I'm learning hypothesis testing, and one thing doesn't make sense to me. I've learned that if there's a REALLY low p value (at an end of the normal distribution), then we reject the null hypothesis (Ho) for Ha.
My reasoning is: once I get the p value for the alternative hypothesis that I want to test, if there's a large "chunk" of the curve (thus a high p val), to the right for example, that would mean there's a good probability that the mean is > than the null hypo. Apparently it's the contrary. My professor said that it's because the p value is the possibility that it happens by chance, and thus it should be low to accept Ha. However, I'm having trouble visualizing that concept when I'm thinking in terms of the bell curve.
Could somebody please explain why this is so?
I appreciate it!