Left-tailed test

we want to know if the population mean difference $\mu_1-\mu_2$ is significantly less than 0 .

confidence interval method: If the lower limit of the confidence interval for the population mean difference $\mu_1-\mu_2$ (cell D21) is smaller than 0 (the population mean difference when the null hypothesis is true), accept the null hypothesis. Otherwise, reject it (the appropriate conclusion appears in cell D23).