Abstract:

This paper discusses issues relating to the representativeness of data obtained from an audience measurement panel. The paper gives a broad overview of most factors which have a key influence on representativeness but concentrates on the practical activities which must be carried out on a regular basis in order to maintain the quality of the data. The paper reviews relevant statistical issues and describes the current methods employed for motivating panel members on the UK measurement panel, including a description of the tangible rewards and communicative procedures. However, the most important part of the paper describes and examines the role of input quality control procedures. These procedures are applied to the data registered by individual panel members in such a way as to identify any system/meter malfunctions and, in particular, any deterioration in a panel member's performance in button pressing activity. In the UK these input’ procedures embrace a number of specific checks, including extreme viewing, nil viewing, unallocated button usage and uncovered viewing. When an individual fails a certain input procedure this initiates either a telephone call to the panel member or a special meter validation check. This process itself forms an important integrated part of the panel maintenance procedures described in the early part of this paper. Because this integrated approach is so important to the overall accuracy of the data the parameters used in the input procedures are critical. The paper examines the reasons for choosing the currently used parameters and presents details of the number of checks made and their outcomes in terms of panel member errors, genuine behaviour, and system malfunctions. In summary, the paper demonstrates how important panel maintenance and input’ quality control procedures are in ensuring that an audience measurement panel continues to provide consistent representative data.