"This is Sparsa" for SEM
Is it possible to do SEM in JASP if in the data set (spreadsheet in JASP) I use there are gaps in my data/spreadsheet? To illustrate: mostly all nice data for all the variables: but if I have data in a box for one variable, but in the corresponding box on the same row for a different variable this box is empty, will this cause a problem?
My data is a bit sparse for some variables in some places of the data. Can JASP handle this? Or it will render the result spurious?
Thank you.
Comments
For SEM, assuming you're treating your variables as numeric (scale/continuous), this has to do with full-information maximum likelihood (FIML) estimation, which is the default in JASP. If variables are so sparse that it doesn't run, you'll know.
The alternative to FIML estimation with numeric data is maximum likelihood (ML) estimation, which uses listwise deletion (i.e., you would lose all the rows in which there is at least one missing data somewhere in the columns that you use in your model). But again, the default in JASP is to use FIML estimation.