Portal de Periódicos da CAPES

Media Supply, Audience Demand, and the Geography of News Consumption in the United States

2009; Taylor & Francis; Volume: 26; Issue: 3 Linguagem: Inglês

10.1080/10584600903053361

1091-7675

Scott L. Althaus, Anne Cizmar, James G. Gimpel,

Media Studies and Communication

Abstract The choice to seek out political information is a function both of the individual traits of consumers and of the supply of news content in particular media markets, but previous research has tended to focus only on the individual-level correlates of news exposure. This article explores how the size and complexity of local information markets influence levels of exposure to local, network, and cable television news, as well as exposure to talk radio, online news sources, and daily newspapers. Using multilevel modeling and spatial lag regression, our analysis shows that the structure and demographic tendencies of local news markets are strongly correlated with patterns of individual-level news exposure even after controlling for individual-level demographic characteristics known to predict information seeking behavior. Moreover, we find consistent evidence of regional information cultures that influence demand for news beyond the impact of demographic and market-level factors. Keywords: news exposuremedia consumptionmedia marketsspatial analysisnews consumptiongeographyspatial datamedia supply Notes 1. However, if local news products deliver large lead-in audiences to national programs, then the size of local news markets could be positively related to local variations in the size of television audiences for network news broadcasts (e.g., CitationMcDowell & Sutherland, 2000). Nonetheless, such a relationship would be limited to local and network television audiences, and could not affect the relative size of cable news audiences. 2. DMAs are drawn for television audiences, but the audience structure for radio is quite similar. Arbitron chooses to structure radio markets by metropolitan statistical areas instead of DMAs, but for purposes of simplicity, we aggregated up the number of radio stations within the boundaries of television DMAs. 3. For counties with unusual distributions of audience attention, Nielsen occasionally splits a county into two different markets. But this is a fairly rare occurrence. 4. The number of valid cases was further reduced in our multilevel analysis by the approximately 16,000 respondents who refused to answer the income question. 5. The talk radio question was altered in 2004 to capture use of National Public Radio separately from other talk radio programming. To create a talk radio variable comparable to that in 2000, we used the larger of the 2 days per week estimates given for NPR and non-NPR talk radio. Correlation analysis at the DMA level shows that this measure yields almost identical estimates of talk radio use as came from the 2000 data. 6. Spatial dependency is not strictly a technical problem in econometrics because it commonly reflects an underlying substantive process. In our case, it may be anchored in the common social interactions that occur within and across such arbitrary designations as jurisdictional or market boundaries. Commuting across such boundaries is a daily occurrence, for example, as human settlement and interaction is hardly confined to such uncontrolled spaces. In addition to broadcast signals, neighboring markets share many common population characteristics and environmental features that structure cross-market relationships. These spatial commonalities produce a synchrony among observations that declines with distance. 7. The formulation for Moran's I is as follows:, where n is the total number of observations, xi and xj are pairs of observations, is the mean for all observations (DMAs) in the nation, and wij is the weights matrix of 1s and 0s specifying the neighbor relationships for pairs of observations. 8. GeoDa is available for free download, with complete documentation, at https://www.geoda.asu.edu. 9. The NAES question for online news asks respondents to report how many days per week they use online sources for news of the presidential campaign; questions for the other five news media simply ask for the number of days per week. 10. Average Democratic vote share is the average percentage of the DMA's presidential vote that went to Democratic candidates in 1992 and 1996; political competition is 100 minus the absolute value of the average difference between the Democratic and Republican percentage of the presidential vote in 1992 and 1996. Higher values of the political competition variable indicate a closer contest between the major parties.