Professional Documents
Culture Documents
Sample Survey
Learning Objectives
1.
Learn what a sample survey is and how it differs from an experiment as a method of collecting data.
2.
3.
4.
Learn about four sample designs: (1) simple random sampling, (2) stratified simple random
sampling, (3) cluster sampling, and (4) systematic sampling.
5.
Lean how to estimate a population mean, a population total, and a population proportion using the
above sample designs.
6.
7.
Learn how to choose the appropriate sample size using stratified and simple random sampling.
8.
Learn how to allocate the total sample to the various strata using stratified simple random sampling.
21 - 1
Chapter 21
Solutions:
1.
2.
a.
b.
sx =
c.
a.
b.
20
50
800 50
= 2.7386
800
8 400 80
Nsx = 400
= 320
400
80
c.
3.
a.
b.
= .0437
1000 99
c.
4.
B = 15
n=
(70) 2
4900
=
= 72.9830
2
2
(15)
(70)
67.1389
+
4
450
A sample size of 73 will provide an approximate 95% confidence interval of width 30.
5.
a.
= 10,040.83
771 50
approximate 95% confidence interval
149,670 2(10,040.83)
or
$129,588.34 to $169,751.66
b.
X = N x = 771(149,670) = 115,395,570
sx$ = N sx = 771(10,040.83) = 7,741,479.93
21 - 2
Sample Survey
771 50 (.36)(.64)
p = 18/50 = 0.36 and s p =
= .0663
49
771
6.
B = 5000/2 = 2500 Use the value of s for the previous year in the formula to determine the
necessary sample size.
n=
(31.3) 2
979.69
=
= 336.0051
2
2
(2.5)
(31.3)
2.9157
+
4
724
A sample size of 337 will provide an approximate 95% confidence interval of width no larger than
$5000.
7.
a.
Stratum 1: = 138
Stratum 2: x2 = 103
Stratum 3: x3 = 210
b.
Stratum 1
x1 = 138
30 200 20
sx1 =
= 6.3640
200
20
138 2(6.3640)
or
125.272 to 150.728
Stratum 2
x2 = 103
25 250 30
sx2 =
= 4.2817
250
30
21 - 3
Chapter 21
103 2(4.2817)
or
94.4366 to 111.5634
Stratum 3
x3 = 210
50 100 25
sx3 =
= 8.6603
100
25
210 2(8.6603)
or
192.6794 to 227.3206
c.
200
250
100
xst =
138 + 550 103 + 550 210
550
2
20
30
25
(550)
1
=
2
(550)
3,515,833.3 = 3.4092
8.
a.
b.
c.
21 - 4
Sample Survey
9.
a.
Stratum 1
p1 = .50
200 20 (.50)(.50)
s p1 =
= .1088
19
200
50 2(.1088)
or
.2824 to .7176
Stratum 2
p2 = .78
250 30 (.78)(.22)
s p2 =
= .0722
29
250
.78 2(.0722)
or
.6356 to .9244
Stratum 3
p3 = .21
100 25 (.21)(.79)
s p3 =
= .0720
24
100
.21 2(.0720)
or
.066 to .354
200
250
100
(.50) +
(.78) +
(.21) = .5745
550
550
550
b.
pst =
c.
1
(.5)(.5)
(.78)(.22)
(.21)(.79)
s pst =
200(180)
+ 250(220)
+ 100(75)
2
(550)
19
29
24
1
=
2
(550)
d.
21 - 5
Chapter 21
10. a.
n=
(20) 2
2
2
(1400) 2
+ 300(150) + 600(75) + 500(100)
2
(140, 000) 2
= 92.8359
196, 000, 000 + 15,125, 000
With B = 10, the first term in the denominator in the formula for n changes.
n=
(140, 000) 2
(140, 000) 2
=
= 305.6530
2
49, 000, 000 + 15,125,000
2 (10)
(1400)
+ 15,125, 000
4
Rounding up, we see that a sample size of 306 is needed to provide this level of precision.
300(150)
n1 = 306
= 98
140, 000
600(75)
n2 = 306
= 98
140, 000
500(100)
n3 = 306
= 109
140,000
Due to rounding, the total of the allocations to each strata only add to 305. Note that even though
the sample size is larger, the proportion allocated to each stratum has not changed.
n=
(140, 000) 2
(140, 000) 2
=
= 274.6060
(15, 000) 2
56, 250, 000 + 15,125, 000
+ 15,125, 000
4
Rounding up, we see that a sample size of 275 will provide the desired level of precision.
The allocations to the strata are in the same proportion as for parts a and b.
300(150)
n1 = 275
= 98
140, 000
600(75)
n2 = 275
= 88
140, 000
21 - 6
Sample Survey
500(100)
n3 = 275
= 98
140, 000
Again, due to rounding, the stratum allocations do not add to the total sample size. Another item
could be sampled from, say, stratum 3 if desired.
11. a.
b.
x1 = 29.5333
x2 = 64.775
x3 = 45.2125
x4 = 53.0300
Indianapolis
13.3603 38 6
29.533 2
38
6
29.533 10.9086(.9177)
or
19.5222 to 39.5438
Louisville
25.0666 45 8
64.775 2
45
8
64.775 17.7248(.9068)
or
48.7022 to 80.8478
St. Louis
19.4084 80 8
45.2125 2
80
8
70
10
53.0300 18.7719(.9258)
or
35.6510 to 70.4090
c.
38 1 45 5 80 3 70 5
pst =
+
+
+
= .4269
233 6 233 8 233 8 233 10
21 - 7
Chapter 21
d.
1 5
6 6
p (1 p1 )
N1 ( N1 n1 ) 1
= 38(32) = 33.7778
n1 1
5
5 3
8 8
p2 (1 p2 )
N 2 ( N 2 n2 )
= 45(37) = 55.7478
n2 1
7
3 5
8 8
p3 (1 p3 )
N 3 ( N 3 n3 )
= 80(72) = 192.8571
n3 1
7
5 5
10 10
p4 (1 p4 )
N 4 ( N 4 n4 )
= 70(60) = 116.6667
n4 1
9
1
s pst =
33.7778 + 55.7478 + 192.8571 + 116.6667 ] =
2 [
(233)
1
(399.0494) = .0857
(233) 2
b.
c.
38
45
80
70
xst =
29.5333 +
64.775 +
45.2125 +
53.0300 = 48.7821
233
233
233
233
N1 ( N1 n1 )
s12
(13.3603) 2
= 38(32)
= 36,175.517
n1
6
N 2 ( N 2 n2 )
s22
(25.0666) 2
= 45(37)
= 130, 772.1
n2
8
N 3 ( N 3 n3 )
s32
(19.4084) 2
= 80(72)
= 271, 213.91
n3
8
N 4 ( N 4 n4 )
s42
(29.6810) 2
= 70(60)
= 370, 003.94
n4
10
21 - 8
Sample Survey
1
sxst =
36,175.517 + 130, 772.1 + 271, 213.91 + 370, 003.94]
2 [
(233)
=
1
(808,165.47) = 3.8583
(233) 2
approximate 95% confidence interval
xst 2sxst
48.7821 2(3.8583)
or
41.0655 to 56.4987
n=
(11, 275) 2
= 27.3394
3, 404, 025 + 1, 245,875
(30)2
2
2
2
(123) 2
+ 50(80) + 38(150) + 35(45)
4
Rounding up we see that a sample size of 28 is necessary to obtain the desired precision.
13.
50(80)
n1 = 28
= 10
11, 275
38(150)
n2 = 28
= 14
11, 275
35(45)
n3 = 28
=4
11, 275
b.
n=
2
2
2
+ 50(100) + 38(100) + 35(100)
50(100)
n1 = 33
= 13
12,300
21 - 9
[ 123(100)]
= 33
Chapter 21
38(100)
n2 = 33
= 10
12,300
35(100)
n3 = 33
=9
12,300
This is the same as proportional allocation . Note that for each stratum
N
nh = n h
N
14. a.
xc =
xi
750
=
= 15
Mi
50
X = M x = 300(15) = 4500
c
pc =
b.
ai
15
=
= .30
M i 50
( xi xc M i ) 2
25 4 4450
sxc =
= 1.4708
2
(25)(4)(12) 3
s X = Ms xc = 300(1.4708) = 441.24
(ai pc M i )2
25 4 4.82
s pc =
= .0484
2
(25)(4)(12) 3
c.
d.
21 - 10
Sample Survey
e.
15. a.
xc =
10,400
= 80
130
X = M x = 600(80) = 48,000
c
pc =
b.
13
= .10
130
( xi xc M i ) 2
886,150
30 6
sxc =
= 7.6861
2
(30)(6)(20)
5
s X = 600(7.6861) = 4611.66
approximate 95% confidence
Interval for Population Total:
48,000 2(4611.66)
or
38,776.68 to 57,223.32
(ai pc M i ) 2
30 6
s pc =
2
(30)(6)(20)
6.68
= .0211
5
21 - 11
Chapter 21
xc =
2000
= 40
50
pc =
35
= .70
50
( xi xc M i ) 2
7292
120 10
sxc =
= 2.0683
2
(120)(10)(50
/12)
9
approximate 95% confidence
Interval for Mean age:
40 2(2.0683)
or
35.8634 to 44.1366
d.
(ai pc M i ) 2
= [ 8 - .7 (12) ]2 + + [ 12 - .7 (13) ]2
= (-.4)2 + (-.7)2 + (-.4)2 + (.3)2 + (-1.2)2 + (-.1)2 + (-1.4)2 + (.3)2
+ (.7)2 + (2.9)2
= 13.3
120 10
s pc =
2
(120)(10)(50 /12)
13.3
= .0883
9
xc =
21 - 12
Sample Survey
b.
c.
( xi xc M i ) 2
(150)(8)(40)
approximate 95% confidence
Interval for Mean Age of Agents:
36.9737 2(2.2394)
or
32.4949 to 41.4525
d.
(ai pc M i ) 2
150 8
s pc =
2
(150)(8)(40)
141.0989
= .0386
7
(ai pc M i ) 2
150 8
s pc =
2
(150)(8)(40)
68.8787
= .0270
7
21 - 13
Chapter 21
18. a.
p = 0.19
sp =
(0.19)(0.81)
= 0.0206
363
p = 0.31
sp =
(0.31)(0.69)
= 0.0243
363
p = 0.17
sp =
(0.17)(0.83)
= 0.0197
373
d.
(0.5)(0.5)
= 0.0262
363
21 - 14
Sample Survey
19. a.
Assume (N - n) / N 1
p = .55
sp =
b.
(0.55)(0.45)
= 0.0222
504
p = .31
sp =
(0.31)(0.69)
= 0.0206
504
c.
The estimate of the standard error in part (a) is larger because p is closer to .50.
d.
e.
20. a.
sx =
c.
p = .73
3000 200 (.73)(.27)
sp =
= .0304
3000 199
21 - 15
Chapter 21
Approximate 95% Confidence Interval for Proportion that are Generally Satisfied:
.73 2(.0304)
or
.6692 to .7908
d. If management administered the questionnaire and anonymity was not guaranteed we would expect
a definite upward bias in the percent reporting they were generally satisfied with their job. A
procedure for guaranteeing anonymity should reduce the bias.
21. a.
p = 1/3
= .0840
29
380
c.
p = 19 / 45 = .4222
= .0722
44
760
1
29
h
+760(715)
4545.1892 = .0482
21 - 16
Sample Survey
b.
[ ph (1 ph )]
nh 1
= (380) (380 - 30) (9 / 30) (21 / 30) / 29 + (760) (760 - 45) (12 / 45) (33 / 45) / 44 +
(260) (260 - 25)(11 / 25) (14 / 25) / 24
= 4005.5079
1
s pst =
4005.5079 = .0452
2
(1400)
Approximate 95% Confidence Interval:
.3079 2(.0452)
or
.2175 to .3983
c.
[ ph (1 ph )]
nh 1
= (380) (380 - 30) (21 / 30) (9 / 30) / 29 + (760) (760 - 45) (34 / 45) (11 / 45) / 44 +
(260) (260 - 25) (15 / 25) (10 / 25) / 24
= 3855.0417
1
s pst =
3855.0417 = .0443
2
(1400)
Approximate 95% Confidence Interval:
.7116 2(.0443)
or
.6230 to .8002
21 - 17
Chapter 21
d.
23. a.
n=
2
2
2
2
2
+ 3000(80) + 600(150) + 250(220) + 100(700) + 50(3000)
Rounding up, we need a sample size of 171 for the desired precision.
3000(80)
n1 = 171
= 68
605, 000
600(150)
n2 = 171
= 25
605, 000
250(220)
n3 = 171
= 16
605,000
100(700)
n4 = 171
= 20
605, 000
50(3000)
n5 = 171
= 42
605, 000
24. a.
xc =
pc =
12 + 2 + 30 + 8 + 10 + 22
84
=
= .35
14 + 7 + 96 + 23 + 71 + 29 240
(ai pc M i ) 2
100 6
s pc =
2
(100)(6)(48)
424.52
= .0760
5
21 - 18
Sample Survey
21 - 19