Health stories tend to move in waves: a single study lands, headlines amplify it, and a week later a different study seems to contradict the first. In reality, individual studies almost never settle questions on their own. They contribute evidence that researchers integrate over years.
You do not need to read studies the way a researcher does to avoid being misled. A short checklist — what kind of study, how many people, who funded it, and what the authors actually claim — is enough to navigate most health news responsibly. This article walks through that checklist.
Know what kind of study you are reading
Different study designs answer different questions. A randomized controlled trial (RCT) can test whether an intervention causes an effect, because participants are randomly assigned to groups. An observational study can identify associations but cannot, by itself, establish cause. A case report is a single patient's experience and is the weakest form of evidence for general claims.
Reviews and meta-analyses pool many studies together. When done carefully, they are the strongest evidence on a topic, because they reflect the weight of many independent results rather than one team's findings.
Sample size and population
A trial of 30 people can produce a striking result that does not survive replication in larger samples. Look for the number of participants, and look at who they were: a study in 50 elite athletes may not generalize to a 60-year-old with chronic conditions. Reputable health writing names the sample and the population in the first few paragraphs.
Sample size alone is not enough. A well-designed small RCT can outweigh a sprawling but poorly controlled observational study. The point is to ask, not to memorize a threshold.
Effect size versus statistical significance
An effect can be statistically significant — meaning it is unlikely to be due to chance — and still be too small to matter in everyday life. A 2% relative reduction in a rare outcome, even if 'significant,' may not be a reason to change behavior. Good health journalism reports the absolute effect, not only the relative one.
Ask yourself: out of 1,000 people similar to me, how many would benefit, and by how much? If the article does not let you answer that question, look for one that does.
Statistical significance tells you a result is unlikely by chance. It does not tell you the result matters.
Funding and conflict of interest
Industry funding does not automatically discredit a study, but it is a relevant signal, especially when an investigator's career or company depends on a particular outcome. Reputable journals require disclosures, and reputable journalism reproduces them.
When a study is press-released by the company that paid for it, treat the press release as a marketing document. The study itself, with its disclosures, methods, and limitations sections, is the real document.