In common with many other countries, there are concerns in the UK about the length of the media list measured by the National Readership Survey. Some users are concerned that the list of over 300 titles is too long, and that respondent fatigue may impact upon the quality of the estimates concerned. On the other hand, some users would welcome the opportunity to extend the list of titles surveyed, or the number of questions asked about each title. There are a number of possible solutions to reduce the number of titles any particular respondent is asked about. One solution is to split the media list into a number of parts, and then show each respondent a selection of those parts rather than the entire list.
This, we believe, is the basis of the experiment ag.ma have been carrying out in Germany, which we will hear about at this Symposium. In the German test the media list has been split into three parts. Each respondent is then asked about two of the three parts. Great care is taken in deciding how to split the media list into these three parts, but the parts will be allocated to respondents at random. So, for instance, a respondent who reads motoring titles may or may not happen to be asked about the part of the media list that includes motoring titles. Another feature of the ag.ma test design is that the overall sample size has been increased by a third so that the base to analyse results for any particular title remains unchanged.
Symposium: 2005: Prague, Session 5 - Technical Problems and Solutions
Authors: Page, Katherine
Organisations: Ipsos MediaCT
Topics: Passive Measurement, Questionnaire Design