How can I calculate the effect sizes or find the minimal detectable change at the 90% confidence for variables of three comparison groups?!!!!
I do not understand your question. Clinical significance and statistical significance are not the same and do not go hand in hand. You can have a difference between two groups that is statistically significant but is not clinically significant. On the other hand, in my opinion, you cannot have a difference that is clinically significant unless it is also statistically significant.
How can I calculate the effect sizes for variables if I have three independent groups?
I already have statistical significance but I need to confirm the clinical significance too if the effect size is high. So how can I calculate the cohen's d effect size for three independent groups ?
If you want to use Cohen's d, how about reporting the maximum Cohen's d across all pairwise comparisons ? Or list them for each pairwise comparison ?
Very good questions! So, is there any other way to calculate the effect size for more than two groups?
Mariam Ameer , usually if you have a design with multiple groups, the effect size measurement is something like r-squared. This could be r-squared, eta-squared, partial eta-squared, omega-squared, and so on. These range from 0 and 1, and indicate the amount of the variability in the dependent variable explained by the model. They are different if what information they convey than Cohen's d.
- Badges
- Science topic
- Similar topics
- Medical Science Nursing
- Dissertations