If I measure the happiness of six people and then feed them only chocolate for two weeks and measure their happiness again, what stats test should I use to test the hypothesis that a chocolate diet increases happiness? (assume happiness is measured as some continuous variable that is normally distributed)