An assortment of comments and questions regarding stats for undergrads:
I'm surprised Gotelli and Ellison's "A Primer of Ecological Statistics" did not make the list. Most grad students seem to have a copy of this kicking it around. Do people feel its written at a level appropriate for undergrads?
Also, Dalgaard seems like a rough book to use as a supplemental text, unless its just for getting code. Has anyone used this in an undergrad course?
Regarding, code, I believe there are programs that can act as a more user-friendly front-end to R, but are not familiar with them.
The Journal of Applied Ecology has been publishing some well-written articles on good statistical practice. There may be some good articles for supplemental reading, http://www3.interscience.wiley.com/journal/122683826/abstract , but may still be too advanced for undergrads to really digest.
I've considering working on a statistics-teaching project for my graduate teaching minor and have been thinking a lot about what I would've liked to have known about stats as an undergrad. A couple of ideas 1)would it be useful to expose undergrad to some of the statistics peer-reviewed literature, such as statistics special features in Ecology? Not that they need to be able to understand everything, but so they can at least see how statistics is presented in the literature (and how often it is incomprehensible and needs to be slowly picked apart)
2) would it be useful to expose student to the different ways statistical models are presented, such as typical equation form (y = beta1*X1 + beta2*x2 + beta3*x1*x2) and also expose them to matrix notation (Y = Xb). I had no clue about matrix algebra when I started taking advanced stats classes and was totally bewildered by references in journals to vectors, covariance matrices, etc. Undergrads don't need to be able to use matrix algebra, but it seems like it would be good to briefly expose them to it.
3)it seems like starting a class off in excel and then transitioning to another program might be a route for building up skills and understanding. You can bust out regressions, t-tests, and one-way ANOVAs in Excel really easily. You could then introduce multiple regression and the software that can handle it by asking "ok, so what if we have TWO independent variables we want to include in a regression?"
4)Would it be useful to provide students a bibliography of articles they should read if/when they go to grad school, such as Hurlbert's "Pseudoreplication and the Design of Ecological Field Experiments." This could perhaps be paired with some philosophical discussion of about issues like pseudoreplication, co-linearity, autocorrelation, and other sneaky problems that they may encounter if they use statistics later in their career.
