Message/Author |
|
Tino Nsenene posted on Monday, September 19, 2016 - 12:50 am
|
|
|
Hello, Table 1.15 in the new Mplus-book shows how to do a Linear regression with heteroscedasticity modeling. I'd like to do something similar for a type = twolevel analysis. But it seems that the 'constraint' command is not available for type = twolevel. Is there any other syntax option to model heteroscedasticity when using type = twolevel in Mplus? Will 'constraint' for twolevel analysis become available in Mplus 8? Best Tino |
|
|
Q1. No. The closest you can get is to have a multiple-group analysis where groups are based on categories of the variable that influences the heteroscedasticity. Q2. Probably not in Version 8.0 because the wish list is long. |
|
|
Hi, On page 45 of the book, it is said that a heteroscedasticity model such as (1.64) can be used in situation where the b parameter is expected to have a negative value. What are the options available in MPLUS when the b parameter is expected to have a positive value? Best regards, Patrick |
|
|
The same approach can be used. |
|
Back to top |