JMP Script: Capability Ananysis in JMP 15

jan_solo_ff — Mon, 09 Mar 2020 16:08:26 GMT

Hi Everyone,

We recently upgraded from JMP 14 to 15.

One of our scripts is currently showing a different result when a capability analysis is done.

In JMP 14 we had the following result:

In JMP 15, the exact same expression suddenly shows the PPx values (note - this is from a different analysis, so different limits):

The code we use to generate the chart:

Distribution(
	Uniform Scaling( 1 ),
	Stack( 1 ),
	Arrange in Rows( 6 ),
	Automatic Recalc( 1 ),
	Continuous Distribution(
		Column( Eval( Column( i ) ) ), // i = loopcounter
		Quantiles( 0 ),
		Horizontal Layout( 1 ),
		Histogram( 0 ),
		Vertical( 0 ),
		Axes on Left( 1 ),
		Outlier Box Plot( 0 ),
		Customize Summary Statistics(
			Std Err Mean( 0 ),
			Upper Mean Confidence Interval( 0 ),
			Lower Mean Confidence Interval( 0 ),
			Maximum( 1 )
		), 
		Capability Analysis(
			// Column # in Limits table doesn't match columns # in result table
			USL( Num( Eval( dtDistLimits:_USL[i - 1] ) ) ),
			LSL( Num( Eval( dtDistLimits:_LSL[i - 1] ) ) )
		)
	),
	SendToReport(
		Dispatch( {}, "Distributions", OutlineBox, {Set Title( "Distributions" )} ),
		Dispatch(
			{},
			"Mean(Result",
			OutlineBox,
			{Set Title(
				sComponentName || " Specification: (LSL " || sLslDisplayValue || ", USL " || sUslDisplayValue
				 || ")"
			)}
		),
		Dispatch(
			{"Distributions", "Mean(Result", "Summary Statistics"},
			"",
			NumberColBox,
			{Set Format( 9, 3 )}
		),
		Dispatch( {"Distributions", "Mean(Result"}, "Capability Analysis", OutlineBox, {Set Title( "" )} ),
		Dispatch(
			{"Distributions", "Mean(Result", "Capability Analysis"},
			"Specification",
			StringColBox,
			{Name( "Hide/Unhide" )(1)}
		),
		Dispatch(
			{"Distributions", "Mean(Result", "Capability Analysis"},
			"Value",
			NumberColBox,
			{Name( "Hide/Unhide" )(1)}
		),
		Dispatch(
			{"Distributions", "Mean(Result", "Capability Analysis"},
			"Portion",
			StringColBox,
			{Name( "Hide/Unhide" )(1)}
		),
		Dispatch(
			{"Distributions", "Mean(Result", "Capability Analysis"},
			"% Actual",
			NumberColBox,
			{Name( "Hide/Unhide" )(1)}
		),
		Dispatch(
			{"Distributions", "Mean(Result", "Capability Analysis", "Long Term Sigma"},
			"1",
			ScaleBox,
			{Min( Eval( iMinScale ) ), Max( Eval( iMaxScale ) ), Inc( 0.5 ), Minor Ticks( 5 )}
		),
		Dispatch(
			{"Distributions", "Mean(Result", "Capability Analysis", "Long Term Sigma"},
			"Lower CI",
			NumberColBox,
			{Name( "Hide/Unhide" )(1)}
		),
		Dispatch(
			{"Distributions", "Mean(Result", "Capability Analysis", "Long Term Sigma"},
			"Upper CI",
			NumberColBox,
			{Name( "Hide/Unhide" )(1)}
		),
		Dispatch(
			{"Distributions", "Mean(Result", "Capability Analysis", "Long Term Sigma"},
			"Portion",
			StringColBox,
			{Name( "Hide/Unhide" )(1)}
		), 
								// THIS DISPATCH GENERATES ERRORS
		Dispatch(
			{"Distributions", "Mean(RESULT", "Capability Analysis", "Long Term Sigma"},
			"Percent",
			NumberColBox,
			{Name( "Hide/Unhide" )(1)}
		), 
								// THIS DISPATCH GENERATES ERRORS
		Dispatch(
			{"Distributions", "Mean(RESULT", "Capability Analysis", "Long Term Sigma"},
			"PPM",
			NumberColBox,
			{Name( "Hide/Unhide" )(1)}
		),
		Dispatch(
			{"Distributions", "Mean(RESULT", "Capability Analysis", "Long Term Sigma"},
			"Sigma Quality",
			NumberColBox,
			{Name( "Hide/Unhide" )(1)}
		)
	)
); // Distribution

s there something I can do get the correct (old) results again?

Thank You,

Jan

Re: JMP Script: Capability Ananysis in JMP 15

David_Burnham — Mon, 09 Mar 2020 16:52:00 GMT

You can turn off the Ppk labelling using the PkK Capability Labeling message.

Either

Names Default To Here( 1 );
dt = Open( "$SAMPLE_DATA/Quality Control/Coating.jmp" );
obj = dt << Distribution( Column( :Weight ) );
obj << PpK Capability Labeling( 0 );
obj << Capability Analysis( LSL( 16 ), USL( 24 ), Target( 20 ) );

Names Default To Here( 1 );
dt = Open( "$SAMPLE_DATA/Quality Control/Coating.jmp" );
obj = dt << Distribution( Column( :Weight ),
	PpK Capability Labeling( 0 ),
	Capability Analysis( LSL( 16 ), USL( 24 ), Target( 20 ) )
 );

Blame AAIG for the mess which is modern day capability analysis.

My favourite quotes:

The relatively new introduction of the notation Pp and Ppk and the redefinition of Cp and Cpk without stakeholder consensus have caused considerable confusion. In our experience, many auditors and practitioners are unaware that what the AIAG calls Ppk has been in use in industry for about 30 years as Cpk.

http://www.sematech.org/docubase/document/Semiconductor_Industry_SPC_Practices.pdf

Or, put another way

The process performance indicesd Pp and Ppk are a step backward. They are a waste of engineering effort and management effort - they tell you nothing.

- D.C Montgomery

Or, put more strongly

The mandated use of Pp and Ppk through quality standards or industry guidelines ias "statistical terrorism"

- Kotz and Lovelace

Luckily we have a switch and can turn them off :)

topic JMP Script: Capability Ananysis in JMP 15 in Discussions

JMP Script: Capability Ananysis in JMP 15

Re: JMP Script: Capability Ananysis in JMP 15