## When Can We Say That Something Doesnâ€™t Work?

When are we allowed to conclude that an intervention doesn’t work? Is it when a study or several studies fail to find significant differences?

When are we allowed to conclude that an intervention doesn’t work? Is it when a study or several studies fail to find significant differences?

A review of Erich Lehmann’s last book, Fisher, Neyman, and the Creation of Classical Statistics.

An R package that computes thousands of compatibility intervals and their corresponding P- and S-values and plots them to produce a function.

A discussion about what P-values are including common misinterpretations and how something called an S-value can better help us interpret P-values.

Discussions praising the efficiency of randomized trials are widespread, however, few of these discussions take a close look at some of the common assumptions that individuals hold regarding randomized trials. And unfortunately, these common assumptions may be based on outdated evidence and simplistic ideas.

Another misinterpretation of what statistical power is and how trial results should be reported in a popular journal.

The Bradford Hill Criteria are commonly used as a checklist to argue for causality when randomized trials aren’t possible. However, the originator of these viewpoints never intended for them to be used this way. In this post, I examine the shortcomings of using these viewpoints as a checklist in the real world.

Analytic statistics are commonly used to make inferences from the data. However, they are often misused because of misconceptions about what they do. In this post, I discuss how standard error is commonly misused in clinical trials.

Most studies aim to achieve high statistical power and precision by increasing sample sizes. Many researchers will conclude that there is no effect if they get a nonsignificant result in a high-powered study. In this post, I discuss why this is incorrect.

A look at a time when Bayesian data analysis went off the rails.