Professional Documents
Culture Documents
SamplingandSamplingDistributions
LearningObjectives
1.
Understandtheimportanceofsamplingandhowresultsfromsamplescanbeusedtoprovide
estimatesofpopulationcharacteristicssuchasthepopulationmean,thepopulationstandard
deviationand/orthepopulationproportion.
2.
Knowwhatsimplerandomsamplingisandhowsimplerandomsamplesareselected.
3.
Understandtheconceptofasamplingdistribution.
4.
Understandthecentrallimittheoremandtheimportantroleitplaysinsampling.
5.
Specificallyknowthecharacteristicsofthesamplingdistributionofthesamplemean( x )andthe
samplingdistributionofthesampleproportion( p ).
6.
Learnaboutavarietyofsamplingmethodsincludingstratifiedrandomsampling,clustersampling,
systematicsampling,conveniencesamplingandjudgmentsampling.
7.
Knowthedefinitionofthefollowingterms:
parameter
sampledpopulation
samplestatistic
simplerandomsampling
samplingwithoutreplacement
samplingwithreplacement
pointestimator
pointestimate
targetpopulation
samplingdistribution
finitepopulationcorrectionfactor
standarderror
centrallimittheorem
unbiased
relativeefficiency
consistency
71
Chapter7
Solutions:
1.
a.
AB,AC,AD,AE,BC,BD,BE,CD,CE,DE
b.
With10samples,eachhasa1/10probability.
c.
EandCbecause8and0donotapply.;5identifiesE;7doesnotapply;5isskippedsinceEis
alreadyinthesample;3identifiesC;2isnotneededsincethesampleofsize2iscomplete.
2.
Usingthelast3digitsofeach5digitgroupingprovidestherandomnumbers:
601,022,448,147,229,553,147,289,209
Numbersgreaterthan350donotapplyandthe147canonlybeusedonce.Thus,the
simplerandomsampleoffourincludes22,147,229,and289.
3.
4.
459,147,385,113,340,401,215,2,33,348
a.
5,0,5,8
BellSouth,LSILogic,GeneralElectric
b.
N!
10!
3,628,800
120
n !( N n)! 3!(10 3)! (6)(5040)
5.
283,610,39,254,568,353,602,421,638,164
6.
2782,493,825,1807,289
7.
108,290,201,292,322,9,244,249,226,125,(continuingatthetopofcolumn9)147,and113.
8.
Randomnumbersused:13,8,27,23,25,18
Thesecondoccurrenceoftherandomnumber13isignored.
Companiesselected:ExxonMobil,Chevron,Travelers,Microsoft,Pfizer,andIntel
9.
102,115,122,290,447,351,157,498,55,165,528,25
10. a.
Finitepopulation.AframecouldbeconstructedobtainingalistoflicenseddriversfromtheNew
YorkStatedriverslicensebureau.
b.
Infinitepopulation.Samplingfromaprocess.Theprocessistheproductionlineproducingboxesof
cereal.
c.
Infinitepopulation.Samplingfromaprocess.Theprocessisoneofgeneratingarrivalstothe
GoldenGateBridge.
72
SamplingandSamplingDistributions
d.
Finitepopulation.Aframecouldbeconstructedbyobtainingalistingofstudentsenrolledinthe
coursefromtheprofessor.
e.
Infinitepopulation.Samplingfromaprocess.Theprocessisoneofgeneratingordersforthemail
orderfirm.
11. a.
b.
x xi / n
54
9
6
( xi x ) 2
n 1
( xi x ) 2 =(4)2+(1)2+12(2)2+12+52=48
s=
12. a.
b.
13. a.
48
31
.
61
p =75/150=.50
p =55/150=.3667
x xi / n
465
93
5
b.
Totals
s
14. a.
94
100
85
94
92
465
+1
+7
-8
+1
-1
0
( xi x ) 2
1
49
64
1
1
116
Eighteenofthe40fundsinthesampleareloadfunds.Ourpointestimateis
18
.45
40
Sixofthe40fundsinthesamplearehighriskfunds.Ourpointestimateis
p
c.
( xi x )
( xi x ) 2
116
5.39
n 1
4
p
b.
xi
6
.15
40
Thebelowaveragefundratingsarelowandverylow.Twelveofthefundshavearatingoflowand
6havearatingofverylow.Ourpointestimateis
73
Chapter7
15. a.
b.
16. a.
18
.45
40
x xi / n
$45,500
$4,550
10
( xi x ) 2
n 1
9, 068, 620
$1003.80
10 1
Wewouldusethesampleproportionfortheestimate.
p
5
.10
50
(Authors'note:TheactualproportionfromNewYorkis p
b.
52
.104. )
500
ThesampleproportionfromMinnesotais
2
.04
50
OurestimateofthenumberofFortune500companiesfromNewYorkis(.04)500=20.
p
(Authors'note:TheactualnumberfromMinnesotais18.)
c.
Fourteenofthe50inthesamplecomefromthese4states.So36donot.
p
36
.72
50
(Authors'note:TheactualproportionfromMinnesotais p
17. a.
409/999=.41
b.
299/999=.30
c.
291/999=.29
18. a.
366
.732. )
500
E ( x ) 200
b.
x / n 50 / 100 5
c.
NormalwithE( x )=200and x =5
d.
Itshowstheprobabilitydistributionofallpossiblesamplemeansthatcanbeobservedwithrandom
samplesofsize100.Thisdistributioncanbeusedtocomputetheprobabilitythat x iswithina
specifiedfrom
19. a.
Thesamplingdistributionisnormalwith
E ( x ) ==200
74
SamplingandSamplingDistributions
x / n 50 / 100 5
At x = 195, z
P ( z 1) = .8413
x 5
1 P ( z 1) = .1587
x
5
At x = 190, z
x 10
2
x
5
x
10
5
P ( z 2) = .9772
2 P ( z 2) = .0228
x / n
20.
x 25 / 50 3.54
x 25 / 100 2.50
x 25 / 150 2.04
x 25 / 200 1.77
Thestandarderrorofthemeandecreasesasthesamplesizeincreases.
21. a.
b.
x / n 10 / 50 141
.
n/N=50/50,000=.001
Use x / n 10 / 50 141
.
c.
n/N=50/5000=.01
75
Chapter7
Use x / n 10 / 50 141
.
d.
n/N=50/500=.10
Use x
N n
N 1 n
500 50 10
134
.
500 1 50
Note:Onlycase(d)wheren/N=.10requirestheuseofthefinitepopulationcorrectionfactor.
22. a.
x / n 4000 / 60 516.40
x
51,800
E( x )
Thenormaldistributionfor x isbasedontheCentralLimitTheorem.
b.
c.
Asthesamplesizeisincreased,thestandarderrorofthemean, x ,isreduced.Thisappears
logicalfromthepointofviewthatlargersamplesshouldtendtoprovidesamplemeansthatare
closertothepopulationmean.Thus,thevariabilityinthesamplemean,measuredintermsof x ,
shoulddecreaseasthesamplesizeisincreased.
23. a.
Withasampleofsize60 x
At x = 52,300, z
4000
60
516.40
52,300 51,800
.97
516.40
51,300 51,800
.97
516.40
SamplingandSamplingDistributions
b.
4000
120
365.15
At x = 52,300, z
52,300 51,800
1.37
365.15
51,300 51,800
1.37
365.15
Normaldistribution, E ( x ) 17.5
x / n 4 / 50 .57
b.
Within1weekmeans16.5 x 18.5
At x =18.5, z
18.5 17.5
1.75 P(z 1.75) = .9599
.57
18.0 17.5
.88
.57
At x =17.0, z = -.88
25.
below.
a.
512 502
.95
10.54
492 502
.95 P(z < -.95) = .1711
10.54
probability=.8289.1711=.6578
77
Chapter7
b.
525 515
.95
10.54
505 515
.95
10.54
probability=.8289.1711=.6578
Theprobabilityofbeingwithin10ofthemeanontheMathematicsportionofthetestisexactlythe
sameastheprobabilityofbeingwithin10ontheCriticalReadingportionoftheSAT.Thisis
becausethestandarderroristhesameinbothcases.Thefactthatthemeansdifferdoesnotaffect
theprobabilitycalculation.
c. x / n 100 / 100 10.0 Thestandarderrorissmallerherebecausethesamplesizeislarger.
z
504 494
1.00
10.0
484 494
1.00
10.0
probability=.8413.1587=.6826
Theprobabilityislargerherethanitisinparts(a)and(b)becausethelargersamplesizehasmade
thestandarderrorsmaller.
26. a.
x 939
/ n
n=50
n=100
n=400
25
245 / 30
25
245 / 50
.56 P(.56z.56)=.7123.2877=.4246
.72 P(.72z.72)=.7642.2358=.5284
25
245 / 100
25
245 / 400
1.02 P(1.02z1.02)=.8461.1539=.6922
2.04 P(2.04z2.04)=.9793.0207=.9586
b. Alargersampleincreasestheprobabilitythatthesamplemeanwillbewithinaspecified
distanceofthepopulationmean.Intheautomobileinsuranceexample,theprobabilityof
78
SamplingandSamplingDistributions
beingwithin25ofrangesfrom.4246forasampleofsize30to.9586forasampleof
size400.
27. a.
At x =158,000,z = -1.58
P(z < -1.58) = .0571, thus
P(158,000 x 178,000) = .9429 - .0571 = .8858
b.
Inpart(b)wehaveahigherprobabilityofobtainingasamplemeanwithin$10,000ofthe
populationmeanbecausethestandarderrorissmaller.
d.
Thisisagraphofanormaldistributionwith E ( x ) =95and
x / n 14 / 30 2.56
b.
Within3strokesmeans92 x 98
z
98 95
1.17
2.56
92 95
1.17
2.56
P(92 x 98)=P(1.17z1.17)=.8790.1210=.7580
Theprobabilitythesamplemeanswillbewithin3strokesofthepopulationmeanof95is.7580.
c.
x / n 14 / 45 2.09
Within3strokesmeans103 x 109
79
Chapter7
109 106
1.44
2.09
103 106
1.44
2.09
P(103 x 109)=P(1.44z1.44)=.9251.0749=.8502
Theprobabilitythesamplemeanswillbewithin3strokesofthepopulationmeanof106is.8502.
d.
=2.34=.20
29.
a.
n=30
z
x
.03
.82
/ n .20 / 30
P(2.31 x 2.37)=P(.82z.82)=.7939.2061=.5878
b.
n=50
z
/ n
.03
.20 / 50
1.06
P(2.31 x 2.37)=P(1.06z1.06)=.8554.1446=.7108
c.
n=100
z
/ n
.03
.20 / 100
1.50
P(2.31 x 2.37)=P(1.50z1.50)=.9332.0668=.8664
d.
Noneofthesamplesizesinparts(a),(b),and(c)arelargeenough.Atz=1.96wefindP(1.96z
1.96)=.95.So,wemustfindthesamplesizecorrespondingtoz=1.96.Solve
.03
1.96
.20 / n
.20
13.0667
.03
n 1.96
n 170.73
Rounding up, we see that a sample size of 171 will be needed to ensure a probability of .95 that the
sample mean will be within $.03 of the population mean.
30. a.
b.
n/N=40/4000=.01<.05;therefore,thefinitepopulationcorrectionfactorisnotnecessary.
Withthefinitepopulationcorrectionfactor
710
SamplingandSamplingDistributions
N n
N 1 n
4000 40 8.2
129
.
4000 1 40
Withoutthefinitepopulationcorrectionfactor
x / n 130
.
Includingthefinitepopulationcorrectionfactorprovidesonlyaslightlydifferentvaluefor x than
whenthecorrectionfactorisnotused.
c.
x
2
154
.
130
.
130
.
P(z1.54)=.9382
P(z<1.54)=.0618
Probability = .9382 - .0618=.8764
31. a.
E( p )=p=.40
p (1 p )
.40(.60)
.0490
n
100
b.
c.
d.
Itshowstheprobabilitydistributionforthesampleproportion p .
32. a.
E( p )=.40
p (1 p )
.40(.60)
.0346
n
200
Within.03means.37 p .43
z
p p
.03
.87
p
.0346
p p
.05
711
Chapter7
33.
p(1 p)
n
(.55)(.45)
.0497
100
(.55)(.45)
.0352
200
(.55)(.45)
.0222
500
(.55)(.45)
.0157
1000
Thestandarderroroftheproportion, p , decreasesasnincreases
34. a.
(.30)(.70)
.0458
100
Within.04means.26 p .34
z
p p
.04
(.30)(.70)
.0324
200
p p
.04
(.30)(.70)
.0205
500
p p
.04
SamplingandSamplingDistributions
d.
(.30)(.70)
.0145
1000
p p
.04
Withalargersample,thereisahigherprobability p willbewithin.04ofthepopulation
proportionp.
35. a.
p(1 p )
.30(.70)
.0458
n
100
.30
Thenormaldistributionisappropriatebecausenp=100(.30)=30andn(1p)=100(.70)=70are
bothgreaterthan5.
b.
P(.20 p .40)=?
z
.40 .30
2.18 P(z 2.18) = .9854
.0458
P(.25 p .35)=?
z
.35 .30
1.09 P(z 1.09) = .8621
.0458
Thisisagraphofanormaldistributionwithameanof E ( p ) =.66and
713
Chapter7
p
b.
p (1 p )
.66(1 .66)
.0273
n
300
Within.04means.62 p .70
z
.70 .66
1.47
.0273
.62 .66
1.47
.0273
P(.62 p .70)=P(1.47z1.47)=.9292.0708=.8584
c.
p (1 p )
.87(1 .87)
.0194
n
300
Within.04means.83 p .91
z
.91 .87
2.06
.0194
.83 .87
2.06
.0194
P(.83 p .91)=P(2.06z2.06)=.9803.0197=.9606
d.
Yes,theprobabilityofbeingwithin .04ishigherforthesampleofyouthusers.Thisisbecause
thestandarderrorissmallerforthepopulationproportionasitgetscloserto1.
e.
Forn=600, p
.66(1 .66)
.0193
600
Within.04means.62 p .70
z
.70 .66
2.07
.0193
.62 .66
2.07
.0193
P(.62 p .70)=P(2.07z2.07)=.9808.0192=.9616
Theprobabilityislargerthaninpart(b).Thisisbecausethelargersamplesizehasreducedthe
standarderror.
37. a.
Normaldistribution
E ( p ) .12
b.
p(1 p)
(.12)(1 .12)
.0140
540
p p
.03
714
SamplingandSamplingDistributions
P(.09 p .15) = .9838 - .0162 =.9676
c.
p p .015
38. a.
Normaldistribution
E( p )=.56
b.
p(1 p)
(.56)(.44)
.0248
400
.02
.81 P(z .81) = .7910
.0248
.04
1.61 P(z 1.61) = .9463
.0248
b.
p (1 p )
.75(1 .75)
.0204
n
450
p p
.04
p(1 p )
.75(1 .75)
.0306
n
200
715
Chapter7
d.
p p
.75(1 .75)
200
.04
1.31
P(z 1.31) = .9049
.0306
40. a.
Theprobabilityofthesampleproportionbeingwithin.04ofthepopulationmeanwasreducedfrom
.9500to.8098.Sothereisagaininprecisionbyincreasingthesamplesizefrom200to450.Ifthe
extracostofusingthelargersamplesizeisnottoogreat,weshouldprobablydoso.
E( p )=.76
p (1 p )
.76(1 .76)
.0214
n
400
Normaldistributionbecausenp=400(.76)=304andn(1p)=400(.24)=96
b.
.79 .76
1.40 P(z 1.40) = .9192
.0214
p (1 p )
.76(1 .76)
.0156
n
750
.79 .76
1.92 P(z 1.92) = .9726
.0156
E( p )=.17
p (1 p )
(.17)(1 .17)
.0133
800
Distributionisapproximatelynormalbecausenp=800(.17)=136>5
andn(1p)=800(.83)=664>5
b.
.19 .17
1.51 P(z 1.51) = .9345
.0133
SamplingandSamplingDistributions
c.
p (1 p )
(.17)(1 .17)
.0094
1600
.19 .17
2.13 P(z 2.13) = .9834
.0094
Therandomnumberscorrespondingtothefirstsevenuniversitiesselectedare
122,99,25,55,115,102,61
Thethird,fourthandfifthcolumnsofTable7.1wereneededtofind7randomnumbersof133or
lesswithoutduplicatenumbers.
Authorsnote:Theuniversitiesidentifiedare:ClarksonU.(122),U.ofArizona(99),UCLA(25),
U.ofMaryland(55),U.ofNewHampshire(115),FloridaStateU.(102),ClemsonU.(61).
43. a.
Normaldistributionbecausen=50
E( x )=6883
b.
2000
282.84
n
50
/ n
300
2000 / 50
At7500, z
7500 6883
2000 / 50
2.18
Normaldistributionbecauseofcentrallimittheorem(n>30)
E ( x ) =115.50
35
5.53
n
40
717
Chapter7
b.
/ n
10
35 / 40
At x =100, z
100 115.50
35 / 40
2.80
P( x 100)=P(z -2.80)=.0026
Yes,thisisanunusuallylowspendinggroupof40alums.Theprobabilityofspendingthismuchor
lessisonly.0026.
45.
Withn=60thecentrallimittheoremallowsustoconcludethesamplingdistributionis
approximatelynormal.
a.
Thismeans14 x 16
At x =16, z
16 15
4 / 60
1.94
Thismeans14.25 x 15.75
At x =15.75, z
15.75 15
4 / 60
1.45
=27,175=7400
46.
a.
x 7400 / 60 955
b.
x
0
0
x
955
P( x >27,175)=P(z>0)=.50
Note:Thiscouldhavebeenansweredeasilywithoutanycalculations;27,175istheexpectedvalue
ofthesamplingdistributionof x .
c.
x 1000
SamplingandSamplingDistributions
x 1000
N n
N 1 n
N=2000
2000 50 144
2011
.
2000 1 50
x
N=5000
5000 50 144
20.26
5000 1 50
N=10,000
10,000 50 144
20.31
10,000 1 50
Note:Withn/N.05forallthreecases,commonstatisticalpracticewouldbetoignore
144
20.36 foreachcase.
thefinitepopulationcorrectionfactoranduse x
50
b.
N=2000
z
25
1.24
20.11
Chapter7
25
1.23 P(z 1.23) = .8907
20.26
25
1.23 P(z 1.23) = .8907
20.31
500
20
n
n
2
n =500/20=25andn=(25) =625
b.
For25,
z
25
1.25 P(z 1.25) = .8944
20
Samplingdistributionof x
0.05
n
30
0.05
1.9
x
2.1
1.9 + 2.1 = 2
=
2
Theareabelow x =2.1mustbe1.05=.95.Anareaof.95inthestandardnormaltableshows
z=1.645.
720
SamplingandSamplingDistributions
Thus,
z
2.1 2.0
/ 30
1.645
Solvefor
50.
(.1) 30
.33
1.645
p=.28
a.
ThisisthegraphofanormaldistributionwithE( p )=p=.28and
p(1 p)
.28(1 .28)
.0290
n
240
b.Within.04means.24 p .32
z
.32 .28
1.38
.0290
.24 .28
1.38
.0290
P(.24 p .32)=P(1.38z1.38)=.9162.0838=.8324
c. Within.02means.26 p .30
z
.30 .28
.69
.0290
.26 .28
.69
.0290
P(.26 p .30)=P(.69z.69)=.7549.2451=.5098
51.
p (1 p )
(.40)(.60)
.0245
400
P( p .375)=?
z
.375 .40
1.02
.0245
P( p .375)=1.1539=.8461
52. a.
p(1 p)
(.40)(1 .40)
.0251
380
Within.04means.36 p .44
721
Chapter7
.44 .40
1.59
.0251
.36 .40
1.59
.0251
P(.36 p .44)=P(1.59z1.59)=.9441.0559=.8882
b.
WewantP( p .45)
p p .45 .40
z
1.99
p
.0251
P( p
53. a.
NormaldistributionwithE( p )=.15and
p
b.
.45)=P(z 1.99)=1.9767=.0233
p (1 p )
(.15)(.85)
.0292
150
P(.12 p .18)=?
z
.18 .15
1.03 P(z 1.03) = .8485
.0292
p(1 p)
.25(.75)
.0625
n
Solveforn
n
b.
.25(.75)
48
(.0625) 2
c.
P( p .30)=?
z
.30 .25
.80 P(z .80) = .7881
.0625
P( p .30)=1.7881=.2119
722
SamplingandSamplingDistributions
723