Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

Showing results for

Find a Community

- Home
- /
- Analytics
- /
- Stat Procs
- /
- proc mixed parameter dispersion matrix

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

04-12-2015 08:47 AM

The UCLA HLM site has several examples using the HSB dataset on this page: http://www.ats.ucla.edu/stat/hlm/seminars/hlm_mlm/608/mlm_hlm_seminar_v608.htm

The third model enters SES as a group-mean centered variable at level-1. HLM output reports a .908 reliability estimate for the intercept and a .260 reliability estimate associated with SES.

Using the variance components reported for the intercept and the level-1 residual, I’m able to calculate a lambda value for each schoolid (n=160), and then sum across those to obtain an overall reliability estimate that matches the HLM output. However, I’ve been unable to do similar calculations and arrive at the .260 provided for the SES reliability.

It's not clear to me how I can obtain the parameter dispersion matrix which would contain both parameter and error dispersion around each Beta. Essentially, i want to know these two facets about the slope parameter associated with SES. Is there a way to obtain this using ODS? Is there something that can be obtained and further processed inside IML?

Any thoughts or insights are greatly appreciated.

Jason