I ran numerous repeated measures ANOVAs as part of the analysis for my PhD - something that was bought up in my viva was that I had neglected to provide any measure of effect size. I am now rectifying this as part of my corrections but I am very stuck.
The feedback from my examiner was as follows:
"A commonly-used measure of effect-size for within-subjects design is Cohen's d. This is also the default effect size measure for within-subjects effects in G Power, and is easy to calculate (we recommend using the standard deviation of paired differences in the denominator). However, opinions differ as to what the most appropriate measure of effect size is, and so you may use any reasonable measure that you choose, provided that you give a justification for your choice and are consistent across analyses."
When I tried to calculate effect sizes in G Power it only seemed to use Cohen's d when using a t-test? When using a F-test it said 'Effect size f(V)' rather than 'Effect size d' so I was not confident that this was actually calculating Cohen's d? And when I have looked on google for how to calculate Cohen's d for ANOVA I have just found help/calculators for use with t-tests.
When I calculated effect sizes for my ANOVAs in SPSS it was eta squared - is this a justifiable measure of effect size for ANOVA instead of Cohen's d? And if not is someone able to point me to some information on how to calculate Cohen's d for a repeated measures ANOVA?
Thank you
Hi Ruth,
Here are some links that may be useful:
Page 247 -- https://books.google.com/books?id=HFphCwAAQBAJ&pg=PA247&dq=anova+%22cohen%27s+d%22&hl=en&sa=X&ved=0ahUKEwjGrrvHob7aAhXDqFkKHe27C9MQuwUILTAA#v=onepage&q=anova%20%22cohen's%20d%22&f=false
Page 262 -- https://books.google.com/books?id=5vP0CAAAQBAJ&pg=PA762&dq=anova+%22cohen%27s+d%22&hl=en&sa=X&ved=0ahUKEwiTg96uor7aAhUCyFkKHUDDCTQQuwUIMzAB#v=onepage&q=anova%20%22cohen's%20d%22&f=false
Have a great day!
--Adrian
Hi
Please refer to:
1. Landau, S. and Everitt, B. S.(2004). A Handbook of Statistical analyses using SPSS.
2.Howitt, D. and Cramer, D.(2008). Introduction to SPSS.
Regards,
Zuhair
What would you think is a resonable, sensible way to express the sie of an effect you would measure with an ANOVA? If you make up your mind here, one can try to find what statistic would best represent this and how it can be calculated from the data. Eta² is an effect size measure that is used, but do you understand what it is to give a justification why you would provide this mesaure?
Hi Ruth,
Here are some links that may be useful:
Page 247 -- https://books.google.com/books?id=HFphCwAAQBAJ&pg=PA247&dq=anova+%22cohen%27s+d%22&hl=en&sa=X&ved=0ahUKEwjGrrvHob7aAhXDqFkKHe27C9MQuwUILTAA#v=onepage&q=anova%20%22cohen's%20d%22&f=false
Page 262 -- https://books.google.com/books?id=5vP0CAAAQBAJ&pg=PA762&dq=anova+%22cohen%27s+d%22&hl=en&sa=X&ved=0ahUKEwiTg96uor7aAhUCyFkKHUDDCTQQuwUIMzAB#v=onepage&q=anova%20%22cohen's%20d%22&f=false
Have a great day!
--Adrian
Thank you for your help - those reference were particularly useful Adrian and have solved my problem.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Modeling
- Analyses of Variance
- ANOVA