I have created 2 files to demonstrate my problem.
In "Test 1" I have data from 4 people (i.e., AAAA, BBBB, CCCC, DDDD). Each person has 8 responses to an "estimate" question and 8 responses to a "truth" question. Half the responses are from Domain "0" and the other half from Domain "1".
I would like to find the correlation of between the estimate and truth question for each person in each domain. I know that I can do that by going to Analyze / Multivariate Methods / Multivariate and putting the estimate and truth questions in the "Y, Columns" box, and both "ID" and "Domain" in the "By" box.
The problem is that I want the output of this analysis to be in another JMP table so that I can do further analyses on the correlations. For example, I want to know what the average persons' correlation is between estimate and truth in each of the domains. In order to do this, I need a file that looks like the data in the "Test 2" file.
Can you please tell me an automatic way to produce the "Test 2" data from the Test 1 data? I cannot do this manually because the actual dataset that I am working with is huge.
You could explot the 'makeCombinedDataTableMessage'. In JSL, something like:
NamesDefaultToHere(1);
// Data table
dt = DataTable("Test 1.jmp");
// Multivariate
m = dt << Multivariate(Y( :Estimate, :Truth ), By( :ID, :Domain ));
// 'm' is a list because of the 'By' variable: Get the report for the first level
mRep = Report(m[1]);
// Get the correlation matrix into a table dt2 (for all levels)
dt2 = mRep[MatrixBox(1)] << makeCombinedDataTable;
// Manipulate dt2 to remove redundancy
dt2 << setName("Correlation Coefficients For By Groups");
dt2 << deleteColumns({3, 4});
r2del = dt2 << getRowsWhere(:Truth == 1);
dt2 << deleteRows(r2del);
Column(dt2, "Truth") << setName("Correlation Coefficient")
