Hedges' g vs. Cohen's d: Which Effect Size Should You Report?

April 10, 2026 3 min read By Angel Reyes

You're writing your results chapter, and you know you need to report an effect size for your group comparison. You've seen both Cohen's d and Hedges' g mentioned in the literature. They look nearly identical. So which one should you use?

What They Have in Common

Both Cohen's d and Hedges' g measure the standardized mean difference between two groups. They answer the same fundamental question: "How far apart are these two group means, measured in standard deviation units?"

The formula is essentially the same: take the difference between the two means and divide by a measure of spread. The interpretation is the same too — both use Cohen's benchmarks of 0.2 (small), 0.5 (medium), and 0.8 (large).

The Key Difference

Cohen's d divides by the pooled standard deviation and tends to slightly overestimate the true effect size, especially in small samples.

Hedges' g applies a correction factor that adjusts for this small-sample bias. The correction is minimal with large samples but meaningful when your groups are small (roughly n < 20 per group).

Mathematically, Hedges' g multiplies Cohen's d by a correction factor:

g = d × (1 - 3 / (4(n₁ + n₂) - 9))

With 200 total participants, the difference between d and g is negligible — maybe the third decimal place. With 20 total participants, the difference is noticeable.

When to Use Each

Use Hedges' g When:

  • Your sample sizes are small (under 20 per group)
  • You're conducting a meta-analysis (Hedges' g is the standard in meta-analytic work)
  • You want the most accurate, unbiased estimate
  • Your field or journal prefers it

Use Cohen's d When:

  • Your sample sizes are moderate to large
  • You're following a convention established in your field
  • Prior studies you're comparing to reported Cohen's d
  • Your committee specifically requests it

When It Doesn't Matter:

With samples of 40+ per group, the values are virtually identical. If your committee doesn't have a preference and your sample isn't tiny, either measure is defensible.

What Your Committee Expects

Here's a practical tip: look at what measure the studies in your literature review use. If most of your cited sources report Cohen's d, report Cohen's d. If your field leans toward Hedges' g, follow suit. Consistency with your literature makes your results easier to compare and discuss.

If you're conducting a meta-analysis or systematic review, Hedges' g is the standard choice. Nearly all meta-analytic software defaults to it because the bias correction is important when pooling results across many studies, some of which may have small samples.

How to Report Either One

In APA style, the formatting is the same for both. Include the effect size label, value, and interpretation:

"Students in the intervention group scored significantly higher than the control group, t(48) = 2.34, p = .023, d = 0.67, indicating a medium effect."

"The difference between groups was significant, t(18) = 2.12, p = .048, g = 0.72, representing a medium-to-large effect."

How to Calculate Them

Most tools will calculate both:

  • SPSS: Reports Cohen's d when you check the effect size option in newer versions
  • R: The effectsize package calculates both with cohens_d() and hedges_g()
  • G*Power: Uses Cohen's d for power analysis
  • Online calculators: Many free calculators offer both options

The Bottom Line

Don't lose sleep over this choice. The important thing is that you report an effect size — many students forget entirely, and that's a much bigger problem than choosing between two very similar measures. Pick the one that fits your sample size and field conventions, report it clearly, and move on to interpreting what it means for your research.