Cohen's D vs. Hedges' G

Introduction

• Purpose of Video: Clarify the distinction between Cohen's D and Hedges' G.
• Reason for Discussion: Both are often reported, and there's confusion regarding their differences.
• Recommended Reading: A paper by Rosnow and Rosenthal on effect size estimates.

Key Differences

• Formula for Hedges' G:
  • Difference between two means divided by a pooled standard deviation (sample-based).
• Formula for Cohen's D:
  • Difference between two means divided by a pooled Sigma (population-level standard deviation).

Conceptual Differences

• Standard Deviation:
  • Hedges' G uses the sample standard deviation (n - 1).
  • Cohen's D uses the population standard deviation (n).

Reporting Recommendations

• Preferred Reporting: Recommend reporting Hedges' G because typically, only sample standard deviations are accessible.
• Discrepancy in Use: Cohen's D is more frequently reported in literature despite the lack of access to population standard deviation.

Popularity and Misconceptions

• Frequency of Reporting:
  • Hedges' G: 250 times in 2017 (Google Scholar).
  • Cohen's D: 5,240 times in 2017.
• Potential Misreporting:
  • Most studies likely do not have access to population data.
  • Cohen's D is conceptualized in the context of power, relevant for population effect size.

Conclusion

• Recommendation: For reporting the standardized difference between two means, use Hedges' G unless population standard deviation is genuinely available.
• Reviewer's Perspective: Some reviewers may prefer Hedges' G over Cohen's D due to the typical availability of data.

Remember to check the paper by Rosnow and Rosenthal for more in-depth understanding of effect size estimates.