|
|
 |
| Loading specific factors on covariate |
 
|
| Message/Author |
|
|
| Wenshan Li posted on Thursday, July 30, 2020 - 12:52 pm
|
|
|
Hello,
I'm running a ESEM with 5 factors (F1-F5), the outcome "hb", and a covariate "wealth" that will load onto hb. I want to add paths from only 2 of the 5 factors on the covariate (as suggested by Modification Indices). But when I do I get an error message. See below for the (simplified) codes + error message (marked by ***):
ANALYSIS:
TYPE = COMPLEX;
MODEL:
F1-F5 BY item1 item2..... (*1);
hb ON F1-F5 wealth;
F1-F2 on wealth;
*** ERROR in MODEL command
EFA factors in the same set as F1 must have the same set of regressions.
Problem with:
F1 ON WEALTH
F3 ON WEALTH (not specified or fixed)
F4 ON WEALTH (not specified or fixed)
F5 ON WEALTH (not specified or fixed)
***
This error doesn't occur if I add all 5 factors to Wealth.
Based on the error message, it seems like I have to fix the other 3 factors... but is it the variance I have to fix? Or factor loading? And can I fix it to just any number (eg. 1)?
Thank you very much and I look forward to reading your advice! |
|
|
|
|
You should say
f1-f5 on wealth;
that is, also specify f3-f5 on wealth (they are not specified as the message says). |
|
|
| Wenshan Li posted on Thursday, July 30, 2020 - 3:40 pm
|
|
|
Thank you Bengt for your reply. I was actually wondering if it's possible to ONLY have F1-F2 on Wealth, without F3-F5. If the MI indices only indicate that adding F1 and F2 on wealth is beneficial, then shouldn't those be the only paths added?
In other words, is it possible for F1-F2 ON Wealth to work somehow, or would I always have to bundle all factors together?
Thank you very much. |
|
|
|
|
| All 5 ESEM factors need to have the same covariates. |
|
|
| Wenshan Li posted on Thursday, July 30, 2020 - 4:03 pm
|
|
|
| I see, thank you very much! |
|
|
| Back to top |
|
|
