Elena Llaudet points us to this new textbook for complete beginners. Here’s the table of contents:
1. Introduction [mostly on basic operations in R]
2. Estimating Causal Effects with Randomized Experiments [goes through an experiment on class sizes and test scores]
3. Inferring Population Characteristics via Survey Research [goes through the example of who in U.K. supported Brexit]
4. Predicting Outcomes Using Linear Regression [example of predicting GDP from night-time light emissions]
5. Estimating Causal Effects with Observational Data [example of estimating effects of Russian TV on Ukrainians’ voting behavior]
6. Probability [distributions, law of large numbers, central limit theorem]
7. Quantifying Uncertainty [estimation, confidence intervals, hypothesis testing]
And the whole thing is less than 250 pages! I haven’t looked at the whole book, but what I’ve seen is very impressive. Also, it’s refreshing to see an intro book proceeding from an entirely new perspective rather than just presenting the same old sequence of topics. There are lots of intro statistics books out there, with prices ranging from $0 to $198.47. This one is different, in a good way.
Seems like a great first book on statistics, especially for the social sciences—but not really limited to the social sciences either, as the general concepts of measurement, comparison, and variation arise in all application areas.
After you read, or teach out of, Llaudet and Imai’s new book, I recommend our own Regression and Other Stories, which I love so much that Aki and I have almost finished an entirely new book full of stories, activities, and demonstrations that can be used when teaching that material—but Regression and Other Stories is a lot for students who don’t have previous statistical background, so it’s good to see this new book as a starting point. As the title says, it’s a friendly and practical introduction!
Kosuke — with various coauthors — has written four different flavors of a quantitative social science textbook with those concepts. The original one used R code but few R packages besides the ones that come with R. Then there was one with Stata, a version with the tidyverse R packages, and now this one that is even more introductory. I have used the original one and now the tidyverse one mostly to create homework problems for master’s and Ph.D. classes, even though I would say the material is undergraduate.
Ben:
I saw Kosuke’s earlier book and liked it. It seemed to me to be an excellent presentation of standard material, with more of the good stuff and less of the bad stuff compared to other intro statistics books that I’ve seen. This new one seems closer to revolutionary to me in how cleanly it is structured. It makes sense that when writing a series of similar books, Kosuke and his colleagues have been improving!
OK, but who actually reads books anymore?