(download pdf) Biostatistics For The Biological And Health Sciences 1st Edition Triola Solutions Manual full chapter

Biostatistics For The Biological And
Health Sciences 1st Edition Triola

Solutions Manual
Go to download the full and correct content document:
https://testbankfan.com/product/biostatistics-for-the-biological-and-health-sciences-1s
t-edition-triola-solutions-manual/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...
Biostatistics for the Biological and Health Sciences

1st Edition Triola Test Bank
https://testbankfan.com/product/biostatistics-for-the-biological-
and-health-sciences-1st-edition-triola-test-bank/

2nd Edition Triola Solutions Manual
and-health-sciences-2nd-edition-triola-solutions-manual/

2nd Edition Triola Test Bank
and-health-sciences-2nd-edition-triola-test-bank/
Biostatistics An Applied Introduction for the Public

Health Practitioner 1st Edition Bush Test Bank
https://testbankfan.com/product/biostatistics-an-applied-
introduction-for-the-public-health-practitioner-1st-edition-bush-
test-bank/
Essential Statistics for the Behavioral Sciences 1st
Edition Privitera Solutions Manual
https://testbankfan.com/product/essential-statistics-for-the-
behavioral-sciences-1st-edition-privitera-solutions-manual/
Biocalculus Calculus Probability and Statistics for the

Life Sciences 1st Edition Stewart Solutions Manual
https://testbankfan.com/product/biocalculus-calculus-probability-
and-statistics-for-the-life-sciences-1st-edition-stewart-
solutions-manual/
Laboratory Manual and Workbook for Biological

Anthropology 1st Edition Soluri Solutions Manual
https://testbankfan.com/product/laboratory-manual-and-workbook-
for-biological-anthropology-1st-edition-soluri-solutions-manual/
Statistics For Engineering And The Sciences 5th Edition

Mendenhall Solutions Manual
https://testbankfan.com/product/statistics-for-engineering-and-
the-sciences-5th-edition-mendenhall-solutions-manual/
Anatomy and Physiology Foundations for the Health

Professions 1st Edition Deborah Roiger Solutions Manual
https://testbankfan.com/product/anatomy-and-physiology-
foundations-for-the-health-professions-1st-edition-deborah-
roiger-solutions-manual/
Chapter 6: Estimates and Samples with One Sample 167
Chapter 6. Estimates and Samples with One Sample

6-2 Estimating a Population Proportion
In exercises 1–4, find the critical value zα / 2 , that corresponds to the given confidence level.
1. By inspecting the table on p. 261, we see that the critical value is 2.575.
2. By inspecting the table on p. 261, we see that the critical value is 1.645.
3. For a 98% confidence interval, α=.02, and so α/2=.01. The area to the left is then 1–.01=.99. Referring to
Table A-2 we find that the area .99 corresponds to the z-score 2.33, which is the critical value.
The formulas used to compute pˆ and E in exercises 5-7 are
pˆ = [(upper confidence limit) + (lower confidence limit)]/2, and

E = [(upper confidence limit) – (lower confidence limit)]/2
(see page 268)
5. The upper limit of the interval is 0.280 and the lower limit of the interval is 0.220. Using the formulas above,
pˆ is 0.250 and the margin of error is 0.030. So we express the confidence interval as 0.250 ±0.030.
8. The value of pˆ is 0.742 and the margin of error E = 0.030. The lower limit of the confidence interval is then
pˆ − E which is 0.712 and the upper limit is then pˆ + E which is 0.772. The confidence interval is 0.712 < p <
0.772.
In Exercises 9-12, use the given confidence interval limits to find the point estimate pˆ and the margin of error E.
The formulas used to compute pˆ and E in exercises 9-12 are
pˆ = (upper confidence limit) + (lower confidence limit)/2, and

E = (upper confidence limit) – (lower confidence limit)/2
(see page 268)
pˆ = 0.464 and the margin of error E = 0.020.
168 Chapter 6: Estimates and Samples with One Sample
In Exercises 13-16, assume that a sample is used to estimate a population proportion p. Find the margin of error
E that corresponds to the given statistics and confidence level.
We use the following formula, provided on page 261, to calculate the margin of error, E. The formula for pˆ is
pˆ = x/n. The formula for qˆ is 1- pˆ .
pˆ qˆ
E = zα
2 n
13. On page 261, the table of confidence levels and critical values indicates that for a confidence level of 95%, the
ˆ =200/800=0.250 and qˆ =1–0.250=0.750. So the margin
critical value is 1.96. Also, n = 800, x = 200, and so p
of error is
pˆ qˆ (0.250)(0.750)
E = zα = (1.96) = (1.96)(0.015) = 0.030
2 n 800
ˆ =400/1200=0.333 and qˆ =1–0.333=0.667. So the
critical value is 2.575. Also, n = 1200, x = 400, and so p
margin of error is
pˆ qˆ (0.333)(0.667)
E = zα = (2.575) = (2.575)(0.014) = 0.035
2 n 1200
critical value is 2.575. The sample size n=1000. The percent of successes is 45%, and so p ˆ =0.450 and
qˆ =0.550. So the margin of error is
pˆ qˆ (0.450)(0.550)
E = zα = (2.575) = (2.575)(0.016) = 0.041
2 n 1000
critical value is 1.96. The sample size n=500. The percent of successes is 80% and so p ˆ =0.800 and qˆ =0.200.
So the margin of error is
pˆ qˆ (0.800)(0.200)
E = zα = (1.96) = (1.96)(0.006) = 0.011
2 n 500
In Exercises 17-20, use the sample data and confidence level to construct the confidence interval estimate of the
population proportion p.
We use the formula for confidence intervals for the population proportion p, found on page 262, to calculate
the confidence intervals.
pˆ − E < p < pˆ + E
We also use the following formula, provided on page 261, to calculate the margin of error, E. The formula for pˆ is
pˆ = x/n. The formula for qˆ is 1- pˆ .
pˆ qˆ
E = zα
2 n
critical value is 1.96. The sample size n=400 and number of successes x=300. So p ˆ =300/400=0.750 and
qˆ =1–0.750=0.250. The margin of error is
pˆ qˆ (0.750)(0.250)
E = zα = (1.96) = (1.96)(0.022) = 0.042
2 n 400
This means the confidence interval is
pˆ − E < p < pˆ + E
0.750 − 0.042 < p < 0.750 + 0.042
0.708 < p < 0.792
critical value is 2.575. The sample size is n=1200 and the number of successes is 200, so
pˆ = x / n = 200 /1200 = 0.167 and qˆ = 1 − p = 0.833. So the margin of error is
pˆ qˆ (0.167)(0.833)
E = zα = (2.575) = (2.575)(0.011) = 0.028
2 n 1200
Table A-2 we find that the area .99 corresponds to the z-score 2.33, which is the critical value. The sample size
is n=1655 and the number of successes is 176, so pˆ = x / n = 176 /1655 = 0.106 and qˆ = 1 − p = 0.894. So the
margin of error is
pˆ qˆ (0.106)(0.894)
E = zα = (2.33) = (2.33)(0.008) = 0.018
2 n 1655
critical value is 1.645. The sample size is n=2001 and the number of successes is 1776, so
pˆ = x / n = 1776 / 2001 = 0.888 and qˆ = 1 − p = 0.112. So the margin of error is
pˆ qˆ (0.888)(0.112)
E = zα = (1.645) = (1.645)(0.007) = 0.012
2 n 2001
In Exercises 21-24, use the given data to find the minimum sample size required to estimate a population
proportion or percentage.
We use the formula for sample size calculation for population proportion confidence intervals, found on page
266, to calculate the minimum sample size required.
[zα / 2 ]2 pˆ qˆ
When an estimate for pˆ is known: n =
E2
[z ]2 ⋅ 0.25
When an estimate for pˆ is unknown: n = α / 2 2
E
ˆ . The margin of error is 0.060. Using the formula, we find
critical value is 2.575. No estimate is known for p
that
[z ]2 ⋅ 0.25 [ 2.575]2 ⋅ 0.25
n = α /2 2 = = 460.46
E (0.060) 2
We round up, so the minimum required sample size is n=461.
critical value is 1.96. No estimate is known for pˆ . The margin of error is 0.038. Using the formula, we find
that
[z ]2 ⋅ 0.25 [1.96]2 ⋅ 0.25
n = α /2 2 = = 665.10
E (0.038) 2
critical value is 1.96. The estimate for pˆ is 0.185, which makes qˆ =0.815. The margin of error is 0.050. Using
the formula, we find that
[z ]2 ⋅ pˆ qˆ [1.96]2 ⋅ 0.185⋅ 0.815
n = α /2 2 = = 231.69
E (0.050) 2
critical value is 1.645. The estimate for p ˆ is 0.080, which makes qˆ =0.920. The margin of error is 0.030.
Using the formula, we find that
[z ]2 ⋅ pˆ qˆ [1.645]2 ⋅ 0.080 ⋅ 0.920
n = α /2 2 = = 221.29
E (0.030) 2
25. Interpreting Calculator Display

a. We are 95% confident that the interval 0.04891 to 0.05308 actually does contain the true value of p, the
proportion of males aged 18=20 who drove while impaired in the last month.
b. Alcohol-impaired driving does appear to be a problem for males aged 18-20, as the data suggests that
approximately 1 in 20 such males drive while impaired each month.
c. It would be conservative to use the upper limit of the confidence interval, and so it would be prudent to use
the value 5.3%.
26. Interpreting Calculator Display

a. We are 99% confident that the interval 0.93053 to 0.94926 actually does contain the true value of p, the
proportion of households that have a telephone.
b. It appears that the percentage of households that do not have a telephone is about 6%. Pollsters should be
worried about this, as that 6% of households represents a segment of the population that should be
represented in a sample.
27. Mendelian Genetics

a. The sample size n=705+224=929, and there were x=705 with red flowers. The requirements for this
procedure are satisfied, as npˆ = x = 705 and nqˆ = n − x = 224. pˆ =705/929=.759 and qˆ =1– pˆ =.241. On
page 261, the table of confidence levels and critical values indicates that for a confidence level of 95%, the
critical value is 1.96. The margin of error is given by
pˆ qˆ (0.759)(0.241)
E = zα = (1.96) = (1.96)(0.014) = 0.028.
2 n 929
So the confidence interval is ( pˆ − E, pˆ + E) = (0.759 − 0.028,0.759 + 0.028) = (0.731,0.787)
b. Though the results are not precisely 75%, the results are well within the margin of error, and so do not
contradict the theory of Mendel.
28. Drug Testing

a. The sample size is n=221, p ˆ =0.032 and qˆ =1– pˆ =0.968. The requirements for this procedure are satisfied,
as np ˆ = 221* 0.032 = 7.072 and nqˆ = 221* 0.968 = 213.928 .On page 261, the table of confidence
levels and critical values indicates that for a confidence level of 99%, the critical value is 2.575. The
margin of error is given by
pˆ qˆ (0.032)(0.968)
E = zα = (2.575) = (2.575)(0.012) = 0.030.
2 n 221
So the confidence interval is
( pˆ − E, pˆ + E) = (0.032 − 0.030,0.032 + 0.030) = (0.002,0.062)
b. It would appear that, due to the margin of error, that dizziness as a side effect is no more likely than
dizziness in the general public.
29. Smoking and College Education

a. The sample size is n=785, pˆ =0.183 and qˆ =1– pˆ =0.817. The requirements for this procedure are satisfied,
as npˆ = 785 * 0.183 = 143.655 and nqˆ = 785 * 0.817 = 641.345. For a 98% confidence interval, α=.02, and
so α/2=.01. The area to the left is then 1–.01=.99. Referring to Table A-2 we find that the area .99
corresponds to the z-score 2.33, which is the critical value. The margin of error is given by
pˆ qˆ (0.183)(0.817)
E = zα = (2.33) = (2.33)(0.014) = 0.032.
2 n 785
( pˆ − E, pˆ + E) = (0.183− 0.032,0.183+ 0.032) = (0.151,0.215)
b. We are 98% confident that the true percentage of college students who smoke is between 15.1% and
21.5%, and so we are 98% confident that that percentage is lower than the 27% rate for the general public.
This would imply that the smoking rate for college students is substantially different than the rate for the
general public.
30. Sample Size for Left-Handed Golfers

a. On page 261, the table of confidence levels and critical values indicates that for a confidence level of 99%,
the critical value is 2.575. No estimate is known for pˆ . The margin of error is 0.025. Using the formula for
sample size from p. 266, we find that
[z ]2 ⋅ 0.25 [2.575]2 ⋅ 0.25
n = α /2 2 = = 2652.25
E (0.025) 2
b. On page 261, the table of confidence levels and critical values indicates that for a confidence level of 99%,
the critical value is 2.575. The estimate for pˆ is 0.150, which makes qˆ =0.850. The margin of error is
0.025. Using the formula, we find that
[z ]2 ⋅ pˆ qˆ [2.575]2 ⋅ 0.150 ⋅ 0.850
n = α /2 2 = = 1352.648
E (0.025) 2
c. This would make the sample a self-selected sample, and people with a strong interest would be inclined to
participate. It may occur that those who are left-handed would then be overrepresented, inflating the value
of pˆ .
31. Sample Size for Plant Growers

a. For a 94% confidence interval, α=.06, and so α/2=.03. The area to the left is then 1–.03=.97. Referring to
Table A-2 we find that the area .97 corresponds to the z-score 1.88, which is the critical value. The estimate
for pˆ is 0.860, which makes qˆ =0.140. The margin of error is 0.030. Using the formula for sample size
from page 266, we find that
[z ]2 ⋅ pˆ qˆ [1.88]2 ⋅ 0.860 ⋅ 0.140
n = α /2 2 = = 472.824
E (0.030) 2
b. Again, the critical value is 1.88. No estimate is known for pˆ . The margin of error is 0.030. Using the
formula for sample size from p. 266, we find that
[z ]2 ⋅ 0.25 [1.88]2 ⋅ 0.25
n = α /2 = = 981.778
E2 (0.030) 2
c. This would make the sample a self-selected sample, and people with a strong interest would be inclined to
participate. It may occur that those who like to grow plants in their homes would then be overrepresented,
inflating the value of pˆ .
32. Color Blindness

a. The sample size n=80, and there were x=7 with red/green color blindness. The requirements for this
procedure are satisfied, as np ˆ = x = 7 and nqˆ = n − x = 73. pˆ =7/80=0.088 and qˆ =1– pˆ =0.912. On
pˆ qˆ (0.088)(0.912)
E = zα = (1.645) = (1.645)(0.032) = 0.052.
2 n 80
( pˆ − E, pˆ + E) = (0.088 − 0.052,0.088 + 0.052) = (0.036,0.140)
b. For a 96% confidence interval, α=.04, and so α/2=.02. The area to the left is then 1–.02=.98. Referring to
Table A-2 we find that the area .98 corresponds to the z-score 2.05 which is the critical value. The estimate
for p ˆ is 0.088 as calculated in part (a), and qˆ =0.912. The margin of error is 0.030. Using the formula for
sample size from page 266, we find that
[z ]2 ⋅ pˆ qˆ [ 2.05]2 ⋅ 0.088 ⋅ 0.912
n = α /2 2 = = 374.751
E (0.030) 2
c. We are 96% confident that the true percentage of men who are red/green color blind is between 3.6% and
14.0%, and so we are 96% confident that that percentage is higher than the .25% rate of red/green color-
blindness in women. Thus, we can safely conclude that the red/green color-blindness for women is lower
than the rate for men.
33. Cell Phones and Cancer

a. The sample size n=420,095, and there were x=135 cell phone users who developed brain or nervous system
cancer. The requirements for this procedure are satisfied, as npˆ = x = 135 and
nqˆ = n − x = 419,960. pˆ =135/420,095=0.000321 and qˆ =1– pˆ =0.999679. On page 261, the table of
confidence levels and critical values indicates that for a confidence level of 95%, the critical value is 1.96.
The margin of error is given by
pˆ qˆ (0.000321)(0.999679)
E = zα = (1.96) = (1.96)(0.0000276) = 0.0000542.
2 n 420,095
( pˆ − E, pˆ + E) = (0.000321− 0.000054,0.000321+ 0.000054) = (0.000267,0.000375)
b. We are 95% confident that the true percentage of cell phone users who develop brain or nervous system
cancer is between 0.0267% and 0.0375%. The rate of such cancer in those who do not use cell phones is
0.034%. This percentage is within the confidence interval, so it does not appear that the rate of such cancer
in cell phone users is different than the rate in those who do not use cell phones.
34. Gender Selection

a. The sample size n=325, and there were x=295 of the births using XSORT were girls. The requirements for
this procedure are satisfied, as np ˆ = x = 295 and nqˆ = n − x = 30 . pˆ =295/325=0.908 and qˆ =1–
pˆ =0.092. On page 261, the table of confidence levels and critical values indicates that for a confidence
level of 99%, the critical value is 2.575. The margin of error is given by
pˆ qˆ (0.908)(0.092)
E = zα = (2.575) = (2.575)(0.016) = 0.041.
2 n 325
( pˆ − E, pˆ + E) = (0.908 − 0.041,0.908 + 0.041) = (0.867,0.949)
b. We are 99% confident that the true percentage of female births following the use of XSORT is between
86.7% and 94.9%. Assuming that girls comprise 50% of births, we are 99% confident that the birthrate of
girls following the use of the XSORT method is higher than 50%, and so it would appear that the XSORT
method is effective.
35. Gender Selection

a. The sample size n=51, and there were x=39 of the births using YSORT were boys. The requirements for
this procedure are satisfied, as npˆ = x = 39 and nqˆ = n − x = 12 . p
ˆ =39/51=0.765 and qˆ =1– pˆ =0.235. On
pˆ qˆ (0.765)(0.235)
E = zα = (2.575) = (2.575)(0.059) = 0.153.
2 n 51
( pˆ − E, pˆ + E) = (0.765− 0.153,0.765+ 0.153) = (0.612,0.918)
b. We are 99% confident that the true percentage of female births following the use of YSORT is between
61.2% and 91.8%. Assuming that boys comprise 50% of births, we are 99% confident that the birthrate of
boys following the use of the YSORT method is higher than 50%, and so it would appear that the YSORT
36. Pilot Fatalities

a. The sample size n=8411,and pilots dies in 5.2% of the crashes, so pˆ = 0.052 and qˆ =0.948. The
requirements for this procedure are satisfied, as npˆ = 8411× 0.052 = 437.372 and
nqˆ = 8411× 0.948 = 7973.628. On page 261, the table of confidence levels and critical values indicates that
for a confidence level of 95%, the critical value is 1.96. The margin of error is given by
pˆ qˆ (0.052)(0.948)
E = zα = (1.96) = (1.96)(0.008) = 0.015.
2 n 8411
( pˆ − E, pˆ + E) = (0.052 − 0.015,0.052 + 0.015) = (0.037,0.067)
b. We are 99% confident that the true percentage of female births following the use of YSORT is between
61.2% and 91.8%. Assuming that boys comprise 50% of births, we are 99% confident that the birthrate of
boys following the use of the YSORT method is higher than 50%, and so it would appear that the YSORT
37. Wearing Hunter Orange

a. The sample size n=123, and there were x=6 hunters who were wearing orange at the time of the injury. The
requirements for this procedure are satisfied, as npˆ = x = 6 and nqˆ = n − x = 117. pˆ =6/123=0.049 and
qˆ =1– pˆ =0.951. On page 261, the table of confidence levels and critical values indicates that for a
confidence level of 95%, the critical value is 1.96. The margin of error is given by
pˆ qˆ (0.049)(0.951)
E = zα = (1.96) = (1.96)(0.019) = 0.038 .
2 n 123
( pˆ − E, pˆ + E) = (0.049 − 0.038,0.049 + 0.038) = (0.011,0.087)
b. The sample size n=1115, and there were x=811 hunters who routinely wear orange. The requirements for
this procedure are satisfied, as npˆ = x = 811 and nqˆ = n − x = 304. pˆ =811/1115=0.727 and qˆ =1– pˆ =0.273.
On page 261, the table of confidence levels and critical values indicates that for a confidence level of 95%,
the critical value is 1.96. The margin of error is given by
pˆ qˆ (0.727)(0.273)
E = zα = (1.96) = (1.96)(0.013) = 0.026.
2 n 1115
( pˆ − E, pˆ + E) = (0.727 − 0.026,0.727 + 0.026) = (0.701,0.753)
c. If hunters who wear orange were as likely to get injured as those who did not wear orange, then the
percentage of injuries sustained by hunters who wore orange should be approximately equal to the
percentage of hunters who wear orange. The confidence interval for the percentage of hunters who were
wearing orange when they were injured is significantly lower than the confidence interval for the
percentage of hunters who routinely wear orange, with no overlap. The results do then seem to indicate that
hunters who wear orange are les likely to be injured because of being mistaken for game.
38. Effects of Sample Size and Confidence Level

a. The confidence interval will get narrower, as shown here. If the sample size n=10,000, and x=2,000 then
pˆ =0.2 and the critical value is 1.96. The margin of error is given by
pˆ qˆ (0.2)(0.8)
E = zα = (1.96) = (1.96)(0.004) = 0.008 .
2 n 10,000
( pˆ − E, pˆ + E) = (0.2 − 0.008,0.2 + 0.008) = (0.192,0.208)
In general, this will always occur, as the only change in the margin of error, E, is that the denominator in
the radicand gets larger, thus decreasing the value of E.
b. If the confidence level is increased, the critical value gets larger. This increases the margin of error, E.
39. Using Finite Population Correction Factor

The population size is N=10,000 households. As in Exercise 31, the critical value is 1.88. No estimate is
known for pˆ , so we allow pˆ qˆ = 0.25 in the formula. The margin of error is 0.030. Using the formula from
Exercise 39 on page 272, we find that
Npˆ qˆ[zα / 2 ]2 10,000 ⋅ 0.25[1.88]2 8836
n= = = = 894.088
pˆ qˆ[zα / 2 ] + (N − 1)E
2 2
0.25[1.88] + (10,000 − 1)(0.030)
2 2
9.8827
6-3 Estimating a Population Mean: σ Known

In Exercises 1–4, find the critical value zα / 2 , that corresponds to the given confidence level.
2. By inspecting the table on p. 261, we se that the critical value is 1.96.
4. For a 99.5% confidence interval, α=.005, and so α/2=.0025. The area to the left is then 1–.0025=.9975.
Referring to Table A-2 we find that the area .9975 corresponds to the z-score 2.81, which is the critical value.
In Exercises 5–8, determine whether the given conditions justify using the margin of error E = zα / 2σ / n when
finding a confidence interval estimate of the population mean µ.
5. Since the sample size is greater than 30, using E = zα / 2σ / n is justified.
6. Since the sample size is less than 30, σ is unknown, and the population’s distribution is unknown, using
E = zα / 2σ / n is not justified.
7. Since the original population is normal, and σ is known, using E = zα / 2σ / n is justified.
8. Though the original population is normally distributed and so the sample size does not matter, σ is unknown,
and so using E = zα / 2σ / n is not justified.
In Exercises 9–12, use the given confidence level and sample data to find (a) the margin of error E and (b) a
confidence interval estimate for the population mean µ.
The formulas to compute margin of error and the confidence interval are
E = zα / 2σ / n and x − E < µ < x + E
critical value is 1.96. The sample size is n=100, the sample mean is x = 95,000 and the population standard
deviation is σ = 12,345. The margin of error is
E = zα / 2σ / n = (1.96) ⋅ (12,345) / 100 = 2419.62
x−E< µ < x+E
95,000 − 2419.62 < µ < 95,000 + 2419.62
92,580.38 < µ < 97, 419.62
critical value is 2.575. The sample size is n=50, the sample mean is x = 80.5 and the population standard
deviation is σ = 4.6. The margin of error is
E = zα / 2σ / n = (2.575) ⋅ (4.6) / 50 = 1.675
x−E<µ < x+E
80.5−1.675 < µ < 80.5+1.675
78.825 < µ < 82.175
critical value is 1.645. The sample size is n=25, the sample mean is x = 5.24 and the population standard
deviation is σ = 2.50. The margin of error is
E = zα / 2σ / n = (1.645) ⋅ (2.50) / 25 = .823
x−E<µ < x+E
5.24 − .823 < µ < 5.24 + .823
4.417 < µ < 6.063
critical value is 1.96. The sample size is n=28, the sample mean is x = 45,678 and the population standard
deviation is σ = 9,900 . The margin of error is
E = zα / 2σ / n = (1.96) ⋅ (9900) / 28 = 3667.01
x−E<µ < x+E
45,678 − 3667.01 < µ < 45,678 + 3667.01
42,010.99 < µ < 49,345.01
In Exercises 13–16, use the given margin of error, confidence level, and population standard deviation σ to find
the minimum sample size required to estimate the population mean µ.
We use the formula for sample size calculation for population mean confidence intervals, found on page 279,
to calculate the minimum sample size required.
zα / 2σ
2
n=
E
13. The margin of error, E, is 125 and the population standard deviation σ = 500. On page 261, the table of
confidence levels and critical values indicates that for a confidence level of 95%, the critical value is 1.96.
Using the formula,
z σ
2 2
1.96 ⋅ 500
n = α /2 = = [7.84]2 = 61.466
E 125
14. The margin of error, E, is 3 and the population standard deviation σ = 15. On page 261, the table of confidence
levels and critical values indicates that for a confidence level of 99%, the critical value is 2.575. Using the
formula,
zα / 2σ
2 2
2.575⋅15
n= = = [12.875]2 = 165.766
E 3
15. The margin of error, E, is 5 and the population standard deviation σ = 48. On page 261, the table of confidence
levels and critical values indicates that for a confidence level of 90%, the critical value is 1.645. Using the
formula,
z σ
2 2
1.645⋅ 48
n = α /2 = = [15.792]2 = 249.387
E 5
16. The margin of error, E, is 500 and the population standard deviation σ = 9877. For a 94% confidence interval,
α=.06, and so α/2=.03. We find the area to the left, which is 1–.03=.97. Referring to Table A-2 we find that
the area .97 corresponds to the z-score 1.88, which is the critical value. Using the formula,
z σ
2 2
1.88 ⋅ 9877
n = α /2 = = [37.138]2 = 1379.195
E 500
In Exercises 17–20, refer to the accompanying TI-83/84 Plus calculator display of a 95% confidence interval
generated by using the methods of this section. The sample display results from using a sample of 80 measured
cholesterol levels of randomly selected adults.
17. The point estimate for the population mean µ is x = 318.1.
18. The confidence interval is 262.09 < µ < 374.11
19. Since the lower confidence interval limit is x − E , we find that

E = 318.1− 262.09 = 56.01. So the confidence interval may be expressed as 318.1± 56.01
20. We are 95% confident that the interval from 262.09 to 374.11 actually does contain the true population mean,
µ, cholesterol level of adults.
21. Everglades Temperatures

The requirements are met, as the 61 days are randomly selected, it is assumed that σ = 1.7, and the sample size
is n = 61 which is >30. On page 261, the table of confidence levels and critical values indicates that for a
confidence level of 95%, the critical value is 1.96. The sample mean is x = 30.4 . The margin of error is
E = zα / 2σ / n = (1.96) ⋅ (1.7) / 61 = 0.427
x−E<µ < x+E
30.4 − 0.427 < µ < 30.4 + 0.427
30.0 < µ < 30.8
It was not realistic to assume that σ is known.
22. Weights of Bears

The requirements are met, as the 54 bears are randomly selected, it is assumed that σ = 121.8, and the sample
size is n = 54 which is greater than 30. On page 261, the table of confidence levels and critical values indicates
that for a confidence level of 99%, the critical value is 2.575. The sample mean is x = 182.9. The margin of
error is
E = zα / 2σ / n = (2.575) ⋅ (121.8) / 54 = 42.680
x−E< µ < x+E
182.9 − 42.680 < µ < 182.9 + 42.680
140.2 < µ < 225.6
23. Cotinine Levels of Smokers

The requirements are met, as the 40 smokers are randomly selected, it is assumed that σ = 119.5, and the
sample size is n = 40 which is greater than 30. On page 261, the table of confidence levels and critical values
indicates that for a confidence level of 90%, the critical value is 1.645. The sample mean is x = 172.5. The
margin of error is
E = zα / 2σ / n = (1.645) ⋅ (119.5) / 40 = 31.082
x−E< µ < x+E
172.5− 31.082 < µ < 172.5+ 31.082
141.4 < µ < 203.6
24. Head Circumferences

The requirements are met, as the 100 two month old babies are randomly selected, it is assumed that σ = 1.6,
and the sample size is n = 100 which is greater than 30. On page 261, the table of confidence levels and critical
values indicates that for a confidence level of 99%, the critical value is 2.575. The sample mean is x = 40.6.
The margin of error is
E = zα / 2σ / n = (2.575) ⋅ (1.6) / 100 = 0.412
x−E< µ < x+E
40.6 − 0.412 < µ < 40.6 + 0.412
40.2 < µ < 41.0
25. Sample Size for Mean IQ of Biology Majors

The margin of error, E, is 2 and the population standard deviation is conservatively estimated to be σ = 15. On
critical value is 1.96. Using the formula,
z σ
2 2
1.96 ⋅15
n = α /2 = = [14.7]2 = 216.09
E 2
26. Sample Size for Estimating Income

The margin of error, E, is 500 and the population standard deviation is conservatively estimated to be σ =
6250. On page 261, the table of confidence levels and critical values indicates that for a confidence level of
95%, the critical value is 1.96. Using the formula,
z σ
2 2
1.96 ⋅ 6250
n = α /2 = = [24.5]2 = 600.25
E 500
27. Sample Size using Range Rule of Thumb

Estimates for the minimum and maximum ages for typical textbooks currently used in colleges will vary.
Suppose that the minimum age is 0 and the maximum age is 10. The range is then 10. The range rule of thumb
estimates σ = range/4. In this example, the estimated value of σ = 2.5.
The margin of error, E, is 0.25 and the population standard deviation is estimated, by the range rule of thumb,
to be σ = 2.5. On page 261, the table of confidence levels and critical values indicates that for a confidence
level of 90%, the critical value is 1.645. Using the formula,
z σ
2 2
1.645⋅ 2.5
n = α /2 = = [16.45]2 = 270.603
E 0.25
28. Sample Size Using Sample Data

The minimum and maximum pulse rates for males, as found in Appendix B, Data Set 1, are 56 and 96. The
range is then 40. The range rule of thumb estimates σ = range/4. In this example, the estimated value of σ =
10.
The margin of error, E, is 2 and the population standard deviation is estimated, by the range rule of thumb, to
be σ = 10. On page 261, the table of confidence levels and critical values indicates that for a confidence level
of 95%, the critical value is 1.96. Using the formula,
z σ
2 2
1.96 ⋅10
n = α /2 = = [9.8]2 = 96.04
E 2
If instead the sample standard deviation, s = 11.297 is used in place of the known population standard
deviation, σ, we instead get the following:
z σ
2 2
1.96 ⋅11.297
n = α /2 = = [11.071]2 = 122.568
E 2
The sample sizes found are fairly different (actual difference = 26). It is more likely that the sample size n=123
is closer to the correct sample size.
29. Sample Size Using Sample Data

The minimum and maximum diastolic blood pressure for females, as found in Appendix B, Data Set 1, are 41
and 102. The range is then 61. The range rule of thumb estimates σ = range/4. In this example, the estimated
value of σ = 15.25.
The margin of error, E, is 3 and the population standard deviation is estimated, by the range rule of thumb, to
be σ = 10. On page 261, the table of confidence levels and critical values indicates that for a confidence level
of 95%, the critical value is 1.96. Using the formula,
z σ
2 2
1.96 ⋅15.25
n = α /2 = = [9.963]2 = 99.268
E 3
If instead the sample standard deviation, s = 11.626 is used in place of the known population standard
deviation, σ, we instead get the following:
zα / 2σ
2 2
1.96 ⋅11.626
n= = = [7.596]2 = 57.694
E 3

The sample sizes found are fairly different (actual difference = 43). It is more likely that the sample size n=123
is closer to the correct sample size.
30. Confidence Interval with Finite Population Correction Factor

The requirements are met, as the 35 IQ scores are randomly selected, it is assumed that σ = 15, and the sample
size is n = 35 which is greater than 30. On page 261, the table of confidence levels and critical values indicates
that for a confidence level of 95%, the critical value is 1.96. The sample mean is x = 110 . The margin of error
is found using the finite population correction factor, since the population size is N = 250. Note that 35 > .05N
= 12.5, which indicates the correction is necessary.
σ N −n 15 250 − 35
E = zα / 2 = (1.96) ⋅ 4.970 × 0.929 = 4.618
n N −1 35 250 −1
x−E< µ < x+E
110 − 4.618 < µ < 110 + 4.618
105 < µ < 115
31. Sample Size with Finite Population Correction Factor

The margin of error, E, is 2 and the population standard deviation is conservatively estimated to be σ = 15. On
critical value is 1.96. The population from which the sample is to be taken is finite, N = 200, and so the
formula which incorporates the finite population correction factor should be used.
Nσ 2 (zα / 2 ) 2 200 ⋅ (15) 2 (1.96) 2 172872
n= = = = 104.117
(N − 1)E + σ (zα / 2 )
2 2 2
199(2) 2 + (15) 2 (1.96) 2 1660.36
6-4 Estimating a Population Mean: σ Not Known

In Exercises 1–8, do one of the following, as appropriate: (a) Find the critical value zα / 2 , (b) find the critical
value tα / 2 , (c) state that neither the normal nor the t distribution applies.
1. Since the population appears to be normal, and σ is unknown, it is appropriate to use the t distribution. For a
95% confidence interval, α=.05, and so α/2=.025. Since n=5, the degrees of freedom df=n–1=4. Using Table
A-3, we see that t .025 = 2.776 .
A-3, we see that t .025 = 2.262 .
3. Since the population appears to be very skewed, and the sample size is less than 30 (n=15), neither the normal
nor the t distribution applies.
4. Since the sample size is greater than 30 (n=45) and σ is known, it is appropriate to use the normal distribution.
On page 261, the table of confidence levels and critical values indicates that for a confidence level of 99%, the
critical value, zα / 2 , is 2.575.
90% confidence interval, α=.10, and so α/2=.05. Since n=92, the degrees of freedom df=n–1=91. Table A-3
does not include 91 degrees of freedom, so we use the closest number of degrees of freedom, 90. From Table
A-3, we see that t .05 = 1.662.
6. Since the population appears to be very skewed, and the sample size is less than 30 (n=9), neither the normal
nor the t distribution applies.
7. Since the population appears to be normal and σ is known, it is appropriate to use the normal distribution. For
a 98% confidence interval, α=.02, and so α/2=.01. The area to the left is then 1–.01=.99. Referring to Table A-
2 we find that the area .99 corresponds to the z-score 2.33, which is the critical value, zα / 2 .
A-3, we see that t .01 = 2.434 .
In Exercises 9 and 10, use the given confidence level and sample data to find (a) the margin of error and (b) the
confidence interval for the population mean µ. Assume that the population has a normal distribution.
The formulas to compute margin of error and the confidence interval are
s
E = tα / 2 and x − E < µ < x + E
n
9. For a 95% confidence interval, α=.05, and so α/2=.025. Since n=15, the degrees of freedom df=n–1=14. Using
Table A-3, we see that t .025 = 2.145.
s 108
a. The margin of error is E = t .025 = 2.145 = 59.81
n 15
b. The confidence interval is
x−E< µ < x+E
496 − 59.81 < µ < 496 + 59.81
436.19 < µ < 555.81
10. For a 99% confidence interval, α=.01, and so α/2=.005. Since n=32, the degrees of freedom df=n–1=31. Using
Table A-3, we see that t .005 = 2.744 .
s 0.70
a The margin of error is E = t .005 = 2.744 = 0.340
n 32
b The confidence interval is
x−E< µ < x+E
14.50 − 0.340 < µ < 14.50 + 0.340
14.160 < µ < 14.840
In Exercises 11 and 12, use the given data and the corresponding TI-83/84 Plus calculator display to express the
confidence interval in the format of x − E < µ < x + E . Also write a statement that interprets the confidence
interval.
The formula used to compute E in exercises 11 and 12is

E = [(upper confidence limit) – (lower confidence limit)]/2
(see page 293)
11. The lower confidence limit is 112.84 and the upper confidence limit is 121.56. This makes the margin of error
E = [121.56–112.84]/2 = 4.36. The sample mean x =117.2. The confidence interval is
x−E<µ < x+E
117.2 − 4.36 < µ < 117.2 + 4.36
112.84 < µ < 121.56
On the basis of the sample data, we are 95% confident that the limits of 112.84 and 121.56 do actually contain
the population mean IQ score of biology students.
12. The lower confidence limit is 75.297 and the upper confidence limit is 80.453. This makes the margin of error
E = [80.453–75.297]/2 = 2.578. The sample mean x =77.875. The confidence interval is
x−E<µ < x+E
77.875− 2.578 < µ < 77.875+ 2.578
75.297 < µ < 80.453
On the basis of the sample data, we are 99% confident that the limits of 75.297 and 80.453 do actually contain
the population mean height of NBA players.
In Exercises 13-24, construct the indicated confidence intervals.
13. Historical Corn Data

a. We assume that the data is normally distributed. For a 95% confidence interval, α=.05, and so α/2=.025.
Since n=11, the degrees of freedom df=n–1=10. Using Table A-3, we see that t .025 = 2.228 . Calculating x
from the sample data gives
x = Σx / n = 20627 /11 = 1875.182
Calculating the sample standard deviation from the data gives
nΣ(x 2 ) − (Σx) 2 11⋅ 39787267 − (20627) 2
s= = = 332.850
n(n −1) 11⋅10
s 332.850
The margin of error is E = t .025 = 2.228 = 223.598
n 11
The confidence interval is
x−E<µ < x+E
1875.182 − 223.598 < µ < 1875.182 + 223.598
1651.584 < µ < 2098.780
b. The intervals overlap, and so it is possible that the mean yields of kiln-dried and non-kiln dried corn seeds
are equal. There is no evidence that the method of kiln-drying the seeds changes the yield of the seeds.
14. Crash Hospital Costs

a. It appears that the data is bell shaped, and so it is assumed to be normally distributed. For a 99% confidence
interval, α=.01, and so α/2=.005. Since n=20, the degrees of freedom df=n–1=19. Using Table A-3, we see
that t .005 = 2.861. The sample mean x =9004 with a sample standard deviation s=5629.
s 5629
n 20
x−E< µ < x+E
9,004 − 3,601.091 < µ < 9,004 + 3,601.091
5,402.909 < µ < 12,605.091
b. It is 99% likely that the true mean hospital cost is between $5,402.91 and $12,605.09. To be conservative,
you should use the value $12,605.09.
15. Heights of Parents

a. It appears that the data is not far from normal. For a 99% confidence interval, α=.01, and so α/2=.005.
Since n=20, the degrees of freedom df=n–1=19. Using Table A-3, we see that t .005 = 2.861. The sample
mean x =4.4 with a sample standard deviation s=4.2.
s 4.2
n 20
x−E< µ < x+E
4.4 − 2.687 < µ < 4.4 + 2.687
1.713 < µ < 7.087
b. It is 99% likely that the true mean difference in mother’s and father’s height is between 1.713 inches and
7.087 inches. The confidence interval does not contain 0, so the confidence interval supports the
sociologist’s claim that women tend to marry men who are taller than themselves.
16. Monitoring Lead in Air

We assume that the data is normally distributed. For a 95% confidence interval, α=.05, and so α/2=.025. Since
n=6, the degrees of freedom df=n–1=5. Using Table A-3, we see that t .025 = 2.571. Calculating x from the
sample data gives
x = Σx / n = 9.23/ 6 = 1.538
Calculating the sample standard deviation from the data gives
nΣ(x 2 ) − (Σx) 2 6 ⋅ 32.520 − (9.23) 2
s= = = 1.914
n(n −1) 6⋅5
s 1.914
n 6
x−E< µ < x+E
1.538 − 2.009 < µ < 1.538 + 2.009
−0.471 < µ < 3.547
Yes, it would seem the confidence interval is not very good. One of the data values, 5.40, is extremely large,
suggesting a skewed population or an outlier. Because of this, the assumption that the data is normal is not
realistic.
17. Shoveling Heart Rates

a. We will assume that the data is not far from normal. For a 95% confidence interval, α=.05, and so
α/2=.025. Since n=10, the degrees of freedom df=n–1=9. Using Table A-3, we see that t .005 = 2.262 . The
sample mean x =175 with a sample standard deviation s=15.
s 15
n 10
x−E< µ < x+E
175−10.730 < µ < 175+10.730
164.270 < µ < 185.730
b. We will assume that the data is not far from normal. For a 95% confidence interval, α=.05, and so
α/2=.025. Since n=10, the degrees of freedom df=n–1=9. Using Table A-3, we see that t .005 = 2.262 . The
sample mean x =124 with a sample standard deviation s=18.
s 18
n 10
x−E< µ < x+E
124 −12.876 < µ < 124 +12.876
111.124 < µ < 136.876
c. The value 185 would be of great concern, as it is a value from the interval, and is very high. It may be that,
on average, those who shovel manually achieve this high heart rate.
d. The confidence interval for heart rates for those shoveling manually and those using a snow thrower do not
overlap and are significantly far apart. This is strong evidence that the mean heart rates for the two groups
are significantly different.
18. Pulse Rates

a. Since the sample size is n=40 and σ is unknown, the t interval is indicated. For a 95% confidence interval,
α=.05, and so α/2=.025. Since n=40, the degrees of freedom df=n–1=39. The degrees of freedom 39 is not
included in Table A-3, so we use the degrees of freedom closest to 39, which is 38. Using Table A-3, we
see that, for df=38, t .025 = 2.024 . The sample mean x =69.4 with a sample standard deviation s=11.3.
s 11.3
n 40
x−E<µ < x+E
69.4 − 3.616 < µ < 69.4 + 3.616
65.784 < µ < 73.016
b. There is some overlap between the two confidence intervals, and so we cannot conclude that the population
means for males and females are different.
19. Skull Breadths

First, we find a 95% confidence interval for the skulls from 4000B.C. and then find a 95% confidence interval
for the skulls from 150 A.D. We will assume both samples come from populations that appear normally
distributed, which indicates the use of t intervals is appropriate. For a 95% confidence interval, α=.05, and so
α/2=.025. Since n=12 for both samples, each interval will be based on the t distribution with degrees of
freedom df=n–1=11. Using Table A-3, we see that t .025 = 2.201. Calculating x for the sample of skulls from
4000 B.C. gives
x = Σx / n = 1544 /12 = 128.7
Calculating the sample standard deviation for the sample of skulls from 4000 B.C. gives
nΣ(x 2 ) − (Σx) 2 12 ⋅198898 − (1544) 2
s= = = 4.6
n(n −1) 12 ⋅11
s 4.6
n 12
x−E< µ < x+E
128.7 − 2.9 < µ < 128.7 + 2.9
125.8 < µ < 131.6
We are 95% confident that the true mean skull breadth for male Egyptian skulls from 4000B.C. is between
125.8 and 131.6.
For the skulls form 150 A.D, we do the same. Calculating x for the sample of skulls from 150 A.D gives
x = Σx / n = 1600 /12 = 133.3
Calculating the sample standard deviation for the sample of skulls from 150 A.D gives
nΣ(x 2 ) − (Σx) 2 12 ⋅ 213610 − (1600) 2
s= = = 5.0
n(n −1) 12 ⋅11
s 5.0
n 12
x−E< µ < x+E
133.3− 3.2 < µ < 133.3+ 3.2
130.1 < µ < 136.5
We are 95% confident that the true mean skull breadth for male Egyptian skulls from 150 A.D is between
130.1 and 136.5.
There is some overlap in the two confidence intervals, and so it can not be concluded that there is a difference
in the mean skull breadths. So there is not enough evidence to say that the head sizes does not seem to have
changed.
20. Head Circumferences

First, we find a 95% confidence interval for the head circumferences for two-year old males and then find a
95% confidence interval for the head circumferences for two-year old females. Both samples have sample size
n=50, which indicates the use of t intervals is appropriate. For a 95% confidence interval, α=.05, and so
α/2=.025. Since n=50 for both samples, each interval will be based on the t distribution with degrees of
freedom df=n–1=49. The degrees of freedom 49 is not included in Table A-3, so we use the degrees of
freedom closest to 49, which is 50. Using Table A-3, we see that t .025 = 2.009 . Calculating x for the head
circumferences for two-year old males gives
x = Σx / n = 2054.9 / 50 = 41.10
Calculating the sample standard deviation for the sample of head circumferences for two-year old males
nΣ(x 2 ) − (Σx) 2 50 ⋅84562.2 − (2054.9) 2
s= = = 1.498
n(n −1) 50 ⋅ 49
s 1.498
n 50
x−E< µ < x+E
41.10 − 0.43 < µ < 41.10 + 0.43
40.67 < µ < 41.53
We are 95% confident that the true mean head circumference for two-year old males is between 40.77 cm and
41.53 cm.
For the head circumferences for two-year old females, we do the same. Calculating x for the sample of the
head circumferences for two-year old females gives
x = Σx / n = 2002.4 / 50 = 40.05
Calculating the sample standard deviation for the sample of the head circumferences for two-year old females
gives
nΣ(x 2 ) − (Σx) 2 50 ⋅80323.8 − (2002.4) 2
s= = = 1.640
n(n −1) 50 ⋅ 49
s 1.640
n 50
x−E< µ < x+E
40.05− 0.47 < µ < 40.05+ 0.47
39.58 < µ < 40.52
We are 95% confident that the true mean for the head circumferences for two-year old females is between
39.58 cm and 40.52 cm.
There is no overlap in the two confidence intervals, and so it can be concluded that there is a difference in the
mean head circumferences of male and female two year olds. So there is enough evidence to say that the head
circumferences seem to be different.
21. Body Mass Index

a. Since the sample size is n=40 and σ is unknown, the t interval is indicated. For a 99% confidence interval,
see that, for df=38, t .005 = 2.712 . Calculating x for the BMI for males gives
x = Σx / n = 1039.9 / 40 = 26.00
Calculating the sample standard deviation for the sample of BMI for males
nΣ(x 2 ) − (Σx) 2 40 ⋅ 27493.8 − (1039.9) 2
s= = = 3.431
n(n −1) 40 ⋅ 39
s 3.431
n 40
x−E<µ < x+E
26.00 −1.47 < µ < 26.00 +1.47
24.53 < µ < 27.47
We are 99% confident that the true mean BMI for males is between 24.53 and 27.47.
b. Since the sample size is n=40 and σ is unknown, the t interval is indicated. For a 99% confidence interval,
see that, for df=38, t .005 = 2.712 . Calculating x for the BMI for females gives
x = Σx / n = 1029.6 / 40 = 25.74
Calculating the sample standard deviation for the sample of BMI for females
nΣ(x 2 ) − (Σx) 2 40 ⋅ 27984.5− (1029.6) 2
s= = = 6.166
n(n −1) 40 ⋅ 39
s 6.166
n 40
x−E< µ < x+E
25.74 − 2.64 < µ < 25.74 + 2.64
23.10 < µ < 28.38
We are 99% confident that the true mean BMI for females is between 23.10 and 28.38.
c. We are 99% confident that the true mean BMI for males is between 24.53 and 27.47, and we are 99%
confident that the true mean BMI for females is between 23.10 and 28.38. Since there is much overlap
between the two intervals, it is possible that the mean BMI for males is not larger than that for females.
22. Petal Lengths of Irises

We find a 95% confidence interval for the petal lengths of each class. Both samples have sample size n=50,
which indicates the use of t intervals is appropriate. For a 95% confidence interval, α=.05, and so α/2=.025.
Since n=50 for both samples, each interval will be based on the t distribution with degrees of freedom df=n–
1=49. The degrees of freedom 49 is not included in Table A-3, so we use the degrees of freedom closest to 39,
which is 50. Using Table A-3, we see that t .025 = 2.009 . Calculating x for the petal length of class setosa gives
x = Σx / n = 73.2 / 50 = 1.46
Calculating the sample standard deviation for the petal length of class setosa
nΣ(x 2 ) − (Σx) 2 50 ⋅108.64 − (73.2) 2
s= = = 0.174
n(n −1) 50 ⋅ 49
s 0.174
n 50
x−E< µ < x+E
1.46 − .05 < µ < 1.46 + .05
1.41 < µ < 1.51
We are 95% confident that the true mean petal length of class setosa is between 1.41 mm and 1.51 mm.
For the petal length of class versicolor, we do the same. Calculating x for the sample of the petal length of
class versicolor gives
x = Σx / n = 213/ 50 = 4.26
Calculating the sample standard deviation for the sample of the petal length of class versicolor gives
nΣ(x 2 ) − (Σx) 2 50 ⋅ 918.2 − (213) 2
s= = = 0.470
n(n −1) 50 ⋅ 49
s 0.470
n 50
x−E<µ < x+E
4.26 − 0.13 < µ < 4.26 + 0.13
4.13 < µ < 4.39
We are 95% confident that the true mean petal length of class versicolor is between 4.13 mm and 4.39 mm.
For the petal length of class virginica, we do the same. Calculating x for the sample of the petal length of
class virginica gives
x = Σx / n = 277.6 / 50 = 5.55
Calculating the sample standard deviation for the sample of the petal length of class virginica gives
nΣ(x 2 ) − (Σx) 2 50 ⋅1556.16 − (277.6) 2

s= = = 0.552
n(n −1) 50 ⋅ 49
s 0.552
n 50
x−E<µ < x+E
5.55− 0.16 < µ < 5.55+ 0.16
5.39 < µ < 5.71
We are 95% confident that the true mean petal length of class virginica is between 5.39 mm and 5.71 mm.
No pair of the intervals overlap, and each is more than 1 mm separated. This would indicate dramatic
differences.
23. Sepal Widths of Irises

We find a 95% confidence interval for the sepal widths of each class. Both samples have sample size n=50,
which indicates the use of t intervals is appropriate. For a 95% confidence interval, α=.05, and so α/2=.025.
Since n=50 for both samples, each interval will be based on the t distribution with degrees of freedom df=n–
1=49. The degrees of freedom 49 is not included in Table A-3, so we use the degrees of freedom closest to 49,
which is 50. Using Table A-3, we see that t .025 = 2.009 . Calculating x for the sepal widths of class setosa
gives
x = Σx / n = 170.9 / 50 = 3.42
Calculating the sample standard deviation for the sepal widths of class setosa
nΣ(x 2 ) − (Σx) 2 50 ⋅ 591.25− (170.9) 2
s= = = 0.381
n(n −1) 50 ⋅ 49
s 0.381
n 50
x−E<µ < x+E
3.42 − 0.11 < µ < 3.42 + 0.11
3.31 < µ < 3.53
We are 95% confident that the true mean sepal widths of class setosa is between 3.31 mm and 3.53 mm.
For the sepal widths of class versicolor, we do the same. Calculating x for the sample of the sepal widths of
class versicolor gives
x = Σx / n = 138.5/ 50 = 2.77
Calculating the sample standard deviation for the sample of the sepal widths of class versicolor gives
nΣ(x 2 ) − (Σx) 2 50 ⋅ 388.47 − (138.5) 2
s= = = 0.314
n(n −1) 50 ⋅ 49
s 0.314
n 50
x−E< µ < x+E
2.77 − 0.09 < µ < 2.77 + 0.09
2.68 < µ < 2.86
We are 95% confident that the true mean sepal widths of class versicolor is between 2.68 mm and 2.86 mm.
For the sepal widths of class virginica, we do the same. Calculating x for the sample of the sepal widths of
class virginica gives
x = Σx / n = 148.7 / 50 = 2.97
Calculating the sample standard deviation for the sample of the sepal widths of class virginica gives
nΣ(x 2 ) − (Σx) 2 50 ⋅ 447.33− (148.7) 2
s= = = 0.322
n(n −1) 50 ⋅ 49
s 0.322
n 50
x−E< µ < x+E
2.97 − 0.09 < µ < 2.97 + 0.09
2.88 < µ < 3.06
We are 95% confident that the true mean sepal widths of class virginica is between 2.88 mm and 3.06 mm.
No pair of the intervals overlap, but the confidence intervals are rather close to one another. This would
indicate differences, but dramatic may be too strong a word.
24. Hemoglobin Counts

First, we find a 95% confidence interval for the hemoglobin counts for males and then find a 95% confidence
interval for the hemoglobin counts for females. We assume that each population is normal, which indicates the
use of t intervals is appropriate. For a 95% confidence interval, α=.05, and so α/2=.025. Since n=23 for the
men, the degrees of freedom df=n–1=22. Using Table A-3, we see that t .025 = 2.074 . Calculating x for the
hemoglobin counts for males gives
x = Σx / n = 343/ 23 = 14.913
Calculating the sample standard deviation for the sample of hemoglobin counts for males
nΣ(x 2 ) − (Σx) 2 23⋅ 5130.025− (343) 2
s= = = 0.822
n(n −1) 23⋅ 22
s 0.822
n 23
x−E<µ < x+E
14.913− 0.355 < µ < 14.913+ 0.355
14.558 < µ < 15.268
We are 95% confident that the true mean hemoglobin counts for males is between 14.558 and 15.268.
For the hemoglobin counts for females, we do the same. For a 95% confidence interval, α=.05, and so
α/2=.025. Since n=27 for the females, the degrees of freedom df=n–1=26. Using Table A-3, we see that
t .025 = 2.056 . Calculating x for the sample of the hemoglobin counts for females gives
x = Σx / n = 352.3/ 27 = 13.048
Calculating the sample standard deviation for the sample of the hemoglobin counts for females gives
nΣ(x 2 ) − (Σx) 2 27 ⋅ 4637.76 − (352.3) 2
s= = = 1.254
n(n −1) 27 ⋅ 26
s 1.254
n 27
x−E< µ < x+E
13.048 − 0.496 < µ < 13.048 + 0.496
12.552 < µ < 13.544
We are 95% confident that the true mean for the hemoglobin counts for females is between 12.552 and 13.544.
mean hemoglobin counts for males and females. So there is enough evidence to say that the hemoglobin counts
seem to be different.
25. Yeast Cell Counts

The sample size for the yeast cell counts is 400, so a t interval is indicated. For a 95% confidence interval,
included in Table A-3, so we use the degrees of freedom closest to 399, which is 400. Using Table A-3, we see
that t .025 = 1.966 . Calculating x for the yeast cell counts gives
x = Σx / n = 1872 / 400 = 4.68
Calculating the sample standard deviation for the sample of hemoglobin counts for males
nΣ(x 2 ) − (Σx) 2 400 ⋅10544 − (1872) 2
s= = = 2.114
n(n −1) 400 ⋅ 399
s 2.114
n 400
x−E< µ < x+E
4.68 − 0.208 < µ < 4.68 + 0.208
4.472 < µ < 4.888
We are 95% confident that the true mean yeast cell count is between 4.472 and 4.888 cells, which is entirely
contained within the required cell count range. The sample data then appear to be acceptable.
26. Effects of Exercise and Stress

First, we find a 95% confidence interval for the systolic blood pressure pre-exercise with no stress and then
find a 95% confidence interval for the systolic blood pressure pre-exercise with stress from the speech test. We
assume that each population is normal, which indicates the use of t intervals is appropriate. For a 95%
confidence interval, α=.05, and so α/2=.025. Since n=24 for both samples, each interval will be based on the t
distribution with degrees of freedom df=n–1=23. Using Table A-3, we see that t .025 = 2.069 . Calculating x for
the systolic blood pressure pre-exercise with no stress gives
x = Σx / n = 2729.33/ 24 = 113.72
Calculating the sample standard deviation for the systolic blood pressure pre-exercise with no stress
nΣ(x 2 ) − (Σx) 2 24 ⋅ 313620.22 − (2729.33) 2
s= = = 11.859
n(n −1) 24 ⋅ 23
s 11.859
n 24
x−E< µ < x+E
113.72 − 5.008 < µ < 113.72 + 5.008
108.712 < µ < 118.728
We are 95% confident that the true mean systolic blood pressure pre-exercise with no stress is between
108.712 and 118.728.
For the systolic blood pressure pre-exercise with stress from the speech test, we do the same. Calculating x
for the sample of the systolic blood pressure pre-exercise with stress from the speech test gives
x = Σx / n = 3297.50 / 24 = 137.40
Calculating the sample standard deviation for the sample of the systolic blood pressure pre-exercise with stress
from the speech test gives
nΣ(x 2 ) − (Σx) 2 24 ⋅ 458340.75− (3297.50) 2
s= = = 15.149
n(n −1) 24 ⋅ 23
s 15.149
n 24
x−E< µ < x+E
137.40 − 6.398 < µ < 137.40 + 6.398
131.002 < µ < 143.798
We are 95% confident that the true mean for the systolic blood pressure pre-exercise with stress from the
speech test is between 131.102 and 143.698.
mean systolic blood pressure pre-exercise with no stress and pre-exercise with stress from the speech test. It
suggests that speech stress is significant.
27. Using the Wrong Distribution

The normal distribution’s critical values are smaller, so the resulting confidence interval would be narrower
than it should be.
6-5 Estimating a Population Variance
In Exercises 1–4, find the critical values χ L2 and χ 2R that correspond to the given confidence level and
sample size.
1. For a 95% confidence interval, α=0.05, and so α/2=0.025 and 1-α/2=0.975. The degrees of freedom is df=n-
1=15. Referring to Table A-4 we find that the critical values are χ L2 = χ .975
2
= 6.262 and χ 2R = χ .025
2
= 27.488 .
2
= 32.357 and χ 2R = χ .025
2
= 71.420 .
1=79. Since df=79 is not on the table, we use the closest value, df=80. Referring to Table A-4 we find that the
critical values are χ L2 = χ .995
2
= 51.172 and χ 2R = χ .005
2
= 116.321 .
1=39. Since df=39 is not on the table, we use the closest value, df=40. Referring to Table A-4 we find that the
2
= 26.509 and χ 2R = χ .05
2
= 55.758 .
In Exercises 5-8, use the given confidence level and sample data to find a confidence interval for the population
standard deviation σ. In each case, assume that a simple random sample has been selected from a population that
has a normal distribution.
2
= 8.907 and χ 2R = χ .025
2
= 32.852.
(n −1)s 2 (n −1)s 2
< σ <
χR2
χL
2
19 ⋅12,3452 19 ⋅12,3452
<σ <
32.852 8.907
9,388.3 < σ < 18,030.3
2
= 11.160 and χ 2R = χ .005
2
= 48.290.
(n −1)s 2 (n − 1)s 2
<σ <
χ 2
R
χ L2
26 ⋅ 4.6 2 26 ⋅ 4.6 2
<σ <
48.290 11.160
3.38 < σ < 7.02
2
= 17.708 and χ 2R = χ 052 = 42.557.
(n −1)s 2 (n −1)s 2
<σ <
χ 2
R
χ L2
29 ⋅ 2.50 2 29 ⋅ 2.50 2
<σ <
42.557 17.708
2.064 < σ < 3.199
2
= 32.357 and χ 2R = χ .025
2
= 71.420.
(n −1)s 2 (n −1)s 2
<σ <
χR2
χL
2
50 ⋅ 9,900 2 50 ⋅ 9, 900 2
<σ <
71.420 32.357
8,283.4 < σ < 12,306.5
In Exercises 9-12,assume that each sample is a simple random sample obtained from a normally distributed
population.
9. We are to be 95% confident we are within 10% of σ. We use Table 6-2 from page 305 to see that a sample size
of 191 is necessary.
10. We are to be 95% confident we are within 30% of σ. We use Table 6-2 from page 305 to see that a sample size
of 20 is necessary.
11. We are to be 99% confident we are within 1% of the population variance, σ2. We use Table 6-2 from page 305
to see that a sample size of 133,448 is necessary. This sample size is not practical in most cases.
12. We are to be 95% confident we are within 20% of the population variance, σ2. We use Table 6-2 from page
305 to see that a sample size of 210 is necessary.
In Exercises 13-20, assume that each sample is a simple random sample from a population with a normal
distribution.
We use the following formula, found on page 302, in each exercise to find the confidence interval for the
population standard deviation.
(n −1)s 2 (n − 1)s 2
< σ <
χR
2
χL
2

We assume that the data is normally distributed and the sample is a simple random sample. For a 95%
confidence interval, α=0.05, and so α/2=0.025 and 1-α/2=0.975. The degrees of freedom is df=n-1=10.
Referring to Table A-4 we find that the critical values are χ L2 = χ .975
2
= 3.247 and χ 2R = χ .025
2
= 20.483.
Calculating the sample standard variance from the data gives
nΣ(x 2 ) − (Σx) 2 11 ⋅ 38475192 − (20256) 2
s2 = = = 117,468.873
n(n −1) 11⋅10
(n −1)s 2 (n −1)s 2
<σ <
χ 2R χ L2
10 ⋅117,468.873 10 ⋅117,468.873
<σ <
20.483 3.247
239.5 < σ < 601.5

a. We assume that the data is normally distributed and the sample is a simple random sample. For a 95%
2
= 3.247 and χ 2R = χ .025
2
= 20.483.
Calculating the sample variance, from the data gives
nΣ(x 2 ) − (Σx) 2 11⋅ 39787267 − (20627) 2
s2 = = = 110, 789.164
n(n −1) 11⋅10
(n −1)s 2 (n −1)s 2
<σ <
2
χR 2
χL
10 ⋅110,789.164 10 ⋅110,789.164
<σ <
20.483 3.247
232.6 < σ < 584.1
b. There is significant overlap between the two intervals. There is no reason to believe that the two population
standard deviations are different.
15. Monitoring Lead in Air

We assume that the data is normally distributed and the sample is a simple random sample. For a 95%
2
= 0.831 and χ 2R = χ .025
2
= 12.833.
Calculating the sample variance, we have
nΣ(x 2 ) − (Σx) 2 6 ⋅ 32.520 − (9.23) 2
s2 = = = 3.664
n(n −1) 6⋅5
(n −1)s 2 (n − 1)s 2
<σ <
χR
2 2
χL
5⋅ 3.664 5⋅ 3.664
<σ <
12.833 0.831
1.195 < σ < 4.695
The data is skewed to the right, and so may not be from a normally distributed population.
16. Shoveling Heart Rates

2
= 2.700 and χ 2R = χ .025
2
= 19.023. The
sample standard deviation for maximum heart rates while shoveling is s=15.
(n −1)s 2 (n − 1)s 2
<σ <
χ 2
R
χ L2
9 ⋅152 9 ⋅152
<σ <
19.023 2.700
10.3 < σ < 27.4
b. The critical values are as above. The sample standard deviation for maximum heart rates using a
snowblower is s=18.
(n −1)s 2 (n − 1)s 2
< σ <
χR
2 2
χL
9 ⋅18 2
9 ⋅18 2
<σ <
19.023 2.700
12.4 < σ < 32.9
c. The confidence intervals overlap significantly, so it does not appear that the variances for heart rates
following manual shoveling and after using a snowblower are different.
17. Pulse Rates

Since df=39 is not on the table, we use the closest value, df=40. Referring to Table A-4 we find that the
2
= 24.433 and χ 2R = χ .025
2
= 59.342. The sample standard deviation for male
pulse rates is s=11.3.
(n −1)s 2 (n −1)s 2
<σ <
χ 2R χ L2
39 ⋅11.32 39 ⋅11.32
<σ <
59.342 24.433
9.16 < σ < 14.28
b. The critical values are as above. The sample standard deviation for female pulse rates is s=12.5.
(n −1)s 2 (n −1)s 2
<σ <
χR
2 2
χL
39 ⋅12.5 2
39 ⋅12.52
<σ <
59.342 24.433
10.13 < σ < 15.79
c. The confidence intervals overlap significantly, so it does not appear that the variances for pulse rates in
men and women are different.
18. Patient Waiting Times

2
= 2.700 and χ 2R = χ .025
2
= 19.023.
Calculating the sample variance for waiting in the single line, we have
nΣ(x 2 ) − (Σx) 2 10 ⋅ 51,327 − (715) 2
s2 = = = 22.722
n(n −1) 10 ⋅ 9
(n −1)s 2 (n −1)s 2
< σ <
χR
2 2
χL
9 ⋅ 22.722 9 ⋅ 22.722
<σ <
19.023 2.700
3.3 < σ < 8.7
b. The critical values are as above. Calculating the sample variance for waiting in the multiple lines, we have
nΣ(x 2 ) − (Σx) 2 10 ⋅ 54,109 − (715) 2
s2 = = = 331.833
n(n −1) 10 ⋅ 9

(n −1)s 2 (n −1)s 2
< σ <
2
χR 2
χL
9 ⋅ 331.833 9 ⋅ 331.833
<σ <
19.023 2.700
12.5 < σ < 33.3
c. The confidence intervals do not overlap. So it appears that the population variances for waiting in a single
line and multiple lines are different.
19. Body Mass Index

Since df=39 is not on the table, we use the closest value, df=40. Referring to Table A-4 we find that the
2
= 20.707 and χ 2R = χ .005
2
= 66.766. Calculating the sample standard deviation
for the sample of BMI for males
nΣ(x 2 ) − (Σx) 2 40 ⋅ 27493.8 − (1039.9) 2
s= = = 3.431
n(n −1) 40 ⋅ 39
(n −1)s 2 (n −1)s 2
<σ <
χ 2
R
χ L2
39 ⋅ 3.4312 39 ⋅ 3.4312
<σ <
66.766 20.707
2.6 < σ < 4.7
b. The critical values are as above. Calculating the sample standard deviation for the sample of BMI for
females
nΣ(x 2 ) − (Σx) 2 40 ⋅ 27984.5− (1029.6) 2
s= = = 6.166
n(n −1) 40 ⋅ 39
(n −1)s 2 (n −1)s 2
<σ <
χR2
χL2
39 ⋅ 6.166 2 39 ⋅ 6.166 2
<σ <
66.766 20.707
4.7 < σ < 8.5
c. The confidence intervals are adjacent, but do not overlap. So it appears that the population variances for
BMI for males and females are different.
20. Finding Critical Values

It is reasonable to assume that heights of men are normally distributed and the sample is a simple random
sample. Since the sample size is larger than 100 (n=772), we use the alternate formula for critical values as
presented in the problem. To do so, we need the critical values from Table A-2. For a 95% confidence interval,
α=0.05, and so α/2=0.025 and 1-α/2=0.975. By inspecting the table on p. 261, we see that the critical value is
1.96. We calculate the critical values for a χ 2 distribution, with degrees of freedom k=n-1=771.
[
1
] [1
]
2 2
χ L2 = −zα / 2 + 2k − 1 = −1.96 + 2 ⋅ 771−1 = 695.48, and
2 2
[
1
] [1
]
2 2
χ R = zα / 2 + 2k −1 = 1.96 + 2 ⋅ 771−1 = 849.362, and
2
2 2
Using the critical values found above, and the sample standard deviation s=2.8, we find the confidence interval
(n −1)s 2 (n −1)s 2
<σ <
χ 2
R
χ L2
771⋅ 2.8 2 771⋅ 2.8 2
<σ <
849.362 695.48
2.67 < σ < 2.95
Review Exercises
1. Mendel’s Genetics Experiments
a. We are finding the point estimate, pˆ , for a population proportion, and will convert to percentages when we
find the estimate. The sample size, n=787+277=1064. The number of short stem offspring is x=277.
Calculating pˆ = x / n = 277 /1064 = 0.260. A point estimate for the percentage is then 26.0%.
b. We are finding the confidence interval for a population proportion, and will convert to percentages when
we find the interval. The sample size, n=787+277=1064. The number of short stem offspring is x=277.
Calculating pˆ = x / n = 277 /1064 = 0.260. The requirements for this procedure are satisfied, as
npˆ = x = 277 and nqˆ = n − x = 787. pˆ =0.260 and qˆ =1– pˆ =0.740. On page 261, the table of confidence
levels and critical values indicates that for a confidence level of 95%, the critical value is 1.96. The margin
of error is given by
pˆ qˆ (0.260)(0.740)
E = zα = (1.96) = (1.96)(0.013) = 0.026.
2 n 1064
( pˆ − E, pˆ + E) = (0.240 − 0.026,0.240 + 0.026) = (0.214,0.266) .
A confidence interval for the percentage of offspring that are short stemmed would then be (21.4%,26.6%).
c. Here, we are asked for a sample size. On page 261, the table of confidence levels and critical values
indicates that for a confidence level of 99%, the critical value is 2.575. No estimate is known for pˆ . The
margin of error is 2.5 percentage points. Converted to decimal, the margin of error is to be 0.025. Using the
formula from page 266, we find that
[z ]2 ⋅ 0.25 [2.575]2 ⋅ 0.25
n = α /2 2 = = 2652.25
E (0.025) 2
2. Poplar Trees
a. We are finding a confidence interval for the population mean, µ, when the population standard deviation, σ,
is unknown. We will assume that the data is not far from normal, and so a t interval is indicated. For a 95%
confidence interval, α=.05, and so α/2=.025. Since n=20, the degrees of freedom df=n–1=19. Using Table
A-3, we see that t .025 = 2.093. Calculating x for the heights of trees in the control group gives
x = Σx / n = 99 / 20 = 4.95
Calculating the sample standard deviation for the sample of heights of trees in the control group
nΣ(x 2 ) − (Σx) 2 20 ⋅ 528.42 − (99) 2
s= = = 1.421
n(n −1) 20 ⋅19
s 1.421
n 20
The confidence interval for heights of trees in the control group is
x−E<µ < x+E
4.95− 0.67 < µ < 4.95+ 0.67
4.28 < µ < 5.62
b. We are finding a confidence interval for the population mean, µ, when the population standard deviation, σ,
is unknown. We will assume that the data is not far from normal, and so a t interval is indicated. For a 95%
confidence interval, α=.05, and so α/2=.025. Since n=20, the degrees of freedom df=n–1=19. Using Table
A-3, we see that t .025 = 2.093. Calculating x for the heights of trees in the irrigation treatment group gives
x = Σx / n = 89.6 / 20 = 4.48
Calculating the sample standard deviation for the sample of heights of trees in the irrigation treatment
group
nΣ(x 2 ) − (Σx) 2 20 ⋅ 461.84 − (89.6) 2
s= = = 1.783
n(n −1) 20 ⋅19
s 1.783
n 20
The confidence interval for heights of trees in the irrigation treatment group is
x−E<µ < x+E
4.48 − 0.83 < µ < 4.48 + 0.83
3.65 < µ < 5.31
c. The two intervals overlap, and so there does not appear to be a significant difference in the mean heights of
trees in the control group versus the irrigation treatment group.
3. Estimating Variation
a. We are finding a confidence interval for the standard deviation of the heights of trees in the control group.
For a 95% confidence interval, α=0.05, and so α/2=0.025 and 1-α/2=0.975. The degrees of freedom is
df=n-1=19. Referring to Table A-4 we find that the critical values are χ L2 = χ .975
2
= 8.907 and
χ R = χ .025 = 32.852. Calculating the sample standard deviation for the sample of heights of trees in the
2 2
control group
nΣ(x 2 ) − (Σx) 2 20 ⋅ 528.42 − (99) 2
s= = = 1.421
n(n −1) 20 ⋅19
(n −1)s 2 (n −1)s 2
<σ <
χ 2
R
χ L2
19 ⋅1.4212 19 ⋅1.4212
<σ <
32.852 8.907
1.08 < σ < 2.08
b. We are finding a confidence interval for the standard deviation of the heights of trees in the irrigation
treatment group. For a 95% confidence interval, α=0.05, and so α/2=0.025 and 1-α/2=0.975. The degrees
of freedom is df=n-1=19. Referring to Table A-4 we find that the critical values are χ L2 = χ .975
2
= 8.907 and
χ R = χ .025 = 32.852. Calculating the sample standard deviation for the heights of trees in the irrigation
2 2
treatment group
nΣ(x 2 ) − (Σx) 2 20 ⋅ 461.84 − (89.6) 2
s= = = 1.783
n(n −1) 20 ⋅19
(n −1)s 2 (n −1)s 2
<σ <
χR 2
χL 2
19 ⋅1.783 2
19 ⋅1.7832
<σ <
32.852 8.907
1.36 < σ < 2.60
c. The two intervals overlap, and so there does not seem to be a significant difference in the standard
deviations of the heights of trees in the control group and the irrigation treatment group.
4. Estimates from Voter Surveys

a. We are finding the point estimate, pˆ , for a population proportion of people who voted for the winner, and
will convert to percentages when we find the estimate. The sample size is n=611. The number of voters
who say they voted for the winner was x=308. Calculating pˆ = x / n = 308 / 611 = 0.504. A point estimate
for the percentage of voters who voted for the winner is then 50.4%.
b. We are finding the confidence interval for a population proportion, and will convert to percentages when
we find the interval. The sample size is n=611. The number of voters who say they voted for the winner
was x=308. Calculating pˆ = x / n = 308 / 611 = 0.504. The requirements for this procedure are satisfied, as
npˆ = x = 308 and nqˆ = n − x = 303. pˆ =0.504 and qˆ =1– pˆ =0.496. For a 98% confidence interval, α=.02,
and so α/2=.01. The area to the left is then 1–.01=.99. Referring to Table A-2 we find that the area .99
corresponds to the z-score 2.33, which is the critical value. The margin of error is given by
pˆ qˆ (0.504)(0.496)
E = zα = (2.33) = (2.33)(0.014) = 0.047.
2 n 611
( pˆ − E, pˆ + E) = (0.504 − 0.047,0.504 + 0.047) = (0.457,0.551) .
A confidence interval for the percentage of voters who voted for the winner would then be (45.7%,55.1%).
c. This result is not consistent with the results. 43% is not within the 98% confidence interval as determined
by survey. A possible cause for this discrepancy is that people may not be entirely honest, and want to
appear as if they voted for the winner.
5. Determining Sample Size

We need to find the sample size required for a confidence interval for a population proportion. For a 97%
confidence interval, α=.003, and so α/2=0.015. The area to the left is then 1–0.015=.985. Referring to Table
A-2 we find that the area .985 corresponds to the z-score 2.17, which is the critical value. No estimate is
known for pˆ . The margin of error is 0.020. Using the formula found on page 266, we find that
[z ]2 ⋅ 0.25 [2.17]2 ⋅ 0.25
n = α /2 2 = = 2,943.063
E (0.020) 2
We round up, so the minimum required sample size of biology majors to insure a maximum of a 2% error in
estimating the percentage of US biology majors who get grades of B or higher is n=2,944.
6. Alcohol Service Policy: Determining Sample Size

We need to find the sample size required for a confidence interval for a population proportion. For a 98%
confidence interval, α=.02, and so α/2=.01. The area to the left is then 1–.01=.99. Referring to Table A-2 we
find that the area .99 corresponds to the z-score 2.33, which is the critical value. We use the percentage from
the previous poll, 93%, as an estimate for pˆ , so we use pˆ =0.930 and qˆ =0.070. The margin of error is 0.040.
Using the formula found on page 266, we find that
[z ]2 ⋅ pˆ qˆ [2.33]2 ⋅ 0.930 ⋅ 0.070
n = α /2 2 = = 220.888
E (0.040) 2
We round up, so the minimum required sample size of randomly selected adults to insure a maximum of a 4%
error in estimating the percentage of adults who think restaurants and bars should refuse service to patrons who
have had too much to drink is n=221.
7. Investigating Effect of Exercise

First, we find a 95% confidence interval for the systolic blood pressure pre-exercise with stress from math test
and then find a 95% confidence interval for the systolic blood pressure post-exercise with stress from the math
test. We assume that each population is normal, and the population standard deviations for the two populations
are unknown, which indicates the use of t intervals is appropriate. For a 95% confidence interval, α=.05, and
so α/2=.025. Since n=24 for both samples, each interval will be based on the t distribution with degrees of
freedom df=n–1=23. Using Table A-3, we see that t .025 = 2.069 . Calculating x for the systolic blood pressure
pre-exercise with stress from math test gives
x = Σx / n = 3,194.33/ 24 = 133.10
Calculating the sample standard deviation for the systolic blood pressure pre-exercise with stress from math
test
nΣ(x 2 ) − (Σx) 2 24 ⋅ 430,963.44 − (3,194.33) 2
s= = = 15.889
n(n −1) 24 ⋅ 23
s 15.889
n 24
x−E< µ < x+E

133.10 − 6.710 < µ < 133.10 + 6.710
126.39 < µ < 139.81
We are 95% confident that the true mean systolic blood pressure pre-exercise with stress from math testis
between 126.39 and 139.81.
For the systolic blood pressure post-exercise with stress from math test, we do the same. Calculating x for the
sample of the systolic blood pressure post-exercise with stress from math test gives
x = Σx / n = 3045.67 / 24 = 126.90
Calculating the sample standard deviation for the sample of the systolic blood pressure post-exercise with
stress from math test gives
nΣ(x 2 ) − (Σx) 2 24 ⋅ 390,825.89 − (3045.67) 2
s= = = 13.709
n(n −1) 24 ⋅ 23
s 13.709
n 24
x−E< µ < x+E
126.90 − 5.790 < µ < 126.90 + 5.790
121.110 < µ < 132.690
We are 95% confident that the true mean for the systolic blood pressure post-exercise with stress from math
test is between 121.110 and 132.690.
There is overlap in the two confidence intervals, and so it can not be concluded that there is a difference in the
mean systolic blood pressure pre-exercise with stress from math test and post-exercise with stress from math
test.
Cumulative Review Exercises

1. Analyzing Weights of Supermodels
Here is the data listed in order, and the necessary sums.
Σx = 1089 Σx 2 = 132223
a. The mean is x = Σx / n = 1089 / 9 = 121
b. The median is 123
c. The modes are 119 & 128
d. The midrange is (105+128)/2 = 116.5
e. The range is (128-105) = 23
nΣ(x 2 ) − (Σx) 2 9 ⋅132223− (1089) 2
f. The variance is s 2 = = = 56.75
n(n −1) 9 ⋅8
nΣ(x 2 ) − (Σx) 2 9 ⋅132223− (1089) 2
g. The standard deviation is s = = = 7.533
n(n −1) 9 ⋅8
h. Q1 = 119
i. Q2 = 123
j. Q3 = 127
k. The data is ratio.
l. The boxplot is below.
130
120
110
100
1
m. We assume that the weights of supermodels are normally distributed. Since σ is unknown, the t interval is
indicated. For a 99% confidence interval, α=.01, and so α/2=.005. Since n=9, the degrees of freedom df=n–
1=8. Using Table A-3, we see that, for df=9, t.005 = 3.355 . Earlier in the exercise, we found the sample
mean and standard deviation, x = 121 and s = 7.533
s 7.533
n 9
x−E<µ < x+E
121− 8.424 < µ < 121+ 8.424
112.6 < µ < 129.4
We are 99%confident that the true mean weight for supermodels is between 112.6 and 129.4 pounds.
n. We are finding a confidence interval for the standard deviation of the weights of supermodels. For a 99%
2
= 1.344 and χ 2R = χ .005
2
= 21.955.
Earlier in the exercise we found the sample standard deviation, s = 7.533.
(n −1)s 2 (n − 1)s 2
<σ <
χR2
χL 2
8 ⋅ 7.5332 8 ⋅ 7.5332
<σ <
21.955 1.344
4.5 < σ < 18.4
o. The margin of error, E, is 2 and the population standard deviation is approximated by the sample standard
deviation, s = 7.533 . On page 261, the table of confidence levels and critical values indicates that for a
confidence level of 99%, the critical value is 2.575. Using the formula,
zα / 2σ
2 2
2.575⋅ 7.533
n= = = [9.699]2 = 94.07
E 2
p. The upper limit of the 99% confidence interval for supermodels is 129.4, which is less than one standard
deviation below the mean weight of women in general. So we cannot say that the weights of supermodels is
substantially less than the weights of randomly selected women.
2. X-Linked Recessive Disorders

a. We will use the normal approximation to the binomial. The conditions are satisfied, since
np = 200 × 0.25 = 50 and n(1− p) = 200 × 0.75 = 150 . The mean for the binomial is µ = np = 50 with
standard deviation
σ = np(1− p) = 200 ⋅ 0.25⋅ 0.75 = 6.124.
The probability that at least 65 have the X-linked recessive disorder would be the same as the area to the
right, under the standard normal curve, of
x − µ 64.5− 50
z= = = 2.37. Using Table A-2, this area is 1–.9911 = .0089, which is the desired
σ 6.124
probability.
b. The sample size n=200, and there were x=65 births with X-linked recessive disorder. The requirements for
this procedure are satisfied, as npˆ = x = 65 and nqˆ = n − x = 135. pˆ =65/200=0.325 and qˆ =1– pˆ =0.675. On
pˆ qˆ (0.325)(0.675)
E = zα = (1.96) = (1.96)(0.033) = 0.065.
2 n 200
( pˆ − E, pˆ + E) = (0.325− 0.065,0.325+ 0.065) = (0.260,0.390) .
c. The expert’s determination of 0.25 does not fall in the 95% confidence interval, which indicates that there
is at least a 95% chance that the determination is wrong. It appears that the expert was incorrect.
3. Analyzing Survey Results

a. The sample size was 1059 with 413 responding yes, so the percentage is 39.0%
b. The sample size n=413+646=1059, and there were x=413 who answered yes. The requirements for this
procedure are satisfied, as npˆ = x = 413 and nqˆ = n − x = 646. pˆ =413/1059=0.390 and qˆ =1– pˆ =0.610. On
pˆ qˆ (0.390)(0.610)
E = zα = (1.96) = (1.96)(0.015) = 0.029.
2 n 1059
( pˆ − E, pˆ + E) = (0.390 − 0.029,0.390 + 0.029) = (0.361,0.419)
c. Yes. The value 0.50 is not within the 95% confidence interval, so we are at least 95% confident that the
percentage who would answer yes is different from 50%.
d. A sensible response could be that the procedure takes that into account. At the end, there is a 5% chance
that the procedure produced an incorrect estimate, that is, that the true percentage that would answer yes to
the question lies outside the confidence interval.
Another random document with
no related content on Scribd:
possibly to Karinhall, Göring’s palatial estate near Berlin. And then
again it might have been flown out of the country altogether—to one
of the neutral countries, Spain or Switzerland.
Captain Posey and Private Lincoln Kirstein picked up additional
rumors from museum directors in Luxembourg. They had heard that
the altarpiece was in a salt mine, but they had also been told that it
was in the vaults of the Berlin Reichsbank. It was impossible to
reconcile these conflicting pieces of information. Finally, near Trier,
Posey and Kirstein tracked down a young German scholar who had
been in France during the occupation. Lincoln told me later that it
was hard to believe that this unassuming fellow had been high in the
confidence of Göring and other members of the Nazi inner circle.
From him they learned that the altarpiece had been taken to Alt
Aussee.
Then followed the rapid advance across Germany. To Posey and
Kirstein it was a period of agonizing suspense. They couldn’t be sure
that Third Army would move into the area in which the mine lay. Just
as their hopes began to fade, occupancy of the cherished area did
fall to Third Army. Tactical troops were alerted to the importance of
the isolated mountain region. It was of no significance as a military
objective and would doubtless otherwise have been left unoccupied
for the moment. They pressed forward through Bad Ischl and the
wild confusion of capitulating German troops to the wilder confusion
of surrendering SS units in the little village of Alt Aussee itself. From
there it was but a mile to the mine.
When they reached the mine, they found it heavily guarded by
men of the 80th Infantry Division, but the mine had been dynamited.
It wasn’t possible to go into the mine chambers. Armed with
acetylene lamps, Posey and Lincoln entered the main tunnel. They
groped their way along the damp passageway for a distance of a
quarter of a mile or more before they reached the debris of the first
block. After assessing the damage they returned to consult the
Austrian mineworkers. The miners said it would take from ten days
to two weeks to clear the passageway. Captain Posey thought that
the Army Engineers could do it in less than a week, perhaps in two
or three days. Both were wrong. They entered the first mine chamber
the next day.
And now, here before us, stood the fabulous panels which they
had found on that May morning a few weeks before. While we
examined them, Sieber pieced out the one gap in the story of the
altarpiece: the Nazis had taken it from Paris to the Castle of
Neuschwanstein where a restorer from Munich worked on the
blisters which had developed on some of the panels. The altarpiece
remained at the castle for two years. It was brought to the mine in
the summer of 1944. Pieces of waxed paper were still affixed to the
surface of the panels, on the places where the blisters had been laid.
The big panel representing St. John had split lengthwise with the
grain of wood. This had happened at the mine. Sieber had repaired
it, and George said he had done a good job. As we were leaving the
Mineral Kabinett, Sieber asked me if Andrew Mellon really had
offered ten million dollars for the altarpiece. People had said so in
Berlin. I hated to tell him that the story was without foundation, so far
as I knew.
When we came out of the mine at noon we found that Steve and
Shrady had finished loading two trucks. They said they had enough
pictures left to fill two more. The afternoon crew would be coming on
at four. In the meantime it was up to Lamont and me to select at
least two hundred paintings, so that the loading could go on without
interruption.
After lunch we returned to the mine with Sieber. This time the trip
on the train was much longer. Our objective was a part of the mine
called the Springerwerke. Owing to the peculiar honeycomb
structure of the mine network, it had been necessary to establish
guard posts at intervals along the tunnels. One of these was at the
entrance to the Springerwerke. Two GIs were on duty. It was a
dismal assignment as well as a cold one. They were bundled up in
fleece-lined coats which had been made for German troops on the
Russian front. George and Steve had obtained about two hundred of
these coats and were using them for packing unframed pictures.
Each time a convoy of empty trucks returned from Munich, they
counted the coats carefully to make sure that none had disappeared.
The Springerwerke contained more than two thousand paintings.
They were arranged in two tiers around three sides and down the
center of a room fifty feet long and thirty feet wide. In one section we
found thirty or forty Italian paintings of the fourteenth and fifteenth
centuries from the well known Lanz Collection of Amsterdam. Next to
them we came upon the group of canvases, which the Germans in
their greedy haste had filched from Bruges when they had made off
with the Michelangelo Madonna. Aside from these two lots, the
pictures had been stored according to size rather than by
provenance. It was a bewildering assortment. Quality was to be our
guide in making this initial selection. And as a kind of corollary, we
were to set aside for shipment all pictures bearing the infamous
“E.R.R.” stencil—the initials of the Rosenberg looting organization.
We had Sieber and four of the gnomes to help. Sieber stood by with
list and flashlight. Two of the gnomes hauled out the pictures for us
to examine. The other two put protective pads of paper filled with
excelsior across the corners of the ones chosen.
By the end of the afternoon we had picked out between a hundred
and fifty and two hundred paintings. The crew which had come on at
four had already gone up with one trainload. We took stock of the lot
waiting to go. We hadn’t done so badly: our selection included works
of Hals, Breughel, Titian, Rembrandt, Tintoretto, Rubens, Van Dyck,
Lancret, Nattier, Reynolds, and a raft of smaller examples of the
seventeenth century Dutch school. Not a dud among them, we
agreed smugly.
Before we knocked off for supper, Sieber showed us an adjoining
room divided into small compartments. Each one contained a
miscellaneous assortment of art objects—pictures, porcelain, bric-a-
brac of various kinds. Each compartment bore a label with the name
of a different family, fifteen or twenty in all. They were the pilfered
possessions of Viennese Jewish families. Our feelings were of both
pathos and disgust. After working with the fruits of looting on a grand
scale, we found these trifles sordid evidences of greedy persecution.
Lamont and I spent the next two days in the Springerwerke. We
worked nights as well. It was a molelike existence. On the third
morning we transferred our base of operations to the Kammergrafen,
the largest of the mine caverns and the most remote. It was three-
quarters of a mile from the mine entrance. Whereas the other mine
chambers were on one level, the great galleries of the
Kammergrafen were on several. Beyond those in which floors had
been laid, there were vast unlighted caves of echoing blackness.
The galleries were so high that those on the first level could
accommodate three tiers of pictures between floor and ceiling, while
those on the second had four tiers.
The records listed six thousand pictures. In addition there were
quantities of sculpture, hundreds of examples of the very finest
eighteenth century French cabinetwork, tapestries and rugs, and the
books and manuscripts of the Biblioteca Herziana in Rome—one of
the greatest historical libraries in the world. Kammergrafen was
quality and quantity combined, for here had been stored the
collections for Linz.
Among the pictures, for example, were canvases from the
Rothschild, Gutmann and Mannheimer collections, the celebrated
tempera panels of the fourteenth century Hohenfurth altarpiece,
Rembrandts and other great Dutch masters from the stock of
Goudstikker, who had been the Duveen of Amsterdam, a collection
of French pictures known as the “Sammlung Berta,” and hundreds of
nineteenth century German paintings, these last the objects of
Hitler’s special veneration.
The sculpture ranged from ancient to modern, with notable
emphasis on examples of the Gothic period. There were Egyptian
tomb figures, Roman portrait busts, Renaissance and Baroque
bronzes, exquisite French marbles of the eighteenth century and
delicate Tanagra figurines. A bewildering hodgepodge of the plastic
arts.
There were tapestries from Cracow, furniture from the Castle at
Posen, rows of inlaid tables and cabinets from the Vienna
Rothschilds, shelves and cases filled with the finest porcelains, prints
and drawings of the sixteenth, seventeenth and eighteenth centuries,
the decorations from the Reichskanzlei in Berlin, and Hitler’s own
purchases from the annual exhibitions of German art in Munich.
Such was the Kammergrafen treasure. And the best of it, as I have
said before, was to have adorned the galleries of the unbuilt Führer
Museum at Linz, the city by the Danube which Hitler aspired to raise
to the dignity of Vienna.
The rarest treasure of that collection was the celebrated Vermeer
Portrait of the Artist in His Studio. This superb work of the
seventeenth Dutch master, by whom there are only some forty
unquestioned examples in the world, had been for years in the
collection of Count Czernin at Vienna. The collection was semipublic;
I had visited it before the war. Known simply as the “Czernin
Vermeer,” the picture had long been coveted by the great collectors.
It had remained for Hitler to succeed where others had failed: he
acquired this masterpiece in 1940 for an alleged price of one million,
four hundred thousand Reichsmarks—part of his earnings from the
sale of Mein Kampf. He boasted at the time that Mr. Mellon had
offered six million dollars for it. Whether the sale was made under
duress is still a matter of controversy. Members of the Czernin family
today contend that it was. The picture has now been returned to
Vienna where the matter will be ultimately decided.
Rivaling the Vermeer in international significance were the fifteen
cases of paintings and sculpture from Monte Cassino. The paintings
included Titian’s Danaë, Raphael’s Madonna of the Divine Love,
Peter Breughel’s Blind Leading the Blind, a Crucifixion by Van Dyck,
an Annunciation by Filippino Lippi, a Sacra Conversazione by Palma
Vecchio, a Landscape by Claude Lorraine, and Sebastiano del
Piombo’s Portrait of Pope Clement VII. Among the sculpture were
antique bronzes of the greatest rarity and importance from
Herculaneum and Pompeii. All had belonged to the Naples Museum.
In 1943 the Italians had placed them, together with one hundred and
seventy-two other cases of objects from the Naples Museum, in the
Abbey of Monte Cassino for safekeeping. The following January,
arrangements were made for all of the cases to be returned to the
Vatican. When they arrived, fifteen were missing. Members of the
Hermann Göring Division had carried them off as a birthday gift for
the Reichsmarschall. Göring was incensed, when he learned of the
arrival of these treasures, and refused to accept them. There is
reason to believe that such was his reaction, for he had striven to
maintain a semblance of legality in his art transactions. Even this
rapacious collector could not have interpreted the behavior of his
loyal officers as “correct,” so far as the Monte Cassino affair was
concerned. After the Reichsmarschall’s refusal of the cases, they
were brought to Alt Aussee for storage, pending their later return to
Italy.
The Springerwerke had been child’s play compared with the task
confronting us in the Kammergrafen. We began arbitrarily with the
big pictures. Titian, Tintoretto, Veronese, Van Dyck, Rubens,
Rembrandt—all were represented in profusion. Many of them were
from private collections in Holland, Belgium and France, and were
unknown to us save through reproduction. It was a great lesson in
connoisseurship, particularly when we had exhausted the “stars” and
come to the lesser masters. The Dutch school of the seventeenth
century was abundantly represented. There were scores, hundreds,
of still lifes and flower paintings. My predilection for them amused
Lamont and Sieber. I had always admired these incredibly deft
creations of the seventeenth century Dutch artists, and here was an
unparalleled opportunity to study them.
There was one peculiar thing about our selections: if a picture
looked good to us down in the mine, it invariably looked better when
we examined it later in the light of day at the mine entrance. This
happened time and again. I remember one instance in particular.
The painting was a large Rembrandt, a study of two dead peacocks.
Down in the mine we had looked at it without much enthusiasm,
though we admired it, and had even hesitated to include it in that first
selection, which was to number only the best of the best. The next
morning, as it was being loaded onto the truck, we were struck by its
distinction.
And I remember the next time I saw that picture: it was at the
Rijksmuseum in Amsterdam. It was included in the small group of
outstanding Dutch masterpieces returned to Holland by special plane
as a gesture of token restitution in the name of General Eisenhower!
Lamont and I liked Sieber, the German restorer. Lamont referred
to him as the “tragic Gilles in white.” Not that there was anything
particularly tragic about him, any more than there was about the
plight in which most Nazis found themselves. He made no bones
about having joined the Party in the early thirties—ironically enough
at the suggestion of one of his clients, a Jewish art dealer, who had
thought it would be a good thing for Sieber’s business. Sieber had
been a restorer of pictures in Berlin and had done a little dealing on
the side. It was the half-mournful expression he perpetually wore,
together with his white costume, that accounted for Lamont’s
appellation. George had described him as a good, run-of-the-mill
restorer, perhaps a little better than average. He had sized him up as
a man ninety-eight per cent preoccupied with his profession and
possibly two per cent concerned with politics. And George, as I think
I have observed before, was a good judge of men. Sieber was a
quiet, willing worker. He was neither fawning on the one hand, nor
arrogant on the other. When you asked him a question, you always
got a considered answer.
One evening we drew Sieber out on the subject of the attempted
destruction of the mine. We had heard several versions of this
fantastic plot and, according to one, Sieber had been instrumental in
foiling the conspirators. The story was as follows: On the tenth,
thirteenth and thirtieth of April 1945, Glinz, the Gauinspektor of Ober-
Donau, had come to the mine with eight great cases marked in black
letters “Marmor—Nicht stürzen,” that is, “Marble—Don’t drop.” He
was acting on explicit instructions from Eigruber, Gauleiter of the
region, to place them at strategic positions in the mine tunnels. Each
case contained a hundred-pound bomb. Had these bombs been
detonated, the entire contents of the mine would have been
destroyed. The resulting cave-ins would have blocked every means
of access. It would have taken months to repair the apparatus which
carried off the water seeping constantly into the mine chambers. By
that time the treasures they contained would have been completely
ruined. It is generally agreed that Eigruber had obtained Hitler’s tacit
consent to this artistic Götterdämmerung, if not his actual approval of
it.
I learned later that Captain Posey found a letter from Martin
Bormann, Hitler’s deputy, stating in the first paragraph that the
contents of the mine must, at all costs, be kept from falling into the
hands of the enemy. And then the second paragraph stated that the
contents of the mine must not be harmed.
Members of the Austrian resistance movement got wind of this
diabolical plan and took Sieber into their confidence. His intimate
knowledge of the mine passageways enabled him to set off small
charges of dynamite here and there along the tunnels without
endangering the contents of the chambers beyond. The resulting
damage was slight and served a twofold purpose: it gave the
impression that the mine had been permanently walled up; and, if
that ruse were discovered, immediate access to the art works
themselves was denied the plotters. Eigruber did discover that his
attempt had been thwarted and in his rage gave orders for the
counterconspirators to be rounded up and shot. But by that time it
was the seventh of May, so the tragedy was mercifully averted.
Sieber told the story in a straightforward, factual way. I don’t think
it mattered to him who got control of the mine, but it was simply
unthinkable that any harm should come to the precious things it
housed. It was very much to his credit that he never capitalized on
the part he played in this affair. The only other reference he made to
it was when he later showed us the places where he had set off the
charges of dynamite.
During all the time we were at the mine, Sieber made only one
request of us. It came at the very end of our stay and was
reasonable enough: he asked if we could expedite his return to
Germany. There seemed to be little prospect of regular employment
for him in Austria, but that was of less concern to him than the
welfare of his wife and young daughter who lived with him in a house
near the mine. Some months later, one of our officers tried to get him
a job in Wiesbaden, but he was not acceptable to the Military
Government authorities there because of his political affiliations. The
last I heard of him, he and his family were still at Alt Aussee, waiting
for permission to go back to Germany. I hope they finally got it.
Our concentrated efforts underground produced the desired
results. Pictures were coming out of the mine at such a prodigious
rate that George called a halt. Enough of a backlog had been
accumulated to make further selection down in the mine
unnecessary for a couple of days. Lamont and I had better help with
the actual loading.
Loading a truck was a specialized operation, and George had
perfected the technique. Lamont, Steve and Shrady were his pupils.
That left me the only neophyte. So far I had had experience only with
the loading of cases and heavier objects such as furniture and
sculpture.
Packing pictures, especially unframed ones—and there were a
great many of the latter at Alt Aussee—was an altogether different
problem. The first step was to place a length of waterproof paper
over the side bars of the truck and spread it smoothly on the floor to
the center of the truck bed. For this we had a large supply of stout,
green, clothlike paper which had been used by the Wehrmacht as
protection against gas attacks. Then a strip of felt was laid over the
paper. The third step was to place “sausages” in two rows, end to
end, on the floor of the truck. The space between the rows would
depend on the size of the pictures to be loaded, for they were
intended to cushion the shock as the trucks rumbled along over
bumpy roads. The “sausages” had been George’s invention. Packing
materials of all kinds were at a premium, and certain types just didn’t
exist. To make up for the lack of the usual packer’s pads, George
had improvised this substitute. In one of the mine chambers he had
found a large supply of ordinary curtain material of machine-made
ecru lace. This had been cut up into yard lengths, eighteen inches
wide. When rolled around a central core of coarser cloth, or
sometimes excelsior, and tied with string, they were a very
satisfactory “ersatz” product. We used to refer to them augustly as
the pads made from Hitler’s window curtains. Their manufacture was
periodically one of the major industries at the mine. George had
trained a crew of the gnomes and they loved to turn them out. It was
easy work. Seated at long benches they resembled a kind of Alpine
“husking bee.”
Once the paper, felt and “sausages” were in place, the pictures
could be brought to the truck. One after another they were placed in
a stack leading from the sideboards of the truck to the center. Pads
and small blankets were inserted between them to prevent rubbing.
To ensure safe packing, all the pictures in a given row had to be
carefully selected as to size, ranging from large to medium in one,
and from medium to small in another. That was the most tedious part
of the entire operation. As soon as a row had been “built,” it required
only a few minutes to bring first the felt and then the green paper
over the top of the row, tuck both down along the sides, and then
lash the whole stack firmly to the side of the truck. By this method it
was possible to load as many as a hundred and fifty medium-sized
canvases on a single truck, for three rows of twenty-five each could
be built up on either side of our big two-and-a-half-ton trucks. A truck
loaded in this way could often accommodate several pieces of
sculpture as well. Carefully padded and swathed in blankets, these
could be placed down the center. The final step was to adjust the
tarpaulin over the bows and to close the tailboard. A truck of the size
we used could normally be loaded in two hours.
We were at the mercy of the weather, as far as loading was
concerned. On rainy days we could work on only one truck at a time,
because there was but one doorway with a protective stoop under
which a single truck could park. Taking advantage of the sunny
mornings, we would divide up into two teams—George and Steve on
one truck, Lamont and I on another. As soon as a truck was filled
and the tarpaulin securely fastened down, it was driven to one side
of the narrow terrace in front of the building. The average convoy
consisted of six trucks.
We had a crew of eighteen Negro drivers. Barboza, the C.O., was
a very starchy lieutenant, Jamaica born. He and his men were
billeted down in the little town of Bad Aussee. They were magnificent
drivers but a bit reckless. Their occasional disregard for their
vehicles was a worry to George. It would have been so with any
drivers, I guess. A breakdown on the steep mountain roads could be
a serious matter. It meant the complete disruption of the convoy
schedule, involving reloading en route. To provide for this
contingency, we made a practice of loading the trucks to three-
quarters of their capacity. The contents of a single truck could thus
be absorbed by the others.
When a convoy was ready to start, either George or I would lead
off in a jeep and escort the six trucks down the precipitous road to Alt
Aussee. Two half-tracks from the 11th Armored Division, as front and
rear guards, would be waiting to accompany the convoy to Munich.
At breakfast one morning George said, “This looks like a good day
to load the gold-seal products.” He meant the Michelangelo
Madonna and the Ghent altarpiece. This was an important event, for
they were unquestionably the two most precious things still at the
mine. Every possible precaution would have to be taken to make this
operation a success. It must go off without a hitch. If anything
happened to either of these masterpieces, the repercussions would
be catastrophic. They would overshadow all the accomplishments of
our MFA&A officers.
For the past several days, George and Steve had been working on
the Madonna. She was now heavily padded and trussed up like a
ham, ready to be brought out of the mine. We all went down to the
Kaiser Josef chamber where we had first seen her. George made a
final inspection of the ropes and pulleys which had been set up to
hoist her onto the waiting train. Then, with a satisfied smile, he said,
“I think we could bounce her from Alp to Alp, all the way to Munich,
without doing her any harm.”
Once the statue was gently loaded on the little flat car, the train
pulled slowly out of the mine chamber and switched back onto the
track of the main tunnel. From there it chugged slowly—George
walking alongside—to the mine entrance where the truck stood
waiting. This truck and the one which was to carry the altarpiece had
been put in perfect condition. And George had put the fear of God
into the two drivers, both of whom he had personally chosen from
our crew. A dozen of the gnomes were waiting to lift the marble onto
the truck. We slid it cautiously to the fore part where boards had
been laid parallel and nailed to the floor. These would prevent
shifting. On either side of the statue, small packing cases about two
feet square were arranged in even rows and lashed firmly to the
sides of the truck. These cases had been stored in a chamber of the
mine called the Kapelle. They contained the coin collection intended
for the Linz Museum and were accordingly marked “Münz Kabinett”
or Coin Room. Blankets were wedged in between the cases and the
statue. The large case containing the Greek sarcophagus from
Salonika was set in place behind the statue and similarly secured to
the floor and at the sides. That done, the truck was ready to go.
As George was putting the finishing touches on the packing, he
said, “Tom, will you go down and get the ‘Lamb’?” To be entrusted
with the removal of the great altarpiece was an exciting assignment.
I wanted to share it with someone who would also get a kick out of
telling his grandchildren that he had actually brought the famous
panels out of their underground hiding place. I called Lamont, and
the two of us, followed by eight of the gnomes, hitched four of the
“dogs” to the little engine and proceeded to the Mineral Kabinett.
One of the “dogs,” especially designed to carry pictures of unusual
height, had a lower bed than the others. We would use this one for
the big central panel of the altarpiece. Otherwise it would not clear a
portion of the mine tunnel where the jagged rocks hung low over the
track.
The panels were now in their cases, and it was a relatively simple
matter to carry them from the storage room to the train. We had to
make two trips down and back in order to get all ten cases up to the
mine entrance. They were much lighter than the statue, but the
loading was a more exacting undertaking. Lashing them upright in
parallel rows, in the truck, and stowing cases on either side for
ballast, took time. We didn’t finish until well after six. It had taken
most of the day to load the “gold-seal products.”
Mary Magdalene by Cranach. Göring was
especially fond of Cranach’s work and owned many
paintings by him.
The Mystic Marriage of Saint Catherine by David was one of the
finest in the Göring Collection.
Diana, an exquisite Boucher acquired by Göring
from the Rothschild Collection, has been returned to
France.
Atalanta and Meleager by Rubens, found in the
Göring Collection, was from the Goudstikker
Collection of Amsterdam.
That night we held a conference in George’s room. He was to go

to Munich the next morning with the convoy to supervise the
unloading of the Madonna and the “Lamb” at the depot. He expected
to come back directly, but might have to go on up to Frankfurt. He
mapped out the work he wanted us to do while he was away. In
addition to the job at the mine, there was a special one down at Bad
Ischl. A series of famous panels by Albrecht Altdorfer, one of the
greatest German painters of the fifteenth century, was stored on the
second floor of a highly combustible old inn. These panels were
among his finest works and belonged to the monastery of St. Florian
outside Linz. Fifty or sixty pieces of sculpture—mostly polychromed
wood figures, fifteenth century Gothic—also the property of the
monastery, were stored there too. George thought we’d better figure
on two trucks. We were to pick up the stuff, bring it back to the mine
and then send it to Munich with the next convoy.
We made notations of what he had told us, and then Steve
produced drinks—something of an occasion, for liquor was hard to
come by at the mine. Somehow he had got hold of a bottle of cognac
and insisted on making “Alpine Specials.” This was a drink consisting
of a jigger of cognac and an equal amount of a pink syrupy liquid
resembling grenadine. Steve prized the syrup. Eder, the chemical
engineer at the mine, had concocted it especially for him. The
mixture made a drink of dubious merit. We drank to the success of
George’s trip to Munich.
The convoy got off early the next morning. Lamont and I went
down with George as far as the village. Two half-tracks were waiting
to escort the trucks to Munich. They were equipped with radios for
intercommunication, in case of delays along the way. Between Alt
Aussee and Salzburg, the road led through isolated country.
Conditions were as yet far from settled. Small bands of SS troops
still lurked in the mountains. The half-tracks weren’t just going along
for the ride.
When we returned to the mine, we found Steve and Shrady in
conversation. They were planning an excursion and asked us to join
them. We expressed pious disapproval of letting up on the job the
minute George’s back was turned. Steve’s answer was, “That’s a
crock. I haven’t taken a day off in two weeks. You can do as you like,
me lads; I’m off to see the wizard.” The two of them climbed into the
sporty Mercedes-Benz convertible which Shrady had recently
acquired, and drove off down the mountain.
“I have an idea,” said Lamont. “They can have their fun today.
We’ll have ours tomorrow, while they pick up the Altdorfer panels at
Bad Ischl.”
Lamont turned to the loading of trucks, while I went down to the
Kammergrafen with Sieber. In the course of the morning I selected
approximately two hundred pictures. We concentrated on those of
small size, which were stored on the top racks, and the work went
rapidly. Among the paintings we chose was one which had a
typewritten label on the back. I read the words “Von dem Führer
noch nicht entschieden”—not yet decided upon by the Führer. I
asked Sieber what this signified. He explained that every picture
intended for the Linz Museum—and this was one of them—had to be
personally approved by Hitler before it could be officially included in
the prospective collection. I could easily understand that the Führer
would have wanted to examine the more important acquisitions, but
that each canvas had to receive his personal approval struck me as
preposterous. Hitler had entrusted the formation of his collections
first to Dr. Hans Posse, a noted scholar, and, after his death, to Dr.
Hermann Voss, director of the Dresden Gallery. This meticulous
procedure, involving the submission of all pictures to the Führer in
Munich, must have been trying to those two luminaries of the
German art world.
When I came up from the mine for lunch I found that Lamont had
completed the loading of two trucks. As the stock of pictures in the
packing room at the mine entrance was almost exhausted, he said
that he would join Sieber and me in making further selections. We
returned to the Kammergrafen and continued with the smaller
pictures.
On one of the top shelves we found a cardboard carton bearing
the name of Dr. Helmut von Hümmel, who had been connected with
the formation of the Linz collections. The label indicated that the
contents had been destined for the museum. On the carton
appeared the word “Sittenbilder.” Lamont and I knew the word
“Bilder” meant pictures, but the other two syllables conveyed
nothing. Sieber knew very little English but tried to explain. He
thought perhaps the word meant “customs,” or something like that. I
thought that I understood him. He probably meant that they were
little scenes from everyday life.
We opened the carton. On top were three small watercolors, with
beautiful gray-blue mounts and carved, gilt frames. If they were not
by François Boucher, they were by a close pupil. The workmanship
was exquisite and they were highly pornographic. So these were
“Sittenbilder.” In our limited German we tried to tell Sieber that they
might be called “scenes from life,” but hardly everyday life.
The rest of the things in the carton were of the same order, some
of them contemporary, all of them licentious. None approached the
first three watercolors in sheer virtuosity of technique. We wondered
just which department of the Linz Museum would have harbored
them.
Later we showed them to Steve. When he looked at the three
watercolors he asked, “Who did those?”
“They look very much like Boucher,” I said without thinking.
“Boucher?” asked Steve incredulously. “Not the fellow who painted
that ‘Holy Family’ I saw this afternoon?”
Lamont said quickly, “Tom means a pupil of Boucher, not Boucher
himself.”
“Well, that’s more like it,” said Steve. “I didn’t see how an artist
who painted anything so beautiful as that big picture could paint
smutty things like these.”
The Holy Family to which he referred was a sensitive and tender
representation. Steve’s point was well taken. The fact that he had
not seen much eighteenth century French painting didn’t alter the
validity of his argument.
On our way up from the Kammergrafen that afternoon we stopped
at the Kapelle. This was one of the mine chambers which I had
visited only long enough to take out some of the cases used in
packing the Bruges Madonna. In addition to the Münz Kabinett
collections, the Kapelle contained the magnificent collection of
Spanish armor—casques, breastplates, full suits of armor, and a
great number of firearms—which had been gathered together at
Schloss Konopischt by the late Archduke Franz Ferdinand. Most of it
was of sixteenth century workmanship, exquisitely inlaid with gold
and silver. Formerly Austrian, it had been the property of the Czech
government since the last World War. Nonetheless, the Germans
had carried it off, using some flimsy nationalistic argument to justify
their action. While the atmosphere of the mine was excellent for
paintings, it was not satisfactory for metal objects. Consequently,
every piece in the Konopischt collection had been heavily coated
with grease to keep it from oxidizing.
Their storage room, the Kapelle, was—as the name indicates—a
chapel, dedicated to the memory of Dollfuss, the Austrian
Chancellor. It had an electrically illuminated altar, hewn from a block
of translucent salt crystal, which was one of the sights of the mine.
That night when Steve and Shrady returned from their outing we
all had hot chocolate and cheese and crackers in the comfortable
kitchen on the ground floor of the main building. They had had a
wonderful day. Maria, Shrady’s interpreter-secretary, had been with
them. They had gone first to St. Wolfgang. There a sentry had tried
to prevent them from driving up the road leading to the little church.
Leopold, the Belgian king, was living near there with his wife, and
motorists were not allowed on that road. But they had got around the
sentry and gone into the church to see the wonderful carved
altarpiece by Michael Pacher. Steve had brought us some colored
photographs of it. Afterward they had had a swim in the lake and a
picnic lunch. And in the afternoon—this had been the high spot of
the day—they had gone over to Bad Ischl and called on Franz Léhar.
The old fellow had been delighted to see them, had played the Merry
Widow waltz for them and given them autographed photographs. It
sounded like fun. Our day at the mine had been very prosaic in
comparison. Mention of Bad Ischl reminded Lamont of his scheme
for the next day. He proposed to Steve and Shrady that they should
call for the Altdorfer panels. They fell in with the suggestion at once,
and before we could explain that we thought we’d take the day off,
Steve was telling us about some of the things we should see in the
neighborhood.
We slept late and when we got up the sky was gray and
threatening. It was no day for an outing. In fact it was so cold that we
decided we’d be warmer down in the mine. There was still one series
of chambers which we had not explored. This was the Mondsberg,
and it took us almost three-quarters of an hour to reach it. The
pictures were arranged on racks as in the Kammergrafen and the
Springerwerke, only in the Mondsberg there were a great many

(download pdf) Biostatistics For The Biological And Health Sciences 1st Edition Triola Solutions Manual full chapter

Uploaded by

Copyright:

Available Formats

You might also like

(download pdf) Biostatistics For The Biological And Health Sciences 1st Edition Triola Solutions Manual full chapter

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

(download pdf) Biostatistics For The Biological And Health Sciences 1st Edition Triola Solutions Manual full chapter

Uploaded by

Copyright:

Available Formats

Biostatistics For The Biological And

Health Sciences 1st Edition Triola

Biostatistics for the Biological and Health Sciences

Biostatistics for the Biological and Health Sciences

Biostatistics for the Biological and Health Sciences

Biostatistics An Applied Introduction for the Public

Biocalculus Calculus Probability and Statistics for the

Laboratory Manual and Workbook for Biological

Statistics For Engineering And The Sciences 5th Edition

Anatomy and Physiology Foundations for the Health

Chapter 6. Estimates and Samples with One Sample

The formulas used to compute pˆ and E in exercises 5-7 are

pˆ = [(upper confidence limit) + (lower confidence limit)]/2, and

The formulas used to compute pˆ and E in exercises 9-12 are

pˆ = (upper confidence limit) + (lower confidence limit)/2, and

25. Interpreting Calculator Display

26. Interpreting Calculator Display

27. Mendelian Genetics

28. Drug Testing

29. Smoking and College Education

30. Sample Size for Left-Handed Golfers

31. Sample Size for Plant Growers

32. Color Blindness

33. Cell Phones and Cancer

34. Gender Selection

35. Gender Selection

36. Pilot Fatalities

37. Wearing Hunter Orange

38. Effects of Sample Size and Confidence Level

39. Using Finite Population Correction Factor

6-3 Estimating a Population Mean: σ Known

2. By inspecting the table on p. 261, we se that the critical value is 1.96.

5. Since the sample size is greater than 30, using E = zα / 2σ / n is justified.

7. Since the original population is normal, and σ is known, using E = zα / 2σ / n is justified.

E = zα / 2σ / n and x − E < µ < x + E

17. The point estimate for the population mean µ is x = 318.1.

18. The confidence interval is 262.09 < µ < 374.11

19. Since the lower confidence interval limit is x − E , we find that

21. Everglades Temperatures

22. Weights of Bears

23. Cotinine Levels of Smokers

24. Head Circumferences

25. Sample Size for Mean IQ of Biology Majors

26. Sample Size for Estimating Income

27. Sample Size using Range Rule of Thumb

28. Sample Size Using Sample Data

29. Sample Size Using Sample Data

We round up, so the minimum required sample size is n=58.

30. Confidence Interval with Finite Population Correction Factor

31. Sample Size with Finite Population Correction Factor

6-4 Estimating a Population Mean: σ Not Known

The formula used to compute E in exercises 11 and 12is

In Exercises 13-24, construct the indicated confidence intervals.

13. Historical Corn Data

14. Crash Hospital Costs

15. Heights of Parents

16. Monitoring Lead in Air

17. Shoveling Heart Rates

18. Pulse Rates