Research Gold
ServicesPricingHow It WorksSamplesAboutFAQ
LoginGet Started
Research Gold

Professional evidence synthesis support for researchers, clinicians, and academic institutions worldwide.

6801 Gaylord Pkwy
Frisco, TX 75034, USA

Company

  • About
  • Blog
  • Careers

Services

  • Systematic Review
  • Scoping Review
  • Meta-Analysis
  • Pricing

Resources

  • PRISMA Guide
  • Samples
  • FAQ
  • How It Works

Legal

  • Privacy Policy
  • Terms of Service
  • NDA & Confidentiality

© 2026 Research Gold. All rights reserved.

PrivacyTerms
All Resources

Heterogeneity & Sample Size Calculator

Free

Assess heterogeneity in your meta-analysis with I², Cochran's Q, τ², and prediction intervals. Calculate the minimum number of studies needed to detect a given effect size.

How to Use

Heterogeneity tab: Enter the effect size and standard error for each study in your meta-analysis. The tool calculates Cochran's Q, I², τ², and a 95% prediction interval using the DerSimonian-Laird method.
Power tab: Enter your expected effect size, significance level, and desired power to estimate the minimum number of studies needed.

Study Data

#Effect Size (y)Standard Error (SE)
1
2
3
Enter effect sizes and SEs for at least 2 studies.

Interpreting Heterogeneity

I² = 0–40%

Low heterogeneity. Studies are reasonably consistent. A fixed-effect model may be appropriate.

I² = 40–75%

Moderate heterogeneity. Consider exploring sources via subgroup or meta-regression analyses.

I² = 75–100%

High heterogeneity. Results should be interpreted with caution. Investigate sources before pooling.

Frequently Asked Questions

What is I² (I-squared)?

I² describes the percentage of variability in effect estimates that is due to heterogeneity rather than sampling error. An I² of 0% means all variability is due to chance; 100% means all variability reflects true differences between studies. It was proposed by Higgins & Thompson (2002).

What is τ² (tau-squared)?

τ² is the between-study variance in a random-effects meta-analysis. Unlike I², τ² is on the scale of the effect size, making it useful for calculating prediction intervals. It's estimated using methods like DerSimonian-Laird or REML.

When should I use a random-effects model?

Use random-effects when you expect genuine variation between studies (different populations, interventions, settings). If I² > 0 or the Q test is significant, a random-effects model is generally more appropriate than fixed-effect.

How many studies do I need for a meta-analysis?

Technically, you can pool 2+ studies. However, with fewer than 5 studies, estimates of between-study variance (τ²) are imprecise, and tests for heterogeneity have low power. Most methodologists recommend at least 5-10 studies for reliable random-effects results.

Need a Full Meta-Analysis?

From data extraction to forest plots, our team handles the entire quantitative synthesis with rigorous statistical methodology.

Explore Services