You are on page 1of 19

Practice of Statistics for Business and

Economics 4th Edition Moore Solutions


Manual
Visit to download the full and correct content document: https://testbankdeal.com/dow
nload/practice-of-statistics-for-business-and-economics-4th-edition-moore-solutions-
manual/
Chapter 9: Inference for Categorical Data
9.1 (a) Gender is the explanatory and Commercial preference is the response because we are interested in
how gender influences commercial preference. (b) r × c = 2 × 2. Gender is the column variable because it
is the explanatory variable; commercial preference is the row variable because it is the response variable.

Gender
Preference F M
A 44 79
B 56 61

(c) 4 cells. (d)

Gender
Preference F M Total
A 44 79 123
B 56 61 117
Total 100 140 240

9.2

Age
Terminated 40orless over40 Total
No 645 760 1405
Yes 30 90 120
Total 675 850 1525

9.3 40.98%. 60%.


9.4 Parallel. 59.02%.
9.5 Answers will vary. It seems appropriate to look at commercial preference for each gender.

Commercial preference by
Gender
100
80
60
40
20
0
F M
gender

A B

205
206 Chapter 9 Inference for Categorical Data

9.6 Answers will vary. It seems appropriate to look at termination for each age group.

Termination by Age
800
700
600
500
400
300
200
100
0
40orless over40
age

no yes

9.7 14. 40.98%. 22.95% of companies use adaptive flexibility, and of that 22.95%, 40.98% of them
should be highly competitive assuming no association between flexibility and competitiveness. From the
formula, (25)(14)/(61) = 5.74.
9.8 (25)(0.2295) = 5.74. Of the 25 companies that are high competitiveness, we expect 22.95% to use
adaptive flexibility or 5.74 companies.
9.9 df = (5 – 1)(3 – 1) = 8.
9.10 0.05 < P-value < 0.10.

9.11 (a) X 2 = 23.74 ≈ z2 = 23.72. (b) (z*)2 = 3.2912 = 10.83 = X2*. (c) The z test null hypothesis indicates
that the proportions are equal or that small and large companies use social media equally; in other words,
the size of the company doesn’t matter in determining social media use—there is no relation between
company size and social media use, the null hypothesis for the X2 test.
9.12 This is due to rounding error.
Instructor’s Guide With Solutions 207

9.13 CA: 269.524, HI: 256.988, IN: 294.596, NV: 109.690, OH: 471.667.
9.14 CA: 0.5820, HI: 0.0000, IN: 0.0196, NV: 0.0660, OH: 0.2264. 0.0369 + 0.5820 + 0.0000 + 0.0196 +
0.0660 + 0.2264 = 0.9308.

 61  61.75  59  66.5  49  61.75  77  95 141 114  88  76 


2 2 2 2 2 2

9.15  2     
 15.19. df = 
61.75 66.5 61.75 95 114 76
5, 0.005 < P-value < 0.01. The data provide evidence that the bag is different from the percents stated by
the M&M Mars Company.
9.16 (a) The null hypothesis is that the coin matches the theoretical proportions of 50/50, 50% heads and
50% tails, and is fair. The alternative is that the coin does not match the theoretical 50/50 and is biased.
 5067  5000  4933  5000
2 2

(b)  2    1.7956. df = 1, 0.15 < P-value < 0.20. The data do not provide
5000 5000
evidence that the coin is biased.
9.17 (a) Answers will vary. The percent that tip for each shirt color are Black 30.99%, White 36.76%,
Red 57.97%, Yellow 43.06%, Blue 37.31%, Green 38.57%. (b) H0: There is no association between
whether a male customer tips and shirt color of the server. Hα: There is an association between whether a
male customer tips and shirt color of the server. (c) From software, X 2 = 12.3482, df = 5, P-value =
0.0303 (Using Table F, 0.025 < P-value < 0.05.) (d) A conditional graph by ShirtColor is shown.

Tip by ShirtColor
80
70
60
50
40
30
20
10
0
Black White Red Yellow Blue Green
ShirtColor

Yes No

(e) The data provide evidence of an association between whether or not the male customer left a tip and
shirt color worn by the server. Red-shirted servers got the most tips!
9.18 (a) Answers will vary. The percent that tip for each shirt color are: Black 35.29%, White 33.33%,
Red 28.30%, Yellow 38.00%, Blue 34.04%, Green 32.73%. (b) H0: There is no association between
whether a female customer tips and shirt color of the server. Hα: There is an association between whether
a female customer tips and shirt color of the server. (c) From software, X 2 = 1.1913, df = 5, P-value =
0.9457 (Using Table F, P-value > 0.25.) (d) A conditional graph by ShirtColor is shown.
208 Chapter 9 Inference for Categorical Data

Tip by ShirtColor
80
70
60
50
40
30
20
10
0
Black White Red Yellow Blue Green
ShirtColor

Yes No

(e) The data do not provide evidence of an association between whether or not the female customer left a
tip and shirt color worn by the server.
9.19 (a)

anxiety
option low medium high
A 8 11 27
B 52 49 33

(b) Answers will vary. The percent that choose the higher-priced option for each math anxiety group are
Low 13.33%, Moderate 18.33%, High 45%. (c) H0: There is no association between level of math anxiety
and which rental option is chosen. Hα: There is an association between level of math anxiety and which
rental option is chosen. (d) From software, X 2 = 18.2803, df = 2, P-value = 0.0001 (Using Table F, P-
value < 0.0005.) (e) The data provide evidence of an association between level of math anxiety and which
rental option is chosen. The higher the math anxiety, the more likely someone will be to choose the
higher-priced rental option.
9.20 H0: There is no association between intensity and portrait. Hα: There is an association between
intensity and portrait. From software, X 2 = 0.0751, df = 1, P-value = 0.7841(Using Table F, P-value >
0.25.) The data do not provide evidence of an association between intensity and portrait. A huge percent,
91.45%, classified Nivea as feminine, while only 8.55% classified it as masculine. As shown in the graph,
it was about equally split 50/50 between high and low intensity regardless of sex.
Instructor’s Guide With Solutions 209

Portrait
1
0.8
0.6
0.4
0.2
0
Female Male

Intensity by Portrait
60
50
40
30
20
10
0
Female Male
Portrait

High Low

9.21 H0: There is no association between intensity and portrait. Hα: There is an association between
intensity and portrait. From software, X 2 = 12.5248, df = 1, P-value = 0.0004 (Using Table F, P-value <
0.0005.) The data provide evidence of an association between intensity and portrait. A huge percent,
91.08%, classified Audi as masculine, but additionally, a huge percent of those also classified it as high
intensity, 88.57%. While of the 8.92% that classified Audi as feminine, only 62.5% of those saw it as
high intensity; hence the association between portrait and intensity.

Portrait
1
0.8
0.6
0.4
0.2
0
Female Male
210 Chapter 9 Inference for Categorical Data

Intensity by Portrait
100
80
60
40
20
0
Female Male
Portrait

High Low

9.22 H0: There is no association between intensity and portrait. Hα: There is an association between
intensity and portrait. From software, X 2 = 107.2007, df = 1, P-value < 0.0001 (Using Table F, P-value <
0.0005.) The data provide evidence of an association between intensity and portrait. A large percent,
71.59%, classified H&M as feminine, but additionally, a large percent of those also classified it as high
intensity, 86.08%. While of the 28.41% that classified H&M as masculine, most of these, 79.22%, saw it
as low intensity, hence the association between portrait and intensity.

Portrait
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
Female Male

Intensity by Portrait
100
80
60
40
20
0
Female Male
Portrait

High Low
Instructor’s Guide With Solutions 211

9.23 Nivea is viewed as feminine but about equally likely to be high or low intensity. Audi is viewed
masculine but, additionally, is primarily high intensity—especially so among those who also thought it
was masculine. H&M, however, showed a unique characteristic with a higher percent viewing it as
feminine but primarily of high intensity; while those who viewed it as masculine primarily viewed it of
low intensity, a reversal of what we saw in the other group.
9.24 (a) Answers will vary. It seems appropriate to look at response by institution type. (b) The presidents
for the public institutions were split, almost 50/50, with slightly more answering Yes than No. However,
among the presidents of the private institutions we see a very different result for the 2-year and 4-year
institutes. For the 2-year institutes, many more said they believed that online courses offer an equal
educational value, whereas for the 4-year institutes, many more said No, that the online courses didn’t
offer an equal educational value.

Response by institution type


70
60
50
40
30
20
10
0
Priv4 Pub4 Pri2 Pub2
Type

Yes No

9.25 (a) H0: There is no association between response and institution type. Hα: There is an association
between response and institution type. (b) From software, X 2 = 17.2974, df = 3, P-value = 0.0006 (Using
Table F, 0.0005 < P-value < 0.001.) (c) The data provide evidence of an association between whether or
not the president believes that online courses offer an equal educational value as classroom courses and
what institution type the president is from.
9.26 (a) 206 responded “Yes.” (b)

who
response public president
yes 621 206
no 1521 189

(c) No, combining the public responses with the presidents’ doesn’t give a clear overall group so the
marginal totals or percents for this table are not meaningful. (d) From software, X 2 = 81.4186, df = 1, P-
value < 0.0001 (Using Table F, P-value < 0.0005.) The data provide evidence of an association between
who was asked and whether or not they believe that online courses offer an equal educational value as
classroom courses. The presidents were much more likely (52.15%) to answer yes than the public
(28.99%).
9.27 (a) The percent of banks for each asset size that offer RDC are Under $100: 16.94%, $101 to $200:
30.89%, $201 or more: 56.85%.
212 Chapter 9 Inference for Categorical Data

OfferRDC by AssetSize
90
80
70
60
50
40
30
20
10
0
Under100 101To200 201OrMore

Yes No

(b) From software, X 2 = 96.3054, df = 2, P-value < 0.0001 (Using Table F, P-value < 0.0005.) The data
provide evidence of an association between the size of the bank and whether or not it offers RDC.
Generally speaking, the small size banks, as measured by assets, are less likely to offer RDC.
9.28 (a) The percent of banks in each region that offer RDC are: Northeast: 42.42%, Southeast: 48.31%,
Central: 38.69%, Midwest: 25.82%, Southwest: 34.62%, West: 44.53%.

OfferRDC by Region
80
70
60
50
40
30
20
10
0
Northeast Southeast Central Midwest Southwest West

OfferRDC OfferRDC

(b) From software, X 2 = 24.0586, df = 5. (c) P-value < 0.0005.


Instructor’s Guide With Solutions 213

(d) The data provide evidence of an association between the region of the bank and whether or not it
offers RDC. Certain regions, like the Southeast, West, and Northeast, have a higher percent of banks that
offer RDC.
9.29 (a)

Gender
TrustEssential Male Female Total
Agree 9097 10,935 20,032
Disagree 685 423 1,108
Total 9782 11,358 21,140

(b) 93% of males and 96.28% of females felt trust and honesty were essential. (c) A higher percent of
females than males feel that trust and honesty were essential in business and the workplace. (d) From
software, X 2 = 113.7358, df = 1, P-value < 0.0001 (Using Table F, P-value < 0.0005.) The data provide
evidence of an association between gender and whether or not they thought trust and honesty were
essential in business and the workplace.

TrustEssential by Gender
100
90
80
70
60
50
40
30
20
10
0
Male Female
Gender

Agree Disagree

9.30 (a)
214 Chapter 9 Inference for Categorical Data

Lied Male Female Total


Yes 6,057 5,966 12,023
No 4,145 5,719 9,864
Total 10,202 11,685 21,887

(b) Answers will vary. 59.25% of males and 51.06% of females said that they lied to a teacher at least
once during the past year.
Conditional distribution of Lied given Gender
Lied Male Female
Yes 59.25 51.06
No 40.75 48.94
Total 100 100

(c) For the males, about 59% admitted that they lied whereas for the females, 51% admitted that they had
lied. Males are more willing to admit that they lied than females. (d) From software, X 2 = 147.9720, df =
1, P-value < 0.0001 (Using Table F, P-value < 0.0005.) The data provide evidence of an association
between gender and whether or not they said that they lied to a teacher at least once during the past year.

Lied by Gender
70
60
50
40
30
20
10
0
Male Female

Yes No

9.31 (a) 59.17%. (b) Larger business have higher nonresponse rates. 79.5% of large, 60% of medium, and
38% of small businesses did not respond. (c)
Instructor’s Guide With Solutions 215

Response by Size
90
80
70
60
50
40
30
20
10
0
Small Medium Large
Size

Yes No

(d) H0: There is no association between size of business and response rate. Hα: There is an association
between size of business and response rate. From software, X 2 = 71.3722, df = 2, P-value < 0.0001
(Using Table F, P-value < 0.0005.) The data provide evidence of an association between size of business
and response rate.
9.32 (a) For younger than 40: 6.6% were hired, 93.4% were not. For 40 or older: 1.18% were hired,
98.82% were not. (b)

Hired by Age
100
90
80
70
60
50
40
30
20
10
0
YoungerThan40 40orOlder
age

Hired NotHired

(c) The company appears to hire a lot more from the younger than 40 age group than from the 40 or older
age group. (d) The qualifications of the candidates is a lurking variable. (e) H0: There is no association
between age and whether or not an applicant is hired. Hα: There is an association between age and
whether or not an applicant is hired. From software, X 2 = 7.8678, df = 1, P-value = 0.0050 (Using Table
F, 0.005 < P-value < 0.01.) The data provide evidence of an association between age and whether or not
an applicant is hired.
9.33 Answers will vary.
216 Chapter 9 Inference for Categorical Data

Smokers Overweight Not


Early Death 120 1040
No 440 5600

Non Smokers Overweight Not


Early Death 400 400
No 4000 4800

Combined Overweight Not


Early Death 520 1440
No 4440 10400

(b) For smokers: X 2 = 12.7037, df = 1, P-value = 0.0004. There is an association between early death and
being overweight: 21.43% of overweight smokers die early while only 15.66% of those not overweight
die early. For non-smokers: X 2 = 6.1030, df = 1, P-value = 0.0135. There is an association between early
death and being overweight: 9.09% of overweight non-smokers die early while only 7.69% of those not
overweight die early. For the combined table: X 2 = 9.5540, df = 1, P-value = 0.0020. There is an
association between early death and being overweight: 10.48% of the total overweight people die early
while a greater percent, 12.16%, of those not overweight die early.
9.34 (a)

Gender Admit Deny Total


Male 490 210 700
Female 280 220 500
Total 770 430 1200

(b) 70% of males and 56% of females overall are admitted. (c) For Business: 80% of males and 90% of
females are admitted. For Law: 10% of males and 33.33% of females are admitted. (d) Because a huge
number of males apply to business school and the business school has a higher overall admittance rate
(82.5%) than law school (27.5%), this pulls up the overall percent of males that are admitted to both
schools. Similarly a large number of females apply to law school, but with low admittance rates; this
makes the overall percent of females that are admitted to both schools look bad, when within each school
females are admitted with a higher percent than males. (e) H0: There is no association between gender and
whether or not an applicant is admitted. Hα: There is an association between gender and whether or not an
applicant is admitted. From software, X 2 = 24.8626, df = 1, P-value < 0.0001 (Using Table F, P-value <
0.0005.) The data provide evidence of an association between gender and whether or not an applicant is
admitted. (f) H0: There is no association between gender and whether or not an applicant is admitted. Hα:
There is an association between gender and whether or not an applicant is admitted. X 2 = 10.3896, df = 1,
P-value = 0.0013 (Using Table F, 0.001 < P-value < 0.0025.) The data provide evidence of an association
between gender and whether or not an applicant is admitted. (g) H0: There is no association between
gender and whether or not an applicant is admitted. Hα: There is an association between gender and
whether or not an applicant is admitted. X 2 = 20.4807, df = 1, P-value < 0.0001 (Using Table F, P-value
< 0.0005.) The data provide evidence of an association between gender and whether or not an applicant is
admitted. (h) The results for the two schools are similar.
Instructor’s Guide With Solutions 217

9.35 (a) A P-value cannot be negative. (b) Expected cell counts are computed under the assumption that
the null hypothesis is true, not the alternative. (c) The alternative hypothesis should be that there is an
association between two categorical variables.
9.36 (a) 50%. 50%. All cells have counts of 50. (b) 50%, 55%, 60%, 65%, 70%, 75%. (c)

a Percent of zeros X2 P-value


0 50 0 1
5 55 0.50 0.4795
10 60 2.02 0.1552
15 65 4.60 0.0320
20 70 8.33 0.0039
25 75 13.33 0.0003

(d) In summary, as the percent of zeros goes up, the X 2 value gets larger decreasing the P-value and
making the data more significant.

Chi-square value
14
12
10
8
6
4
2
0
50 55 60 65 70 75

P-value
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
50 55 60 65 70 75

9.37 (a) U = 0: 41.7%. U = 1: 62.5%. X 2 = 0.8333, P-value > 0.25. (b)


218 Chapter 9 Inference for Categorical Data

multiplier X2 P-value
x1 0.8333 0.3613
x2 1.6667 0.1967
x4 3.3333 0.0679
x6 5.0000 0.0253
x8 6.6667 0.0098

(c) In relation to sample size, collecting twice as much data that demonstrate the same association doubles
the X 2 value and makes the data more significant. Similarly, collecting four times as much data that
portray the same association quadruples the X 2 value and makes the data even more significant, etc.

P-value
0.4000
0.3500
0.3000
0.2500
0.2000
0.1500
0.1000
0.0500
0.0000
0 2 4 6 8 10

9.38 (a)

Year
Broadband 2001 2005 2009 2013 Total
Yes 135 743 1418 1575 3871
No 2115 1507 832 675 5129
Total 2250 2250 2250 2250 9000

(b) H0: There is no association between year and Internet access using broadband. Hα: There is an
association between year and Internet access using broadband. From software, X 2 = 2385.2, df = 3, P-
value < 0.0001 (Using Table F, P-value < 0.0005.) The data provide evidence of an association between
year and Internet access using broadband. Broadband usage has increased steadily from 6% in 2001 to
70% in 2013.
9.39 (a)

Year
Dialup 2001 2005 2009 2013 Total
Yes 923 630 158 68 1779
No 1327 1620 2092 2182 7221
Total 2250 2250 2250 2250 9000
Instructor’s Guide With Solutions 219

(b) H0: There is no association between year and Internet access using dial-up. Hα: There is an association
between year and Internet access using dial-up. From software, X 2 = 1365.3, df = 3, P-value < 0.0001
(Using Table F, P-value < 0.0005.) The data provide evidence of an association between year and Internet
access using dial-up. Dialup usage has decreased steadily from 41% in 2001 to 3% in 2013. (c) Since
2001, broadband usage has increased dramatically, from 6% in 2001 to 70% in 2013; at the same time,
use of dial-up access has plummeted, going from 41% in 2001 to only 3% in 2013.

broadband
80.00%
70.00%
60.00%
50.00%
40.00%
30.00%
20.00%
10.00%
0.00%
2001 2005 2009 2013

dialup
45.00%
40.00%
35.00%
30.00%
25.00%
20.00%
15.00%
10.00%
5.00%
0.00%
2001 2005 2009 2013

9.40 Answers will vary.


9.41 (a) df = 2, P-value < 0.0005 (0.00004 from software). (b) df = 3, P-value < 0.0005 (0.00015 from
software). (c) df = 2, P-value < 0.0005 (0.00004 from software). (d) df = 4, P-value < 0.0005 (0.00044
from software).
9.42 (a)

Stratum
Claim Small Medium Large
Allowed 51 12 4
Not 6 5 1
220 Chapter 9 Inference for Categorical Data

(b) 10.53% of Small, 29.41% of Medium, and 20% of Large were not allowed. (c) H0: There is no
association between the size of the claim and whether or not it is allowed. Hα: There is an association
between the size of the claim and whether or not it is allowed. (d) From software, X 2 = 3.7212, df = 2, P-
value = 0.1556 (Using Table F, 0.15 < P-value < 0.20.) The data do not provide evidence of an
association between the size of the claim and whether or not it is allowed.

9.43 (a) For small: 6  3342   351.8, so 352. For medium: 5  246   72.35, so 73. For large:
57 17
(6 57)(1  6 57) 
 58  11.6, so 12. (b) For small: m  1.96 (3342)2 
1
  266.26. For medium:
5  57 
 (5 17)(1  5 17)  2  (1 5)(1  1 5) 
m  1.96 (246)2    53.28. For large: m  1.96 (58)    20.34.
 17   5 

9.44 Answers will vary.

A B C Total
Yes 100 200 300 600
No 200 400 600 1200
Total 300 600 900 1800

9.45 From software, X 2 = 852.4330, df = 1, P-value < 0.0001 (Using Table F, P-value < 0.0005.) Z2 = (–
29.2)2 = 852.64 = X 2 with rounding error.

Juror
Person Juror Not Total
MA 339 143,272 143,611
NotMA 531 37,393 37,924
Total 870 180,665 181,535

9.46 (a)

Class
Response First-year Senior Total
Positive 9,697 13,514 23,211
Negative 3,474 3,483 6,957
Total 13,171 16,997 30,168

(b) H0: There is no association between the class and response. Hα: There is an association between class
and response. (c) From software, X 2 = 144.8132, df = 1, P-value < 0.0001 (Using Table F, P-value <
0.0005.) The data provide evidence of an association between the class and response, a higher percent of
seniors responded positively than first-year students. (d) Answers may vary. 79.51% of seniors responded
positively while only 73.62% of first-year students responded positively. So while the percent is
somewhat higher for seniors, there doesn’t seem to be a drastic difference between the two groups.
9.47 (a) 5.21% of those over 40 while only 1.57% of those under 40 were laid off. Yes, it looks like a
higher percent of over 40 were laid off than under 40.
Instructor’s Guide With Solutions 221

Over40
Released No Yes Total
Yes 8 42 50
No 503 764 1267
Total 511 806 1317

(b) From software, X 2 = 11.3783, df = 1, P-value = 0.0007 (Using Table F, 0.0005 < P-value < 0.001).
The data provide evidence of an association between age group and being laid off.
9.48 The two-way table is shown below. From software, X 2 = 48.6980, df = 2, P-value < 0.0001 (Using
Table F, P-value < 0.0005). The data provide evidence of an association between age group and
performance appraisal. As shown in the graph, those over 40 are less likely to get usually exceeds
expectations or continually exceeds expectations than those under 40.

Over40
Appraisal No Yes Total
PartOrFull 86 233 319
Usually 352 493 845
Continually 64 35 99
Total 502 761 1263

Appraisal by Age
80
70
60
50
40
30
20
10
0
No Yes
Over40

PartOrFull Usually Continually

9.49 9.27 and 9.47 are based on a single sample (the second model); 9.31 and 9.42 are based on separate
samples (the first model).
9.50 (a) The expected counts are shown in the table below.

OfferRDC
AssetSize Yes No
Under100 114.54 257.46
101To200 58.808 132.19
201OrMor 60.655 136.34

(b) From software, X 2 = 96.3054. (c) df = 2. (d) P-value < 0.0005.


222 Chapter 9 Inference for Categorical Data

9.51 (a) 80.19% of men and 28.44% of women died. From software, X 2 = 332.2054, df = 1, P-value <
0.0001 (Using Table F, P-value < 0.0005.) The data provide evidence that a higher proportion of men
died than women. Answers will vary. (b) Among the women, 4.55% of Highest, 12.62% of Middle, and
51.44% of Lowest died. From software, X 2 = 103.7665, df = 2, P-value < 0.0001 (Using Table F, P-value
< 0.0005.) The data provide evidence of an association between death and economic status for the
women. (c) Among the men, 64.53% of Highest, 87.21% of Middle, and 83.13% of Lowest died. From
software, X 2 = 34.6206, df = 2, P-value < 0.0001 (Using Table F, P-value < 0.0005.) The data provide
evidence of an association between death and economic status for the men.
9.52 From Table A, the probabilities are: (1) 0.2743. (2) 0.4602 – 0.2743 = 0.1859. (3) 0.5398 – 0.4602 =
0.0796. (4) 0.7257 – 0.5398 = 0.1859. (5) 1 – 0.7257 = 0.2743.

139  137.15 102  92.95  41  39.8   78  92.95  140 137.15 


2 2 2 2 2

 
2
     3.4061.
137.15 92.95 39.8 92.95 137.15
df = 4, P-value > 0.25 (0.4923 from software). The data do not provide evidence that the data are different
from a standard Normal distribution.
9.53 Answers will vary. Most results will give a fairly decent randomization and should fail to reject the
null hypothesis. Changing the interval likely will not change the result and should still fail to reject the
null hypothesis.
9.54 The probability for each interval is 0.2. The expected number for each interval is 100.

114  100   92  100  108 100  101 100  85 100 


2 2 2 2 2

 
2
   
 5.5. df = 4, 0.20 < P-value <
100 100 100 100 100
0.25 (0.2397 from software). The data do not provide evidence that the data are different from a Uniform
(0, 1) distribution.
9.55 Answers will vary. Most results will give a fairly decent randomization and should fail to reject the
null hypothesis. Changing the interval likely will not change the result and should still fail to reject the
null hypothesis.
9.56 It is highly unlikely to get a P-value so high; the random sample would have to produce an almost
perfectly uniform result with an equal number (~100) in each of the 5 groups. Most random samples will
vary slightly and getting a result so exact would be extremely rare. If the class had 2000 students, it is
plausible that one (some) of the 2000 students might get an extremely rare result which provides an
unusually high P-value.

You might also like