|
|
 |
| Accounting for types of instruments i... |
 
|
| Message/Author |
|
|
| Guido Biele posted on Thursday, April 05, 2018 - 6:20 am
|
|
|
I would like to do ESEM on data that stems partially from questionnaires and partially from neuropsychological tests.
When I do this in a naive fashion, the resulting factors reflect different functions (e.g., executive functions (EF), IQ, emotional lability), but also the type of instrument. E.g., I find one factor for EF measured with a parent questionnaire and one factor for EF measured with neuropsychological tests.
Is there a way I can set up an analysis to control for the effect of instrument type?
The model I have in mind would have a latent variable for each participant that captures the ability to perform tests and would, which in addition to the standard latent variables of an EFA account for performance in all tests.
Is there an example for such or a similar model, or is there a specific chapter of the documentation I should read for this? |
|
|
|
|
| You may want to check the literature on MTMM. Also, ask on SEMNET. |
|
|
|
|
Thanks for the pointer to MTMM, though I am not sure if MTMM is applicable for my problem because I am trying to do an exploratory analysis.
I'll ask on SEMNET. |
|
|
|
|
| You might be able to do that using bi-factor rotation (ROTATION = BI-GEOMIN;) or within the ESEM framework use EFA factors in combination with CFA factors. |
|
|
| Back to top |
|
|
