Abstract:

The ratio of the estimated readership of a newspaper or magazine to its net sale or circulation - 'readers-per- copy' - remains a controversial statistic, principally because many people believe, erroneously, either that similar publications should, necessarily, have the same numbers of readers-per-copy or that readership should increase or decrease proportionally with circulation. This paper develops a simple and robust, exponential model of the readership/circulation relationship, which generalises across publication types, readership measurement methods, countries, time periods and regions. Whilst the model is, intentionally, only descriptive, it can already assist in pointing towards comparative bias in readership estimation, between titles or measurement methods. In a final section, improvement of the predictive accuracy of the model is discussed, alongside the possible inclusion of other variables - notably the source of copies and the degree of interest in a newspaper or magazine shown by potential secondary readers.