- I am testing the difference in carotenoids of 5 groups of diferent sex or development stage animals with one way ANOVA. My samples are from 13 the smallest to 44 the greatest. Do I use the eta square value?
So I use the Kruskall-Wallis test to compare the groups when the variables were not normally distributed, and the post-hoc Dunn test.
Cohen suggested that d=0.2 be considered a 'small' effect size, 0.5 represents a 'medium' effect size and 0.8 a 'large' effect size. This means that if two groups' means don't differ by 0.2 standard deviations or more, the difference is trivial, even if it is statistically significant. Cohen’s d is used to describe the standardized mean difference of an effect. This value can be used to compare effects across studies, even when the dependent variables are measured in different ways, for example when one study uses 7-point scales to measure dependent variables, while the other study uses 9-point scales, or even when completely different measures are used, such as when one study uses self-report measures, and another study used physiological measurements. Please see the attached document for details: https://www.frontiersin.org/articles/10.3389/fpsyg.2013.00863/full
Cohen suggested that d=0.2 be considered a 'small' effect size, 0.5 represents a 'medium' effect size and 0.8 a 'large' effect size. This means that if two groups' means don't differ by 0.2 standard deviations or more, the difference is trivial, even if it is statistically significant. Cohen’s d is used to describe the standardized mean difference of an effect. This value can be used to compare effects across studies, even when the dependent variables are measured in different ways, for example when one study uses 7-point scales to measure dependent variables, while the other study uses 9-point scales, or even when completely different measures are used, such as when one study uses self-report measures, and another study used physiological measurements. Please see the attached document for details: https://www.frontiersin.org/articles/10.3389/fpsyg.2013.00863/full
Hello Jaime,
How to compute:
Cohen's d = (Mean1 - Mean2 ) / SDpooled
where Mean1 is the mean for batch "A" and Mean 2 is the mean for batch "B" that you wish to compare. The SDpooled is the pooled within-groups standard deviation.
For a one-way ANOVA with 5 groups, there will be 10 possible d values. I would suggest using the square root of MSE (the error MS) as the pooled SD throughout. You could, of course, compute the specific pooled SD for each of the 10 comparisons.
How to interpret:
Attaching qualitative meaning to the magnitude of Cohen's d depends on the specific variables, the population, and the audience. Cohen would be the first to acknowledge that his suggested guidelines (e.g., 0.2, 0.5, 0.8) are just that, guidelines, and do not substitute for intimate understanding of context. In your field, differences that other researchers might consider noteworthy could vary by orders of magnitude from Cohen's suggestions. You might compare your results to those of other, similar studies in your field as a basis for judging the magnitude of the effect. Alternatively, you could poll other experts in the field.
Good luck with your work.
Hello Jaime Färber-Lorda. I suggest that you also take a look at this article by Thom Baguley:
Article Standardized or simple effect size: What should be reported?
HTH.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Modeling
- Analyses of Variance
- ANOVA