Repeat Viewing & Viewing Intensity of TV Programs

Television audience research can provide much useful information to broadcasters and advertisers to better understand the audience viewing behavior and improve the effectiveness of programming and advertising. Among all topics studied, repeat viewing is a major concern for academic researchers. It is to study the extent to which same people view different episodes of regular programs. Viewers not only differ greatly in the frequency they watch TV but also in the viewing intensity. Thus it is also important to examine the variation in the amount of time people spend on watching television. However, most studies on these two topics use very simple methods which are far from enough to model the TV viewing behavior of a heterogeneous audience. This paper develops more advanced viewer-based statistical models: Random-Coefficient Beta Binomial Distribution (BBD) Regression Model for repeat viewing and Random-Coefficient Beta Model for viewing intensity. Empirical results show that the models fit viewing data in Hong Kong very well. Based on the results, viewers can be segmented into different groups. Both the parameters estimated and the groups classified can give management many useful insights in the audience viewing behavior.