01
Fall, 2009
1
488
484
2
478
478
3
480
492
4
426
444
5
440
436
6
410
398
7
458
464
8
460
476
(a) According to the data, can you conclude, at the significance level of 0.10, that the corneal
d 0
sd
40
10.744
1.053
Since t t 81, 0.05 1.895 , do not reject H 0 at 0.10 , and conclude that the
average corneal thicknesses are unaffected by glaucoma.
(b) A 90% CI for 1 2 is given by
d t n 1,
sd
n 4 1.895 10.744
8 [ 11.198,3.198]
2A (for AMS PhD students). Suppose we have two independent random samples from
2
2
two normal populations: X 1 , X 2 , , X n1 ~ N 1 , , and Y1 , Y2 , , Yn2 ~ N 2 , .
Please derive the pooled-variance t-test using the pivotal quantity method. Please make
sure that you include the following key steps.
2t
M t
k/2
(c) Then we found, from the theorem of sampling from the normal population, and
the independence properties of the random samples, that Z and W are
independent, and therefore, by the definition of the t-distribution, we have
X Y 1 2 ~ t
obtained our pivotal quantity: T
n n 2 , where
S p2
n1 1 S12 n2 1 S 22
n1 n 2 2
S p 1 / n1 1 / n 2
2B (for all except-AMS PhD students) . Let X1, X2, , Xn be a random sample from a
normal population N(, 2). Furthermore, the population variance 2 is known.
(1). For a 2-sided test of H0: = 0 versus Ha: 0, at the significance level , please
derive the one-sample z-test using the pivotal quantity method. (* Please include the
derivation of the pivotal quantity, the proof of its distribution, and the derivation of the
rejection region for full credit.)
(2) Please derive the 100(1-)% confidence interval for using the pivotal quantity
method.
Solution: [1] Since this is inference on one population mean, we start with its point
estimator the sample mean, X . The distribution of X is N , 2 / n . The distribution
of X is not entirely known because is unknown. By taking the linear transformation
Z
X
, however, we can see that Z is a function of the sample statistics and the
/ n
tZ
/ n
/ n / n
MZ t E e E e
e
E e
n
t
t
t
n
i1 X i
et* i1 X i
/ n
e / n E e n
e
E
t
/ n
t
/ n
t
in1 E et X i e
n t *
in1 e
t*
2 t*
* 2
t
/ n
t
/ n
n
2
n 2
e2
Therefore Z ~ N 0,1 .
[3] Next we derive the one-sample z-test and its rejection region.
For a 2-sided test of H0: = 0 versus Ha: 0, the test statistic is the pivotal quantity
X 0
at = 0, that is, Z 0
. Intuitively, we would reject H0 in favor of Ha if Z 0 c .
/ n
The problem is how to determine c. By the definition of the significance level, we have
Preject _ H0 | H0 P Z0 c | H0 2PZ0 c | H0
Z 0 Z / 2
3 (for all). A new method of making concrete blocks has been proposed. To test whether
or not the new method increases the compressive strength, five sample blocks are
made by each method. The compressive strengths in 10 pounds per square inch are
listed here:
New Method 15 14 13 15 16
Old Method 13 15 13 12 14
(a). Please construct a 95 % percent confidence interval for the mean difference between
the compressive strengths by two methods.
(b). At the significance level .05, can you conclude that the new method increases the
compressive strength?
(c). What assumptions do you need for the inference in parts (a) and (b)?
(d). Please write up the entire SAS program necessary to answer questions raised in (a),
(b) and (c). Please include the data step as well as tests for testing for various
assumptions.
Solution:
1
1
1.14
n1 n2 2
55 2
Therefore 95% C.I. is [-0.46, 2.86].
(b) Using t-test with hypotheses H 0 : 1 2 v.s. H a : 1 2 ,
X X 2 0 14.6 13.4
t0 1
Thus, we cannot reject H0. There isn't enough evidence to conclude that the new
method increases the strength.
(c) (1) Both populations are normally distributed
2
2
(2) 1 2
4. (for all students in class). John Pauzke, president of Cereals Unlimited, Inc., wants to
be very certain that the mean weight of packages satisfies the package label weight of
16 ounces. The packages are filled by a machine that is set to fill each package to a
specified weight. However, the machine has random variability measured by 2. John
would like to have strong evidence that the mean package weight is above 16 ounces.
George Williams, quality control manager, advises him to examine a random sample of
25 packages of cereal. From his past experience, George knew that the weight of the
cereal packages follows a normal distribution with standard deviation 0.4 ounce. At the
significance level =.05,
(1) What is the decision rule (rejection region) in terms of the sample mean x ? Please
derive the general formula using the concept of Type I error rate.
(2) What is the power of the test when =16.2 ounces? Please derive the general formula
for power calculation first.
(3) What is the sample size necessary to ensure a power of 80% when =16.2 ounces?
Please derive the general formula for sample size calculation based on the Type I and II
error rates first.
Solution:
(1)
H 0 : 16
H a : 16
Z0
X 0
X 16
.
/ n 0.4 / 25
P( Z 0 C | H 0 )
X 16
0.4
z X 16
z0.05
5
0.4 / 25
0.4
1.645 16.1316 .
5
(2)
H 0 : 16 0
H a : 16.2 a 16
Power P (reject H 0 | H a )
P ( Z 0 z0.05 | 16.2)
X 0
z0.05 | 16.2)
/ n
P ( X 16.1316 | 16.2)
P(
| 16.2)
0.4 / 5
0.4 / 5
P( Z 0.855)
0.805
P(
(3)
H 0 : 16 0
H a : 16.2 a 16
power 0.8 0.2. 0.4, 0.05 .
( z z ) 2 2
( a 0 ) 2
(16.2 16) 2
0.22
0.992016
24.8 25
0.04
Hence, about 25 packages of cereal should be sampled to achieve a power of 80% when
=16.2 ounces.