IBM Support

Can SPSS produce an estimated Cohen's d value for data with two independent samples?

Troubleshooting


Problem

I need to compute a "post-hoc" estimate of effect size in the form of Cohen's d, the mean difference divided by the pooled standard deviation. Can SPSS produce this?

Resolving The Problem

SPSS currently can produce partial eta^2 values in several procedures that have ANOVA capabilities, but none produce d. Following are commands for producing an estimate of d in SPSS based on a given sample. The same cautions about positively biased estimates of effect sizes resulting from "post-hoc" computations that apply to results from SPSS procedures that provide partial eta^2 values should be applied here as well.

*Compute's Cohen's d for two independent samples, using observed means and standard deviations.
*Replace groupvar and depvar with appropriate variable names in AGGREGATE command.
*If path c:\temp\temp.sav will not work, replace it in AGGREGATE and the three GET statements
*inside MATRIX.

AGGREGATE
/OUTFILE='c:\temp\temp.sav'
/BREAK=groupvar
/mean=MEAN(depvar)
/sd=SD(depvar)
/n=N(depvar).

MATRIX.
GET means /file='c:\temp\temp.sav' /var=mean.
GET sds /file='c:\temp\temp.sav' /var=sd.
GET ns /file='c:\temp\temp.sav' /var=n.
COMPUTE d=(mmax(means)-mmin(means))/sqrt(((ns(1)-1)*sds(1)**2+(ns(2)-1)*sds(2)**2)/(ns(1)+ns(2)-2)).
PRINT d.
END MATRIX.

[{"Product":{"code":"SSLVMB","label":"IBM SPSS Statistics"},"Business Unit":{"code":"BU059","label":"IBM Software w\/o TPS"},"Component":"Not Applicable","Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"Not Applicable","Edition":"","Line of Business":{"code":"LOB10","label":"Data and AI"}}]

Historical Number

77292

Document Information

Modified date:
16 April 2020

UID

swg21480222