Media Effects Revisited: The Relationship between News Media and Public Opinion

• In Event: Media Effects

Abstract

Almost all theory on the relationship between news media and public opinion approaches it with an assumption of “media effects”— that news media influence what the public knows and thinks about topics in a variety of ways. However, U.S. media, being mostly commercial and competitive, might also be expected to cater to public tastes. If this is true, then public preferences might affect the content and sentiment of news. Hardly any research, and even less theory, addresses the potential reciprocal effects of mass media and public opinion.

The present study addresses this idea longitudinally in the context of media coverage of, and public opinion toward, gays and gay rights. News sentiment toward gay rights is measured using an automated content analysis of hundreds of New York Times articles dealing with gays and gay rights, aggregated into a yearly index, between 1972 and 2018. Collective preferences toward gay rights are measured based on hundreds of survey question marginals, also aggregated into a yearly index, using Stimson’s dyad ratios algorithm, over the same time frame.

Then, the indexes of NYT sentiment and public opinion are analyzed using time series analysis. Recent research shows that the often-used single-equation error correction model can be sensitive to misspecification and, consequently, produce misleading results (Grant and Lebo 2016). To avoid this we use the more robust ARDL model, and in particular, Pesaran et al.’s (2001) bounds testing approach to testing for cointegration. Using this approach to test the relationship between NYT sentiment and public opinion toward gay rights, the results show that collective preferences drive news sentiment, but that news sentiment does not affect public opinion. Based on prior research, this is the opposite of what one would expect. These results call for further study, reconsideration, and possibly reformulation, of theory on the relationship between the content of mass media and public sentiment toward issues.

Grant, Taylor, and Matthew J. Lebo. 2016. “Error Correction Methods with Political Time Series." Political Analysis, 24(1):3-30.

Pesaran, Hashem, Yongcheol Shin, and Richard J. Smith, Richard J. 2001. “Bounds Testing Approaches to the Analysis of Level Relationships." Journal of Applied Econometrics 16(3):289-326.

Authors

• Ken Mulligan, Southern Illinois University, Carbondale

• James Wilson, Southern Illinois University