Automatic segmentation for reach/frequency estimation of newspaper sections and internet papers

This paper will present a new way of estimating reach and frequency without asking the frequency question or conducting double interviewing. Instead, the sample is segmented automatically by a CHAID-analysis, maximising the differences in reading probabilities among the segments. Typically, many segments are created, individualising the reading probabilities more than when using frequency groups.
Two examples are presented: First, an experiment in which heavy users of the Internet are sampled on the Internet itself. The readers of each “Internet paper� are segmented by variables on their use of the Internet including types of sites visited, on their use of other media, on attitudes and on demography. Next, the Gallup Denmark’s new section readership measuring, being a separate telephone survey calculating reading probabilities of the sections to the readers of the paper in the national readership survey. Finally, the question is raised: Is the frequency question needed at all in readership measuring

Download Paper (PDF)