David Kim writes,
I chanced upon your blog about six months ago . . . I’d like to learn more about the statistical techniques and principles I’m reading about, but have always found it difficult to do so solely from a book. Of course, the fact that my last math class was more than 15 years ago doesn’t help. Short of kidnapping a graduate student and forcing him to teach me R, Bayesian probability, hierarchical regression, etc. for food, I’m wondering what “continuing education” options might be open to me (formal or informal).
My reply: I don’t know! Maybe there’s a local university and you could take a class? Also, local American Statistical Association chapters sometimes offer short courses. Beyond that, I don’t know what’s available.
I think that Jim Albert's recent book, "Bayesian Computation with R," would be a very good and approachable introduction to the subject. You'd have to download R and LearnBayes to your computer, but following the examples you'd learn a lot.
How about MIT's open courseware:
http://ocw.mit.edu/OcwWeb/web/home/home/index.htm
Several statistics classes listed, might be worth a try.
I have a two-step suggestion, although I don't know if it will help much if you have a mental block when learning from a book. I am currently in graduate school and I am taking the methods sequence (currently in linear models) and my last math class was algebra I in undergrad. The first step of my suggestion is to look at the syllabi (?) of a couple of people teaching whatever analysis you want to learn, and buy one of the textbooks. If you haven't had math for a while (like me) you will find yourself retreating to "Statistics for dummies" or some other introductory text to understand the basic concepts before you move on. Although there is a lot of information out there that explains concepts fairly well, you may still find yourself in a situation thinking "ok I know this now how do I apply it" and for that you may need to be able to consult with someone. One thing that I have found helpful with understanding the concepts 'beyond the books'– although I don't know if your interest is this involved–is to look for articles that you find interesting that apply the type of analysis have studied and force yourself to work through the methodological approach of the author (and replicate the work if at all possible–look for articles where the data is readily available). This will be painful the first couple of times you do it, but it can really make a lot of what some studies do more accessible. Hope that helps (and is not too rambling).
Hello all. Thanks for the suggestions, I'll definitely look into them. Although I'm sure grad students are far too busy with their own studies these days, if there are any in the NYC/CT area that are interested in some pocket money from tutuoring (like back in high school), please let me know.
Thanks again guys.
Interesting about MIT's open courseware. Berkeley offers web courses, too.
What's the value of having a real teacher these days? And why do we need MIT and Berkeley serving these courses when they could be offered from Bangalore?
I think know some of the reasons, but I'd love to hear Andrew's perspective on this.
Andy,
A course from Bangalore might be fine. Regarding the Berkeley course, it depends what you want to learn. They have a very traditional approach (one lecture on the average and the median, a lecture on the sd and another on "rms error", many lectures on se's and hypothesis testing). I focus more on computation, applications, and how to understand models.
On the other hand, most university stat courses are probably like the Berkeley class. I think you might do better to take a class in a dept of psychology, econ, or poli sci, where there is more interest in the ultimate application of the methods. Also you may get an instructor who is actively involved in using statistical methods in applied research.
Beyond this, with the live human instructor–and, maybe just as important, live human classmates–you get more opportunity for active learning (see here, for example). Again, though, you won't get this in all classes.
David,
Yes, we have MA students who could help, I think. Drop me an email.
Statistics.com offers many short courses on specific topics for a very reasonable price. They are 100% online and you can do them at your own pace, with the advantage that there is a prof and other students to discuss the problems with. I have used this to fill in the gaps throughout grad school.