What if we told you that nearly a century ago, someone came up with a “convenient” but nothing at all special way to see if there’s a 1 in 20 chance that the experiment results you’re seeing are due to pure luck?
Sounds random, doesn’t it? You’d say to yourself, “Well, I guess that could be useful data analysis tool in science and medicine.” It could be a first stop toward figuring out if, say, a drug actually did the thing it was designed to do — rather than it working by happenstance.
But what if we then told you that this little test — and getting a very specific number back from it — rules scientific careers and costs (or makes) companies millions of dollars?
Wouldn’t you wonder why?
The podcast is produced by Jocelyn Gonzales.
About the Authors Reprints
Yes, tricky. Alas, a p-value is NOT the “chance that the experimental results you’re seeing are due to pure luck” but the chance the results MIGHT have been the result of chance. E.g., a coin, possibly with heads on both sides, has a p-value for 2 heads of .25, altho the chance of “pure luck” in the result is zero.
The MIGHT analysis is RANDOM itself.