Media Use, Environmental Mediators, and Pro-environmental Behaviors across and within Countries

ABSTRACT

This study develops a model of direct association of exposure to environmental media content, and indirect association through environmental attitude and environmental efficacy, with pro-environmental behaviors. It also considers secondary media roles of exposure to general news media and involvement in mediated civic activism. The model and hypotheses are tested through Hayes Process mediation models, using secondary, cross-sectional survey data from 11,000 respondents across 11 countries. The model is well-supported overall and within countries, and the secondary media variables have generally consistent effects within countries. Socio-demographic covariates have varying relationships with environmental attitude, environmental efficacy, and pro-environmental behaviors, overall, and within countries. In line with social cognitive theory, these results suggest that media use related to environmental issues does not have to raise individuals’ pro-environmental attitude or efficacy (though it does) to increase engagement in pro-environmental behavior.

KEYWORDS:

• Environmental media content
• mediated civic activism
• pro-environmental behaviors
• mediation model
• multiple countries

Acknowledgements

We greatly appreciate access to data from the National Geographic Research Center, and collaboration with and advice from Dr. Abel Gustafson of the Department of Communication at U. Cincinnati, and Dr. Matt Goldberg of the Yale Program on Climate Change of Yale University. Dr. Rice also acknowledges support from the Arthur N. Rupe Foundation, in his role as the Arthur N. Rupe Professor of Social Effects of Mass Communication. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 HC3 (heteroscedasticity-consistent standard error estimator, form 3) corrects (i.e. uses robust standard errors) for heteroscedasticity (error variance differs across observations). Heteroscedasticity is common in cross-sectional data, and may be significant across groupings (such as, here, countries). Hayes and Cai (2007) conclude that researchers should routinely use a heteroscedasticity-consistent standard error estimator for OLS regression tests. We use the HC3 version implemented in Process.