Ken Milne is back to discuss statistical significance to celebrate our 150th podcast. With the different pearls and pitfalls regarding statistical significance, Ken came back on to try and help us navigate through the process and allow us to better understand what this term actually means.
We have enjoyed Ken Milne being on for our 50th and 100th podcasts. He has talked aboutinterpreting FOAMed and EBM as well as examples when medicine got ahead of the evidence.
This discussion began after Ken shared a recent Nature article discussing statistical significance. In science, the person making the claim is responsible for providing the burden of proof. To help understand where this has led to statistical and how we do not use this correctly, we should first understand p-values better.
Originally, Fisher helped introduce the p-value was a "sniff test" to see if the findings seemed reasonable (a statistically significant outcome or SOO). However, it does not directly apply for patient oriented outcomes (POO) for example. What matters even more is about the ability to replicate these findings. The improper use of p-values has become so significant that professional organizations have warned against its misuse. Additionally, a recent publication dived into more detail on the issue.
As we dive more into the discussion, Ken posted just prior to the podcast recording to get information from statisticians. It made a nice little thread and you can read more on Twitter by following the tweet below.
What matters more than the p-value? Well, there are many options like the ones from the Tweet and the papers discuss. Importantly, we need to pay attention to the ways in which the study was conducted and appraise the papers we read appropriately. For example, we are checking for how the paper was funded, if the study was blinded, and other potential limitations that could affect the validity of the study and its relevant findings.
Let us know what you think by giving us feedback here in the comments section or contacting us on Twitter or Facebook. Remember to look us up on Libsyn and on iTunes. If you have any questions you can also comment below, email at firstname.lastname@example.org, or send a message from the page. We hope to talk to everyone again soon. Until then, continue to provide total care everywhere.
Libsyn and iTunes