For those who have taken 707 may take a look of a new approach called "boolean statistics" to deal with complex interactions;
For those who are taking 707, I hope the picture and the slides stimulate you to embrace R.
Best, Frank Paul Johnson wrote:
In case the pols 707 email list is not working for everybody, I'm sending this message individually to all of the people in the class roster as well as to the list. If you don't get 2 messages, it means you are not signed up properly.News of note:1. Last week some of you pointed out a typo to me in the R assignment that I announced to this email list last week. That has been corrected.In case you did not get that assignment, I suggest you go look in the online archive for the ps707-l list, which you can access from my ps707 web page. And fix your email list subscription.2. Although there was no class last Monday, I still expect everybody is working to learn R and make the computer do the right things. If you spent 10 hours last week and invest 10 hours this week, you should be proficient in input/output. By this time, it should becoming easy & direct for you to start Emacs, open a file of commands, run it through R, save results in text and/or a graphic format.If you did not get to that point yet, I urge you to invest your time wisely in developing these fundamental skills.I will for sure be in the lab Tuesday at 3:30-5:30. On Friday, there is a brief meeting scheduled for the public policy prelim, so I will not be in the lab at 3:30. I may get there by 4:15 or so and will remain until 5.3. Next monday, I plan to address many topics, one of which is the idea of a statistical distribution. You probably already know about the Normal, but there are plenty more, many of which are listed in the reading materials on the syllabus, some of which are not.In that regard, I learned about a new one today while reading the r-help email list. The "skew normal" distro is pretty neat, and you can follow up on it herehttp://tango.stat.unipd.it/SN/Intro/intro.htmland there's an R package for it, which I've not tried yet. I don't know if the package is in CRAN yet, but you can get it directly from the authors and build it if it is not.http://tango.stat.unipd.it/SN/I'll make sure the lab pcs have that package installed, whether it is in CRAN or not.pj
-- Cheng-shan Frank Liu Dept. of Political Science E-mail: ashan_AT_ku.edu University of Kansas Web: http://lark.cc.ku.edu/~ashan 1541 Lilac Lane, Room 504 Office: (785) 864-9019 Lawrence, Kansas 66044-3177 FAX: (785) 864-5700
Attachment:
APSAsc15.pdf
Description: Adobe PDF document