Each week, we publish lay abstracts of new articles from our prestigious portfolio of journals in statistics. The aim is to highlight the latest research to a broader audience in an accessible format.

The article featured today (CoxKnockoff: Controlled feature selection for the Cox model using knockoffs by Daoji Li, Jinzhao Yu & Hui Zhao) is from Stat with the full article now available to read here.

Li, D., Yu, J., & Zhao, H. (2023). CoxKnockoff: Controlled feature selection for the Cox model using knockoffs. Stat, 12(1), e607. https://doi.org/10.1002/sta4.607

In this paper, we propose a new statistical method for medical research and public health. Imagine you observe a large number of factors and want to find out which factors affect how long people live. This method helps researchers choose the right factors. The existing methods have a problem because they sometimes made mistakes by saying something was important when it wasn’t. Think of it like finding fake treasures when you’re hunting for real ones. That’s not good in science. So, in this paper, we came up with a new way to pick the right factors without making mistakes. It’s like having a better treasure map. We found that this new method works well, no matter how many factors they’re looking at, and it doesn’t make as many mistakes as the old methods. We also showed that this new method is powerful enough to find important factors when we have a lot of data points. We tested our method using computer simulations and real data, and it worked great. So, this new method is a helpful tool for scientists to figure out which factors are really important. You can find more details in the full article if you’re interested.