Abstract

Nationwide Newspaper Coverage of the 1998 Tobacco
Master Settlement Agreement: A Community Structure Approach

The 206 billion dollars the tobacco industry agreed to pay 46 states in November, 1998, was unprecedented. Previously, although sued several times, the tobacco industry was almost always victorious. When former tobacco company employees became willing to testify, these outcomes changed. The tobacco industry was shown to have intentionally stalled scientific and health awareness regarding smoking throughout the 50s. States sued to recoup health care money spent on ill smokers. Ultimately the Master Settlement Agreement (MSA) was developed as a legal remedy, signed by 46 states, joining Florida, Minnesota, Mississippi and Texas, which came to similar settlements a few months earlier.

This study explores the connection between city demographics and nationwide news coverage of this event, specifically using a "community structure approach" initiated in Minnesota by Tichenor, Donohue & Olien (1973, 1980) and elaborated in nationwide studies by Pollock and colleagues (1977,1978, 1994-2001). This approach suggests that particular demographic structures of a community are systematically linked to newspaper reporting perspectives on critical issues.

A sample of 281 newspaper articles, all those over 150 words published between 6/1/97 and 4/1/00 in a cross-section of 14 cities nationwide, was obtained through the DIALOG newspaper database. Articles were coded for: placement, headline size, story length and presence of photos. Article "direction" was also coded as favorable, unfavorable or balanced/neutral regarding the MSA. Borrowing the Janis Fadner Coefficient of Imbalance from psychology, a single "Media Vector" was calculated for each newspaper. These scores, along with city demographics, were compared using Pearson correlations, regression analysis and factor analysis.

Significant positive correlations with favorable coverage of the MSA were found for measures of "privilege": percent family income over $100,000 (r= .634, p= .007) and percent with at least a college education (r= .574, p= .016). This finding illustrates Pollock's "violated buffer" hypothesis: the greater the proportion of privilege in a city, the more negative the coverage of a biological or lifestyle threat (in this case, tobacco). Using regression analysis, three variables accounted for 68 percent of the variance: "privilege" (high income); partisanship (percent voting Republican) and percent African American. Both Republican voting and percent African American were linked to less positive or negative coverage of the MSA. Factor analysis of city characteristics and factor regression yielded three factors accounting for 57 percent of the variance: "privilege" (composed of college educated, percent professionals and high income); "vulnerability" (% unemployed); and "stakeholders/healthcare access" (% African American population, % city government finances spent on health, hospitals). Confirming the "violated buffer" hypothesis is congruent with Olien, Donohue & Tichenor's "guard dog" hypothesis, predicting media criticism of industries/sectors openly violating the public trust.