• Communications: research methods in communication

What is content analysis?

Written with AI (How we create content)Target level: Grade 12Last updated: Report an error
Quick answer

Content analysis is a research method that counts and classifies what media texts say, such as headlines, news items, advertisements or posts, by clear rules, so that another researcher using the same rules would reach the same counts. Bernard Berelson defined it in 1952 as the objective, systematic and quantitative description of the manifest content of communication.

The topic explained

Counting what the media say

Content analysis studies the messages themselves, not their audiences. Instead of guessing that the news shows young people mostly as a problem, a researcher counts: of all the items about young people in a sample, how many are about crime, how many about education, how many about sport? In his book of 1952 Bernard Berelson defined content analysis as the objective, systematic and quantitative description of the manifest content of communication. Objective means that the rules, not the researcher's mood, decide; systematic means that every item in the sample is treated the same way; quantitative means that the results are counts and percentages.

Manifest content is what is plainly there, a word, a topic, a picture. Latent content is the hidden meaning, such as whether a story is sympathetic. Critics of Berelson's definition argue that latent meanings can also be analyzed if the rules are clear; this is harder and needs careful definitions. Content analysis answered Lasswell's question “says what?”, and it supplied the counts of television violence in George Gerbner's cultivation research.

The steps

A study starts with a research question, for example: how does a made-up news site present young people? Next comes a sample: which texts, from which dates, how many. Then the unit of analysis, the thing being counted, such as a headline, an article or a photograph, and the categories it is sorted into, written in a codebook with a definition and examples for each. The categories must be mutually exclusive, so that each unit fits only one, and exhaustive, so that every unit fits somewhere, which is why most systems have an “other” category.

Finally the units are coded, usually by two people working separately, and the results are counted and turned into percentages. The common mistake is to start counting before the categories are defined: the counts then change with the coder's judgment.

The five steps of content analysis from top to bottom: a research question; a sample of texts; a unit of analysis and categories; coding by two coders; counting, percentages and interpretation.A research questionA sample: which textsUnit of analysis, categoriesCoding by two codersCount, percent, interpret
The steps of a quantitative content analysis.

A worked example

Suppose a researcher codes 50 made-up headlines about young people from a made-up news site. The counts are crime 18, education 12, sport 10, culture 6 and other 4, which add up to 50. Each count divided by 50 and multiplied by 100 gives the percentages: crime 18 ÷ 50 × 100 = 36%, education 24%, sport 20%, culture 12% and other 8%, which add up to 100%. The finding is that more than a third of the headlines link young people with crime.

To check the coding, a second coder codes the same 50 headlines without seeing the first coder's work. If they agree on 45 of them, the intercoder agreement is 45 ÷ 50 × 100 = 90%. Disagreements, such as a youth band at a festival coded culture by one coder and other by the second, show where the codebook needs a clearer definition.

Bar chart of 50 made-up headlines about young people by topic: Crime 36%, Education 24%, Sport 20%, Culture 12%, Other 8%.Crime36%Education24%Sport20%Culture12%Other8%
Made-up data: 50 headlines from a made-up news site.

What the numbers can and cannot show

Content analysis shows what is in the media, reliably and on a large scale. It cannot show by itself why the content is so or what it does to audiences: 36% crime headlines do not prove that readers fear young people, and the sample may not represent other outlets. A strong exam answer defines the method with Berelson's terms, describes the steps, calculates percentages and agreement, and states the limits.

Coding table of three made-up headlines: teen held after a fight, both coders coded crime; pupils win a contest, both education; a youth band at a festival, coder A culture and coder B other, a disagreement.Made-up headlineCoder ACoder BTeen held after fightCrimeCrimePupils win contestEducationEducationYouth band at festivalCultureOther
The last row is a disagreement to be settled in the codebook.
Free printable worksheet

Practice: content analysis

For 12th grade communications students. Every headline, site and number here is made up. Answer from what is printed on the page, then check yourself against the answers.

Free, no sign-up. The answer key prints on the second page. Regenerating a new worksheet on the same topic needs a free account.

Free printable worksheetlessoons.com
Practice: content analysisName:
  1. Which step of content analysis is missing in the diagram, and what does it include?

    The five steps of content analysis from top to bottom: a research question; a sample of texts; a unit of analysis and categories; coding by two coders; counting, percentages and interpretation.A research questionA sample: which texts?Coding by two codersCount, percent, interpret
    The steps of a quantitative content analysis.
  2. In the chart, what share of the 50 headlines are about crime? Show how it is calculated.

    Bar chart of 50 made-up headlines about young people by topic: Crime 36%, Education 24%, Sport 20%, Culture 12%, Other 8%.Crime36%Education24%Sport20%Culture12%Other8%
    Made-up data: 50 headlines from a made-up news site.
  3. A second made-up sample has 40 items, of which 10 are about crime. What percentage is that? Is it more or less than in the first sample?

  4. Two coders agree on 45 of 50 headlines. What is their intercoder agreement, and why is it calculated?

  5. In the coding table, which headline do the coders disagree on, and how should the codebook fix it?

    Coding table of three made-up headlines: teen held after a fight, both coders coded crime; pupils win a contest, both education; a youth band at a festival, coder A culture and coder B other, a disagreement.Made-up headlineCoder ACoder BTeen held after fightCrimeCrimePupils win contestEducationEducationYouth band at festivalCultureOther
    The last row is a disagreement to be settled in the codebook.
  6. What is the difference between manifest and latent content? Give an example of each from a news photo.

  7. A student writes: “36% of the headlines are about crime, so readers of the site are afraid of young people.” What is wrong with this conclusion?

  8. Plan a small content analysis of made-up advertisements for toys: write a research question, a unit of analysis and four categories.

More topics and worksheets at lessoons.com© 2026 Lessons

Common questions

What is a codebook in content analysis?

The document that defines the unit of analysis and every category, with clear rules and examples, so that every coder sorts the material the same way. It is what makes the method objective and repeatable.

What is intercoder reliability?

The degree to which two or more coders, working separately, put the same units into the same categories. A simple measure is the percentage of units they agree on; researchers also use statistics that correct for chance agreement.

Is content analysis quantitative or qualitative?

Classic content analysis, as Berelson defined it, is quantitative: it counts. There is also qualitative content analysis, which interprets themes and meanings in depth, usually on smaller samples.

What can content analysis not show?

It cannot show by itself why the content looks as it does or how it affects audiences. Those questions need other methods, such as interviews with journalists or surveys and experiments with audiences.

What is a unit of analysis?

The single thing that is counted and coded, such as a headline, an article, a television scene, an advertisement or a post. Choosing it is one of the first decisions in a study, because it changes what the counts mean.

Still have a question? Ask Lessi

Any question about this subject. Lessi answers right here, step by step.

Try asking

Asking needs a free account; your question is kept while you sign up. Answers are written by AI and can contain mistakes. Lessi doesn’t give medical or legal advice.

Learn it with a teacher

Learn this topic live with Lessi

Lessi, the AI teacher, explains it out loud on a board, with short quizzes, at your pace. The first 60 minutes are free.