Abstract:

In February 1992 Nielsen Media Research announced a change in the existing People Meter methodology: specialists (called People Meter Representatives, or PMRs) were to be added to the field force to provide intensified training in the use of the people meter in households with children and teenagers. Research and pragmatic considerations were successfully integrated during the development, gradual implementation and ongoing evaluation of this methodological enhancement. Our goal in malting this change was to improve the accuracy of audience estimates produced by the people meter system. Through the use of intermethod comparisons (including set meter, diary and coincidental data), we developed clear criteria of accuracy. This empirical grounding was enriched with customer input, experience-based knowledge and consultant perspectives. We then formulated hypotheses to define objective goals and to guide the development of specific procedures designed to improve accuracy. A preliminary field implementation stage was intentionally limited in its scope and included qualitative evaluation Since this early implementation was conducted in the live panel, we monitored the process continuously and were prepared to halt the roll out or even retract the change. As we moved to the full-scale roll out, we maintained continuous evaluation of effects. Although this process change did not result in measurable differences in the overall parameters of television viewing, many components of the viewing measure that were predicted to change in the direction of greater accuracy showed statistically significant improvement. By its nature, a change in the initial treatment of panel members will be implemented gradually as the panel turns over. This provides a transition paradigm Although this gradual transition cushions any potential discontinuity in the media marketplace, there are still key issues and sensitivities around the definition of "truth" and the valuation of audience during a long transition cycle. The paper concludes with a discussion of theoretical limits to file measurement of process change effects and the practical issues of providing a stable platform for the television industry while continuously improving the measurement standard.