Visual Complexity in Television News: a Time Series Analysis of Audience Evaluations of an Electronically Estimated Form

VISUAL COMPLEXITY IN TELEVISION NEWS: A TIME SERIES ANALYSIS OF AUDIENCE EVALUATIONS OF AN ELECTRONICALLY ESTIMA ED FORM COMPLEXITY V BLE Tony Rimmer School of Journalism Indiana University Bloomington, IN 47405 Telephone: (812) 335-1717 Office The object of this study was to develop a content-free, form complexity measure of a network TV newscast based on pixel light Intensity, and to test audience reactions over time to this measure. The form complexity variable, referred to here as static complexity, was measured throughout the newscast. It constituted the independent variable. The dependent variable, responses from 110 subjects throughout the same newscast, was measured with bipolar, adjectival scales. Three hypotheses were tested using regression techniques. The first hypothesis proposed that subject ratings of TV news would show a curvilinear relationship with the static complexity of the newscast image. This hypothesis, tested in cross-sectional form I only, was not supported. The second hypothesis proposed that static complexity had a cumulative effect on subject responses, such that a lagged form offered a better description of the relationship between the two variables than did the cross-sectional form. A simple, crosssectional model, accounted for just six percent of the variance in subject responses. An autoregressive model, lagged back 150 seconds, improved the variance accounted for in the subject responses to 31 percent. The third hypothesis proposed that subjects would be able to anticipate variations in the form complexity variable through the time series. Subjects showed significant variation in their responses to the newscast material up to 120 seconds In advance, of associated changes in static complexity. It was concluded that the results of this study may have implications for research in TV news which confines itself only to content-based variables at the expense of form variables, and which assumes only a cross-sectional relationship between newscast material and audience responses. It is suggested that both form ' variables and longitudinal relationships should be considered in the study of audience reactions to TV news. Introduction: Conventional wisdom in television news production suggests that "talking heads" on television are boring. What viewers are said to want is action pictures. This production convention, apparently derived from the movie industry, has become part of the folklore of television news. This study operationalizes "action footage" as a physical measure of image complexity and considers the idea that this image complexity is an important component of the appeal of television news. This appeal, it is suggested, is grounded in a basic drive humans have to process stimuli, the kind of stimuli television news provides. Taking the idea a stage further, the more complex the stimuli (i.e., the more complex the TV news image), the more appeal it. will have to audiences. Ipso facto, action pictures are desired, "talking heads" are'not. The former has a complex image, the latter is not complex enough. In order to explore this image complexity idea, two time series measures were obtained using a regular half hour broadcast of a "CBS Evening News" program and a'group of viewers. The two measures were: 1. An objective, machine-based measure of the complexity of the television image across space. This neasure involved computer-based readings, at five-second intervals, of the variations in light intensity of the pixels which make up the image on the television screen. The measure is referred to in this