Pooled Variance Calculator

Compute pooled variance, pooled standard deviation, standard error, and Cohen's d effect size for two independent samples. Interactive visualization of overlapping distributions.

Group 1 (Sample 1)
Group 2 (Sample 2)
? Psychology: Treatment vs Control (n=30/32, means 15.2/18.6, sds 3.4/3.8)
? Education: Pre-test vs Post-test (independent) (n=35,35; means 68/74; sds 5.2/6.1)
? Clinical: Drug vs Placebo (n=40/38; means 22.5/19.1; sds 4.2/4.5)
? Small sample (n=12/15; means 5.3/7.0; sds 1.2/1.4)
Privacy-first: All calculations run locally in your browser. No data is uploaded or stored.

What is Pooled Variance?

Pooled variance is a weighted average of the variances from two independent samples, assuming the populations share a common variance (homogeneity of variance assumption). It is the cornerstone of the independent two-sample t-test (equal variance version) and is used to calculate the pooled standard deviation, standard error of the mean difference, and Cohen's d effect size. By pooling information, we obtain a more stable estimate of the population variance than using either sample alone.

Pooled Variance Formula (sₚ²):

sₚ² = [ (n₁ - 1) s₁² + (n₂ - 1) s₂² ] / (n₁ + n₂ - 2)

Pooled Standard Deviation: sₚ = √(sₚ²)

Standard Error: SE = sₚ · √(1/n₁ + 1/n₂)

Cohen's d = (x̄₂ - x̄₁) / sₚ

Why Use This Calculator? — Real Applications

  • A/B Testing: Compare conversion rates (continuous metrics) across two landing page variants.
  • Clinical Research: Evaluate treatment efficacy by pooling variances from control and experimental groups.
  • Educational Assessment: Compare test scores from two different instructional methods while respecting variance equality.
  • Meta-Analysis: Aggregate effect sizes (Cohen's d) from independent studies using pooled variances.

Step-by-Step Derivation & Intuition

The formula resembles a weighted average that accounts for sample size and degrees of freedom. Instead of simply averaging variances, we give more weight to larger samples because they provide more reliable estimates. The denominator (n₁ + n₂ - 2) reflects the total degrees of freedom (each variance loses one degree of freedom). The resulting pooled variance is then used to calculate the standard error of the difference between sample means, forming the basis of the t-statistic: t = (x̄₁ - x̄₂) / SE. This calculator also computes Cohen's d, a standardized effect size that expresses the magnitude of the mean difference in pooled standard deviation units: thresholds |0.2| small, |0.5| medium, |0.8| large (Cohen, 1988).

Case Study: Agricultural Yield Comparison

An agronomist tests two fertilizers on crop yield (kg/plot). Group1 (n₁=30, mean=54.2, sd=4.1); Group2 (n₂=28, mean=58.6, sd=5.2). The pooled variance = ((29*16.81)+(27*27.04))/56 = (487.49+730.08)/56 = 21.74, pooled SD = 4.66. Standard error = 4.66 * sqrt(1/30+1/28)=1.23. Cohen's d = (58.6 - 54.2) / 4.66 = 0.94 (large effect). This magnitude shows that Fertilizer B substantially outperforms A. Our interactive graph reveals the overlap in yield distributions and supports the decision.

Assumptions & Cautions

  • Independence: Observations within and between groups must be independent.
  • Normality (robust with moderate n): For moderate sample sizes, the t-test is fairly robust to non-normality.
  • Equality of variances (homoscedasticity): Pooled variance assumes the two populations have equal variance. If violated, consider Welch's t-test.
  • Outliers: Extreme values can distort means and variances — check your data.

Our calculator automatically checks for valid inputs (n>1, positive SDs) and warns about potential issues. Use this tool as part of comprehensive statistical analysis.

Frequently Asked Questions

Use pooled variance when you have reason to believe the two populations have equal variances (e.g., similar spread). If variances are very different, use Welch's t-test (unpooled). Many statistical packages include Levene's test for homogeneity of variance.

Yes, but this calculator accepts summary statistics (means, SDs, sample sizes) directly — ideal when full raw data is not available, typical in meta-analysis or literature review.

Pooled SD represents the common within-group standard deviation, providing a standardized metric across different studies. It makes effect sizes comparable even when group sizes differ.

Conventional benchmarks: small = 0.2, medium = 0.5, large = 0.8. However, interpretation depends on domain: psychology and education often use these; medical studies may consider smaller effects clinically relevant.

Degrees of freedom (df) = n₁ + n₂ - 2. This reflects the combined sample information minus the two estimated means.
References & Further Reading: Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences. Wikipedia: Pooled variance; Statistics How To; Cumming, G. (2012). Understanding The New Statistics.