topic Re: Fit least square ==> asking for control level. in Discussions

Fit least square ==> asking for control level.

yanee — Sun, 11 Jun 2023 11:10:16 GMT

Hi,

I wrote below script (simple fit least square but the effects are categorical) and it worked fine on my system (JMP 15.1). When my college try this script on his computer (JMP 15.0).

When he ran this script, it showed pop up menu asking for control level - see picture.

I manually plotted the least square on his computer and it showed the same message - asking for control level.

If I clicked cancel, then the fit least square showed up.

How do we turn off this control level thing ?

for (i=1, i <= n items (temp_list), i++,
		vlb1 << append(
		fit = Fit Model(
		Y( temp_list[i] ),
		by(:class),
		Effects( :Name( "~DEVICE TYPE" ), :Name( "~STAGE" ) ),
		Personality( "Standard Least Squares" ),
		Emphasis( "Minimal Report" ),
		Run(
			Profiler(
				1,
				Confidence Intervals( 1 )
			)
		)
		
		);
		fit << prediction formula;
		)	
);

Re: Fit least square ==> asking for control level.

jules — Fri, 22 Jan 2021 15:22:42 GMT

Hi @yanee,

It's hard to know for sure why this is happening, but I suspect your colleague has set a preference to perform Dunnett's test automatically, which requires the specification of a control level. That preference can be found under Preferences > Platforms > Effect Fit.

If you would like to ignore platform preferences when running your script, you can use Ignore Platform Preferences( 1 ) in your call to Fit Model(). Here's what that looks like:

for (i=1, i <= n items (temp_list), i++,
		vlb1 << append(
		fit = Fit Model(
		Ignore Platform Preferences( 1 ),
		Y( temp_list[i] ),
		by(:class),
		Effects( :Name( "~DEVICE TYPE" ), :Name( "~STAGE" ) ),
		Personality( "Standard Least Squares" ),
		Emphasis( "Minimal Report" ),
		Run(
			Profiler(
				1,
				Confidence Intervals( 1 )
			)
		)
		
		);
		fit << prediction formula;
		)	
);

I hope this helps!

@jules