3 ms·
The p-value is the probability that you would observe an effect of this size if the hypothesis were not correct. So you want them to be small. In most fields, y
by sp332 12y ago
The p-value is the probability that you would observe an effect of this size if the hypothesis were not correct. So you want them to be small. In most fields, you want to get at least p < 0.05 or 95% certainty, which is about two standard deviations from the mean in a normal distribution. That means that your result is far enough away from what you would get by random chance. https://en.wikipedia.org/wiki/Standard_deviation#mediaviewer/File:Standard_deviation_diagram.svg https://en.wikipedia.org/wiki/Standard_deviation#mediaviewer...
- Yen 12y agoOk, so their observed effect ("significantly increased overall symptoms") is almost certainly due to gluten. Those are indeed good probabilities. I'm interested in how large an effect "significantly increased" means. I suspect that this is defined a little more rigorously in the actual text, but it doesn't seem to be in the abstract.
- pseudonom- 12y ago> almost certainly due to gluten Even if you ignore all the messiness of actually existing scientific practice, the math alone doesn't quite say that. It depends on your prior[0] and the power[1] of the study. [0]: https://en.wikipedia.org/wiki/Bayes'_theorem https://en.wikipedia.org/wiki/Bayes'_theorem [1]: https://en.wikipedia.org/wiki/Statistical_power https://en.wikipedia.org/wiki/Statistical_power
- TeMPOraL 12y agoKeep in mind that p < 0.05 is actually quite a weak standard, and hard sciences tend to aim for a stronger p < 0.01. The problem with p < 0.05 and p-numbers in general is that current scientific practice allows to get whatever results you want with p < 0.05 by doing various tricks with statistics. Unfortunately this practice of "p-number hacking" is being fought nowhere near as it should be.
- sp332 12y agoIt's true that it's easy to abuse or misinterpret p-values. https://xkcd.com/882/ https://xkcd.com/882/ And one journal just banned significance tests! https://news.ycombinator.com/item?id=9109801 https://news.ycombinator.com/item?id=9109801