How We Read Studies
Nutrition headlines routinely overclaim. This page explains how our team actually reads and weighs nutrition research.
Study types matter enormously
Randomised controlled trials (RCTs) are the closest thing to a gold standard for causality, but even RCTs vary wildly in quality. Observational studies are useful for spotting associations, but they cannot prove causation. Systematic reviews and meta-analyses sit near the top of the evidence pyramid when well conducted.
What we look for first
- Funding and conflicts: Industry-funded studies can skew results.
- Sample and setting: Is the study in humans?
- Effect size: A statistically significant result may not translate into a meaningful health difference.
- Replication: Has the finding been reproduced independently?
The null result problem
Studies showing no effect are less likely to be published, which can mislead.
What we will not do
- Present animal or in-vitro findings as directly applicable to human nutrition.
- Upgrade associations to causes.
- Omit uncertainty to make an article more shareable.
