Abstract
That the economy influences support for the government is a cornerstone of electoral research, but how this comes about is less understood. Largely due to data limitations, researchers have had difficulty estimating how much of the economy’s effect on the vote is direct and how much is driven by media coverage. New data and automated text analysis make possible the first media-oriented mediation analysis of the economic vote. Using 3 million news articles from 15 developed countries in six languages, we demonstrate that about 30% of the effect of growth is mediated through news sentiment. In contrast, the effect of unemployment on elections is direct. Deviations of economic reporting from the actual economy correlate with election outcomes, though not as much as is sometimes presumed, as we detail in two election examples in which media reporting likely did and did not determine the outcome.
| Original language | English |
|---|---|
| Pages (from-to) | 267-281 |
| Number of pages | 15 |
| Journal | Journal of Politics |
| Volume | 88 |
| Issue number | 1 |
| DOIs | |
| State | Published - Jan 2026 |
Keywords
- economic news
- economic vote
- media
Fingerprint
Dive into the research topics of 'Do Voters Respond to the Economy or to News Reporting on the Economy? A Mediation Analysis'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver