Category: monte carlo simulations
-
The Linear Probability Model Ain’t So Bad
Although it feels almost sacrilegious, I have recently come around to the linear probability model (LPM). It is easy to understand, its problems are surmountable, and it actually outperforms logistic regression under certain conditions. As I often point out, I have a soft spot for “ham sandwiches”, which is basically…
-
Simulating the Null: Applications in Studying Case Outcomes
The article discusses the application of the Monte Carlo method in simulating null distributions for small-n case studies, specifically in the context of criminology and criminal justice research. The author explores the use of conjunctive analysis of case configurations (CACC) and Fisher’s exact test to analyze sentencing outcomes in death…
-
Monte Carlo Mondays: The Rule of Ten
In a study of the “rule of ten” for logistic regression models, a Monte Carlo experiment was conducted with varying base rates and effect sizes. Results showed that the rule of ten is a useful heuristic for hypothesis testing with strong effects but less reliable for estimating coefficients. Despite limitations,…
West Design & Analysis 