|
|
 |
| Samling weights and categorical data |
 
|
| Message/Author |
|
|
| Markus Riek posted on Tuesday, October 01, 2013 - 4:14 am
|
|
|
Dear Mplus community,
I’d like to know whether Mplus is able to perform the following analysis before purchase it. After reading the package descriptions on the website I think it should be, but my case combines multiple problems, so finally I’m not completely sure.
The case: I have a big sample (n > 15.000) of complex survey data (with sampling weights per case) and missing values.
I want to perform a single level SEM analysis for whole sample. Four Latent Variables based on ordinal indicators (Likert-scales, mostly non-normal distributed) and one dependent ‘latent’ variable based on a binary indicator should be analyzed.
The most important thing is the simultaneous inclusion of the sampling weights and the use of a suitable estimation method for categorical data. The handling of the missing values is a nice to have, as I already did work on that manually and can also use a sample without missing values.
Thanks in advance,
Markus |
|
|
|
|
| This can all be done in Mplus. Please note that categorical data methdology can handle Likert scale variables with floor or ceiling effects. |
|
|
| Markus Riek posted on Tuesday, October 01, 2013 - 7:47 am
|
|
|
| Great, thanks for your fast reply Linda. |
|
|
| Back to top |
|
|
