Quantification of Patterns of Media Exposure : an Analysis of Radio Listening Data

Abstract

The main aim of this paper is to develop and describe a mathematical model for audience exposure patterns. So far, the indices of reach, frequency and GRP have been used by many researchers to measure audience exposure patterns. However, it is insufficient to use these indices for modelling the process of audience exposures. Many researchers of media schedules also developed several reach-estimating models which were applied to the real data. It is true that these models are effective measures for predicting reach of exposure, but they are not adequate as models of describing and understanding audience exposure patterns. In this paper, two basic patterns of individual exposures, fixed exposure pattern and random exposure pattern, are proposed as extreme exposure patterns. First, fixed exposure pattern has a set of consistent and invariable exposures. It implies that people habitually continue to choose a program or never do so. Second, random exposure pattern has a set of unstable exposures chosen according to independent probability. It implies that people choose a program at random each time. These two patterns are, so to say, 'ideal types' which are not found in a pure state. These two discrete patterns can be synthesized by the mathematical model proposed by Coleman. It is based on stochastic process theory which continuously places the two extreme exposure patterns by a parameter. In this analysis, Coleman's model is tested by actual radio audience listening data of a metropolitan area, and the result demonstrates its general accuracy in fitting the data. Then some of the important consequences of the model as an analytical tool are mentioned.