If you want to understand the modern information environment, start with something boring: your cable lineup.
On paper, channel numbers are administrative trivia. In practice, they are part of the informational infrastructure that shapes what people encounter when they flip through television, alongside many other design choices that rarely attract attention. Some places have Fox News in the low 20s. Other places have it near 100. Those differences are often historical accidents, not reflections of local ideology. And yet they predictably change who watches what, which then ripples outward into politics, policy, and behavior.
In a series of papers with coauthors, I have studied what happens when individual exposure to partisan media shifts for reasons that have nothing to do with voters’ prior beliefs. The headline is simple and unsettling: small, plausibly random differences in media exposure can scale into measurable differences in democratic choices and public life.
This research helps us understand how media persuasion works. Partisan news can shape which issues feel salient, which interpretations feel natural, and which tradeoffs seem reasonable. When those shifts occur at scale, they do not stop at individual beliefs. They echo outward through elections, policy choices, institutional behavior, cultural norms, and even everyday actions. While these concerns do not justify restrictions on speech or policing of viewpoints, they do underscore the importance of building credible and trustworthy institutions that can defend and sustain evidence-based policy.
A natural experiment hiding in your remote control
If you compare two counties that are otherwise similar, but where Fox News appears earlier in the channel lineup in one county than in the other, the first county will watch slightly more Fox. That extra viewing is not huge in any single household. But when it accumulates across millions of people and many years, the political and policy consequences become visible. This “channel position” design runs through multiple projects in this research agenda. It lets us ask a clean question: what does partisan media exposure do when the exposure itself is shifted by something close to chance?
Why should channel numbers matter at all? Because people surf. Channel surfers typically start low and click upward until they find something that holds their attention. If a news channel sits in the low 20s rather than around 100, it gets more accidental exposure and more habitual viewing.
Counties […] where Fox News appears earlier in the channel lineup […] will watch slightly more […]. That extra viewing is not huge in any single household. But when it accumulates across millions of people and many years, the political and policy consequences become visible.
The key identifying claim is that channel positions are not systematically chosen to match local politics. They are often inherited from historical quirks of cable systems. Across our papers, we show that channel position is unrelated to pre-existing markers of conservative politics such as historical vote shares or past fiscal policy. This doesn’t mean the channel lineup is perfectly random. Nothing in social science is. But it does mean we can move much closer to making a causal inference about a channel’s political impact than is possible with the usual comparison between people who choose to watch partisan media and those who do not.
From viewers to voters
The first link is political identity, explored in our paper “From Viewers to Voters,” co-authored with Sergio Galletta, Matteo Pinna, and Christopher Warshaw. Using survey data covering 2000–2020, we find that higher exposure to Fox News increases the share of Americans who identify as Republicans and pushes individuals’ policy ideology to the right.
This matters because democratic systems are not only moved by campaign persuasion in the final weeks before an election. They are moved by longer-run shifts in identity and ideology, shifts that change how people interpret events and evaluate parties, candidates, and institutions. It also matters because Fox News has a wide audience that includes many people who are not staunch partisans. We present evidence that many people consume “light” amounts, on the order of minutes per day. That is exactly the kind of exposure that a small nudge in channel position can plausibly move.
Once you shift political identity and ideology, votes move. In presidential elections, counties with a lower Fox News channel position show a higher Republican two-party vote share, which has increased over time. The magnitudes are not that large. But elections are often decided at the margin, and modern American politics has had many margin elections. Electoral College tipping-point states in 2016 and 2020 were won by margins of 0.7% and 0.6%. When the political system is that close, a media effect of around half a percentage point is not a curiosity. It can be pivotal.
We find that higher exposure to Fox News increases the share of Americans who identify as Republicans and pushes individuals’ policy ideology to the right.
Further, the pattern is not confined to the presidency. For Senate elections, we find a positive effect that becomes statistically significant starting in 2014. We also see pro-Republican effects in House races and governor races in later years. The broader takeaway is that partisan cable news, despite its national focus, has measurable effects up and down the ballot.
From elections to policy: local government budgets
If partisan media changed only voting, the story would already be important. But the deeper question is what happens after votes translate into governance. Local government is a useful stress test because Fox News is not a “local politics” channel. It is overwhelmingly national in topic and tone. And yet we find that it reshapes local fiscal policy in a distinctly ideological direction.
In the paper “How Cable News Reshaped Local Government,” co-authored with Sergio Galletta, we estimate that higher Fox News exposure reduces both local revenues and local expenditures per capita. A movement in channel positioning that induces about 45 minutes of additional Fox watching per household per month corresponds to about a 3.4% reduction in revenues and a 3.8% reduction in expenditures, which we translate to roughly $150 per capita per year. Those are big numbers in the context of local public finance. They speak to something beyond “political vibes.” They speak to the size of government.
A movement in channel positioning that induces about 45 minutes of additional Fox watching per household per month corresponds to about a 3.4% reduction in revenues and a 3.8% reduction in expenditures.
The composition of spending shifts too. We see decreases in public services like education, health, and welfare, but not in protective services such as police and corrections. On the revenue side, the largest decreases are in fees for government services, consistent with lower service provision or privatization. And privatization is not just rhetoric. We find increases in private-sector firms providing education and security services, and the share of students in private schools rises with Fox exposure.
We explore three reinforcing routes by which this change occurs:
- Selection into office and candidacy. In a sample of mayoral elections, we find evidence that Fox exposure reduces the probability of Democrats running and increases the probability of a Republican win.
- Changes in political agendas. Higher Fox exposure causes Democrats to talk more about fiscal issues and less about public welfare issues.
- Shifts in policy preferences. In Texas ballot referenda related to tax collections, higher Fox exposure increases support for fiscally conservative positions.
The third channel is particularly important. It suggests that media effects are not only about party loyalty. They can be about policy preferences themselves.
From policy preferences to institutional behavior: sentencing
The next step is to ask whether partisan media can move decisions inside institutions that are not usually treated as political. Criminal sentencing is a high-stakes outcome, and judges are supposed to be insulated from day-to-day political currents. In practice, insulation varies. Many judges are elected.
In “Conservative News Media and Criminal Justice,” co-authored with Michael Poyker, we find that higher Fox News viewership increases a county’s average criminal sentencing length. Even judges, who are supposed to be neutral and unbiased, can be moved by media. We look for evidence that Fox exposure changes charging behavior or crime rates, which would suggest policing or offender channels. We find no effect on the number of offenses or on predicted sentence length based on charges, and no effect on crime rates. The evidence points toward the sentencing decision itself.
The institutional detail matters here. When we split by selection systems, the effect appears in states with elected judges, and it is absent in states with appointed judges. That pattern is consistent with electoral pressure as a key mechanism.
To connect media messaging to behavior, we analyze Fox transcripts and measure attention to crime. The effect of Fox viewership on sentencing is stronger in months when Fox mentions crime and drugs more frequently, consistent with the messaging about crime driving the effect, rather than generic conservatism. That is: specific messaging primes specific concerns, which then operate through electoral accountability to influence judges.
Partisan media can leave a lasting imprint by altering the incentives and behavior of other institutions that mediate political information.
A crisis stress test: COVID behavior and beliefs
Elections and sentencing matter, but they are not daily experiences for most people. The pandemic was.
In “The Effect of Fox News on Health Behavior during COVID-19,” co-authored with Sergio Galletta, Dominik Hangartner, Yotam Margalit, and Matteo Pinna, we examine how partisan media exposure shaped beliefs and behavior during the early stages of COVID-19, when accurate information was unusually valuable and individual incentives to respond were strong.
The analysis combines two types of evidence. First, using the same channel-position instrument, we estimate the causal effect of Fox News exposure on actual social distancing behavior, measured using high-frequency mobility data. In areas with greater exogenous exposure to Fox News, people reduced their mobility significantly less in the early phase of the pandemic, indicating slower adoption of distancing and other protective behaviors. Second, we link these behavioral patterns to individual-level survey data on beliefs and attitudes. Fox viewers were substantially more likely to hold false or misleading beliefs about COVID-19 treatments, including being 12.4 percentage points more likely to believe that hydroxychloroquine was effective. They were also much more likely to prioritize economic activity over public health protections.
The broader implication is sobering. These effects emerged in a setting where people had strong personal incentives to seek accurate information and adjust their behavior accordingly. Even under those conditions, partisan media coverage measurably shaped beliefs and behavior. The “marketplace of ideas” did not automatically neutralize biased or misleading information when trust was thin and uncertainty was high.
Feedback loops: when TV changes newspapers
Partisan media does not operate in a vacuum. It sits inside a local ecosystem that includes newspapers, radio, online outlets, and interpersonal discussion. If Fox shifts local demand for political content, or shifts the political equilibrium of a place, other media may adjust too.
In “Media Slant Is Contagious,” co-authored with Philine Widmer, Clementine Abed Meraim, and Sergio Galletta, we study this possibility by focusing on local newspapers. We construct a measure of newspaper slant using text analysis trained on the language of Democratic and Republican members of Congress, allowing us to place each outlet on a common ideological scale. We then ask whether exogenous variation in local Fox News viewership, again driven by channel position, predicts changes in newspaper slant over time.
When political identity substitutes for evidence, policy disagreements harden into conflicts that are difficult to resolve through democratic deliberation.
The answer is yes. Local newspapers become more right-leaning in areas with greater Fox News exposure. These changes emerge gradually, consistent with adjustment rather than an immediate shock, and they are strongest for newspapers operating in more competitive local media markets.
This finding matters because it speaks to persistence. Even if individual viewing habits change, or even if a particular channel’s influence wanes, the informational environment may already have shifted. Partisan media can leave a lasting imprint by altering the incentives and behavior of other institutions that mediate political information.
Democratic resilience in an age of partisan media
Taken together, these projects point to a single theme. Partisan media can move outcomes at multiple layers of the democratic system. Political preferences and identities shift. Election results shift, including in down-ballot races, with real governing consequences. Policy shifts, even at the local level where the media content is not directly about local issues. Institutional behavior shifts, including settings like criminal sentencing where democratic accountability can transmit public mood into formal decision-making. Crisis behavior and beliefs shift, even when accuracy is personally valuable. And the broader media ecosystem shifts, amplifying and extending the original shock. That is a great deal of movement for something as banal as channel placement.
The larger point is not about any one network. It is about structural vulnerability. When political identity is tight and trust is thin, small informational nudges can propagate through the system, compounding into durable differences in governance and lived experience.
It would be a mistake to read these findings as an argument for censorship or government control of the media, however. In open societies, empowering the state to police political content or adjudicate permissible viewpoints risks doing more harm than good. The cure can easily become worse than the disease.
What the evidence does point to is something both more modest and more demanding. Democratic resilience depends on the credibility of institutions. When people do not trust public health authorities, statistical agencies, courts, regulators, or local government, they become more reliant on partisan cues for basic factual orientation, especially in moments of uncertainty. And when political identity substitutes for evidence, policy disagreements harden into conflicts that are difficult to resolve through democratic deliberation.
The constructive lesson is to invest in the foundations of trust rather than in coercive information control. That means institutions that are competent and transparent, that explain uncertainty honestly, that correct mistakes visibly, and that can demonstrate fair treatment across groups. It means evidence that is publicly legible, not only technically correct. And it means defending the norms that make disagreement productive: open debate, adversarial testing of claims, broad democratic legitimacy, and inclusive prosperity that gives citizens a stake in a shared reality.
Partisan media can shift behavior and policy at scale. The response is not to silence voices, but to build a society in which trustworthy institutions and robust democratic norms make persuasion harder for rhetoric that is untethered from evidence.
Elliott Ash is an associate professor at ETH Zurich whose research sits at the intersection of political economy, law, and data science. He is widely known for work on the political economy of media, including how news coverage, platforms, and information shape political behavior and accountability, as well as for research on judges, courts, and legal institutions. More recently, his work has expanded to artificial intelligence, studying legal reasoning, judicial decision-making, and the economic and institutional implications of AI systems.
