0
$\begingroup$

I have participants who have taken an intervention and are being measured at two different time-points. It is expected that the intervention will improve their knowledge, attitude, confidence etc. and I have created a survey to measure this. I would like to conduct a factor analysis to test whether each of these different factors (e.g. knowledge, attitude, confidence) emerge from the data. The presence of these different factors in creating behaviour change in based on theory. I would then like to understand if participants have improved in these areas at time point 2 compared to time point 1 and whether the factor scores differ between participants with multiple different jobs. How can I compare factor scores across two time points and for different groups of participants? To be clear, I don't want to compare the factor structure but rather the factor scores themselves to see if participants are scoring higher in time point 2.

Could I run a CFA at both time points and for multiple different jobs and then compare the factor scores using an ANOVA? Or would I better to use structural equation modelling and test whether the factor scores have changed? Or is there another way?

Thanks very much.

Improve this question
$\endgroup$
1
  • $\begingroup$ Thanks Christian. Is multigroup and longitudinal CFA an extension of SEM? This is quite advanced for me. Would it be incorrect to use an easier method of factor computation (an unrefined method such as weight sum scores) for both time points and job role groups and the compare the factor scores with an ANOVA? I appreciate that this is less precise as it assumes that all the groups with have the same factor structure which ideally should be tested, but an ANOVA would be much easier and quicker. Is it possible to do an ANOVA on factor scores? Thanks $\endgroup$
    – statshelpneeded
    Commented Sep 19, 2023 at 15:56

1 Answer 1

Reset to default
0
$\begingroup$

You can use longitudinal (and potentially multigroup) CFA to compare factor means (and other parameters) across time and groups. Longitudinal and/or multigroup CFA allows you to also test for measurement equivalence (measurement invariance) across time and/or groups.

Improve this answer
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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