New features for Distribution in JMP 16

[#1]

The Distribution platform can be used to explore the distribution of a single variable. This is a rich platform that allows many different visuals, statistics, and tests on variables. This blog post details the new features to the Distribution platform in JMP 16.

Student's t Distribution

The Student’s t distribution has been added to the list of available continuous distributions that can be fit to variables. The Student’s t distribution is a robust option that spans the space between a normal distribution and a Cauchy distribution. As the degrees of freedom in the Student’s t distribution approach infinity, the distribution is equivalent to the normal. When the degrees of freedom in the Student’s t distribution equals 1, the distribution is equivalent to the Cauchy. The default behavior is to estimate the degrees of freedom for the Student's t distribution. As an alternative, you can choose to use the Fix Parameters option to specify a particular DF (degrees of freedom).

As an example, run the following JSL:

dt=Open("$SAMPLE_DATA/Body Fat.jmp");
obj=dt<<Distribution( Continuous Distribution( Column( :"Neck circumference (cm)"n ) ) );

This code opens the Body Fat sample data table and performs a distribution analysis on Neck circumference (cm). From the red triangle next to Neck circumference (cm), select Continuous Fit->Fit Student’s t.

Better broadcast functionality for prediction intervals, tolerance intervals, and custom quantiles

In previous versions when prediction intervals, tolerance intervals, and custom quantiles were broadcast across responses and by groups, you would get multiple dialogs for every response and grouping.  Beginning with JMP 16, you get a single dialog that applies to all groups.

 As an example, run the following code:

dt=Open("$SAMPLE_DATA/Cities.jmp");
obj=dt<<Distribution(
	Continuous Distribution( Column( :POP ) ),
	Continuous Distribution( Column( :OZONE ) ),
	By( :Region )
);

This code opens Cities.jmp. It performs a distribution analysis for POP and OZONE by Region. Hold down the Ctrl key while clicking on the red triangle next to POP and select Prediction Interval. Click OK. In previous versions of JMP, you would have been presented with 12 dialogs (one for each response and by level combination). You would have had to click OK on each dialog. Now, the single dialog applies to all 12 combinations.

In addition to prediction intervals, this functionality was added to tolerance intervals and custom quantiles.

Show Within Capability option and preference

Users requested the ability to obtain a capability analysis in Distribution that by default does not show the Within Sigma Capability report. In JMP 16, a Show Within Capability check box was added to the Process Capability dialog and a Show Within Capability option was added to Distribution preferences.

To see an example, run the following code:

dt=Open("$SAMPLE_DATA/Cities.jmp");
obj=dt<<Distribution(
	Continuous Distribution(
		Column( :OZONE ),
		Fit Normal
	)
);

This code opens the Cities sample data table and fits a normal distribution to OZONE. From the red triangle next to Fitted Normal Distribution, choose Process Capability.

Enter 0.075 for LSL.  Enter 0.15 for Target.  Enter 0.25 for USL. Click on the open/closure icon next to Process Capability Options. Uncheck the box for Show Within Capability. Click OK.

Capability analysis is added to the report window, but the Within Sigma Capability section of the report is not shown.

JMP 16 introduced three main new features for the Distribution platform:

1. The Student’s t distribution.
2. Better broadcast functionality for prediction intervals, tolerance intervals, and custom quantiles.
3. Options to get a capability analysis with no within sigma capability statistics by default.