Eli5: What is a null hypothesis and how do type 1 and type 2 errors work.
In: Mathematics
The null hypothesis, in statistics, is the idea that there is no significant difference between two populations.
If, for example, you’re testing whether your company’s fancy new insect killing chemical is more effective than the competition, you might run some tests in two groups: one using a reference chemical (or nothing at all), and one using your new one.
The null hypothesis is “This chemical is no more effective than what was used in the other group.” The *alternative* hypothesis is that it *is* more effective.
“Type I” and “Type II” errors can be thought of a lot easier by their colloquial names: “false positive” and “false negative.” The false positive is where you erroneously believe your chemical *is* more effective (when it actually isn’t), and the false negative is where you erroneously believe your chemical is *not* more effective (when it actually is).
In statistics, these errors occur due to improperly selected cutoff values (how are we judging effectiveness) and confidences (how many tests are we doing).
Latest Answers