Sweet: I’m going to give a couple of introductory statistics tutorials for R (a shell-based software for statistical data analysis and visualization) at the LSA Summer Institute in Stanford, July 2007. The classes will take place towards the end of the presession.
The first class will introduce ways to do standard data analysis for continuous dependent variables and balanced designs (the oh-so-beloved ANOVA, a.k.a analysis of variance; as well as its relation to linear regression). The goal is to learn how R generally thinks and how to analyze your classical 2×2 design in it. The second class will deal with categorical data analysis, that is cases for which the dependent variable is categorical (i.e. speakers’ choice between two ways to realize the VP of an alternating ditransitive verb). We will discuss why ANOVA over proportions are problematic and how logistic regression and its modern relatives (mixed logit models) can helps us overcome the problems with ANOVA for categorical dependent variables. Among other things, modern methods also allow us to move away from balanced designs (though while being cautious).
The classes should be announced soon. Some will be given more than once. Come, join the R party. Some of the classes during the regular session will also be taught using R. You can take my classes to get a little bit of an intro before the other classes start.