People often say a news source is “biased.” That word covers several different things, and separating them makes it easier to think clearly. Every account of events involves choices, so no report can be perfectly neutral. The real questions are which choices were made, how fairly, and whether they distort the picture.

Kinds of bias

• Selection bias: which stories are covered at all, and which are ignored.
• Placement: whether a story runs at the top of the page or deep inside.
• Source selection: whose voices are quoted, and whose are not.
• Omission: leaving out facts that would change the reader's understanding.
• Commercial and structural pressures: a preference for conflict, drama and novelty that comes from the need to attract audiences, rather than from political views.

Not all bias is political. A pull toward negative and dramatic stories affects outlets of every political leaning.

Framing

Framing is the way a story is presented: which aspects are emphasized, how the problem is defined, what is suggested as its cause and what solutions are implied. Consider a factory closing. One frame focuses on the workers who lost their jobs; another on the company's struggle with foreign competition; another on the town's future. Each can be accurate, yet each leads readers to think differently about what happened and what should be done.

A useful distinction is between episodic frames, which tell a story through one person or event, and thematic frames, which place it in a broader context of trends and causes. Research suggests episodic stories tend to lead audiences to blame individuals, while thematic stories lead them to look at wider causes.

Loaded language

Word choice carries judgments. A crowd can be described as “demonstrators,” “protesters” or “a mob.” A government can be a “government” or a “regime.” A person can have “said,” “claimed” or “admitted” something. Each choice nudges the reader. Watch for adjectives and verbs that tell you how to feel before you have seen the evidence.

Balance and false balance

Fair reporting usually tries to include relevant perspectives on contested questions. This is balance, and it matters when reasonable people disagree, for example over values or uncertain predictions.

False balance, sometimes called false equivalence, happens when a story gives equal weight to two sides even though the evidence strongly favors one. A classic historical example is smoking. By the 1950s and 1960s, evidence linking cigarettes to lung cancer was mounting, culminating in the US Surgeon General's report of 1964. Tobacco companies funded efforts to stress “doubt” and “controversy.” Coverage that presented the question as an evenly matched debate gave the public a misleading picture of the science for years.

The test is not “did they quote both sides?” but “does the coverage accurately reflect the weight of evidence?” Judging that weight takes care, though. Some questions that seem settled are later reopened by new evidence, and critics warn that calling a live debate settled can shut out legitimate dissent. Good reporting explains what the evidence shows and how confident experts are.

The hostile media effect

Our own perspective shapes how we see bias. In a well-known 1985 study, researchers showed the same television news coverage of a 1982 Middle East conflict to viewers with opposing sympathies. Each group judged the coverage to be biased against its own side. This is the hostile media effect: strong partisans tend to see neutral coverage as slanted against them.

That does not mean bias is never real. It means that feeling a story is unfair is not, by itself, enough to prove it is. A better test is to look for specific evidence: missing facts, unrepresentative sources, loaded words or a headline the story does not support.

Worked example: one event, three stories

An invented city council votes 5 to 4 to raise bus fares by 25 cents to cover a budget shortfall. Here are three possible openings.

Version A: “The city council voted 5 to 4 on Monday to raise bus fares by 25 cents, citing a projected shortfall in the transit budget.”

Version B: “Struggling commuters will pay more to get to work after a divided council pushed through a fare hike Monday, despite protests from riders.”

Version C: “Facing a growing deficit, city leaders made the tough but responsible choice Monday to keep the buses running with a modest fare adjustment.”

All three could be factually accurate. But compare them:

• Framing: Version B frames the story around riders as victims; Version C around leaders solving a problem. Version A stays closest to the basic facts.
• Loaded language: “Pushed through” and “fare hike” suggest something forced on the public; “tough but responsible choice” and “modest fare adjustment” suggest prudent leadership.
• Selection: Version B mentions protests; Version C mentions keeping buses running. Each adds a fact that supports its frame.
• Episodic versus thematic: A follow-up story built around one single mother's commute would be episodic. A story on how transit funding has changed across the region over a decade would be thematic.

A reader who supports the fare increase might call Version B biased; a reader who opposes it might call Version C biased. Both would have specific evidence to point to. Version A would likely draw complaints from both sides, which, as the hostile media effect suggests, is not necessarily a sign that it is unfair.