Re: Experimental Research In Education: The Most Exciting Talk at the 2005 Joint Statistical Meetings

comments:

Since it is impossible to randomly assign to race, any analysis would have to be quasi-experimental (aka non-experimental, aka observational),

In quasi-experimental work a great deal of the work is to rule out plausible rival hypotheses that would be ruled out by a random assignment to treatment. The link below identifies many of these plausible rival hypotheses.

http://appliedpersonnelresearch.com/papers/adimpact.pdf

Art Kendall
Social Research Consultants

Donald Macnaughton wrote:
Referring to my February 11 post, Michael Granaas wrote
(in EdStat)

I am wondering why this is being circulated. The opening comments don't provide me with an adequate context to judge whether I should just be reading for information or if there is a desire for comments.

I would be very interested in and appreciative of any comments that readers have about the ideas. The full post is at
http://www.matstat.com/teach/p0048.htm

I'm especially interested in comments about the discussion of whether human performance or behavior can be predicted from a person's race. That discussion is available at http://www.matstat.com/teach/p0048.htm#AppC


Don Macnaughton
.

Loading