Meta-analysis

Definition

A quantitative synthesis of comparable studies, used to estimate an overall effect and its variability.

People say “research shows” when they mean a tweet. A meta-analysis is a specific method. Hofmann et al. (2012) reviewed meta-analyses of CBT: robust effects for many problems in clinical trials, with variation by condition and comparison — and trial populations are not website users.

In simple terms

It is a careful average of many tests, with caveats. It is not “science says you should.”

Scientific explanation

Quality depends on which studies were included, heterogeneity, and publication bias. A review of meta-analyses sits one level further out and inherits those limits.

What research shows

In our library: CBT has substantial trial support across several conditions, with important variation.

What the research does not prove

That you should start a protocol from this page, or that non-CBT approaches are worthless.

What remains uncertain

Moderator effects (who benefits) are often the scientifically interesting part and the easiest to overclaim.

Practical implications

When a claim cites “a meta-analysis,” ask: of what, in whom, versus what, with what limits?

Educational only — not a diagnosis, not a treatment plan, not a crisis service.

Related mechanisms

• study pooling
• effect size
• heterogeneity