Observational study: ask what researchers watched and measured

When reading a report, start by locating what the investigators actually did. Did they observe an existing situation, or assign a change as part of the study? Write down the answer from the methods rather than guessing from the headline. This entry uses a nonmedical reading example; it does not explain any individual person's health.

Counting visitors, without rearranging the room

In an invented community-hall report, volunteers record the number of visitors and occupied chairs on five Saturday mornings. They leave the room layout as it is. A newsletter later says, “More chairs bring more visitors.”

Nico, who is helping organize a family craft day, checks the report. It records what happened on those mornings. It does not describe the volunteers assigning extra chairs to some mornings and withholding them on others. Nico puts the newsletter's sentence beside the methods and circles the word “bring.”

Nico's sister notices that the busiest morning also included a plant swap. Their father wonders whether extra chairs were put out because organizers expected more visitors. They record these as questions, not as competing explanations they have proved. The report may contain useful counts while leaving the newsletter's stronger wording unestablished.

Make the observation concrete

For another article, try finding answers to a few plain questions:

• Who or what was included, and who chose to take part?
• What did the researchers record: a direct count, a questionnaire answer, a record, or something else?
• When did they collect the information?
• What, if anything, did they assign or change?
• Which explanation does the headline add beyond the recorded comparison?

It is fine to stop after one important discrepancy. A family planning an event does not need to produce a complete scientific appraisal just to avoid repeating an unsupported headline.

Different questions belong to nearby terms

For the hall example, observational study directs attention to how the information was collected. Correlation directs attention to the reported relationship between measurements. Causation asks whether the chairs made a difference to attendance. Confounding invites a closer look at a possible shared explanation, such as the event being advertised. Keep these questions distinct instead of letting one label answer all four.

The randomized-controlled-trial entry offers a different reading question: where does the report describe assignment by chance? Do not add that feature to the hall story just because the numbers were entered into a spreadsheet.

Leave room for different household interests

A child may care whether there will be space for a craft project, while an adult is checking the seating claim. A sibling need not gather visitor counts to earn a say in the room layout. Partners can agree to discuss the actual floor space and leave the larger research question open.

Use the report for the information you can identify in it. If an important passage is inaccessible or unclear, name that limit. This page's invented questions are not a research protocol or a promise that every uncertainty can be resolved at home.

Source

[1] National Cancer Institute. Observational study. NCI Dictionary of Cancer Terms. Definition paragraph; publication/update date unknown; accessed September 5, 2026. Supports [1]; clinical terminology only, with no study result adopted.

Related guides

• Randomized controlled trial
• Correlation
• Causation
• Confounding