There is a version of media criticism that never gets anywhere, and it goes like this: the news is biased because it told me what to believe, and I disagree. It rarely survives contact with an actual news story, because most reporting does not tell you what to believe. It reports a flood, a bill, a shooting, a rate cut. Nobody is instructed to feel anything. And yet you walk away from a week of news with a firm sense of which three problems facing the country are the urgent ones — a sense you did not have last month, about issues that did not change last month. That gap has a name and sixty years of research behind it. It is called agenda-setting. * The one-line version comes from political scientist Bernard C. Cohen, writing in The Press and Foreign Policy (Princeton University Press, 1963, p. 13). The press, he wrote, "may not be successful much of the time in telling people what to think, but it is stunningly successful in telling its readers what to think about." That is the whole theory. Coverage does not usually change your opinion on an issue. It changes which issue you have an opinion about — which is frequently the more consequential thing to control. * The study that turned Cohen's aphorism into a measurable finding is small, old, and unusually clean. In 1968, two University of North Carolina journalism professors, Maxwell McCombs and Donald Shaw, ran a survey in Chapel Hill, North Carolina. Between September 18 and October 6 they completed 100 interviews with registered voters drawn from five precincts chosen to be economically, socially, and racially representative — and crucially, they interviewed only people who had not yet decided how to vote, on the logic that the undecided are the most open to campaign information. Each respondent was asked, in effect, what the government should be concentrating on, explicitly regardless of what any politician was saying. * At the same time, the researchers content-analyzed everything those voters were actually reading and watching: the Durham Morning Herald, the Durham Sun, the Raleigh News and Observer, the Raleigh Times, The New York Times, Time, Newsweek, and the NBC and CBS evening news broadcasts, covering September 12 to October 6. Stories were sorted into 15 issue categories and split into "major" and "minor" by how much space, time, and display each got — a front-page lead or a 45-second-plus TV story counted as major. * Then they lined the two lists up. The correlation between the media's major-item emphasis and what voters independently named as the important issues was +.967. For minor items it was +.979. On a scale where 1.0 is a perfect match, the public's private ranking of what mattered was very nearly a photocopy of the press's ranking. "The Agenda-Setting Function of Mass Media" ran in Public Opinion Quarterly in 1972 and became one of the most cited papers in the field; by the theory's own later accounting, more than 400 empirical studies have followed it across dozens of countries. * Two honest caveats, because the paper itself makes them. First, this was a correlation, not proof of causation — a near-perfect match tells you the agendas moved together, not which one moved first. Second, and this is the detail we enjoy: a footnote on the first page notes the study was partially supported by a grant from the National Association of Broadcasters. The founding study of "the media decide what you worry about" was part-funded by the broadcasters' trade association. * The causation question was settled ten years later in a lab. Shanto Iyengar, Mark Peters, and Donald Kinder ran two experiments — published as "Experimental Demonstrations of the 'Not-So-Minimal' Consequences of Television News Programs" in the American Political Science Review, Vol. 76, December 1982, pp. 848–858 — in which participants watched evening newscasts that had been quietly edited to include more coverage of a particular problem. Viewers came out rating that problem as more important than viewers who had not seen the extra segments. Same people, same week, one edit. They also found something sharper, now called priming: the issues a newscast emphasises become the yardstick people use to judge a president's overall performance. Cover the economy heavily and the president gets graded on the economy. * Notice what this means for the usual bias argument. To move an audience, nobody has to write a single slanted sentence. Story selection alone does the work — which events become stories, which stories run long, which get the top of the hour and which get 20 seconds at 6:52. Every one of those choices can be defended individually as pure news judgment, and the aggregate still adds up to a ranking of national problems that the audience will adopt as its own. Agenda-setting is the reason "we only report the facts" is a much weaker defence than it sounds. * Researchers later added a second level to the theory. First-level agenda-setting is which objects get attention — immigration, inflation, crime. Second-level agenda-setting is which attributes of those objects get attention: crime covered as a policing story versus a poverty story, immigration covered as a border story versus a labour story. The second level is where "what to think about" quietly starts leaning on "what to think," and it does it without a single editorial. * The 2026 wrinkle is that the agenda is no longer set solely by nine outlets in one town. Ranking algorithms, aggregators, and AI summaries now sit between the newsroom and you, doing their own round of selection on top of the newsroom's. That does not repeal agenda-setting; it adds a second gatekeeper whose ranking criteria are not published anywhere and cannot be written to. Meanwhile the reverse channel is real too: what trends online now visibly reshapes what gets assigned the next morning. * The practical takeaway is a habit, not an opinion. Once a week, ask what you would name as the country's most important problems, then ask honestly where that ranking came from — whether anything in your own life changed to justify it, or whether the change was in the volume of coverage. Then go looking for the story that was true all month and never became a story: the ongoing one with no fresh event to hang it on. That is not paranoia about the press. It is just knowing that a list arrived in your head, and that somebody else made the list.