The Science of upcoming elections
Word spam again.You know how the causation fallacy works no....it's easier to disprove the negative.
If I claim masks work, but then cases are skyrocketing in the place where I chose as an example, it's a clear example of how masks don't work, at least under said broad definition, so now I have to explain away why.
If I claim cases are low because of masks, however, that's a fallacy because something else might have caused it.
Similarly, if I claim treats work to keep dogs happy, if my dog is grumpy after I've given him a treat, it goes to show that treats do not in fact work to keep dogs happy because there is a clear exception...I now have to qualify my statement somehow (in the afternoon, after a nap, only corgis etc).
But if I claim my dog is happy because I gave him a treat, that is not necessarily the case because my dog may be happy for other reasons, and not just the treat. Now, you can argue that the most obvious reason why the dog is happy is because I just gave him a treat (using Occam's razor) but I can also argue he's happy because I just walked through the door.
A single data point can only dismiss a single cause hypothesis.
It fails miserably when there are multiple causes.
Example:
H0: avoiding donuts helps you lose weight.
Ha: avoiding donuts has no impact on weight.
Evidence: yesterday, I ate a pound cake instead of my usual donuts. I still gained weight.
Bad conclusion: donuts have no impact on weight
Correct conclusion;: donuts are not the only cause of weight gain.