/* Simulate some test data */.
SET SEED=1234567.
MATRIX.
SAVE MAKE(100,2,0) /OUTFILE * /VARIABLES DepVar GrpVar.
END MATRIX.
COMPUTE GrpVar=RV.BERNOULLI(.5).
COMPUTE DepVar=RV.NORMAL(0,1) + GrpVar * 2 .
DESCRIPTIVES VARIABLES DepVar GrpVar /STATISTICS Min MAX MEAN.
T-TEST GROUPS GrpVar(0,1) /VARIABLE DepVar.
UNIANOVA DepVar BY GrpVar /PRINT HOMOGENEITY .
REGRESSION DEPENDENT DepVar/METHOD ENTER GrpVar.
COMPUTE GrpVar=NOT(GrpVar).
REGRESSION DEPENDENT DepVar/METHOD ENTER GrpVar.
Bruce Weaver wrote
Hear hear! The OP might also benefit from estimating the regression model twice, once using an indicator variable for male sex and again using an indicator variable for female sex.
To the OP, if this is not clear, create a variable called Male such that males score 1 and females 0; then create another variable called Female such that females score 1 and males 0. Then estimate the linear regression model twice, first using Male as the lone predictor variable, and then again using Female as the lone predictor variable. Pay attention to the B-values in the table of regression coefficients, and see how they relate to values you see in the output from the t-test and ANOVA carried out on the same data.
HTH.
David Marso wrote
Run them both and also a linear regression.
You will discover some interesting patterns.
SQRT(F) = t (for 2 group case)
The regression will replicate the ANOVA WRT SS, df etc and you will see the regression coefficients are directly related to the group means and their difference.
---
Elona Hasmujaj wrote
Hello!
I need to find the differences on internet addiction level (scale), between two genders (male and female).
I should use Anova or t-test?
Thank you,
Elona
Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email me.
---
"Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis."
Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?"
Locked
