Create the new column in your data table. Make sure you go into the Col Info, and set the format for the column to "PValue".
Then open the Formula Editor for the new column and paste the below column into it
dt = Current Data Table();
// change the Index value for the starting column number of the columns to be analyzed
theRow = dt[Row(), Index( 5, N Col( dt ) - 1 )];
// Create a temporary working data table to calculate the normality on
dtTemp = New Table( "temp", private, New Column( "x", values( theRow ) ) );
// Run the Distribution Platform to calculate the normality
dis = Distribution( invisible, Continuous Distribution( Column( :x ), Fit Normal( Goodness of Fit( 1 ) ) ) );
// Retrieve the P Value of the Normality statistic
pval = (Report( dis )["Goodness-of-Fit Test", Number Col Box( 2 )] << get)[1];
// Clean up the behind the scenes working table and display
dis << close window;
Close( dtTemp, nosave );
// Apply the pvalue to the current row
pval;
This formula will take several minutes to run on a 2000 row table.
The comments in the above formula will not be saved as part of the formula(it;s a JMP efficiency, to keep only the executable part of the formula).
Normally, JMP data is formatted 90 degrees from the way your data are structured. The column contain the data to be analyzed, and the rows represent the different instances for the given measurment.
Jim
