In a complex design where we want to look at the effect of 2 variables, entered as demeaned EV's I also want to look at the interaction of these two.
What is the correct order:
1st: Take the value of Demeaned EV1 x value of Demeaned EV2 => Value EV3 and then demean these values of EV3 again?
'raw' value for EV1 x raw value for EV2 => value for the interaction effect (EV3) and then Demean this value?
It gives different values: the first procedure leads to 4 different values, the second to only 2 values for example.
EV1 EV2 both EV's are demeaned.
subj1 -1 1.7
subj2 1 -1.1
subj3 -1 1.7
subj4 1 -1.1
subj5 1 -1.1
subj6 -1 -1.1 etc.
Originally they were just: variable1 (belonging to EV1) 14 x value 1 and 14 x value -1
and Variable2: 17 x value -1 and 11 x value +1 => leading to the above EV's schedule.
Thanks a lot....
Drs. Anda van Stegeren,
University of Amsterdam
Department of Clinical Psychology
1018 WB Amsterdam
tel. ++ 31 20 525 6799
fax. ++31 20 639 1369
E-mail: [log in to unmask]