Subscribe Bookmark RSS Feed

Mixed Model with Complex Data (Stratified Sampling)

triunk

Community Trekker

Joined:

Nov 20, 2012

I was wondering if:

a) The new JMP Pro11 Is offering Mixed model or Hierarchical Linear Regression with Complex Data? Allow post-stratification weights?

b) The new JMP Pro 11 allow mixed models with more than two levels?

c) What are the procedures in the new JMP Pro 11 to select predictors of level 1, predictors of level 2 and predictors of level 2 with mixed models

d) Are any examples relevant to a), b) and c) available?


Thank you

1 ACCEPTED SOLUTION

Accepted Solutions
jiancao

Staff

Joined:

Jul 7, 2014

Solution

a). You can fit hierarchical linear models in JMP Pro 11 (Analyze =>Fit Model, and choose Mixed Model Personality).


b). It supports more than two levels (assuming your data is adequate for multi-level modeling). To specify level-1 and level-2 random effects using Nest Random Coefficients button from the Random Effects tab. See JMP blog post for how to use Nest Random Coefficients.


c) After a model is fit t ratios and associated p-values indicate the statistical significance of fixed effects. For estimated variance and covariance parameters, you can examine the reported 95% confidence limits to see if such estimates are statistically different from zero at α=5% or not.

d) Specifying a hierarchical linear model is similar to specifying a random coefficient model. Refer to JMP documentation or the blog post above for details.

2 REPLIES
jiancao

Staff

Joined:

Jul 7, 2014

Solution

a). You can fit hierarchical linear models in JMP Pro 11 (Analyze =>Fit Model, and choose Mixed Model Personality).


b). It supports more than two levels (assuming your data is adequate for multi-level modeling). To specify level-1 and level-2 random effects using Nest Random Coefficients button from the Random Effects tab. See JMP blog post for how to use Nest Random Coefficients.


c) After a model is fit t ratios and associated p-values indicate the statistical significance of fixed effects. For estimated variance and covariance parameters, you can examine the reported 95% confidence limits to see if such estimates are statistically different from zero at α=5% or not.

d) Specifying a hierarchical linear model is similar to specifying a random coefficient model. Refer to JMP documentation or the blog post above for details.

gail_massari

Community Manager

Joined:

Feb 27, 2013

You may find Jian's 3 JMP Pro Linear Regression videos useful.  The first is on random coefficients models.

http://www.jmp.com/about/events/mastering/ondemand_webcast.shtml?reglink=mstrAdvancedLinearMixedMode...