Re: st: xtmixed- testing difference between intercepts and slopes

["Mary E. Mackesy-Amiti" <mmamiti@uic.edu>]

On 3/14/2011 10:55 AM, Bernadette Puckett wrote:
> Dear Stata List,
>
> I am currently conducting a multilevel longitudinal model with time
> point clustered in student and student nested within high school to
> examine the association between high school quality and achievement. I
> have 3 categorical levels of school quality that I am interacting with
> 4 time points in order to test whether the slope significantly differs
> across time point and quality level. I have a host of child-level and
> school-level controls.
>
> How do test whether there is a significant intercept difference
> between quality levels at Time 2? Or at Time 3? Are my coefficients
> displaying differences in slope or intercept? If intercept, then how
> do I test for significant differences between slope of quality at
> quality level 2* time 2, and quality level 3*time2.

It sounds like you don't understand the model you are fitting.  The 
output below shows the fixed part of the mixed model; the regression 
coefficients are interpreted in the same way as an ordinary regression.  
The random effects are reported as variances following the fixed effects.

To test specific effects within the interaction use the -test- and/or 
-lincom- postestimation commands.  Also use -margins- to obtain marginal 
predicted values; this is quite useful for understanding interactions.


> Here is my code:
>
> xtmixed achievement time quality##time vector.of.controls  || school:,
> variance cov(un) || child_id: time, variance cov(un) mle.
>
> And example output:
> achieve       Coef.	Std. Err.	z	P>z	
> 						
> quality
> 2     -1.41589	1.628779	-1.48	0.138	
> 3    -2.989446	1.683687	-1.18	0.237	
>
> time
> 1     15.55104	24.33336	1.05	     0.294	
> 2     23.36034	24.53234	1.77  	0.077	
> 3     56.20899	24.9002	3.06	     0.002	
>
> quality#
> time
> 2 1     2.991244	.986465	4.05	0.000	
> 2 2      1.5299	        1.094144	1.76	0.078	
> 2 3     1.441748	1.256821	1.15	0.251	
> 3 1     2.798601	1.064483	3.76	0.000	
> 3 2     1.127278	1.162627	0.88	0.377	
> 3 3     1.970249	1.316493	1.42	0.155	
>
> ex.
> control-.5441684   .4442959    -1.22   0.221    -1.414972    .3266356
>   _cons |  -34.72625   25.54131    -1.36   0.174    -84.78629    15.33379
>
>
> Thank you,
> Bernadette
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/statalist/faq
> *   http://www.ats.ucla.edu/stat/stata/

--
Mary Ellen Mackesy-Amiti, Ph.D.
Research Assistant Professor
Community Outreach Intervention Projects (COIP)
School of Public Health m/c 923
Division of Epidemiology and Biostatistics
University of Illinois at Chicago
ph. 312-355-4892
fax: 312-996-1450

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/