significance test

a test to see if the results of a statistical test differ from the expected result.

In statistics, results are not usually exactly as predicted. Most people do not have exactly average blood pressure, and are not an average age. Most randomly selected groups are not exactly the same, even if they are chosen from the same population of scores. In inferential statistics some kind of cutoff is chosen. Historically, this has usually been a p value of .05, although smaller p values are sometimes used. This means that if there is less than a 5 percent #probability of the data is assumed to have not occurred by chance.