Interpreting generalized eta squared effect sizes from ezANOVA

[Antonio J. Morales Artacho]

Hello,

I was wondering whether anyone could provide source information regarding the interpretation of generalized eta squared effect sizes. As I read in the paper from Bakeman (2005) cited in the "ez" package (full text here: https://link.springer.com/content/pdf/10.3758%2FBF03192707.pdf): 

Cohen (1988, pp. 413–414), who did not consider repeated measures designs explicitly, defined an η2 (which is not the same as Cohen’s f 2) of .02 as small, one of .13 as medium, and one of .26 as large. It seems appropriate to apply the same guidelines to η2 G as well. Still, within an area of research in which designs do not vary, investigators can develop their own common guidelines for η2 P, and so it can still prove an effective, if local, effect size statistic. 

Could anyone provide further information relating these "rules of thumb" for generalized eta squared effect sizes? 

Thank you!

Antonio