Agenda-Setting

Story

For six weeks they thought they were friends

Picture this scene. A small island in the middle of the ocean, the year 1914. A handful of English, French and German people live there. No telegraph cable reaches the island; the mail steamer calls once every sixty days. In mid-September the steamer finally appears, and the whole colony gathers on the jetty, because everyone is curious about the outcome of a court case they read about in the newspaper.

The captain brings very different news: war broke out in Europe more than six weeks ago. The English and the French are fighting the Germans. And the islanders have spent all that time treating one another like friends.

This scene is not a historical document; it is how the journalist Walter Lippmann opens his book Public Opinion (1922). The lesson: people act not on the real world but on the picture of the world in their heads. The title of the first chapter says as much: “The World Outside and the Pictures in Our Heads.”

Why is it like this? Because, for Lippmann, the real environment is “too big, too complex, and too fleeting for direct acquaintance”. We approach it with maps. Someone else usually draws the map; and on faraway matters such as war, politics and the economy, that someone else is largely the news media.

This lesson follows a single question: who draws the thick lines on the map, and how far do we trust those thick lines?

Card

One sentence, a hypothesis not yet tested

In 1963 the political scientist Bernard Cohen published his book The Press and Foreign Policy. It was based on interviews conducted in 1953–54 and in 1960 with Washington’s foreign correspondents and foreign-policy officials. According to McCombs’s summary, Cohen had looked qualitatively at how news sources and journalists shape each other’s agendas.

The most famous sentence in the book later became the theory’s slogan. According to Cohen, the press may often not succeed in telling people what to think, but it is “stunningly successful in telling its readers what to think about”.

Note: this sentence was not a finding but an observation. In their 1972 article McCombs and Shaw cite it as the most concise statement of the theory and treat it as a hypothesis. The same article quotes a sentence pointing the same way from Kurt and Gladys Engel Lang: “The mass media force attention to certain issues”.

So by the mid-1960s the idea is in the air: Lippmann had spoken of “the pictures in our heads”, and Cohen and the Langs tied it to the steering of attention. The one thing missing was measurement. In 1968 two young associate professors tried exactly that; you will meet them on the next cards.

Card

Who and when: five names and a timeline

The heroes of the story in brief:

• Walter Lippmann: journalist; the idea of “the pictures in our heads”.
• Bernard Cohen: political scientist; the “what to think about” sentence.
• Maxwell McCombs and Donald Shaw: young associate professors of journalism at the University of North Carolina at Chapel Hill; Shaw joined the faculty in 1966, McCombs in 1967.
• Paul Lazarsfeld: pioneer of voter research; the emblem of the tradition that looked for attitude change.

Timeline:

1. 1922: Lippmann, Public Opinion
2. 1963: Cohen, The Press and Foreign Policy
3. Autumn 1968: the Chapel Hill interviews and the analysis of media content
4. 1972: the Public Opinion Quarterly article; in the same year, the three-wave panel study in Charlotte
5. 1973: Funkhouser’s study of the 1960s
6. 1982: Iyengar, Peters and Kinder’s experiments
7. 1997: the second-level study in the Spanish elections

A small footnote: according to Weaver, Shaw used to tease McCombs about always being the first author; he would tell how a student abroad once introduced him as “Professor Et Al.”, and he even admitted that he might have “embellished” these stories a little.

Card

Core claim: the media’s ranking becomes the ranking in your head

Let us state the theory in one sentence: the more space, airtime and priority the news media give to certain issues, the more people tend to see those issues as important. The order of issues in the news is transferred to the order of issues in our heads.

This sentence has three parts:

1. Media agenda: which issues stand out in the news, and how much?
2. Public agenda: which issues do people find important?
3. The relation: the order in the first determines the order in the second; we can call this the transfer of salience.

McCombs and Shaw’s first rationale is very plain: when readers learn about an issue, they also infer how much importance to give it from the amount and position of the coverage. A newspaper’s front page, the first item of a bulletin, works like an “it is important” label.

A small analogy, not evidence, just to make it stick: if a waiter repeats “today’s special” all evening, you start to think that dish matters more than the others on the menu. The dish hasn’t changed, and you didn’t pick it yourself; only the ranking changed.

Also note that the theory is framed in causal language: if the media’s emphasis rises, the salience of the issue among the public rises too; that is, the media influence the public agenda. Measuring whether this really is so, and under what conditions, is the job of the coming cards and lessons.

Myth check

Myth: “The media tell us what to think”

What people assume: agenda-setting means the media instilling whatever idea they like into our minds.

What the evidence says:

• The core of the theory claims not opinions but an order of importance. The 1972 article says so plainly: the media may have little effect on the direction and intensity of attitudes, but by setting the agenda of the campaign they influence the salience of attitudes.
• But the “only what about” slogan is not the whole story either. Later research showed two things: a prominent issue can change the yardstick by which people judge a leader (priming). And which attributes of an issue the media emphasise can also influence thinking about that issue (second level).
• The effect is not the same for everyone and on every issue: some people and issues are affected a lot, others little. You will see this in the “need for orientation” and “issue type” cards.

In short: agenda-setting is not mind control but attention management. But attention management is not a power to sneer at in politics either.

A small self-observation: count the three issues you found “important” last week. How many of them were carried to you by a news feed, a headline or a notification? The theory does not ask which opinion you should adopt; it asks what you take onto your list of things to think about.

Card

Three agendas: media, public and policy

The theory’s key word is agenda. An agenda is a list of issues (or events) arranged in order of importance at a given moment. Researchers speak of three separate agendas:

• Media agenda: how much space and where issues get in newspapers, bulletins and websites. To measure it, content analysis is done: how many stories, how much airtime, which page, was it the first item? The Chapel Hill study used the distinction between “major” and “minor” items for this.
• Public agenda: which issues people find important. To measure it, a survey question is asked (something like “what do you think is the most important problem facing the country?”).
• Policy agenda: the issues that parliament, government and decision makers give priority to.

According to Rogers and Dearing’s observation, studies generally make one of these three agendas the “variable to be explained” and look at how the others influence it. The three-way distinction became firmly established with Dearing and Rogers’ 1996 book.

This lesson concentrates mainly on the first two, that is, the media and public agendas. How far the media influence the policy agenda depends on conditions: issue type, medium, the form of the story and the political context. More on this in the Agenda-Building and Political Communication lesson.

The common word that ties the three together is salience: how far an issue or attribute stands out on an agenda; this is exactly what is transferred (salience).

Card

Related concepts: first level, second level, priming, gatekeeping

Agenda-setting does not stand alone; four concepts orbit it. For now just know what they are; details come in later lessons.

• First level: which issues are important? The Chapel Hill study is at this level.
• Second level: which attributes of an issue or person stand out? In the 1995 Spanish elections, newspapers, television news and political ads influenced voters’ images of the candidates. In another study, the more negative news came out about a country, the more negatively people thought about that country; positive news showed no effect.
• Priming: an area that stands out in the media becomes the yardstick people use to judge a leader. When the Iran–Contra scandal broke in 1986, Central America policy became more heavily weighted in evaluating President Reagan’s performance.
• Gatekeeping: how the media agenda is formed: which story passes through the gate and gets on air? You may remember this from the Gatekeeping lesson. A study examining newspaper coverage of fifty bills in Congress found that in gatekeeping, routine-level forces (for example editors’ news-value judgments) mattered more than the individual characteristics of the writers.

Scheufele and Tewksbury compare framing, agenda-setting and priming as neighbouring but separate models of media effects. For depth: Priming and Second-Level Agenda-Setting and Framing.

Card

The theory’s silent assumptions

Every theory relies on a few things it does not say out loud. Agenda-setting’s four silent assumptions are these:

1. The media are not a mirror of reality. Lippmann’s idea of “the picture in our heads” rests on this. Funkhouser’s study of the 1960s reached a similar point: while press coverage and public interest moved together, neither was in line with real-world indicators; for Vietnam, campus unrest and urban riots, both press coverage and public interest had peaked a year or two before the actual conditions themselves.
2. On distant issues, people have few other sources. The Chapel Hill authors argue that very few people see the presidential candidates in person, and that information arriving through interpersonal talk is also largely passed on from the media.
3. Emphasis is a cue. People infer the importance of an issue from the amount and position of the coverage.
4. Attention is limited. A person can hold only a few issues at the top at the same time; McCombs puts this limit at around five to seven issues and refers to Miller’s “seven plus or minus two”.

One last warning: these assumptions do not hold “always and for everyone”. The answer to the question of who, when and on which issue comes in the next section, on the “need for orientation” card.

Step by step

How it works: the agenda’s five-step journey

Let us follow, step by step, an issue’s journey from the media agenda to your agenda.

1. Material arrives. Events, sources, news agencies and even other media enter the newsroom. The media agenda is fed not only by “the world” but also by one another and by political advertising: in the 1990 Texas gubernatorial election, the newspaper agenda influenced the television agenda, and the political-ad agenda influenced both the television news and the newspaper.
2. Editors select and rank. Which story gets in (gatekeeping), which one goes first, how many seconds or column inches does it get?
3. The audience reads the cues. The headline, the first item, the length, the repetition. These work like an “it is important” signal.
4. The public agenda shifts. The order of issues in your head moves closer to the order of the media you follow. This is what was found in Chapel Hill.
5. Consequences appear. The prominent issue becomes a yardstick for judging leaders (priming); politicians sometimes take the issue onto their own agendas too, but this depends on conditions.

A warning: every step is conditional. If one link in the chain is weak (for example, you never saw that story, or you are living the issue first-hand), the following steps may not happen. You will see these conditions on the next card and in the deep cards.

Lab

Experiment: whose agenda shifts the most? Need for orientation

One thing was clear from the start: agenda-setting does not affect everyone equally. To predict who would be affected more, McCombs and his student David Weaver developed the concept of need for orientation: the combination of personal relevance and uncertainty.

The logic is simple: if an issue matters to you and you are not sure what to think, your need for a “map” grows; you turn back to the news media and adopt its ranking. Weaver measured this in campaign data from Syracuse as the combination of political interest and uncertainty. In the 1972 Charlotte panel, the agendas of those who were very interested in the election and very uncertain about whom to vote for were the most similar to the agendas of the Charlotte Observer newspaper and network television news.

The levels are summarised like this: if interest is low, need for orientation is low; if interest is high but uncertainty is low, it is moderate; if both are high, it is high. A later panel study showed that need for orientation increased the perceived salience of media emphasis. In the Turkish literature the term appears as “yönelim gereksinimi”.

Now you try (three imaginary people). A local election is coming up:

• Ayşe: She is very curious about the election, but has not been able to decide which candidate to choose.
• Berk: He does not care about the election; he does not even open the news.
• Ceren: She cares a great deal about the election and has long since decided whom to vote for.

Whose agenda is most like the news agenda? You will test your guess in the question below.

Story

Autumn 1968, Chapel Hill: a hundred interviews, nine media

September 1968. The US presidential election is weeks away, with Nixon, Humphrey and Wallace on the stage. In the town of Chapel Hill, North Carolina, two young associate professors of journalism want to test a question: do the issues voters regard as “the real matter” overlap with the issues the news pushes forward?

The design was as follows:

• Sample: people picked at random from the registered-voter lists in five of the town’s precincts; the precincts were chosen to represent the town economically, socially and racially. Between 18 September and 6 October, 100 interviews were completed.
• Who? A filter question separated those who had not yet firmly decided whom to vote for; that is, those assumed to be the most open to campaign information.
• Question: Roughly this: “What are you most worried about these days? Whatever the politicians say, what are the two or three main things the government should be dealing with?” The emphasis is on “whatever the politicians say”: so that the answers would be independent of the candidates’ slogans.
• Media: a preliminary study in spring 1968 showed that almost all of Chapel Hill’s political news knowledge came from nine sources: four newspapers from Durham and Raleigh (Durham Morning Herald, Durham Sun, Raleigh News and Observer, Raleigh Times), the New York Times, Time, Newsweek, and the NBC and CBS evening news. Their content between 12 September and 6 October was coded.

Why a single town? The authors say this would keep regional differences and differences in media performance under control.

Card

Results: 0.967 and 0.979, but the really interesting part is the selective-perception test

At first glance the results are striking:

• The correlation between the “major” news emphasis the media gave to the main campaign issues and the importance voters gave to those issues was +0.967.
• For “minor” news the correlation was +0.979.

The closer a correlation gets to 1, the more the two rankings resemble each other. But note: these numbers were computed over a handful of main issue categories and at the aggregate level.

The authors’ real rigour lies elsewhere: they tested the possibility of selective perception. The idea: if voters who lean towards a candidate but have not yet decided (45 people) look only at the news about their own candidate, their ranking should resemble the ranking of that news most. Had that been so, the result would have favoured selective perception rather than agenda-setting. The finding came out the other way round: in 18 of 24 comparisons, voters agreed better not merely with the news about their own candidates but with the composite of all the news. Although the three candidates’ issue emphasis differed greatly, voters’ rankings resembled the composite media ranking.

The authors also wrote one thing themselves: these correlations do not prove agenda-setting; they present a picture consistent with the conditions that should exist if agenda-setting were taking place. This subtlety will be useful in the exam and on the critique card.

Card

Afterwards: replications, cracks and the reality check

After 1972 the research snowballed. Three key links:

Funkhouser (1973): where is reality? He examined the issues of the 1960s. Press coverage and public opinion moved together; but neither was in line with real-world indicators: for Vietnam, campus unrest and urban riots, both press and public interest had peaked a year or two before the actual conditions themselves. The abstract of his second article points the same way: issues that filled the headlines in a period did not necessarily deserve the attention they were getting at the time.

Charlotte 1972: watching over time. In the panel study that followed the same voters over three waves, McCombs found an effect of newspapers on voters’ agendas over time and found no such effect for television; but towards October the television agenda matched voters’ agendas better. His reading: newspapers lead, and towards the end share the stage with television.

McLeod, Becker and Byrnes (1974): looking at the individual. They compared the readers of two newspapers whose content emphasis differed in the 1972 campaign. The result was only moderate support: although “honesty in government” was emphasised in one newspaper, it did not excite the readers of either paper much; the differences appeared mostly among partisans with low interest and motivation who depended on the paper for news.

The total message: agenda-setting lives, but conditionally.

Card

The question of direction: is the media pulling the public, or the public the media?

A correlation does not tell direction: are the media pulling the public, or are newspapers following the reader’s interest? Researchers tried to untie this knot in three ways.

1) Looking at time. The cause must come before the effect. Winter and Eyal found that on civil rights, public interest followed newspaper coverage; the best effect span was four to six weeks.

2) Taking the real world into account. Behr and Iyengar, with long-run data on energy, inflation and unemployment, showed this: television news influences public concern and not the other way round; items presented as the lead story are markedly stronger than ordinary items; and real-world conditions influence public opinion both directly and indirectly, by shaping news coverage.

3) Running an experiment. An experiment is the design that can separate cause from effect: the researcher decides for themselves who is exposed to which news. Iyengar, Peters and Kinder’s two experiments showed that television news profoundly affects which problems viewers find important, and that problems emphasised in the evening news become more heavily weighted in evaluating the president’s performance. The authors present this as a confirmation of Lippmann’s doubt of half a century earlier.

Not every result came out one-directional: for blogs and mainstream media, as we will see later, the relation was two-directional.

Card

A fair critique: what did Chapel Hill not prove?

Agenda-setting is one of the most researched theories; but a good theory knows its limits too. Let us put the criticisms and the defence side by side.

Criticism 1 — Strength of evidence. Chapel Hill was a sample of a hundred people, from a single town, made up of undecided voters; the relations were computed over a handful of main issue categories and at the aggregate level (media total against voter total). The authors said so themselves: comparing aggregate units was enough for a first test, but later research had to move from the societal level to the individual level.

Criticism 2 — The effect may be weak at the individual level. When McLeod and colleagues looked at individuals they found only moderate support; the effect was concentrated among readers with low interest and motivation.

Criticism 3 — Issue type. Zucker’s hypothesis that “the effect is greater on issues not directly experienced” was attractive; but Demers and colleagues could not support this “obtrusiveness” hypothesis.

The defence. The authors answered the objection that the relationship is “spurious” like this: the voter has very limited ways of following the agenda independently of the media, because few people see the candidates in person and interpersonal information is largely passed on from the media. Moreover, although sharp differences between journalists’ news values and readers’ interests are known, a “perfect fit” turned up, which points to the media coming first. Since five of the nine media are national and the other four are not based in Chapel Hill, it is easier to argue that voters fitted their judgment to the media than the reverse.

Lab

Thought experiment: inflation in the headlines and in your head

Imagine: in a fictional country prices are rising fast. Newspapers carry inflation to the headlines; in public opinion polls inflation comes first as “the most important problem”. The two lists look very much alike. Is this agenda-setting?

There are three candidate explanations:

1. The headlines influenced the public. (Agenda-setting.)
2. The public lives the prices, and the newspapers reflect the public’s troubles. (Reverse causation.)
3. Real prices influenced both. (A third variable: real-world cues.)

Now the hard part: can you design data that separates these three explanations? A few hints:

• Time order: did the public’s worry rise after the headlines, or before?
• Control: what would happen if prices stayed the same and only coverage changed (or the reverse)?
• Experiment: what if we had people with the same experience of prices watch different news?

The answer of real studies points this way too: in Behr and Iyengar’s data television news influenced concern, but real conditions also had an effect, both directly and by shaping news coverage. So the first and the third explanation can be true at the same time; what is wrong is to say “only one”.

Card

The agenda in the social media age: the arrow is no longer one-way

In 1968 the agenda was a few newspapers and two television channels. Today everyone is both audience and broadcaster. Has the theory collapsed? Documented research shows not a collapse but a network picture.

• Neuman and colleagues (2014): in 2012 they tracked 29 political issues across traditional and social media with big data. Agenda-setting was not a one-way pattern from traditional media to the masses but a complex and dynamic interaction; the attention rhythms of the two environments were related but moved to significantly different “drummers”.
• Wallsten (2007): in the 2004 US campaign he tracked 35 issues in 10 popular and 50 less-known blogs; on most issues he found not a one-way but a two-way relation between mainstream media and blog agendas.
• Conway, Kenski and Wang (2015): in the 2012 primaries they saw a “symbiotic” relation between the candidates’ Twitter accounts, the parties and the major newspapers; the lags varied by issue: on some issues traditional media followed the candidates, on others Twitter anticipated the agenda.
• Barberá and colleagues (2019): they examined legislators’ and the public’s Twitter issues in the 113th US Congress: legislators follow the public’s agenda rather than steering it, and are more responsive to their own supporters than to the general public.

The common message: the media agenda is still important; but public, politician and journalist feed one another. The answer to “who sets the agenda?” is often “all of them, in turn”.

Story

Whose agenda is the trending list? Twitter in 2009

When people say “agenda” on social media, “trending topics” come to mind. Is this list the public’s own voice, or an echo of the news agenda?

In 2010 Kwak and colleagues published a study for which they crawled the whole of Twitter: 41.7 million user profiles, 4,262 trending topics and 106 million tweets. They classified the trending topics by the period in which they were active and by tweet pattern, and found that more than 85% of the topics were “headline news” or “persistent news”. Headline news: breaking developments. Persistent news: topics that stay constantly on the agenda, such as sports teams, cities and brands. The authors read this as users mostly talking about headline news and reacting to fresh news.

So even on a platform called “everyone’s voice”, the most visible agenda was largely the news agenda. But this is a snapshot of 2009; it cannot be carried directly over to today’s personalised feeds.

Two documents from later years:

• In 2015, among US online news sources, partisan online media played a decisive role; the New York Times and the Washington Post no longer controlled the agenda, they were following partisan media.
• In the 2016 election, fake news sites established a close agenda relationship with fact-based media in Trump coverage, but not in Clinton coverage; according to the authors, rather than an original agenda-setting role, they added noise to an environment that was already sensational.

What about algorithms? In personalised feeds, not everyone sees the same agenda; how this affects the theory is an open question. In 2025 a journal took this question to a special issue on artificial intelligence and chatbots; the results are outside the scope of this lesson. (This last paragraph is reasoning, not a finding.)

Card

Two studies from Turkey: when the agenda is put to the test

Agenda-setting has been tested in Turkey too. Here are two studies whose full texts we read; both are instructive, along with their limits.

1) There is no single media agenda: the coalition talks after 7 June 2015. And Algül examined, by content analysis over a 57-day period, the coalition-talks stories on the front pages of six newspapers (Sabah, Hürriyet, Cumhuriyet, Zaman, Yeni Şafak, Sözcü, all national dailies in Turkey) between 6 July and 31 August 2015, after the 7 June 2015 general election. The newspapers’ agendas differed from one another; the headlines contained commentary; the placement of stories, headline approaches, the actors used and the way the parties were treated varied with editorial policy. In the lesson’s language: there is no single list called “the media agenda”; it can diverge by outlet.

2) The public agenda and the perceived media agenda. Hazar and Sbouh (2020) surveyed 740 communication-faculty students in all, 386 from Ankara Hacı Bayram Veli University and 354 from An-Najah University in Palestine. They compared the students’ own agendas with the issues they thought “the media put most emphasis on”, and wrote that the similarity was low, that is, the theory was not sufficiently effective in this sample. When the same survey asked where they got current information, 75% said social media; television came to 10%, social circle 8.9%, newspapers and magazines 5% and radio 1.1%.

Note: in the second study the media agenda was measured not by content analysis but by the students’ perception; the authors themselves point out that this departs from the classic approach. The sample is only communication students. So this does not mean “the theory has been refuted”, but “no overlap came out in this design”.

Card

How it is asked in the exam: patterns, traps, confused pairs

In university exam-style questions, agenda-setting mostly comes in these five patterns.

1) “What is the theory’s basic claim?” Correct: the media influence what people think about. Trap: “determine what they think”.

2) “Who, when?” McCombs and Shaw: the 1968 election (data), 1972 (article), Chapel Hill. Cohen: 1963. Lippmann: 1922. Traps: mixing up 1968 and 1972; counting Lippmann or Cohen as the founder of the theory (both are the ground of ideas); counting Lazarsfeld as the theory’s precursor (his name stands for the limited-effects tradition).

3) “The Chapel Hill sample?” 100 undecided voters. Traps: “1000 voters”, “decided voters”, “New York”.

4) “Three agendas?” Media, public and policy agenda.

5) “Which is which theory?” Agenda-setting is the order of issues. Framing is how an issue is presented (Framing). Priming is the prominent issue becoming the yardstick of evaluation. Cultivation is long-term television viewing shaping one’s worldview (Cultivation Theory). The spiral of silence is the fear of being left in the minority (The Spiral of Silence).

Names. In the Turkish literature the same concept has been called not only “gündem belirleme” but also names such as “gündem oluşturma”, “gündem kurma”, “gündem hazırlama” and “gündem koyma”; one article lists them with their sources. This topic is in the ninth unit of Polat’s (2016) course book, and agenda-building and political communication are also counted among that unit’s topics. In the exam, write the name the book uses.

Summary

Summary card

• Claim: the more the media emphasise an issue, the more the public tends to see it as important: not what to think, but what to think about.
• Roots: Lippmann (1922), “the pictures in our heads”; Cohen (1963), the “what to think about” sentence.
• Study: McCombs and Shaw, Chapel Hill 1968; 100 undecided voters, nine media; correlations +0.967 (major news) and +0.979 (minor news); in the selective-perception test, 18 of 24 comparisons pointed towards agenda-setting. The authors did not say “we proved it.”
• Three agendas: media, public, policy.
• Conditions: need for orientation = relevance + uncertainty; issue type is disputed; the time lag varies from study to study.
• Tests: Funkhouser’s mismatch with real-world indicators; the Charlotte panel; Behr–Iyengar; the Iyengar–Peters–Kinder experiments.
• Today: network-like, two-way; in the 2009 Twitter data more than 85% of trending topics were news items.
• Turkey: the media agenda diverges by outlet; in a student sample the perceived media agenda and the public agenda overlapped little.
• Exam: what about ≠ what; 1968 data, 1972 article; 100 undecided; three agendas; the distinction between framing, priming and cultivation.