Message/Author |
|
|
Hello, I would like to know under what circumstances and why the intercepts of observed indicators are estimated as the default settings. How do I interpret the intercepts. |
|
|
Starting with Version 5, TYPE=MEANSTRUCTURE is the default. Intercepts are interpreted in line with regression. Most models do not impose a structure on the means of the observed variables in which case model fit is not affected. |
|
Jinseok Kim posted on Friday, September 30, 2011 - 9:51 pm
|
|
|
Hi, I am trying to figure out the parameter default settings written in p408 of user's guide (ver 6), which says "In single group analysis, intercepts and thresholds of observed dependent variables are free." Now, I ran a model like y1 y2 on y3 ; y3 on x1 x2 x3; According to the user's guide, I was expecting NU parameter estimates for y1, y2, and y3, which I could not find in the output. Do I miss anything? Thanks. Jinseok |
|
|
If you are using an older version of Mplus, you should include TYPE=MEANSTRUCTURE; in the ANALYSIS command. If you are using a newer version, including means is the default. If this does not answer your question, please send the relevant files and your license number to support@statmodel.com. |
|
Jinseok Kim posted on Monday, October 03, 2011 - 7:40 pm
|
|
|
Oh... Linda, I think I figure it out. The NU parameters were not estimated maybe because my dataset is "type = stdeviations correlation"? |
|
|
That could do it. You have no means in the data. |
|
Back to top |