Does Ownership Matter? Localism, Content, and the Federal Communications Commission

Abstract

This study examines the relation between local news content and ownership structure in 17 television markets in the United States. It is an extension of the localism research that was conducted by the Federal Communications Commission (FCC) in 2004 and the Local Television News Media Project at the University of Delaware in 2007 (see FCC, 2007). The findings point to the need to consider television markets as the appropriate unit of analysis when examining the effect of ownership on local content. Ownership does matter in the production of news on local broadcasts. When examining only station-level factors, independent stations broadcast more local content on their newscasts than those stations that were either (a) owned-and-operated (O&O) and part of a duopoly, (b) O&O-only, or (c) part of a duopoly-only. However, when examining station-level and market-level factors of television markets, the station-level ownership profiles positively affected local content. Market-level factors that indicated more consolidation negatively affected the proportion of local news presented in the entire designated market area.

Notes

¹Source: Nielsen Media Research. There are 210 designated market areas (DMAs) in the United States. The number of television households in the DMA determines DMA rank. There were approximately 106.7 million TV households in the United States in 2002.

²The specification of whether a story was local or not local was central to the research question. It formed the dependent variable for the regression equations. Therefore, it was absolutely necessary that we were consistent in its specification. Our definition of local was very straightforward: (a) The story had to take place within the designated market area (DMA), and (b) the story had to be seen as “local” by the average viewer in the market. The first part of the definition (the “location” criterion) was readily verifiable, and it was the main criterion for the local-non-local decision. For stories that took place in the core city or any of the larger cities or towns in the DMA, their location was easily identified. For those stories that took place in smaller towns or areas in the DMA where the local-non-local factor was not immediately clear, we consulted maps of the DMA. To make sure that we properly specified the location as local or non-local, we downloaded maps of each DMA with the delineation of all of the counties. Then, we entered the name of the location in the Google™ map program to determine if the location was in any of the counties within the DMA. The location of the story was an objective fact that we could verify by consulting the appropriate sources. Therefore, it was possible to achieve a 100% agreement on that variable. That was crucial to the research. Otherwise, the specification of the dependent variables would have been problematic.

^aAll values are significant at p = .05.

^bDMAs with duopolies with stations in the database.

*p = .05.

*p = .05.

**p = .01.

³All of the independent variables were examined for collinearity using the tolerance value and the variance inflation factor. None was found to be collinear. Further, only two of the days on which the broadcasts were presented were statistically significantly associated with the amount of local news, but the effect (negative) was very small at less than 1%.

*p = .05.

⁴Three variables that appeared in the station-level equation (No. of markets in which owner owns radio stations, No. of TV stations owned by owner, and No. of radio stations owned by owner of TV station) were removed from this analysis because they were highly collinear with each other, with variance inflation factor scores between 5.4 and 14.2.

*p = .05.