In Factor Analysis, all of the variables are considered as Dependent Variables. The Factors extracted are unknown, unmeasured explanatory factors that may be likened to Independent Variables that influence the DVs.
Can I do a factor analysis with one dependent variable and one independent variable? Do I need to have at least two dependent variables to do a factor analysis?
In Factor Analysis, all of the variables are considered as Dependent Variables. The Factors extracted are unknown, unmeasured explanatory factors that may be likened to Independent Variables that influence the DVs.
Thank you for your response. I am sorry for confusing you.
I mean, I have one selection variable and one dependent variable. Can I do a factor analysis with them? in SPSS, I need to put at least two variables in variables section (with one variable in the selection variable section). Looking forward to hearing from you.
The whole purpose of FA is to reduce the number of variables. It makes no sense to run a FA with just 1 variable.
It's like asking to run a cluster analysis with just 1 respondent. That respondent IS the cluster.
Last edited by Injektilo; 07-28-2014 at 05:17 PM.
I agree with Injektilo. You possibly could do this. Why would you want to? Normally you want many (30 plus say) variables before doing EFA (there are actually rules of thumb that sugest how many you need, but they generate signficant numbers regardless).
EFA is a data reduction method and when you have a two variables there is no reason to use it. Do you mean by any chance SEM (structural equation models)? Usually they are not called FACTOR ANALYSIS. Even with that approach I can't see using two variables.
Why not simply do regression?
Last edited by noetsi; 07-28-2014 at 06:40 PM.
"Very few theories have been abandoned because they were found to be invalid on the basis of empirical evidence...." Spanos, 1995
Hi eunkum1,
What is the question that you are actually trying to answer with your analysis? Maybe that way we can suggest an appropriate method.
I agree that for a small number of variables, there is not much point in doing a factor analysis.
Regards,
Londoner
Tweet |