What Is Cohen's d? Formula, Interpretation, and APA Reporting Guide
Direct Summary (TL;DR)
Cohen's d is a standardized effect size measure that quantifies the difference between two group means in units of standard deviation. While a t-test p-value only tells you whether a difference is statistically distinguishable from zero, Cohen's d tells you how large and practically meaningful that difference is, independent of sample size.
Cohen's d = (Mean₁ − Mean₂) / Pooled Standard Deviation
- Small effect: |d| ≈ 0.20
- Medium effect: |d| ≈ 0.50
- Large effect: |d| ≈ 0.80+
Why Cohen's d Is Critical in Psychology Research
In behavioral and social science research, statistical significance (p < .05) is frequently misunderstood. A p-value is a function of both the true effect size and the sample size (N):
- Large samples (N = 10,000) can make a negligible, trivial difference of 0.1 points produce p < .001.
- Small samples (N = 20) can cause a clinically substantial difference to yield p = .12 (a false negative due to low statistical power).
Cohen's d solves this by isolating the magnitude of the difference from sample size, enabling direct comparisons across studies, questionnaires, and meta-analyses.
The Mathematical Formulas
1. Cohen's d Formula (Independent Groups)
d = (M₁ − M₂) / SD_pooled
Where:
- M₁ = Mean of Group 1 (e.g., Experimental condition)
- M₂ = Mean of Group 2 (e.g., Control condition)
- SD_pooled = Weighted average of standard deviations from both groups
2. Pooled Standard Deviation Formula
SD_pooled = √[ ((n₁ − 1)s₁² + (n₂ − 1)s₂²) / (n₁ + n₂ − 2) ]
Where:
- n₁, n₂ = Sample sizes of Group 1 and Group 2
- s₁², s₂² = Variances of Group 1 and Group 2
- n₁ + n₂ − 2 = Degrees of freedom (df) for an independent-samples t-test
Step-by-Step Worked Example
Suppose you conduct an experiment testing an active recall study technique against a passive rereading control group on an exam scored from 0 to 100:
| Group Metric | Active Recall (Group 1) | Passive Rereading (Group 2) |
|---|---|---|
| Sample Size (n) | 32 | 30 |
| Mean Score (M) | 82.40 | 74.80 |
| Standard Deviation (SD) | 9.10 | 9.80 |
Step 1: Calculate the Mean Difference
- Difference = M₁ − M₂ = 82.40 − 74.80 = +7.60
Step 2: Compute the Pooled Standard Deviation
- SS₁ = (32 − 1) × (9.10)² = 31 × 82.81 = 2567.11
- SS₂ = (30 − 1) × (9.80)² = 29 × 96.04 = 2785.16
- df = 32 + 30 − 2 = 60
- SD_pooled = √[ (2567.11 + 2785.16) / 60 ] = √[ 5352.27 / 60 ] = √89.2045 ≈ 9.445
Step 3: Compute Cohen's d
- d = 7.60 / 9.445 ≈ +0.805
Interpretation: Active recall improved exam scores by 0.805 standard deviations, representing a large effect size under Cohen's benchmarks.
Cohen's (1988) Effect Size Benchmarks
Jacob Cohen established standard heuristics for psychological research:
| Absolute Value | Effect Size Tier | Non-Overlap of Distributions | Practical Interpretation |
|---|---|---|---|
| ** | d | < 0.20** | Negligible |
| **0.20 ≤ | d | < 0.50** | Small |
| **0.50 ≤ | d | < 0.80** | Medium |
| ** | d | ≥ 0.80** | Large |
When to Use Hedges' g Instead of Cohen's d
Cohen's d possesses a slight upward positive bias in small samples (N < 40). Hedges' g applies an exact bias-correction factor (J):
g = d × [ 1 − 3 / (4(n₁ + n₂) − 9) ]
If your total sample size is under 40 participants, or if you are preparing data for a meta-analysis, report Hedges' g to ensure unbiased estimation.
How to Report Cohen's d in APA 7th Edition Style
The APA 7th Publication Manual requires effect size reporting alongside statistical tests. Follow these guidelines:
Standard Reporting Template
An independent-samples t-test demonstrated that participants using active recall (M = 82.40, SD = 9.10) scored significantly higher than those using passive rereading (M = 74.80, SD = 9.80), t(60) = 3.16, p = .002, d = 0.80, 95% CI [0.28, 1.33].
Key APA Formatting Rules:
- Italicize statistical abbreviations: M, SD, t, p, d, g, N, n, CI.
- Do not put a leading zero before p-values: Report p = .002, not p = 0.002 (since p cannot exceed 1).
- Include confidence intervals: Always provide a 95% CI for your effect size when possible to indicate precision.
- State the direction: Clearly indicate which group scored higher in your narrative.
Frequently Asked Questions (PAA)
Can Cohen's d be negative?
Yes. The sign indicates directionality: a negative Cohen's d means Group 1 scored lower than Group 2. For instance, in an anxiety reduction study, a negative d signifies that the treatment group experienced fewer symptoms than the control.
Can Cohen's d exceed 1.0 or 2.0?
Yes. Cohen's d has no upper bound. An effect size of d = 2.0 means the group means differ by two full standard deviations (common in physiological or highly controlled cognitive tasks).
What is the difference between Cohen's d and Pearson's r?
Cohen's d measures differences between two discrete groups (independent means), whereas Pearson's r measures the linear relationship between two continuous variables. You can convert between them using the formula: r = d / √(d² + 4).
Key Takeaways
- Cohen's d standardizes mean differences, stripping away sample size bias.
- Cohen's benchmark values are 0.20 (small), 0.50 (medium), and 0.80 (large).
- For small samples (N < 40), utilize Hedges' g for bias-corrected accuracy.
- APA 7th edition mandates reporting effect sizes with confidence intervals for all primary outcome measures.