Political participation, diffuse support and perceptions of political efficacy as predictors of mass media use

Most research on the development of political orientations in children indicates that diffuse support, political efficacy and political participation develop earlier in life than regular and extensive use of television and newspapers for public affairs information. I therefore argue that political orientation variables should be treated as predictor, rather than criterion variables in analyses of political communication.

A test of the model utilizing surveys of 190 Mexican‐American, 176 black and 225 white adults shows that of the independent variables used in the regression models, political participation is the strongest predictor of media use for public affairs information. Respondents from all three ethnic groups who were politically active were also heavier users of TV and newspapers for public affairs information. This tendency can be attributed to the social and practical utility of public affairs information for the politically active citizen, as well as to the reinforcement value of the information.

Results also indicate that two political orientation predictor variables (political participation and political efficacy) used in the path model for whites significantly predicted media use, while two out of three (political participation and diffuse support) were significant for blacks, and one out of three (political participation) was significant for Mexican‐Americans. Thus, although multiple R's from the models for all three ethnic groups were significant, evidence regarding the predicted relationships is less conclusive for blacks and Mexican‐Americans. This can be attributed to the effects of extraneous variables such as relevance of news to ethnic minorities. Because blacks and Mexican‐Americans are not very visible in public affairs nevus, members of these groups may be less interested in these media contents than whites, regardless of political orientations.

I conclude that political orientations to media use path models have some validity, particularly for white samples. Further research should measure the time‐order of the variables more directly, as in longitudinal studies, and use geographically diverse samples.