Cohen's D - getting two different values in Excel and Stata

[Vanessa Ojeda]

Hi all,
Could you let me know why I would get a different Cohen's D value in excel vs. Stata? I have a paired sample (same sample, pre and post survey).
When I calculate Cohens D in excel I get: 2.7 but in Stata I get 2.04.

Any guidance very much appreciated!

mean 1: 3.539065
SD 1: 0.970597
n=139

mean 2: 5.213201
SD 2: 0.636956
n=139



When I calculate cohen's D in excel I get this value:

M1-M2	-1.674136
Pooled SD	0.618228
cohen's D	-2.70796

using this formula:
pooled SD =SQRT(((139-1)*0.970597^2)/(139-1)*0.636956^2) = 0.618228

Cohen's D = -1.674136 / 0.618228 = -2.70796

in Stata, I use the following command:

esizei 139 5.213201 0.636956 139 3.539065 0.970597, cohensd

And my output gives me a Cohen's D value of 2.039377

Effect size based on mean comparison

Obs per group:
Group 1 = 139
Group 2 = 139

Effect size Estimate [95% conf. interval]

Cohen's d 2.039377 1.747979 2.328287

[Joseph Coveney]

It's difficult to read your Excel formula, but maybe it should be more like

=SQRT(  ( (139-1) * 0.970597^2 + (139-1) * 0.636956^2 )  / (139 + 139 - 2)  )

[Vanessa Ojeda]

Hi Joseph Coveney , yes! The updated excel formula you provided was super helpful and now I got the same value. Thank you so much, again!