Factor analysis, correlations and spss

Nina

New Member
#1
Hi,

i just need to clarify...

i extracted 3 factors from a factor analysis which are distinct from each other (ie. they are supposed to measure different 'things'). fo example, factor 1 is about 'personal consciousness' amd factor 2 is about 'public consciousness'. although the intention was to get them to be looking at 'separate' issues,

1) i wonder if at the end of the day, i could just see if there is a relationship between factor 1 and factor 2 to the "overall question", would it be linear regression or just simple correlations will do?

2) So, in spss, am i correct to do this => lets say 3 survey likert-items (qn, a,b,c) loaded on factor 1, and 4 items loaded on factor 2. Do i total the score of the 3 items in factor 1 and also total the 4 items on factor 2 ie. create 2 new columns (of variables) on the spss to check for correlations?

OR, can i just 'insert' the responses of qn a,b,c,(factor 1) straight into the linear regression/correlation for analysis?
 

JohnM

TS Contributor
#2
I'm no expert in factor analysis, so I'm just going with my gut-feeling here.

1) do a multiple linear regression with factor 1 and factor 2 as independent variables, and the "overall question" as the dependent variable - this will allow you to see the amount of explanation each factor contributes, controlling for the other factor

2) use the sums, but also try using the averages, since each factor is comprised of a different number of questions