Subscribe Bookmark RSS Feed

JMP Pro 11 crashes with this Mixed Model

kaylesawyer

Community Trekker

Joined:

Nov 12, 2013

Hi, I'm using JMP Pro 11 with Mac OS X 10.9, and trying to learn how to used the Mixed Model personality. I'm using the Animals.jmp sample dataset, which has 6 subjects, each measured in the fall, winter, spring, and fall. I'm trying to predict miles roamed from species (between subjects) and season (within subjects). JMP crashes ("quit unexpectedly") every time I run this simple Mixed Model:

data = Open ("$SAMPLE_DATA/Animals.jmp");

data << Fit Model(

  Y( :miles ),

  Effects( :species ),

  Center Polynomials( 0 ),

  Personality( Mixed Model ),

  Subject( :subject ),

  Repeated Effects( :season ),

  Repeated Structure( Unstructured ),

  Keep Dialog Open( 1 ),

  Run( Repeated Effects Covariance Parameter Estimates( 0 ) )

);


JMP crashes when I run the script or specify it by hand in the GUI. Am I specifying the model correctly? I also tried adding :season to the Effects:

  Effects( :species, :season ),

but JMP still quits unexpectedly. Should season be nested by subject?


In other software packages (e.g. SPSS), I would structure the data like this:

data << Split(

  Split By( :season ),

  Split( :miles ),

);


In SPSS, you could then specify the repeated measure of 'season' with 4 levels. What would the way to do this with the Mixed Model personality?

Thanks!




1 REPLY
Jeff_Perkinson

Community Manager

Joined:

Jun 23, 2011

Hi Kayle,

I'm sorry that JMP Pro 11 is crashing on you that way. It looks like that might be a bug. I'll get our testing group to look into it.

In the meantime, the correct way to specify the model is to use a column that uniquely identifies each subject. In this case, that's the concatenation of Species and Subject, which doesn't exist in the data currently.

The new transform columns options will allow you to create this concatenation on-the-fly though and specify that as the Subject.

Fit Model(

  Y( :miles ),

  Effects( :species ),

  Center Polynomials( 0 ),

  Personality( Mixed Model ),

  Subject(

  Transform Column(

  "Concatenate[species,subject]",

  Character,

  Formula( :species || " " || Char( :subject ) )

  )

  ),

  Repeated Effects( :season ),

  Repeated Structure( Unstructured ),

  Run( Random Effects Covariance Parameter Estimates( 0 ) )

);


-Jeff

-Jeff