Skip to content

Effective sample size vs Sample size

Terms
Effective Sample Size Sample Statistical Power
Definition

Sample size is how many individuals (or observations) a study actually includes. Effective sample size is a derived quantity that expresses how much *independent* information those individuals provide after accounting for design features such as case–control imbalance, overlap across meta-analytic cohorts, or relatedness—often smaller than the raw head count.

Topics
Probability and inference GWAS

How they differ

Sample size (N) Effective sample size (N_eff)
What it counts Observed participants or rows in the analysis. Information-equivalent size under an idealized independent-samples design (definition varies by context).
Typical GWAS use Cohort description, power intuition, meta-analysis totals. Meta-analysis quality metrics, LD score regression “effective N,” case–control N_eff when combining studies with different case fractions.
When they diverge Still reported as nominal N. Duplicate samples, cryptic relatedness, extreme case–control imbalance, or weighting in meta-analysis reduce effective information per person.

Rule of thumb: Compare studies using effective or harmonized N when the goal is detectable signal or variance of summary statistics; use nominal N for describing who was measured and for consent/traceability.

References

  • Winkler TW, et al. (2014). Quality control and conduct of genome-wide association meta-analyses. Nat Protoc.