Hello.
I am using a Cohen's d effect size in my research and have read several conflicting studies that interpret or classify the size of the effect differently. Some say the following:
d= 0.20 - 0.49 (small effect)
d= 0.50 - 0.79 (moderate effect)
d= ≥ 0.80 (large effect)
Other research states the following:
d= 0 - 0.20 (small effect)
d= 0.20 - 0.50 (moderate effect)
d= ≥ 0.50 (large effect)
Can anyone provide guidance on which classification is correct?
Thanks in advance!
Hello Marcelle,
Jacob Cohen, in his 1988 text, Statistical power analysis for the behavioral sciences (2nd ed.), set the following arbitrary guidelines for his d measure of effect size: "small" = .20; "medium" = .50; "large" = .80. He really didn't declare ranges per se.
Other authors have made different suggestions, or may, in some cases, be giving guidelines for a completely different ES statistic/metric.
Cohen did say that his guidelines were just that: guidelines. Nothing substitutes for intimate knowledge of the specific variable under consideration and the relevant context for that interpretation. As an example, is a 1 cent per gallon increase in gasoline cost a difference of practical importance? To the person who buys 10-20 gallons at the pump, probably not. To the vendor who sells 100,000 gallons a week, definitely so!
Good luck with your work!
- Badges
- Science topic
- Similar topics
- Data Mining
- Classification