Eli5: What is a null hypothesis and how do type 1 and type 2 errors work.
In: Mathematics
The null hypothesis is a hypothesis that there is no statical significance to whatever you are examining or that the data occurs by chance rather than an underlying cause. When you’re trying to prove a hypothesis, you have to prove that the null hypothesis for your hypothesis doesn’t hold; that the evidence of your hypothesis is more than chance.
Type I errors are false positives and Type II are false negatives. Imagine you think you might be pregnant. Your null hypothesis would be that you aren’t pregnant; maybe your morning sickness is just a stomach bug or something. So you take a pregnancy test. If the test says you are pregnant when you aren’t, that’s a false positive, a Type I error. You rejected the true null hypothesis when you shouldn’t have. If you are pregnant and the test says you aren’t that’s a false negative, which is a Type II error. You didn’t reject the false null hypothesis when you should have.
I always remember them because the P in “Positive” has 1 vertical line, like I, while the N in “Negative” has 2 vertical lines, like II.
Latest Answers