given to our statement that the quality of panel data must be improved before these data can be used as a firm basis for decision-making in marketing. To this end the case study is extended to cover data which refer to six brands instead of to one brand. The results are shown in the following section. Then we show the results of a formal analysis in which the differences between actual data and consumer panel data are explained by a number of exogenous variables such as price, advertising expenditures and product characteristics. From this analysis a number of guidelines is obtained which may be used to correct consumer panel data in order to decrease the data bias in consumer panel data to a significant degree. We proceed by showing the improvements in the quality of retail audit data which result from a new way of dealing with non-response in retail audits. Finally, we shall summarise our findings.
- This could also be of interest