Tag Archives: Hypothesis test

Hypothesis Test for Paired Means

This test is sometimes called a Matched Pairs Hypothesis t Test. Matched pair design is when a person is tested against himself or very similar subjects matched in pairs. The randomness happens in which treatment is given first (if the subject is paired with himself) or which treatment is given to who (if contrasting the […]

A Bayesian perspective on interpreting statistical significance

This site illustrates why a P-value cannot be interpreted in a vacuum. Suppose a hypothesis test results in . Even if the null were true, we would still see statistically significant results by chance alone about 5% of the time. But this is relatively infrequent. So if we see a small P-value, we hope that the null […]

Hypothesis Test for Two Independent Means

When you want to find the difference of two means, you should use a hypothesis test for two independent means. You need two independent variables in this test. One must be categorical (i.e. men or women, employed or unemployed) and the other must be quantitative (i.e. temperature, salary). This test has five steps, which are […]