Publications by Justin Esarey
High-Powered Statistical Computing On the iPad
It's winter break… and as any academic knows, breaks are “a good time to get work done.” For the Christmas break, many of us have to travel home to see family members. One of the great privileges of being an academic is that you don't necessarily need to be in your office to get research and administrative work accomplished. But for method...
4300 sym 8 img
Who Survived on the Titanic? Predictive Classification with Parametric and Non-parametric Models
I recently read a really interesting blog post about trying to predict who survived on the Titanic with standard GLM models and two forms of non-parametric classification tree (CART) methodology. The post was featured on R-bloggers, and I think it's worth a closer look. The basic idea was to figure out which of these three model types did a bette...
6222 sym 20 img
How much can we learn from an empirical result? A Bayesian approach to power analysis and the implications for pre-registration.
Just like a lot of political science departments, here at Rice a group of faculty and students meet each week to discuss new research in political methodology. This week, we read a new symposium in Political Analysis about the pre-registration of studies in political science. To briefly summarize, several researchers argued that political scienti...
6883 sym 52 img
Proposed techniques for communicating the amount of information contained in a statistical result
A couple of weeks ago, I posted about how much we can expect to learn about the state of the world on the basis of a statistical significance test. One way of framing this question is: if we’re trying to come to scientific conclusions on the basis of statistical results, how much can we update our belief that some relationship measured by is s...
10679 sym 94 img
How to make a scientific result disappear
Nathan Danneman (a co-author and one of my graduate students from Emory) recently sent me a New Yorker article from 2010 about the “decline effect,” the tendency for initially promising scientific results to get smaller upon replication. Wikipedia can summarize the phenomenon as well as I can: In his article, Lehrer gives several examples whe...
9760 sym R (2612 sym/4 pcs) 60 img
p-values are (possibly biased) estimates of the probability that the null hypothesis is true
Last week, I posted about statisticians’ constant battle against the belief that the p-value associated (for example) with a regression coefficient is equal to the probability that the null hypothesis is true, for a null hypothesis that beta is zero or negative. I argued that (despite our long pedagogical practice) there are, in fact, many si...
6799 sym R (2365 sym/3 pcs) 72 img
Measuring Bias in Published Work
In a series of previous posts, I’ve spent some time looking at the idea that the review and publication process in political science—and specifically, the requirement that a result must be statistically significant in order to be scientifically notable or publishable—produces a very misleading scientific literature. In short, published stud...
8627 sym 88 img 1 tbl