It is difficult to find a scientific paper that does not include p-values in some capacity. Most often, p-values are used to statistically interpret whether the results of a given study are “significant” or if a real effect is being observed. However, a number of recent papers have begun to challenge this paradigm, suggesting that p-values are now used incorrectly and in ways that they were never originally intended. For example, Lewis Halsey (University of Roehampton) and colleagues not only demonstrate that p-values are an arbitrary criterion for forming conclusions, but more worryingly, that they are extremely unreliable.

In this episode of Naturally Speaking, Shaun Killen chats with Lewis Halsey about the controversy surrounding p-values in research, and is joined by statistician Paul Johnson and the stats-savvy James Grecian. They discuss the problems with p-values, the research culture that perpetuates their use, and potential alternatives for interpreting data. Listen to the podcast here!:

Episode 25 – The Problem With P Values

Also, check out this excellent and amusing video by Geoff Cumming illustrating the Dance of the P-Values and how they are not reproducible. Make sure you watch until the part with the dancing icon guys for each p-value and experimental outcome:

Papers mentioned in the episode:

Halsey LG, Curran-Everett D, Vowler SL, Drummond GB (2015) The fickle P value generates irreproducible results. Nature Methods. 12: 179-185. doi. 10.1038/nmeth.3288

Head ML, Holman L, Lanfear R, Kahn AT, Jennions MD (2015) The Extent and Consequences of P-Hacking in Science. PLoS Biol 13(3): e1002106. doi: 10.1371/journal.pbio.1002106

And, the editorial in Basic and Applied Social Psychology explaining that journal’s decision to ban the use of p-values

Intro and outro music sampled from:

“The Curtain Rises” and “Early Riser” Kevin MacLeod (incompetech.com)
Licensed under Creative Commons: By Attribution 3.0
http://creativecommons.org/licenses/by/3.0/