Watch on YouTube

6:17

Retaining or rejecting the null hypothesis

Once the null distribution exists, the decision is close to mechanical. Alpha is the long-run error rate you accept, chosen before the study runs — the share of times you would wrongly reject a true null. The p-value is the probability of a result at least this extreme IF the null were true. Compare them and you retain or reject. A false positive claims an effect that is not there; a false negative misses one that is. Neither disappears, and you decide in advance which you can better afford.

Also covered: alpha · p-value · Type I and Type II error · false positives