7To best answer our original question it might make the most

7.To best answer our original question, it might make the most sense to test for significant correlation between income and drawn nickel size. Incomes (in thousands of $) are shown for each of the 35 samples below:

Income (thousands of $)

Coin size (mm)

29

25

52

20

45

15

48

24

40

19

40

18

32

17

23

26

38

24

42

17

35

20

31

24

50

22

41

23

43

17

73

15

33

25

35

25

24

17

25

19

58

19

35

27

46

22

26

18

39

25

34

18

47

21

83

13

42

25

27

22

20

21

17

29

81

13

34

22

26

20

You can copy the data into Excel by highlighting the data, right-clicking and selecting Copy, then opening Excel, clicking on a blank cell, and selecting Paste from the Edit menu.

Test the claim that there is significant correlation at the 0.01 significance level. Retain at least 3 decimals on all values.

a) If we use L to denote the low income group and H to denote the high income group, identify the correct alternative hypothesis.

There is a linear relationship between family income and drawn coin size

b) The correlation coefficient r is: -0.553    
Hint: You may find it more convenient to use Excel\'s CORREL, SLOPE, and INTERCEPT functions rather than your calculator

For parts c and d, suppose someone tested very similar data to this and found the P-value was 0.00059999999999993

c) Based on this, we

Reject H0

d) Which means

Sample data support that there is a linear relationship between family income and drawn coin size

e) The regression equation (in terms of income x) is:

y

Income (thousands of $)	Coin size (mm)
29	25
52	20
45	15
48	24
40	19
40	18
32	17
23	26
38	24
42	17
35	20
31	24
50	22
41	23
43	17
73	15
33	25
35	25
24	17
25	19
58	19
35	27
46	22
26	18
39	25
34	18
47	21
83	13
42	25
27	22
20	21
17	29
81	13
34	22
26	20

Solution

r= -0.55300305

The regression equation (in terms of income x) is:

y