To develop a model of newspaper readership more closely related to explanation of readership behavior than prediction, a study used a combination of demographic and political participation variables in a multivariate analysis of data obtained in a national survey. Information from interviews of approximately 1,200 adults conducted in 1979 was analyzed to determine level of newspaper use and political activity. Demographic data were also collected. To determine the relationship of the various variables with newspaper readership, a path model that used the log-linear technique originated by L. Goodman (1972) was employed. Results indicated that older (55 and older), wealthy, more educated, politically active people were the most frequent newspaper readers. Overall, the best predictor of readership was political activity, followed by age, followed distantly by education. The findings suggest that since political activity is a good predictor of newspaper readership, and since it is possible to change a person's level of political activity, the newspaper industry should develop campaigns to draw young people into the political process. (FL)