New testing method improves accuracy of beta regression models in small samples
The article tests different methods for analyzing beta regression models with varying dispersion, focusing on small sample sizes. Through numerical analysis, the researchers found that bootstrap tests were the most effective. They also discovered that the score test was more accurate than the likelihood ratio and Wald tests. The study included real data to support their findings.