SAS to JMP, small Script Code conversion

500 Points:  

Question:  How to code the following line in a simple JMP Script.   This is a from a SAS Script.  

In SAS, this line is called: Tests of Hypotheses Using the Type III MS for JUDGE*SAMPLE as an Error Term

How do I apply that one liner to JMP?  

My SAS Code works perfectly:

/* SAS */   /* Calculate P-Values  */    /* Incorporate Panelist Consistency Over Reps */

My incomplete JMP code:  

rfitm = Fit Model(
      Y( :name( "The Fruity Aroma" ) ),
      Effects( :Judge, :Judge * :Sample, :Rep, :Sample ),
      Personality( Standard Least Squares ),
      Emphasis( Minimal Report ),  
      Run Model( :name( "The Fruity Aroma" )
         << {:SAMPLE << {LSMeans Tukey HSD( 0.0500 )}} )
      )<< report;

This JMP code will generate near idential outout as the SAS code, except for the missing lines starting with:   Tests of Hypotheses Using the Type III MS for JUDGE*SAMPLE as an Error Term

Additional information, such as easy example JMP data & script file available if needed.
Who is Participating?
I wear a lot of hats...

"The solutions and answers provided on Experts Exchange have been extremely helpful to me over the last few years. I wear a lot of hats - Developer, Database Administrator, Help Desk, etc., so I know a lot of things but not a lot about one thing. Experts Exchange gives me answers from people who do know a lot about one thing, in a easy to use platform." -Todd S.

JReamAuthor Commented:
My SAS person just told me a that my question should be worded something like, How can I specify the error term that should be used for the F-Test in an ANOVA and the multiple comparison procedure to compare the levels of the effect?  Does JMP have a command equivalent to SASs TEST H=<effect> E=<error> command and its /E=<error> option in PROC GLM?
From the SAS/JMP FAQ:

Why does my output from JMP differ from that of PROC GLM's output?
Differences comparing PROC GLM from SAS to JMP occur when you have unequal cell sizes or random effects. This is due to the different parameterizations and the two algorithms used to calculate the statistics. JMP's algorithm is computationally more efficient, but GLM's is more general. Both are correct, just different. To answer any comparison questions, refer to the secion "Singularities and Missing Cells in Nominal Effects" in Appendix A of the JMP Statistics and Graphics Guide.


I don't use either SAS or JMP, but you have to wonder what the FAQ means when it says the two results are both correct, just different. You pointed out that SAS PROC GLM uses Type III sums of squares, which -- I don't know if you know this, but it's important -- is an appropriate method when you have unequal cell sizes in your data matrix, it uses hierarchical regression to estimate a proper Sum of Squares.

It looks like JMP uses Type I or Type II, I can't tell which because I don't have the user manual.

Your problem centers on this question of which is the appropriate sum of squares for unbalanced cell counts. All I can do is tell you that Type III is my preferred method, and lots of smart people would agree, because it takes the entire data matrix into account, whereas Type I goes to the cell means. The results are equivalent when cell sizes are equal, but they disagree as differences among cell counts get more pronounced.

Here's a bare-bones but useful wiki discussion of the definitions of the various types of error in ANOVA.

Finally, I'm not too optimistic that JMP can do what you're hoping it can do, based on the answer to the FAQ that SAS puts out there. They seem to be explaining it away rather than telling you how to get the results to come out the same. Again, I don't have the JMP manual, but they seem to refer you to an appendix in that manual. Check out your user manual if you have it, and focus on getting JMP to deliver Type III sum of squares (or mean squares, the concept is the same).

Good luck to you, hope this helps.

Experts Exchange Solution brought to you by

Your issues matter to us.

Facing a tech roadblock? Get the help and guidance you need from experienced professionals who care. Ask your question anytime, anywhere, with no hassle.

Start your 7-day free trial
JReamAuthor Commented:
Thanks for the reply jtm.  Quite helpful.  I'd like to leave the question open for a while longer to see if we can get input from any JMP users who hopefully may see the question.

Question is:  With JMP, how can I specify the error term that should be used for the F-Test in an ANOVA and the multiple comparison procedure to compare the levels of the effect?  Does JMP have a command equivalent to SAS's TEST H=<effect> E=<error> command and it's /E=<error> option in PROC GLM?
JReamAuthor Commented:
In closure, it appears that JMP is NOT as capable as SAS.  Which really is not a big surprise considering the scope and cost differences between the two products.  
It's more than this solution.Get answers and train to solve all your tech problems - anytime, anywhere.Try it for free Edge Out The Competitionfor your dream job with proven skills and certifications.Get started today Stand Outas the employee with proven skills.Start learning today for free Move Your Career Forwardwith certification training in the latest technologies.Start your trial today
Programming Theory

From novice to tech pro — start learning today.