POLS Statistics

Below you should see a bunch of directories, in which there are data sets and writeups on statistical topics. It was driving me crazy trying to keep separate directories for separate classes, so I've pooled them and you can take whatever documents suit your interest and aptitude. There are directories pols706 and pols707 that have syllabus and some course-specific material, but the statistical write-ups are tucked into the directories below.

Point and click away. I realize it may seem overwhelming, and I've been trying to coalesce this under a smaller set of headings so it is easier to navigate. If I were you, I'd check out my super writeups on mathematical basis of calculus and matrix algebra that are grouped under Mathematics, statistical distributions in Distributions, as well as my collection of writeups on elementary Regression, MultipleRegression and extensions of it in the family of the GLM Generalized Linear Model. The GLM encompases ordinary regression, logistic regression, count models(Poisson/Negative Binomial) and several other types of models. I still have separate directories on logistic/probit regression and DurationModels because those write-ups are free-standing (and some variants are are not actually members of the GLM family). I also have separate directories on NonlinearRegression, so-called nonparametric regression: Splines and Loess, and TimeSeries.

I usually post files in the pdf format when I have it, and I also post either a tex or lyx file from which that pdf is created. Sometimes I only have text, and I upload that.

You should also see below the newest version of my book project, Stuff Worth Knowing. That book incorporates some of the mathematical writeups that are mentioned above, but it is intended as a more thorough, complete treatment.

The Email list for statistical research at KU is now called stat-l. We encourage all students to join and participate. It can be joined by visiting http://lists.freefaculty.org/listinfo.cgi/sgat-l-freefaculty.org. The archive for that list is also available through a link on that site, which is http://lists.freefaculty.org/pipermail/stat-l-freefaculty.org/

Unix, Linux, R and Sas

I started a centralized "help archive" for computer related tasks in another account. Go here to look it over: Online Guides(FAQ) for Stats and Programming : In there you should find links to various FAQ documents I prepared as well as code samples for SAS and R.

If you are looking for my Rtips page, just Click HERE!

What can I help you with?

SASOriginally, when I started teaching at KU, I preferred SAS, pretty strongly. There are still some problems for which SAS is a preferred tool, especially if one is preparing for a research career in industry or state government, where SAS is still very widely used. SAS is an old framework and, perhaps it is just my "crabby old man" tendency speaking, but SAS does not make a very smooth transition onto the personal computer. If you want to use SAS, you probably have a problem big enough so that you need to use it on a really big computer, such as the "ires" system at KU.

RNow I'm pretty enthusiastic about R. It has all the freedom of SAS, less of the crap, easier programmability, and, best of all, it is free and open source. If you have the money to buy a stat program, I suggest you buy S+. S+ is the same language base as R, but it has a commercial quality graphical interface that it inherited when its company purchased a program called Axum (which I used throughout the 1990s) and put the S+ guts into the Axum graphical interface.

What about other programs? I always thought SPSS was flat-out no good. I not been a frequent Stata user, but I have run it 10 or 20 times. In my view, both SPSS and Stata are handicapped by a major weaknesses, the "one dataset at a time" philosophy.

      Name                    Last modified      Size  Description
Parent Directory - CXTS/ 24-Aug-2004 13:09 - Crosstabs/ 24-Aug-2004 13:11 - Distributions/ 04-Feb-2008 10:42 - DurationModels/ 24-Aug-2004 13:26 - GLM/ 03-Nov-2008 18:52 - LogisticRegression/ 03-Nov-2008 20:41 - Lyx-template-01.lyx 12-Sep-2006 09:40 4.6K MathExer2-curves.pdf 24-Aug-2004 13:23 56K Mathematics/ 03-Nov-2008 20:03 - MaximumLikelihood/ 24-Aug-2004 13:06 - MixedModels/ 10-Feb-2006 21:27 - MultipleRegression/ 03-Nov-2008 20:00 - NonlinearRegression/ 03-Nov-2008 19:21 - R-code/ 03-Nov-2008 19:27 - Regression/ 03-Nov-2008 18:48 - Rtutorial1.txt 30-Aug-2005 08:37 7.7K SAS-Code-snips/ 22-Jan-2001 08:40 - SAStutorial1.txt 29-Mar-2005 09:46 4.6K SamplingDistribution..> 07-Oct-2003 08:37 - Splines/ 03-Feb-2006 00:46 - StuffWorthKnowing.pdf 08-May-2008 10:16 2.1M Systems/ 01-Dec-2006 09:13 - TimeSeries/ 14-Nov-2005 14:14 - WinProgs/ 25-Jan-2008 12:58 - lyxExercise.pdf 29-Nov-2004 11:25 14K ps706/ 08-May-2008 10:44 - ps707/ 28-Jan-2008 19:15 -