The London experiment
We may take the primary objective of the basic readership survey in each country to be to obtain for large numbers of publications estimates of average issue readership that are repeatable and stable, acceptably accurate and free from biases between and within publication categories. The assessment of repeatability and stability for any continuous or repeated survey raises no special methodological problems, but the identification of bias is another matter. By definition, the fundamental source of readership estimates in a country cannot be validated against another comprehensive source since this would merely re-define the second source as fundamental. Evidently total validation of the basic set of measures is an impossibility; the evaluation of primary readership surveys in terms of accuracy and freedom from bias must necessarily depend on validation that is partial in one sense or another.
- This could also be of interest