In: Statistics and Probability
An experimenter was interested in dieting and weight loss among men and women. It was believed that women tend to lose more weight than men in the first 2 weeks of a standard dieting program. A random sample of 15 married coupes were put on the same diet. Their weight loss after 2 weeks is listed below. Conduct a directional t-test of the difference using .05 as your level of significance. Be sure to complete all steps.
Wives (pounds)
2.7,4.4,3.5,3.7,5.6,5.1,3.8,3.5,5.6,4.2,6.3,4.4,3.9,5.1,3.4
Husbands (pounds)
5.0,3.3,4.3,6.1,2.5,1.9,3.2,4.1,4.5,2.7,7.0,1.5,3.7,5.2,1.9
This is a problem of two sample t test, as husband and wife are not correlated.
We are given the weight loss of a random sample of 15 married couples who were put on the same diet for 2 weeks. Samples are independent of each other.
Let us denote the weight loss of wives by A and the weight loss of the husbands by B.
Hypothesis to be tested:
Here we want to test, H0:μA=μB ag. H1: μA≠μB
Test statistic:
The required test statistic is,
T=(XA-XB)-(μA-μB)SA2n+SB2n follows under H0, tn+n-2,here n=15
Where, XA=1ni=1nXAi, XB=1ni=1nXBi,
SA2=1n-1i=1n(XAi-XA)2, SB2=1n-1i=1n(XBi-XB)2
Rejection criterion:
We reject the null hypothesis at 5% level of significance iff the |t|>t28;0.05
Calculations:
Now, XA=4.333 and XB=3.587
SA2=1.0323 and SB2=2.574
Now, the test statistic, t=(4.333-3.587)115(1.0323+2.574)=1.521
The tabulated value of t28;0.05=1.701
Conclusion:
Hence, we reject the null hypothesis H0 at 5% level of significance based on the given sample as |t|>t28;0.05 and conclude that the mean weight loss for the husband and wife is not same.