What Is the Definition of Media Bias in News?

Media bias refers to the systematic slant in how news outlets select, frame, and present information, resulting in coverage that favors particular perspectives, political positions, or interpretations over others. The term covers a surprisingly wide range of behaviors, from the individual word choices a reporter makes to an entire organization’s pattern of deciding which stories deserve coverage in the first place. Researchers have cataloged at least seventeen distinct forms of media bias, which makes the concept far richer and messier than the simple left-versus-right framing most people have in mind.

What Media Bias Actually Includes

Most people think of media bias as a news outlet openly rooting for one political party. That happens, but it is only one slice of a much larger phenomenon. A systematic literature review organized media bias into four major categories: linguistic bias, cognitive bias, text-level context bias, and reporting-level bias.

1arXiv. The Media Bias Taxonomy: A Systematic Literature Review on the Forms and Automated Detection of Media Bias

Linguistic bias lives at the sentence level. It shows up when a writer picks loaded words or subjective phrases that nudge the reader toward a particular interpretation. Calling a protest a “riot” versus a “demonstration,” for instance, changes the emotional color of the story without altering the underlying facts. This is often called framing bias, and researchers define it as the use of subjective words or phrases linked with a particular point of view to sway the meaning of a statement.

2arXiv. The Media Bias Taxonomy: A Systematic Literature Review on the Forms and Automated Detection of Media Bias

Selection bias, sometimes called gatekeeping bias, operates at a higher level. It is not about how a story is told but about which stories get told at all. Editors and writers choose from a vast pool of potential stories every day, and what they leave out shapes public understanding just as powerfully as what they include. A newsroom that covers every scandal involving one political party while ignoring equivalent scandals involving the other is engaging in selection bias even if each individual story is reported accurately.

3arXiv. The Media Bias Taxonomy: A Systematic Literature Review on the Forms and Automated Detection of Media Bias

Beyond these two well-known forms, the taxonomy includes subtler varieties: bias by omission of relevant context, bias through the sources a journalist chooses to quote, bias in headline construction that misrepresents the body of an article, and bias in the visual framing of images and video. A systematic review of the literature identified seventeen distinct forms after merging overlapping definitions across studies.

4Elsevier. A systematic review on media bias detection: What is media bias, how it is expressed, and how to detect it

Why Bias Exists on the Supply Side

One common explanation for media bias is that news outlets are simply giving audiences what they want. Research in economics has explored this demand-driven model, which assumes that consumers prefer to read or watch news that is consistent with their existing beliefs rather than coverage that challenges them.

5Marketing Science. News Consumption and Media Bias Under this theory, a left-leaning audience gravitates toward left-leaning outlets, and a right-leaning audience does the same on the other side, creating a market incentive for slant.

But demand alone does not explain the full picture. Ownership plays a distinct and sometimes dominant role. A study of Swedish newspapers found that outlets held by the same owner tended to offer the same mix of political slant regardless of the local audience’s preferences. Owners were even less inclined to differentiate their outlets’ slant before elections, precisely when the political returns to persuasion are highest.

6Kyklos. Ownership and media slant: Evidence from Swedish newspapers That pattern is hard to square with a purely demand-driven model. If outlets were chasing local readership, you would expect them to tailor their slant to the politics of each circulation area. Instead, the owner’s preferences seemed to override local market signals.

Research across 32 European media markets reinforced this. A study of 211 print and broadcast outlets found that owner influence was greater when ownership of individual outlets was concentrated, when it was held by individuals or families rather than dispersed shareholders, and when an ownership group controlled multiple outlets in the same market.

7European Journal of Communication. Media outlets and their moguls: Why concentrated individual or family ownership is bad for editorial independence

An Australian study offered a particularly clear natural experiment. When News Corp acquired certain newspapers in 2016, those papers shifted in a notably conservative direction. When other acquirers later took over different papers, the slant moved in the opposite direction. The identity of the new owner predicted the shift, not the preferences of the readership.

8PubMed Central. Media ownership and ideological slant: Evidence from Australian newspaper mergers Taken together, these findings suggest that media bias is shaped by both consumer demand and the priorities of the people and companies that own news organizations, with ownership effects sometimes overriding the market.

Structural Bias and the Shape of the News

Not all media bias is ideological. Structural bias refers to patterns built into how journalism operates as an institution, independent of any individual reporter’s or owner’s politics. One well-studied form is the tendency to over-represent government officials and major political institutions in the news. A cross-national study found that the balance of attention given to government and parliament in different countries reflected the power balance in each political system, and that this structural emphasis was further shaped by journalistic cultures that value conflict framing.

9The International Journal of Press/Politics. Structural Bias in Cross-National Perspective

This means that even in a newsroom with no deliberate ideological agenda, the basic routines of newsgathering can distort the picture. Relying heavily on official sources, covering events that fit a conflict narrative, and prioritizing stories with dramatic visuals all introduce bias that has nothing to do with left or right. A story about a government policy announcement gets covered because the press infrastructure is already pointed at the halls of power, while an equally important community-level development goes unreported because no press conference was held.

The Problem of False Balance

One of the more counterintuitive forms of media bias comes from trying too hard to appear balanced. False balance occurs when opposing viewpoints on a scientific or factual issue are presented as though they carry roughly equal weight, even when the evidence overwhelmingly supports one side. Research has shown that balanced messages decrease people’s perception of expert consensus, even when the actual level of scientific agreement is very high.

10Memorial University of Newfoundland. Not every story has two sides: the effect of false balance on perceived scientific consensus

Climate change coverage is the classic example. For decades, many outlets gave roughly equal airtime to climate scientists and skeptics, creating the impression of an active scientific debate when the overwhelming majority of researchers had long agreed on the basics. The journalistic instinct to “show both sides” distorted public understanding of a settled question. This form of bias is tricky because it arises from a norm that most people consider virtuous: fairness. But fairness between two arguments is not the same as accuracy about what the evidence shows, and conflating the two can itself become a source of misinformation.

How Images and Visuals Introduce Bias

Text is not the only vehicle for slant. The photographs, video clips, and graphics that accompany a story carry their own framing effects. Research on coverage of the 2009 Gaza conflict found that visual framing shaped emotional responses and influenced how readers evaluated the actors involved. Stories framed with human-interest visuals elicited stronger emotional reactions and changed who viewers saw as responsible or sympathetic.

11Journalism & Mass Communication Quarterly. Effects of Visual Framing on Emotional Responses and Evaluations of News Stories about the Gaza Conflict 2009

Visual bias is worth noting because it is much harder for the average reader to detect than word choice. You might catch a loaded adjective in a headline, but the decision to illustrate a protest with a photo of burning cars rather than a photo of families marching peacefully shapes your impression before you have read a single word. This is one reason why media literacy efforts increasingly emphasize paying attention to images, not just text.

When Your Own Brain Creates the Bias

Sometimes the bias is not in the news at all. It is in the audience. A landmark 1985 study showed pro-Israeli and pro-Arab partisans identical footage of network television coverage of the Beirut massacre. Both groups rated the very same programs as biased against their own side.

12PubMed. The hostile media phenomenon: biased perception and perceptions of media bias in coverage of the Beirut massacre This phenomenon, known as hostile media perception, has been replicated across different topics and decades. More recent experiments confirm that partisans perceive neutral coverage from outlets opposite to their political leaning as biased, even after accounting for their prior beliefs about the outlet and the news content itself.

13PubMed Central. Hostile media perception affects news bias, but not news sharing intentions

Confirmation bias reinforces this pattern on the consumption side. People tend to seek out and engage with information that supports what they already believe. Research on mobile news consumption found that confirmation bias manifests on smartphones just as it does on desktop computers, meaning the shift to mobile news has not disrupted this tendency.

14Journalism and Media. Mobile Selective Exposure: Confirmation Bias and Impact of Social Cues during Mobile News Consumption in the United States

These psychological dynamics matter for the definition of media bias because they blur the line between bias that exists in a piece of journalism and bias that the audience projects onto it. When someone says “the media is biased,” they may be identifying real slant, reacting to hostile media perception, or doing both at once. Any serious conversation about media bias has to acknowledge that these two layers interact constantly.

How Social Media Amplifies and Distorts Bias

Social media platforms add a layer that traditional definitions of media bias were not designed to handle. Research on Facebook and Twitter has found that users tend to cluster into groups of like-minded people, creating echo chambers where a shared narrative gets reinforced and opposing perspectives rarely break through. A direct comparison of news consumption on Facebook and Reddit found higher segregation on Facebook.

15PubMed Central. The echo chamber effect on social media

Algorithmic recommendation systems play a role in this process. Platforms are designed to maximize engagement, and content that provokes strong reactions tends to generate more clicks, shares, and comments. Research has investigated how these recommendation algorithms amplify polarization by feeding users increasingly homogeneous content.

16Journal of Student Research. The Echo chamber-driven Polarization on Social Media The result is that a person’s news diet can become heavily filtered not just by editorial decisions at news organizations but by platform algorithms tuned to keep them scrolling.

This creates a form of media bias that does not sit neatly with the traditional definition. The original news article might be reasonably balanced, but the algorithmic environment ensures you see that article only if it aligns with your existing views, and that you never see the articles offering a different perspective. The bias lives in the distribution system, not the content.

What Media Bias Does to Trust

The practical consequences of media bias, whether real or perceived, show up most clearly in public trust. Research has found that when citizens perceive news coverage as ideologically slanted, their trust in traditional news media drops.

17The International Journal of Press/Politics. Finding Perceptions of Partisan News Media Bias in an Unlikely Place This creates a feedback loop: as trust declines, people become more skeptical of coverage that challenges their views, which makes them more likely to perceive bias, which further erodes trust.

The hostile media perception research described earlier complicates this further. If partisans on both sides perceive neutral coverage as biased against them, then declining trust in media is not solely a response to actual bias. It is partly a consequence of how partisan minds process information. This does not mean real bias is unimportant, but it does mean that restoring trust requires addressing both the quality of journalism and the psychology of the audience.

Detecting Bias With Computers

Given how many forms bias can take, researchers have increasingly turned to automated methods for detection. A systematic review of the field found that advances in natural language processing are being used to explore automated detection of media bias across its various definitions and operationalizations.

18Social Science Computer Review. Automated Detection of Media Bias Using Artificial Intelligence and Natural Language Processing: A Systematic Review These tools can analyze word choice, sentiment, source selection, and framing patterns at a scale no human reader could match.

Large language models have added new capabilities. Recent work has used them to build datasets of biased news articles and to detect multiple types of bias within individual stories simultaneously.

19arXiv. Decoding News Bias: Multi Bias Detection in News Articles Other approaches draw on social science frameworks to identify bias by word choice and labeling, bridging decades of academic theory with modern computational tools.

20ACL Anthology. Media Bias, the Social Sciences, and NLP: Automating Frame Analyses to Identify Bias by Word Choice and Labeling

The promise of these tools is that they could eventually give readers real-time information about the slant of what they are reading. The challenge is that many forms of bias, particularly selection bias and omission, are about what is absent from a story rather than what is present. Detecting what is missing is a fundamentally harder problem than flagging loaded language in an existing text, and current tools are better at the latter than the former.

What Fact-Checking Can and Cannot Do

Fact-checking organizations have become a major part of the media landscape, positioning themselves as correctives to bias and misinformation. But their effectiveness is not straightforward. An experiment with 859 participants found few differences in effectiveness across different types of fact-checking sources, but did find that sources perceived as more credible were more effective at changing minds. Perceived credibility of the fact-check source mediated how well the correction worked for some source types.

21Communication Research. Checking the Fact-Checkers: The Role of Source Type, Perceived Credibility, and Individual Differences in Fact-Checking Effectiveness

This finding reveals a catch-22. For a fact-check to change someone’s belief, the person has to find the fact-checker credible. But people who already distrust media institutions tend to distrust fact-checkers too, especially when the correction goes against their views. The people who most need fact-checking are often the least receptive to it, which limits how much the practice can do to counteract bias in isolation.

Media Literacy as a Practical Countermeasure

If you cannot eliminate media bias from the supply side, the alternative is to equip audiences to recognize it. Media literacy interventions have shown meaningful effects. A meta-analysis found that these programs generally improve people’s resilience to misinformation, with a moderate overall effect size. The interventions reduced belief in misinformation, improved the ability to distinguish accurate from inaccurate claims, and decreased sharing of false content.

22Communication Research. Media Literacy Interventions Improve Resilience to Misinformation: A Meta-Analytic Investigation of Overall Effect and Moderating Factors

There is an interesting twist. Media literacy training also reduces perception of bias. Researchers found that people exposed to a media literacy presentation were less likely to perceive a story on a controversial issue as biased.

23Newspaper Research Journal. Media Literacy Training Reduces Perception of Bias This could mean that training helps people see past their own hostile media perception, or it could mean that training makes people less critical of media. The interpretation depends on whether the coverage in question was actually biased. Either way, the finding suggests that media literacy changes not just what people believe but how they relate to the news as an institution.

For anyone trying to navigate media bias in practice, a few principles consistently emerge from the research. Consuming news from multiple outlets with different ownership structures exposes you to different selection biases. Paying attention to word choice and image selection helps you catch framing bias. Noticing which sources a story quotes, and which it does not, helps you spot source bias. And being honest about your own political priors helps you distinguish between bias that is in the story and bias that is in your own perception of it. None of these steps eliminates bias, but together they make it considerably harder for any single slant to shape your understanding of the world unchecked.