3 ms·
Not exactly. The headline is a classic example of how people misinterpret "insufficient data to disprove the null hypothesis" as "evidence for the null hypothe
by btilly 16y ago
Not exactly.
The headline is a classic example of how people misinterpret "insufficient data to disprove the null hypothesis" as "evidence for the null hypothesis".
The truth is that from the data presented, we can't say much that is statistically useful. They didn't measure an effect, and didn't disprove it either. The one thing we learned is that the hand gel doesn't give an overwhelming benefit.
However if they had followed 2 groups of 500 people and had a similar success rate, they would have had overwhelming evidence that the gel helped a bit. Or it they followed 2 groups of 500 people, they might have found a smaller difference, and we could have this conversation again only with a smaller possible effect and more data collected.