How to extract Parameter Estimates of Reciprocals (i.e. make into data table) derived from Bivariates Fit Special?

Question

First of all, my Bivariate Analysis works well. Unfortunately "Make Into Data Table"-Reporting doesn't (meaning, jmp does not create any output table).

I applied the following script:

dt = Current Data Table ();

biv = dt << Bivariate(

Y( :Y ),

X( :X ),

Fit Special(

xTran( "Reciprocal" ),

{Line Color( "BlueCyan" ), Line Width( 2 )}

));

//Note: Parameter Estimates are in TableBox3

rep = Report (biv [1]) [TableBox(3)]  << Make Combined Data Table;

Do you know the correct scipt for exporting the parameter estimates to a new table?

I really would appreciate your help so much!

Nick_PV_tech · Answer

Hi,

I am trying the solution proposed by Ian_JMP, but I am unable to extract the parameters from the 'bivRep' object. I am regressing Y by X and I have 5 'By' variables.

My log shows the following:

bivRep = biv << Report;
/*:

List( 5 elements ) assigned.

//:*/
dt2 = bivRep [1] [TableBox(3)] << Make Combined Data Table;
/*:

Cannot subscript Display Box in access or evaluation of 'bivRep[1][Table Box(3)]' , bivRep[1][/*###*/Table Box( 3 )]

In the following script, error marked by /*###*/
dt2 = bivRep[1][/*###*/Table Box( 3 )] << Make Combined Data Table

So I can get a list 'bivRep' with 5 elements (one for each by variable I presume), but the the indexing of the elements within the list doesn't work. What am I doing wrong?

Thanks.

ian_jmp · Answer

You were very close . . .

NamesDefaultToHere(1);

// Data

dt = New Table ("X and Y",

NewColumn("X", Numeric, Continuous, Formula(RandomNormal())),

NewColumn("Y", Numeric, Continuous, Formula(RandomNormal())),

NewColumn("By", Numeric, Nominal, Formula(RandomInteger(1, 3))),

AddRows(100)

);

// Biveriate analysis: biv is a list, one item for each level of the 'By' variable

biv = dt << Bivariate(

Y( :Y ),

X( :X ),

By( :By ),

Fit Special(

xTran( "Reciprocal" ),

{Line Color( "BlueCyan" ), Line Width( 2 )}

)

);

// A reference to the report layer: bivRep is also a list

bivRep = biv << Report;

// 'Make Combined Data Table' avoids having to message to each item in 'bivRep'

dt2 = bivRep [1] [TableBox(3)]  << Make Combined Data Table;