Guidelines for Specifying Zero-Loadin... PreviousNext
Mplus Discussion > Confirmatory Factor Analysis >
 Kasey Stanton posted on Saturday, March 18, 2017 - 9:03 am
I am interested in modeling results from previously reported EFAs of a 48-item measure from two different studies using ESEM, but am not sure of the proper way to specify these models. One of these previous studies reports results from a 3-factor solution, and the other a 4. These data fit CFA models very poorly (I have already tried this), which is why I am giving ESEM a shot.

If I run 3- and 4-factor ESEM analyses wherein every item is allowed to load on every factor, then I am not "recreating" the results of these previous solutions. So, I would like to specify that items cannot load on certain factors based on these previous analyses so that I can represent these 3- and 4-factor EFA models in my ESEM analyses. For example, if item 48 has loadings of .75, .20, and .10 on factors 1-3, respectively, then clearly it is a marker of the first factor. Should it be allowed to load only on the first factor then in my ESEM analysis?

Are there published guidelines or criteria for when it is appropriate to specify that an item should have a zero loading on factors (e.g., all items are assigned zero loadings if they have loadings < .15 on a factor in an EFA)? Unfortunately, the statistical significance of factor loadings is not reported in these previous studies, so I cannot use statistical significance as a criterion for specifying where items do and do not load. Thank you very much.
Back to top
Add Your Message Here
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Options: Enable HTML code in message
Automatically activate URLs in message