You are supposed to work in SMALL groups to pick an example of Simpson's Paradox and present it to the class. You are strongly encouraged to use technology in your presentation, but it is not actually required.
The website of the example we used in class on Thursday is
http://www.cawtech.freeserve.co.uk/simpsons.2.html . An electronic version of the megasearch list is available on Classhomework.com.
Everyone in your group must contribute. Your presentation must include an explanation provided by the weakest member of your group!
Electronic documents can be sent to me at Forensicslime at aol.com or brought in on CD or flashdrive. If you have a question about whether it will work on my machine, send it by 5:00 Tuesday evening. I'll get back to you.
Once you pick your example, post your selection here and identify your class period so no other group in your class picks the same example. First come-first served (and noone is fobbed off with a bad example). :)
CiCi you on Sunday!
Subscribe to:
Post Comments (Atom)
Posts
2 comments:
If you want to incorporate a Geometer's Sketchpad graph into your presentation, please come in early on Tuesday to load the data and build the graph.
Good job, all!
Parabolasolver-
The first point is (0,0). Then, for each of the two subpopulations, use (number of individuals in the first case, number of successes). For instance, the first two points that connect to (0,0) were (number of males who got the flu shot, number of flu-shot guys with the flu) and number of males who did not get the flu shot, number of them who got the flu). That part was straight-forward.
For the next pair of points you have to go a distance in the x and in the y directions based on the other subpopulation results. For instance, you extend out from the (x,y) point you used in the paragraph above like this:
for the NON-flu shot crowd
the next point is (number of men who didn't get the flu shot PLUS the number of women who didn't get the flu shot, number of men who got no shot but got the flu PLUS the corresponding number of women). Of course, in the case where you have only two subpopulations (M/F) this is the same as (total of all who got no shot, total of all who got no shot but got flu). If there are more than two subpopulations, then you have to go with cumulative totals.
Play with it a bit. It's a pretty neat relationship.
Post a Comment