0
$\begingroup$

I'm currently writing my master thesis and therefore want to examine the psychological well-being in a group of patients with a certain disease. I already looked for psychological variables that were found to be increased in that group of patients. I found 16 latent variables such as depression, self-esteem, anxiety, social isolation etc. Now, I would like to find out how those variables are connected and my end goal is to draw a model of the paths. All latent variables such as depression were collected with questionnaires.

Now, I wonder which way would be best to find out about the connections. I am currently struggeling whether I should use an exploratory or a confirmatory factor analysis? In case I would use an exploratory analysis, more questions arise. Is it correct to use the index values of my 16 latent variables and to put those into my intercorrelation matrix? And how do I get from my factor loading matrix to my path model?

Thank you very much for your answers!

$\endgroup$
0
$\begingroup$

In both EFA () and , the goal is to explain correlations among manifest/observed variables via latent/unobserved variables.

Doing whatever factor analysis with the latent variables is, I suspect, not what you want unless you are interested in something called second-order factors. Rather, I suspect you are interested in relationships (paths) among the latent variables, which would turn a into an . Then, those relationships (no relationship vs. correlation vs. path from A to B vs. path from B to A) need to be specified by the researcher.

If you have something like a SEM with only manifest variables (e.g., the factor scores extracted in a previous analysis), that is called a path analysis.

IMHO, 16 latent variables sounds like a lot.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.