Link to Sponsoring Departments Statistical Learning Records, Fall 2005

Dear Habermas Logo and Link to Site Index A Justice Site



Statistics: McCullagh, Richard A. :
Learning Records, Fall 2005

MIRROR SITES: CSUDH - Habermas - UWP
ISSUES AND CONCEPTS: Susan's Archive at UWP
Academic Resources - Daily Site Additions
Lectures - Notes - Texts - Self Tests - Discussions
Visual Sociology - Graduate Exam Study
POST TO: Tutoring - Learning Records - Transform-dom
SEARCH: Topics Index - Site Index - Issue Archives
Google Web Search - Google Site Search

California State University, Dominguez Hills
University of Wisconsin, Parkside
Created: January 10, 2006
Latest Update: January 10, 2006

E-Mail Icon jeannecurran@habermas.org
takata@uwp.edu

Index of Topics on Site Statistical Learning Records, Fall 2005
McCullagh, Richard A. :

Message 991 :

That was a good and simple way to spread information through art. The art work that had to do with sending the message about tests not being worth as much as people value them was a visitor's favorite. Tests may indeed suck because of the way they are formatted to measure the knowledge of a subject, but they are not useless. If presented in the correct manner they are helpful. They let the teacher know what is missing and what needs to be emphasized more. Too much pressure is detrimental to true learning, but a test here or there is not bad if the pressure is off.

I liked learning about patterns through the creation of the box around mid-semester. At first I though it was irrelevant, but I came to realize that, as with other concepts, there is a point to be made involving statistics. Pattern recognition is useful in real life as well. If one can figure out what others are looking for one will accomplish much more than someone just trying to slide by with their heads in the clouds. Love and Paz, Richard McCullagh

Reply:

Good summary, Richard. I like the "heads in the clouds" analogy. One of our objectives in this approach to learning is to capture your attention to make you think deeply about things you've taken for granted.

I also agree with you about tests. Look at the current issue for the Week of December 25. There are five self tests up. Hope you'll look at them. I think at least a couple related conceptually to statistics. Had I not had a reaction to the radiation, and had Schwarzenegger not demanded that bleep special election, I might have gotten up more of those self tests for you. Be sure to follow them as I get them up later in January. But no issue next week. I'm gonna need a break.
jeanne
December 27, 2005

Message 6132:

Everyone's threshold for pain is different. What may be borderline torture to one, may tickle another. The doctor is saying that whether or not the person is treated, there is a chance the pain will return to hurt them again. Therefore regression to the mean states that pain is difficult to measure and the pain, (treated or not) can come back again somewhere down the time line.

Reply:

Not quite, Richard. The doctor is saying that whether there is actual treatment, or a placebo, or nothing, over time the pain level reported will have a tendency to fall more closely to the average measure of pain reported. It has nothing to do with the pain coming back. Just that over time the pain reported will tend to be closer to the average. Look at the normal curve for an understanding of that. Most reports cluster around the average, with fewer and fewer reports 1 standard edeviation out, and even fewer 2 standard deviations out. Check the explanation in the article again.
jeanne
December 22, 2005



Creative Commons License
This work is licensed under a Creative Commons License.
Individual copyrights by other authors may apply.