Letter to the editor: Meta-analysis cannot count the same study six times: Why independence of studies is the first principle of evidence synthesis

Abstract

The foundation of any meta-analysis is that each included study contributes an independent piece of evidence [1–4]. Without this independence, the statistical machinery of pooling breaks down: confidence intervals become artificially narrow, and the illusion of precision replaces genuine inference. This occurs because meta-analytic models assume that each study provides a distinct estimate, with its own sampling error. If the same study is entered more than once, its estimate is no longer independent but duplicated, artificially increasing its weight in the pooled result.

Keywords

Hashtags

This article is published on an external journal. Click below to read the full text.