Chapter 8 PDF
Chapter 8 PDF
CHAPTER 8
TESTING HYPOTHESES:
ONE-SAMPLE TESTS
8-1 We must deal with uncertainty in our decisions because we rarely know the
values for population characteristics. In addition, there are other
circumstances unknown to us in any given situation. Statistical analysis can
reduce uncertainty, but not eliminate it entirely.
8-2 Theoretically, one could toss a coin a large number of times and see if the
proportion of heads was very different from 5. Similarly by recording the
outcomes of many dice rolls, one could see if the proportion of each side
was very different from 1/6. You would need a large number of trials for
each of these samples.
8-3 Yes, it is possible that a false hypothesis will be accepted. If the observed
value does not differ enough from the hypothesized one, the hypothesis will
be accepted. Acceptance of a hypothesis is based on probability and
therefore we can never be absolutely certain that our decision is correct.
8-5 There is always a statistical possibility that a sample does not accurately
represent the population from which it has been drawn.
8-6 We mean that we would not have reasonably expected to find that particular
sample if in fact the hypothesis had been true.
8-7 The level of probability, or certainty, depends upon how accurate our answer
needs to be.
8-9 The z value which leaves (100 - 98)% = 2% in the tails is ± 2.33, so the
interval should be the hypothesized value ± 2.33 standard errors.
147
Since x = 26100 < 27000, Ned should not purchase the Stalwarts. Depending
on how long ago Ned's last purchase was, it may no longer be reasonable to
suppose that = 6000. If has increased sufficiently, Ned's decision could
change.
Since ̅ =2.2 < 2.423, it is unreasonable to see such sample results if really
is 2.5 quarts; the store's claim is not correct.
8-13 H0 : = 45 H1 : > 45
8-14 A null hypothesis represents the hypothesis you are trying to reject.
Alternative hypotheses represent all other possibilities.
8-15 The significance level refers to the probability that the sample statistic lies
outside a specified interval around the hypothesized parameter, assuming the
hypothesis is true. Or, the significance level represents the risk of rejecting
the null hypothesis when it is true.
8-16 Type I: the probability that we will reject the null hypothesis when in
fact it is true.
148
Type II: the probability that we will accept the null hypothesis when in
fact it is false.
8-17 We would prefer to commit a. Type II error and let a. guilty person go free
than to sentence an innocent individual for a crime he or she didn't commit.
8-18 The significance level of a test is the probability of a Type I error, i.e., the
probability that we will reject the null hypothesis when it is, in fact, true.
This is because it indicates the percentage of sample means that fall outside
the limits of what we will accept as confirming the null hypothesis. Hence, if
a sample mean falls outside these limits but is truly from the hypothesized
population, it will lead to a Type I error.
8-19 a)
b)
149
c)
8-21 a) A Type I error (rejecting Ho when it is true) will release a bad battery for
sale, which could have very serious consequences for the patient who buys
it. On the other hand, a Type II error (accepting Ho when it is false) will
cause the company to discard a perfectly good battery, which presumably
does not cost very much to replace Thus it would appear that the engineer
would be reluctant to make Type I errors, but should not be too concerned
about making Type II errors.
150
8-22 A two-tailed test of a hypothesis will reject the null hypothesis if the sample
mean is significantly higher Q! lower than the hypothesized population
mean. Thus a two-tailed test is appropriate when we are testing whether the
population mean is different from some hypothesized value. A one-tailed
test, on the other hand, would be used when we are testing whether the
population mean is lower than or higher than some hypothesized value.
8-24 They should perform a lower-tailed test, with Ho: =3124, HI: < 3124.
8-25 : = 78 : 78 : = 15 : 15
151
8-31 = 16 n = 64 ̅ = 30.3
H0 : = 28 H1 : . 28 = .05
The upper limit of the acceptance region is ZU = 1.64, or
̅ U = z.05 /√ =28 + 1.64(16)/ √64 = 31.28 million dollars
̅ .
Since z = = 1.15, - 1.64(or ̅ 31.28), we do not reject Ho.
/√ /√
The Custom's Commissioner need not be concerned that smuggling has
increased above its historic level.
152
̅ . .
Since z = = -9.30, - 2.33(or ̅ 0.51), we should reject
/√ . /√
Ho' The rate of growth has decreased significantly, and we infer that this
was because of the oil embargo and its consequences.
8-34 From exercise 8-31, we have ( = 16, n = 64, and ̅ CRIT = 31.28.
.
a) P( ̅ 31.28 = 28) =p =p(z 1.64) = .5 - .4495 = .0505
/√
.
b) P( ̅ 31.28 = 29) =p =p(z 1.14) = .5 - .3729 = .1271
/√
.
c) P( ̅ 31.28 = 30) =p =p(z .64) = .5 - .2389 = .2611
/√
153
8-38 n = 85 ̅ = .1412
H0 : p = .19 H1 : p .19 = .04
The lower limit of the acceptance region is zL -1.75 or
. .
̅ L = - z.04 / =.19 – 1.75 = .1155
̅ . .
Since z = = -1.15, - 1.75(or ̅ .1155), we don’t
/ . . /
reject H0, Grant’s Western distribution is not significantly worse than its
eastern distribution.
they should reject Ho' Spray users are significantly less susceptible to
colds.
154
155
156
8-53 Ho : IJ = 28 H1 : IJ < 28
A Type I error would not be serious for the patient. A particular battery
would not be installed even though it would last long enough. A Type II
error could be serious. After installation, this battery might run down before
the scheduled operation, thus endangering the patient's life.
8-54 No, because each of the sample means is equally distant from the
hypothesized mean and, therefore, equally likely to lead to rejection or
acceptance of the null hypothesis.
158
̅ . .
Since z = = 4.19 2.33(or ̅ = .0964),
/ . . /
we reject Ho. Significantly more stocks than usual did set new highs
that day.
8-59
a) 2P(z > 2.15) = 2(.5000 - ,4842) = .0316
b) P(z > 1.6) = (.5000 - .4452) =.0548
c) P{z < -2.33) = (.5000 - ,4901) = .0099
Since x < 86.65· F, we do not reject Ho. The mean temperature is not
significantly higher than 84· F, so the plant should not be cited.
159
8-63 = 68 n = 90 ̅ =218.77
H0 : = 235 H1 : ≠ 235
.
P( ̅ ≤ 218.77 or ̅ ≥ 251.23 H0) = 2p
√
=2p(z ≥ 2.26) = 2( .5- .4881) = .0238
Thus, for all a S; .0238, we would accept Ho and conclude that the
average fare is not significantly different from $235.
8-64 N = 2400 n = 300 ̅ =57/300 = .19
H0 : = .15 H1 : .15 = .05
The upper limit of the acceptance region is Zu = 1.64, or
. .
̅ U = p + z.05 = .15 +1.64 = .1816
̅ . .
Since z = = . .
= 2.07 1.64 (or ̅ .1816), they
160
̅ . .
Since z = = 1.97 2.05(or ̅ = .7760 < .7782),
/ . . /
we do not reject H0. The survey does not support the editor's belief.
8-69 From exercise 8-26, we have u = 5.75, n =25, and xCRIT =42.59.
. .
a) P( ̅ ≤ 42.59 = 41.95) =p
. /√
=p(z ≤ .56) = .5 + .2123 =.7123
. .
b) P( ̅ ≤42.59 = 42.95) =p
. /√
=p(z ≤ -.31) = .5 - .1217 = .3783
. .
c) P( ̅ ≤ 42.59 = 43.95) =p
. /√
=p(z ≤‐1.18) = .5 + .3810 = .1190
161
8-71 From exercise 8-28, we have (j = 2100, n = 25, and xCRIT =:. 13521.4.
.
a) P( ̅ 13521.4 = 14000) =p
/√
=p(z - 1.14) = .5 - .3729 =.1271
.
b) P( ̅ 13521.4 = 13500) =p
/√
=p(z .05) = .5 + .0199 = .5199
.
c) P( ̅ 13521.4 = 13000) =p
/√
=p(z 1.24) = .5 + .3925 = .8925
8-72 n = 60 ̅ = .75
H0 : p = .85 H1 : p .85 α = .04
The lower limit of the acceptance region is zL= - 1.75, or
. .
̅ = p - z.04 = .85 – 1.75 = .7693
̅ . .
Since z = = - 2.17 - 1.75 (or ̅ .7693),
/ . . /
We reject H0. Her accuracy is significantly less than the asserted 85%.
162
. .
c) P( ̅ ≤ 43.06 = 43.95) =p
. /√
=p(z ≤ ‐ .77) = .5 - .2794 = .2206
We do not reject H0. The proportion of parents who redeem the coupon has
not significantly increased. However, there is nothing we can say about the
proportion who use the supplement.
We reject H0. The Company has fallen significantly below its target of a
48% market share.
163