In: Statistics and Probability
10-65. Consider the hypothesis test H0: σ_1^2=σ_2^2 against H1: σ_1^2<σ_2^2. Suppose that the sample sizes are n1 = 6 and n2 = 13, and that s_1^2=23.8 and s_2^2=27.8. Use α = 0.05. Test the hypothesis and explain how the test could be conducted with a confidence interval on σ_1^2/σ_2^2.
The test statistic F = s1^2/s2^2
= 23.8/27.8
= 0.8561
At alpha = 0.05, the critical value is F(, n1 - 1, n2 - 1) = F(0.95, 5, 12) = 0.2138
Since the test statistic value is not less than the critical value, so we should not reject H0.
b) At 95% confidence interval , the critical values are FL = 0.1533
FR = 3.8911
The 95% confidence interval is
= ( 23.8/27.8) * (1/3.8911) < < ( 23.8/27.8) * (1/0.1533)
= 0.22 < < 5.58
Since the interval contains 1, so we should not reject the null hypothesis.