What is it about?
This study looks at a method called "caliper matching," used to compare people who got a treatment with those who didn’t. A common version does the matching in a random order. The researchers found that this random order can lead to very different results just by changing the order or adding a few people. This means the same data can give different answers, which is a problem for scientific reliability.
Featured Image
Photo by Luke Chesser on Unsplash
Why is it important?
Unstable results can lead to misleading conclusions. Researchers might (intentionally or not) pick the result they like best. Even small changes in the data can flip the outcome. To improve trust in science, the authors recommend avoiding random matching, choosing a fixed order, and checking if results are stable before publishing.
Read the Original
This page is a summary of: Instability of estimation results based on caliper matching with propensity scores, PLOS One, June 2025, PLOS,
DOI: 10.1371/journal.pone.0325317.
You can read the full text:
Contributors
The following have contributed to this page
