Dr. McClelland, thank you very much! It is very impressive to see applets like this that visualize an idea. For me this is a communications breakthrough that really gets through a point I have been making for 30 years. I just showed this to a person with no statistics background at all, a lawyer with strong math anxiety, she grasped the concept immediately.
Do the list members know of other visualization applets that clarify statistical ideas? I have found this one to be very useful: http://www.anaesthetist.com/mnm/stats/roc/ Does anyone know of applets that show what happens using other numbers of points to measure a construct? In my excitement, I can think of 3 immediate generalizations of this particular vizualization - - In the survey world: This clearly shows that a 2 point scale throws away a lot of variance. How would 3, 4, 5 points look? I am thinking in terms of "How many points on a response scale? e.g., Likert or extent?" In the psychometrics world: The visualization could also be extended to the minimum number of items in a scale when the item response scale had 2 points (right/wrong), 3, 4,5. In general: It could be extended to other shapes of regression lines. Art [EMAIL PROTECTED] Social Research Consultants University Park, MD USA (301) 864-5570 Gary McClelland wrote: > On 2/4/03 2:25 PM, in article > 3B4D60A9CF29C349B185DC404F988D1002290F13@MAIL2A, "Wuensch, Karl L" > <[EMAIL PROTECTED]> wrote: > > >>Thanks to all who have participated in the discussion on >>dichotomization. I have logged selected parts of the discussion into a >>document for both my students and my colleagues to read. It is available at >>http://core.ecu.edu/psyc/wuenschk/StatHelp/Dichot-Not.doc >><http://core.ecu.edu/psyc/wuenschk/StatHelp/Dichot-Not.doc> . If any of >>you would rather not have your comments included in that document, please >>let me know. >> > > > I've put together a Java applet that illustrates the issues. > > http://psych.colorado.edu/~mcclella/MedianSplit/ > > There is a slider that allows you to move the values of X smoothly from the > continuous analysis to the dichotomized analysis. Doing so you can see that > as the variance of X is thereby reduced, the correlation is reduced, but the > estimate of the slope remains unbiased. The effect of dichotomization is > simply to reduce statistical power. If X is normally distributed, the > effect of dichotomization is equivalent to discarding a random 50% > (approximately) of one's observations. I can't imagine why anyone would > want to do that. > > Note: to use the above applet, PC users will need to have downloaded Sun's > Java plug-in to replace the crippled Java that Microsoft distributes. Mac > users will need to be using OS X. > > gary > . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
