In: Statistics and Probability
Question #2 Hypothesis Testing
A recent issue of AARP Bulletin reported that the average weekly pay for a woman with a high school diploma was $520. Suppose you would like to determine if the average weekly pay for all working women is significantly greater than that for women with a high school diploma. Data providing the weekly pay for a sample of 50 working women are available in the Excel file (Weekly Pay). Formulate the null and alternative hypotheses and conduct the appropriate test. What is your conclusion?
Weekly Pay
582 320 290 542 678 596 760 565 772
333 685 800 619 697 557 804 687 691
759 599 696 950 750 657 675 498
633 753 627 614 569 617 736 712
629 553 679 548 679 1230 565 533
523 641 667 570 598 648 587 424
Solution:
Given: A recent issue of AARP Bulletin reported that the average weekly pay for a woman with a high school diploma was $520.
That is: Population mean =
We have to test if the average weekly pay for all working women is significantly greater than that for women with a high school diploma.
That is we have to test:
We are given that the data of weekly pay for a sample of 50 working women.
Thus we use following steps:
Step 1) State H0 and H1:
Vs
Step 2) Find test statistic:
Since population standard deviation is unknown , we use t test statistic.
where
Thus we need to make following table:
x | x^2 | x | x^2 | x | x^2 | x | x^2 |
582 | 338724 | 800 | 640000 | 750 | 562500 | 736 | 541696 |
333 | 110889 | 696 | 484416 | 569 | 323761 | 565 | 319225 |
759 | 576081 | 627 | 393129 | 679 | 461041 | 587 | 344569 |
633 | 400689 | 679 | 461041 | 598 | 357604 | 565 | 319225 |
629 | 395641 | 667 | 444889 | 596 | 355216 | 687 | 471969 |
523 | 273529 | 542 | 293764 | 557 | 310249 | 498 | 248004 |
320 | 102400 | 619 | 383161 | 657 | 431649 | 712 | 506944 |
685 | 469225 | 950 | 902500 | 617 | 380689 | 533 | 284089 |
599 | 358801 | 614 | 376996 | 1230 | 1512900 | 424 | 179776 |
753 | 567009 | 548 | 300304 | 648 | 419904 | 772 | 595984 |
553 | 305809 | 570 | 324900 | 760 | 577600 | 691 | 477481 |
641 | 410881 | 678 | 459684 | 804 | 646416 | ||
290 | 84100 | 697 | 485809 | 675 | 455625 |
Thus we get:
Thus
Step 3) Find t critical value:
df = n -1 = 50 - 1 = 49
Look in t table for df = 49 or if it is not listed , then look for its previous df
and at one tail area = 0.05 and find corresponding t critical value.
One tail area , since this is one tailed test and level of significance is not given , so we use default level of significance = 0.05.
Thus we get:
df = 49 is not listed so we look for df = 40 and one tail area = 0.05
t critical value = 1.684
Step 4) Decision rule:
Reject H0, if t test statistic value > t critical value , otherwise we fail to reject H0.
Since t test statistic value = 5.617 > t critical value = 1.684
we reject null hypothesis H0.
Step 5) Conclusion:
Since we have rejected H0, there is sufficient evidence to conclude that: the average weekly pay for all working women is significantly greater than that for women with a high school diploma.