This paper deals with new methods of determining target groups in media research. Over the past three decades, media research has essentially developed in response to the advertising industry's interest in information about how large segments of certain target groups could be approached under optimal economic conditions. The definition of target groups, however, has remained at a fairly low level. Many media researchers and users of media analyses assert that further refinement of the demographic definitionsby region, community size, sex, age, social class, and certain occupational groups and of data on buying, consumption and ownership would not really result in an improvement of a media plan's effectiveness. Psychological criteria are seldom applied; imaginative typologies are left to the image advertising activities of the publishing companies themselves. This paper will show that considerable progress in defining target groups is possible, and it will do so using as examples studies conducted by the Allensbach Institute in 1979/81.
Coverage findings can be manipulated depending on the choice of pre-coded response models. The relationship between response and sorting pre-choice alternatives leading to a "yes" or "no" decision has a particular influence here. Internationally, this decision is apparently made quite arbitrarily without any tests using an outside criterion which might at least be termed relatively reliable. The coverage ascertained by original-issue tests is suggested for the print media, and a model of the kind of technical measurements made in blest Germany is suggested for television audience research. As an economic procedure for a transition from media to advertising contact the use of categories concerning using frequency of the media is recommended.
Small details in questionnaire construction, especially pre-coded answer categories, can easily manipulate the measurement of readership figures. Under these circumstances it is a central concern to find a yardstick to measure which readership figures come closest to reality. Is there anything better than testing readership with the editorial interest method, in an operation parallel to the main survey, and then developing the questionnaire construction of the main comparative media survey until the readership figures match the values found with the editorial interest method?
We are facing a dilemma: more information is wanted about as many target groups as possible and the quality of media contacts. We know from a number of field experiments that exact figures regarding media coverage largely depend on the design of the questionnaire. In tracing the influence of the questionnaire on the results the debate concerning the two widely used German intermedia analyses: "Arbeitsgemeinschaft fur Media-Analyse" and "Allensbacher Werbetrager-Analyse", proved particularly productive. The general design of the questionnaire and particularly the wording of questions and response alternatives has a greater influence on the coverage results than the type and quality of the sample. In the following we analyse the influence of different reading behaviour, using the figures of a representative survey conducted by the Allensbach Institut. We will check the correlation between the characteristics of reading behaviour and the reader's contact with editorial text as well as advertisements. In accordance with the findings of other studies the thesis is confirmed: reading frequency can well be considered (and used) as a predictor for advertisement contact, most likely it will even prove more reliable than other characteristics of reading behaviour.