Re: Intersection of two Normal Distribution

From: Richard Ulrich (Rich.Ulrich_at_comcast.net)
Date: 02/28/05 

Date: Mon, 28 Feb 2005 17:15:32 -0500

On Mon, 28 Feb 2005 12:44:46 GMT, "Soumyadip Rak***"
<soumyadiprak***@gmx.net> wrote:

> Hi,
>
> I have two normal distributions and I have found their point of
> intersection. Could cou please tell me how to find the 95% confidence
> interval about this point.

If they have unequal variances, and unequal Ns (or unequal
areas under the curves), then there will be TWO points of
intersection for two normal curves, or there will be none;
there won't be just one. Or do you have in mind some
other intention than "equal density" when you say
"point of intersection"?

That's just one reason why you need to be more specific
about what you have and what you hope to accomplish.

How would you *describe* the 95% CI?

You might be starting with this. -
"The observed 'overlap' between Normals and
Experimentals is 15% above/ below a cutoff, for the
respective groups; what is the 95% CI on that proportion?"

 - For equal Ns and variances, I think you could use the
SE on the difference to put a CI on the observed difference;
that would translate only approximately to a fraction.

SO, what do you want? 

-- 
Rich Ulrich, wpilib@pitt.edu
http://www.pitt.edu/~wpilib/index.html
Quantcast