Andrew Gelman

(I don't think Andrew is teaching this semester, but Val is. She was on his team last semester and is great, you should take her class.) Having had a beast of a Stat teacher in high school (I'm talking about you Mrs. Campbell you ogre, wherever you are!) who I literally learned ZERO from, I was not looking forward to having to take Statistics. At all. To the reviewer below I say the following: the first day of class Andrew told us he would be teaching the class in a different way to the other sections, and that if you wanted a more traditional approach to Stat (in a class that would not be using R) then you should switch to any another section. So, perhaps you should have switched to another section then? I decided to stick with Gelman's class, and am glad that I did! (By the way, I have no computer programming experience whatsoever, and absolutely loathe math.) Andrew, along with Val (who was a HUGE help with the class), Vince the TA, and Daniel the R programming guru, taught statistics in a way that illuminated why the subject is in fact important and exciting as well. Yes the weekly homework assignments sometimes took a long time. Yes the book wasn't always as clear as it could be. But, we were told from the beginning that Andrew was in the process of writing a new book, and that would be the "textbook" that we would use. So, all students knew that the book/class was a work…

I've taken a bunch of statistics classes at the 4000 level and this is the best. Gelman is the best teacher I've encountered at Columbia. He makes enjoyable what could be a truly dry subject. He runs the class with ease and confidence, answering questions with informative digressions and performing calculations on the fly (and showing you how to do these things quickly in your head). It's a very interactive class, without the "fourth wall" that most professors hide behind. You will probably find yourself fully engaged in the class and you might actually enjoy it -- something you probably don't expect from a stats class. There is plenty of work in this class: homework every week, plus 5 minute quizzes at the beginning of most classes to make sure you've actually done the reading, and 10-minute quizzes once a week. Gelman uses his own textbook (not quite yet published), a very handy book with lots of practical tips on how to use statistics, including R code. He also supplies students with a detailed outline for the entire course at the beginning of the semester, so you don't have to spend so much time taking notes during class.

I learned how to fit and build hierarchical multilevel regression models in R.

I think the course has a good intention in introducing regression analysis and applications in political science, but it can raise more confusion than clarity especially in the advanced topics, like Bayesian analysis and nonlinear models. The textbook can be dense in these chapters, and the instructor struggled with giving explanations that are easy to comprehend in class. 1 Columbia University: Arts & Sciences Fall 2024 Course: POLSGU4720_001_2024_3-QUANTMETH1APPLREGCAUSINF : POLSGU4720_001_2024_3 - QUANT METH 1 APPL REG CAUS INF Instructor: Andrew Gelman

how to actually do statistics and pitfalls that statisticians fall into

Learned a lot about regression and its role within social science. Also learned introductory and basic R skills (sort of...).

THE HORROR, THE HORROR… Abandon all hope ye who enter here, for ye will feel like Sisyphus, continually rolling a rock up a hill, only to watch it roll back down. At least Sisyphus though was being punished for what he did. This is more like Kafka. Watch! Just watch! as you are metamorphosed from the primarily positive person into an angry bridge troll, lacking sleep while kicking people in the shins for HAVING THE GALL to smile. In this class, you will learn the fine line that can often exist between tragedy and comedy…you will start to laugh if only to keep from crying. But enough with might seem to the uninitiated like hyperbole, but so very very unfortunately isn’t. It is not as if Mr. Gelman does not know the material, or even doesn’t care (though complaints fall on deaf ears). No. He is just teaching the class wrong. It is not as if the material is difficult—it’s rather that I could learn far better with an Idiot’s Guide to Statistics than with his manner of teaching. His lectures will take a simple concept, and in the interest of making it simpler by providing illustrative examples, actually ends up twisting the material into an unrecognizable glop, leaving you scrambling to ask your friends “What was this class about….? He once made an analogy that learning statistics is like learning a language, where you learn individual phrases and eventually put the rules together.…

One problem set every week (non-trivial), plus short quizzes at the start or end of most classes.

- Practical decision making in applied statistics

Doing another introductory class to statistics might be a better option. 1 Columbia University: Arts & Sciences Fall 2023 Course: POLSGU4720_001_2023_3-QUANTMETH1APPLREGCAUSINF : POLSGU4720_001_2023_3 - QUANT METH 1 APPL REG CAUS INF Instructor: Andrew Gelman