Question

What happens if you test a hypothesis multiple times and the data doesn't support your prediction?

Answer 1

This question is worded in a confusing way. ..

I think you mean you gather data and test a null hypothesis, but fail to reject it. The null hypothesis is generally the opposite of your prediction. For example, if you think a drug might help cure a condition, your null hypothesis is usually that the drug does nothing. Or if you think requiring gun owners to take a gun safety course will reduce accidental firearm injuries, your null hypothesis is usually that the requirement makes no difference to firearm injury rates.

If you gather more data, and still fail to reject, multiple times, it still doesn’t mean your prediction is false. It could just mean you don’t have enough data.

Assuming it’s possible to get more data, and the question is important enough, you keep testing until any plausible effect is too small to matter.

Answer 2

After testing a hypothesis multiple times and data not supporting your prediction, you should form a new hypothesis that is supported by the data you have already collected.