How to read addiction research

Systematic reviews and meta-analyses

Combine many studies on one question using a pre-set method. Cochrane reviews, such as the review of AA and twelve-step programs, are an example.¹

Randomized controlled trials

Randomly assign people to a treatment or comparison, which helps show cause and effect.

Observational studies

Follow people without assigning treatment. Useful, but they can show links that aren't causes.

Laboratory and imaging studies

Reveal mechanisms — such as how near misses affect the brain² — usually in small groups and artificial settings.

If two things occur together — say, gaming and depression — one could cause the other, both could share a cause, or the link could be coincidence. Good research tries to separate these. Headlines often don't.

Consensus

Broad agreement among major health bodies and research reviews.

Strong

Supported by multiple high-quality trials or reviews.

Moderate

Good evidence with some limitations or gaps.

Emerging

Promising early findings that need confirmation.

Debated

Experts disagree or evidence conflicts.