~Cases:
1) s1 and s2 are known where no assumption is made about them. Use menu 3 under STAT, Tests (2-SampZTest).
~Note: Only choose YES under Pooled, if you were assuming that s1=s2.
2) s1 and s2 are unkown. Use menu 4 under STAT, Tests (2-SampTTest).
~Note: Confidence Intervals estimate the difference m1-m2
between two population means. Therefore, the difference of our sample
means is at the center of the interval. Use menu 9 (2-SampZInt), if
s1 and s2 are known or menu 0 (2-SampTInt) when s1 and s2 are unknown. Margin of Error is calculated in a similar way as we did in a one population mean & proportion.
~Procedure for Dependent Samples (Matched Pairs)
~Note: In this case we deal with the
Mean differences of the matched pairs. Therefore, we need to put
these differences in a list, say L3. You can do this after you enter your data into L1 and L2 by:
L1 - L2 --> L3 (2nd, 1, minus, 2nd, 2, stor->, 2nd 3)
then use a T-Test (menu 2, under STAT, Tests)(for a hypothesis test) or
TInterval (menu 8) (for a confidence interval) and proceed to enter the
necessary data.
~Note: For a hypothesis test, select 0 for mo and the appropriate alternate hypothesis.
~Example: A study is done on the ages of married couples.
Suppose that 20 couples are randomly selected and have the ages given
in the following data (H=Husband, W=Wife).
(H,W): (54,60), (33,35), (21,25), (68,60), (38,33), (30,28),
(78,74), (24,41), (71,64), (53,44), (27,20), (62,58), (22,18), (26,20),
(30,35), (44,44), (45,40), (62,63), (38,31), (50,20).
~Does the data suggest that the mean age of married males is greater
than that of married females? (at the 7% significance level).
~Procedure:
1) Enter the data in L1 and L2 then enter their differences in L3.
2) Use a T-Test (menu 2 under STAT. Tests)
3) Enter Data, mo: 0, List: L3, Freq: 1, m>mo, calculate, enter
~Result: All you need is the P-value of .0648 to make a decision.
~Since .0648<.07, the data indicates that the mean age of married
males is greater than the mean age of married females (at the 7% level
of significance).
Conclusion: There is sufficient evidence to conclude that the mean age of married males is greater than that of married females.