1.11 P-Values

A p-value is the probability that your result is a false positive or just a fluke. We usually compare our p-value from our experiment with our alpha level, which is our predetermined cutoff for false positives. The most common alpha is 0.05, which means we only want to accept results that are less than or equal to 5% likely to be a false positive. Said another way, we want to accept results that we are 95% sure are not false positives. While the alpha tells us our cutoff points, the p-value tells us what the likelihood of getting a false positive is for our current experiment. So, if our alpha is 0.05 and our p-value is .04, we assume our result is not a false positive because our p-value is less than our alpha level. If our alpha is 0.05 and our experiment produces a p-value of 0.20, we assume that our result is just a fluke. As long as that p-value is less than our alpha, we can assume that our results are significant.