Our World Statistics Day conversations have been a great reminder of how much statistics can inform our lives. Do you have an example of how statistics has made a difference in your life? Share your story with the Community!
Choose Language Hide Translation Bar
Highlighted
chalpern
Level I

Two questions about Custom two-factor, random block model table

I am new to JMP. I have generated a custom two-factor RBD model table for an ecological field experiment. See attached PDF of screen shots of my Custom Model-building journey.

The experiment consists of 4 spatially defined replicate blocks (Random Block), with two categorical, two-level treatments (+/- Limp and +/- Moist).  The four treatment combinations (+L+M, +L-M, -L+M, -L-M) were randomly assigned to the four plots that comprise each block. n=4 replicates (blocks), # observations=16.


1. Does it look like I've specified the model properly?

2. Why is the response column blank? I can paste in the relevant data, but I am looking for a way to automatically link the model table to the data table.

 

Thanks for any comments or suggestions.

10 REPLIES 10
Highlighted
chalpern
Level I

Re: Two questions about Custom two-factor, random block model table

 

Fit Model(
    Y( :Name( "MY87-Ja88" ) ),
    Random Effects( :Random Block ),
    Effects( :Name( "+/- Limp" ), :Name( "+/- Moist" ), :Name( "+/- Limp" ) * :Name( "+/- Moist" ) ),
    NoBounds( 1 ),
    Personality( "Mixed Model" ),
    Run
)
Highlighted
statman
Level VII

Re: Two questions about Custom two-factor, random block model table

Given the random selection of blocks, your model looks correct.  I don't understand the 2nd question?  Your data table must contain the "factors" (block, L & M) and any response variables. All of the data need to be in one table. Once you have the data table complete you can "run" your model.

Highlighted
chalpern
Level I

Re: Two questions about Custom two-factor, random block model table

I have a data table with factors and response variable(s). Do I need to paste the data from my existing data table into the model table? Or can I get the model script to "point to" that data table?
Highlighted
txnelson
Super User

Re: Two questions about Custom two-factor, random block model table

You need to cut and paste the data into the data table produced by the DOE.  That data table may have different Column Properties that have been set that will be used by the table scripts.

Jim
Highlighted
chalpern
Level I

Re: Two questions about Custom two-factor, random block model table

Thanks for your reply. I'm still confused about how to paste data into Model Table. Specifically, the Custom Design process produces a Model Table in which the sequence of treatment combinations differs from that in my master data table (see excerpts below). Is it OK to re-sort the Model Table to match the sequence of my master data table, so I can paste in the data in the correct order?

DOE  Master table  
Rand Blk+/- Limp+/- Moist#Blk+/- Limp+/- Moist
1
1
1
1
2
2
2
2
+L
+L
-L
-L
-L
-L
+L
+L
+M
-M
+M
-M
-M
+M
+M
-M
1
1
1
1
2
2
2
2
+L
+L
-L
-L
+L
+L
-L
-L
+M
-M
+M
-M
+M
-M
+M
-M

 

Highlighted
statman
Level VII

Re: Two questions about Custom two-factor, random block model table

Yes, just make sure the response variable matches the treatment combinations.

Highlighted
chalpern
Level I

Re: Two questions about Custom two-factor, random block model table

So when I add a covariate in DOE (but not in Fit Model), why do the covariate values and the block x treatment combinations get scrambled?
Highlighted
statman
Level VII

Re: Two questions about Custom two-factor, random block model table

Sorry, but I don't understand your question.  What I don't understand is you say "when I add a covariate in DOE" I don't know what that means?You should have a data table that includes all of the variables manipulated, blocks if you created them a column for each covariate and a column for each response variable.  When you analyze the covariate you should change the factors to nominal and leave the covariate continuous.  You should put the covariate in the model first and use Type I SS to evaluate the covariate.  Use Type III SS to evaluate the other terms in the model.

Highlighted
chalpern
Level I

Re: Two questions about Custom two-factor, random block model table

Thanks for your reply. Never mind the question about DOE! I figured out how to get the right model in Analyze > Fit Model. (script below)

 

However, the sequential tests view doesn't seem available for the Mixed Model personality in Fit Model. Am I missing something?

 

[This script runs on the same data set I described at the beginning of this thread, but with a covariate in another column.]

 

Fit Model(
    Y( :Name( "Ja88-Ap88" ) ),
    Effects( :Ja88, :Name( "+/- Limp" ), :Name( "+/- Moist" ), :Name( "+/- Limp" ) * :Name( "+/- Moist" ) ),
    Random Effects( :Blk# ),
    NoBounds( 1 ),
    Personality( "Mixed Model" ),
    Run( Repeated Effects Covariance Parameter Estimates( 0 ) )
);