Hi Dave, I have a few drills available from http://had.co.nz/stat405 - see the right hand column, about half way down. They seem similar in spirit to what you're thinking of. You might want to look at the "Little Schemer" for a similar approach with a different programming language.
However, I'm not sure how pedagogically useful this approach is. If you break things down too finely, you don't teach the problem solving skills necessary to attack a new problem. Students will try and solve the problems as rapidly, using as little of their brain as possible. I also feel like these small problem fail to invoke any intellectually curiosity - why the heck should I care that mtcars has 32 observations and 11 rows? I'd suggest starting with a big problem that's of interest to the students - how do we detect spam? What determines the price of a used car on ebay? Do soap operas influence baby name trends? Are my facebook friends representative of the university as a whole? Then talk about how you might attack the problem in general, before getting to the concrete tools you'd use in R. Hadley On Fri, Dec 11, 2009 at 10:48 AM, David Kane <d...@kanecap.com> wrote: > Hi, > > I am teaching a one month class in applied statistics and want to > bring my students up to speed in R without devoting much/any lecture > time to R instruction. I think that the best way to do this is to > provide them with a lot of R questions for homework. These questions > would be numerous (there is a lot of material to cover), go from very > simple to somewhat complex, and focus on all the commands and options > that will be useful in applied work. Here are some of my initial > questions: > > ------------------------ > Q: Load the data from the cars data frame into the local workspace. > A: data(cars) > > Q: Find information about the cars data frame. > A: help(cars) > > Q: Calculate the dimensions of the data frame. > A: dim(cars) > > Q: What are the names of the variables? > A: names(cars) > ------------------------ > > Needless to say, the questions will become more complex, including the > writing of simple functions. I also want to provide answers to all the > questions that, in theory, could be used in an automated fashion to > check the students work. My current plan is to load these questions > (somehow) into the quiz module in Moodle (http://moodle.org/). > > Ideally, I would like this system to be usable by very large classes > and even in the context of distance learning. Student goes to a web > page, logs in and is presented with a page of questions (or a single > question). She figures out the answer in her R session and pastes in > the command (or result) into the answer slot on the webpage and pushes > a button (or does it for ten questions first). The server then > determines which questions she got right and which she got wrong. It > might then provide clues to the ones that she has wrong. Once she is > done, the professor gets a list of her results (how many right, how > many wrong, how many required more than one try and so on). > > For now, I am not building that system. (Has anyone already done so?) > Instead, I am just creating the collection of R questions/answers that > might go into such a system. I am aiming for around 1,000 questions. > So: Does anyone know of open sourced collections of R questions like > this which I might use? > > Thanks, > > Dave Kane > Adjunct Instructor, Williams College > > ______________________________________________ > R-help@r-project.org mailing list > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide http://www.R-project.org/posting-guide.html > and provide commented, minimal, self-contained, reproducible code. > -- http://had.co.nz/ ______________________________________________ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.