I need to make overheads to illustrate overlapping distributions with
same/different regression slopes/intercepts in the X-Y plane. Generating
random data, and graphing linear regressions for same/different
slope/intercepts, is of course no problem. The problem is that I usually
don't want to show the points, I want to show ellipses (a single one,
really) for the density of the bivariate data.
I can get a 2D contour plot to show such a thing, but it shows a messy
numeric label for the contour. I presume I can get rid of this if I try.
I can also get just one ellipse to print out, I assume. However, there are
two other problems which the documentation doesn't seem to help me lick.
First, creating a contour plot gives confidence ellipsoids for the random
data, which aren't smooth for extreme regions of the plot. I want
"pretty," ideal pictured ellipses so the student isn't distracted from the
point.
Second, I don't see how to superimpose two plots, one with the regression
line on top of the other, the contour.
Any ideas on how to generate the idea, sampling-error free ellipses? Or do
I just need to grab a computer drawing package and draw lines & ellipses
the way I want them?
William M. Grove | "This is one of those times when we have to
Acting Director, | rise above principle."
CSPR Program | White House Spokesperson
Dept. of Psychology |
University of Minnesota|
http://keys.pgp.com:11371/pks/lookup?op=get&search=0xD8A7B6BB ->RSA key,
ldap://certserver.pgp.com ->D-H key; e-mail me for fingerprint.
-----------------------------------------------------------------------
This message was distributed by s-news@wubios.wustl.edu. To unsubscribe
send e-mail to s-news-request@wubios.wustl.edu with the BODY of the
message: unsubscribe s-news
|