How we grade evidence
Every research-based post shows the strongest kind of evidence it relies on. Here is what each grade means.
The grades
Stronger evidence means we can be more confident in a finding. Weaker evidence is not wrong; it is less certain.
Meta-analysis
Meta-analysis
Pools the results of several studies, usually randomized trials.
Randomized controlled trial
Randomized controlled trial
People are randomly assigned to a treatment or a comparison, which limits bias.
Clinical study
Clinical study
Tests a treatment in patients, but without full randomization or control.
Observational study
Observational study
Looks for links in existing data. Shows association, not cause.
Expert opinion
Expert opinion
Based on clinical experience and guidelines rather than new data.
Anecdotal
Anecdotal
Individual reports or common practice, not formal research.
How a post gets its grade
The grade is the strongest design that supports a claim in the post. A post that cites one small pilot study is graded by that study, however confidently the claim is written.
We also say what the evidence does not show: sample sizes, who was studied, and what was not measured. A "what we still don't know" section appears where the gaps matter.
What the grade is not
The evidence grade is about research, not about products. A testing score on a review grades a product we handled and is shown in a different style, so the two are never confused.