Re: graph of overlapping normal distributions

[Mark4Flies@aol.com]

In a message dated 1/17/06 3:41:30 PM, steve@nmr.mgh.harvard.edu writes:


> Mike,
>
> Check the Sample Scripts directory that came with JMP.   There's a script
> called "demoPlotProb.jsl" that contains the commands for plotting several
> distribution functions (some interactive).
>
> Here's a bare-bones version of the script for plotting a normal
> distribution for two means and SDs (A and B) (i.e. two curves):
>
> (start)
> // Normal Density
> A.mu = 0; A.sigma = 1;
> B.mu = .5; B.sigma = 1.1;
>
> NewWindow("Normal Density",
>     Graph(
>         FrameSize(500,300),
>         XScale(-3,3),
>         YScale(0,1),
>         YFunction(Normal Density((x-A.mu)/A.sigma)/A.sigma,x),
>         YFunction(Normal Density((x-B.mu)/B.sigma)/B.sigma,x)
>     )
> );
> (end)
>
> At least it's a starting point. You'll have to experiment to see if you
> can set line sizes and colors.
>
> Steve
>
>
>
> On Tue, 17 Jan 2006, Michael Bailey wrote:
>
> > In order to demonstrate the concept of effect size to my class, I 
> > want to show them two figures, each of overlapping normal 
> > distributions. (Same mean differences, different variances, to show 
> > why variance is part of effect size). jmp makes it very easy to 
> > generate random normal data, although I haven't figured out yet how 
> > to change the SD. More importantly, I can't figure out how to graph 
> > two distributions on the same axes.
> >
> > Thanks for any help.
> > Mike Bailey
> >
> >
> >


You might also find the animation useful in the DOE > Sample Size, Power calculator for one mean. Note this demonstration correctly applies the non-central t-distribution to the alternative hypothesis.