Business Statistics: A First Course The Normal...

44
Business Statistics: A First Course, 5e © 2009 Prentice-Hall, Inc. Chap 6-1 Chapter 6 The Normal Distribution Business Statistics: A First Course 5 th Edition

Transcript of Business Statistics: A First Course The Normal...

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc.

Chap 6

-1

Ch

ap

ter

6

The N

orm

al D

istr

ibutio

n

Busin

ess S

tatistics:

A F

irst C

ou

rse

5th

Ed

itio

n

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-2

Lea

rnin

g O

bje

ctives

In t

his

ch

ap

ter,

yo

u learn

:

�T

o c

om

pu

te p

rob

ab

ilitie

s f

rom

th

e n

orm

al

dis

trib

utio

n

�T

o u

se

th

e n

orm

al p

rob

ab

ility

plo

t to

de

term

ine

wh

eth

er

a s

et

of da

ta is a

pp

roxim

ate

ly n

orm

ally

dis

trib

ute

d

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-3

Contin

uous P

rob

abili

ty D

istr

ibutions

�A

co

ntin

uo

us r

an

do

m v

ari

ab

leis

a v

ari

ab

le t

ha

t

ca

n a

ssum

e a

ny v

alu

e o

n a

co

ntin

uu

m (

ca

n

assu

me

an

un

co

un

tab

le n

um

be

r o

f va

lue

s)

�th

ickness o

f an ite

m

�tim

e r

equir

ed t

o c

om

ple

te a

task

�te

mpera

ture

of

a s

olu

tion

�he

ight, in inches

�T

he

se

ca

n p

ote

ntia

lly t

ake

on

an

y v

alu

e

de

pe

nd

ing

on

ly o

n t

he

ab

ility

to

pre

cis

ely

an

d

accu

rate

ly m

ea

su

re

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-4

The N

orm

al D

istr

ibution

�‘B

ell S

hap

ed

�S

ym

metr

ical

�M

ean

, M

ed

ian

an

d M

od

eare

Eq

ual

Lo

cati

on

is d

ete

rmin

ed

by t

he

mean

, µ

Sp

read

is d

ete

rmin

ed

by t

he

sta

nd

ard

devia

tio

n, σ

Th

e r

an

do

m v

ari

ab

le h

as a

n

infi

nit

e t

heo

reti

cal

ran

ge:

+∞ ∞∞∞

to

− −−−∞ ∞∞∞

Mean

=

Med

ian

=

Mo

de

X

f(X

)

µ

σ

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-5

The N

orm

al D

istr

ibution

Density F

unction

)(X

21

e2π1

f(X

)

σ

�T

he f

orm

ula

for

the n

orm

al pro

ba

bili

ty d

ensity function is

Wh

ere

e =

th

e m

ath

emat

ical

co

nst

ant

app

rox

imat

ed b

y 2

.718

28

π=

th

e m

ath

emat

ical

co

nst

ant

app

rox

imat

ed b

y 3

.141

59

µ=

th

e pop

ula

tion

mea

n

σ=

th

e pop

ula

tion

sta

nd

ard d

evia

tion

X =

an

y v

alu

e o

f th

e co

nti

nuou

s v

aria

ble

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-6

By v

ary

ing

th

e p

ara

mete

rs µ

an

d σ

, w

e o

bta

in

dif

fere

nt

no

rmal

dis

trib

uti

on

s

Many N

orm

al D

istr

ibution

s

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-7

Th

e N

orm

al D

istr

ibu

tio

n

Sh

ap

e

X

f(X

)

µ

σ

Ch

an

gin

shifts

the

dis

trib

utio

n left o

r right.

Ch

an

gin

g σ

incre

ase

s

or

decre

ases t

he

spre

ad.

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-8

Th

e S

tan

da

rdiz

ed

No

rma

l

�A

ny

no

rma

l d

istr

ibu

tio

n (

with

an

y m

ea

n a

nd

sta

nd

ard

de

via

tio

n c

om

bin

atio

n)

ca

n b

e

tra

nsfo

rme

d in

to t

he

sta

nd

ard

ize

d n

orm

al

dis

trib

utio

n (

Z)

�N

ee

d t

o t

ran

sfo

rm

X

un

its in

to

Z

units

�T

he

sta

nd

ard

ize

d n

orm

al d

istr

ibu

tio

n (

Z)

ha

s a

m

ea

n o

f 0

an

d a

sta

nd

ard

de

via

tio

n o

f 1

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-9

Tra

nsla

tio

n to

th

e S

tan

da

rdiz

ed

N

orm

al D

istr

ibu

tio

n

�T

ran

sla

te f

rom

X t

o t

he

sta

nd

ard

ize

d n

orm

al

(th

e “

Z”

dis

trib

utio

n)

by s

ub

tra

ctin

g t

he

me

an

of

X a

nd

div

idin

g b

y its

sta

nd

ard

de

via

tio

n:

σ

µX

Z−

=

Th

e Z

dis

trib

utio

n a

lwa

ys h

as m

ea

n =

0 a

nd

sta

nd

ard

de

via

tio

n =

1

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-10

The S

tand

ard

ized N

orm

al

Pro

ba

bili

ty D

ensity F

unction

�T

he

fo

rmu

la f

or

the

sta

nd

ard

ize

d n

orm

al

pro

ba

bili

ty d

en

sity fu

nctio

n is

Where

e =

the m

ath

em

atical consta

nt

appro

xim

ate

d b

y 2

.7182

8

π=

the m

ath

em

atical consta

nt

appro

xim

ate

d b

y 3

.1415

9

Z =

an

y v

alu

e o

f th

e s

tandard

ized n

orm

al dis

trib

ution

2(1

/2)Z

e2π1

f(Z

)−

=

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-11

Th

e S

tan

da

rdiz

ed

No

rma

l D

istr

ibu

tio

n

�A

lso k

now

n a

s the “

Z”

dis

trib

ution

�M

ean is 0

�S

tandard

Devia

tion is 1

Z

f(Z

)

0

1

Va

lues a

bo

ve t

he m

ean h

ave p

ositiv

eZ

-valu

es,

valu

es b

elo

w t

he m

ean h

ave n

eg

ative

Z-v

alu

es

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-12

Exam

ple

�If

X

is

dis

trib

ute

d n

orm

ally

with

me

an

of

10

0

an

d s

tan

da

rd d

evia

tio

n o

f 50

, th

e

Z va

lue

fo

r

X =

20

0is

�T

his

sa

ys t

ha

t X

= 2

00

is

tw

o s

tan

da

rd

de

via

tio

ns (

2 in

cre

me

nts

of 5

0 u

nits)

ab

ove

the

me

an

of

10

0.

2.0

501

00

20

0

σ

µX

Z=

−=

−=

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-13

Co

mp

arin

g

X

and

Z

units

Z

10

0

2.0

0

20

0X

No

te t

hat

the s

hap

e o

f th

e d

istr

ibu

tio

n is t

he s

am

e,

on

ly t

he s

cale

has c

han

ged

. W

e c

an

exp

ress t

he

pro

ble

m in

ori

gin

al u

nit

s (

X)

or

in s

tan

dard

ized

u

nit

s (

Z)

(µ=

100, σ

= 5

0)

(µ=

0, σ

= 1

)

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-14

Fin

din

g N

orm

al P

rob

ab

ilitie

s

ab

X

f(X

)P

aX

b(

)≤

Pro

bab

ility

is m

ea

su

red

by th

e a

rea

unde

r th

e c

urv

e

Pa

Xb

()

<<

=

(Note

that

the

pro

ba

bili

ty o

f a

ny

indiv

idu

al valu

e is z

ero

)

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-15

f(X

)

Pro

ba

bili

ty a

s

Are

a U

nd

er

the C

urv

e

0.5

0.5

The t

ota

l are

a u

nd

er

the c

urv

e is 1

.0, and t

he c

urv

e is

sym

metr

ic, so h

alf is a

bove t

he m

ean,

half is b

elo

w

1.0

)X

P(

=∞

<<

−∞

0.5

)X

P(µ

=∞

<<

0.5

µ)

XP

(=

<<

−∞

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-16

Th

e S

tan

da

rdiz

ed

Norm

al T

ab

le

�T

he

Cu

mu

lative

Sta

nd

ard

ize

d N

orm

al ta

ble

in t

he

te

xtb

oo

k (

Ap

pe

nd

ix t

ab

le E

.2)

giv

es t

he

pro

ba

bili

ty le

ss t

han

a d

esir

ed

va

lue

of

Z (

i.e

.,

fro

m n

eg

ative

in

finity to

Z)

Z0

2.0

00.9

772

Exam

ple

:

P(Z

< 2

.00)

= 0

.9772

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-17

Th

e S

tan

da

rdiz

ed

Norm

al T

ab

le

The v

alu

e w

ith

in t

he

table

giv

es t

he

pro

ba

bili

ty f

rom

Z =

− −−−∞ ∞∞∞

up t

o the d

esire

d Z

valu

e

.9772

2.0

P(Z

< 2

.00)

=0.9

772

Th

e r

ow

sho

ws

the v

alu

e o

f Z

to

the first

decim

al po

int

Th

e c

olu

mn

giv

es t

he v

alu

e o

f Z

to the s

econd d

ecim

al po

int

2.0. . .

(co

ntin

ue

d)

Z 0

.00 0

.01

0

.02

0.0

0.1

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-18

Gen

era

l P

rocedure

for

Fin

din

g N

orm

al P

roba

bili

ties

�D

raw

th

e n

orm

al cu

rve

fo

r th

e p

rob

lem

in

term

s o

f X

�T

ran

sla

te X

-va

lue

s t

o Z

-va

lue

s

�U

se

th

e S

tan

da

rdiz

ed

No

rma

l T

ab

le

To

fin

d

P(a

< X

< b

) w

he

n

X

is

dis

trib

ute

d n

orm

ally

:

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-19

Fin

din

g N

orm

al P

roba

bili

ties

�Let X

repre

sent th

e tim

e it ta

kes to

dow

nlo

ad a

n im

age file

fro

m the inte

rnet.

�S

uppose X

is n

orm

al w

ith m

ean 8

.0 a

nd

sta

ndard

devia

tion 5

.0. F

ind P

(X <

8.6

)

X

8.6

8.0

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-20

�Let

X r

epre

sen

t th

e t

ime it

takes t

o d

ow

nlo

ad

an im

age f

ile f

rom

the

inte

rnet.

�S

uppose X

is n

orm

al w

ith m

ea

n 8

.0 a

nd s

tan

dard

devia

tio

n 5

.0.

Fin

d

P(X

< 8

.6)

Z0.1

20

X8.6

8

µ=

8

σ=

10

µ=

0

σ=

1

(co

ntin

ue

d)

Fin

din

g N

orm

al P

roba

bili

ties

0.1

25

.0

8.0

8.6

σ

µX

Z=

−=

−=

P(X

< 8

.6)

P(Z

< 0

.12)

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-21

Z

0.1

2

Z.0

0.0

1

0.0

.50

00

.50

40

.50

80

.53

98

.54

38

0.2

.57

93

.58

32

.58

71

0.3

.61

79

.62

17

.62

55

So

lutio

n: F

ind

ing

P(Z

< 0

.12

)

.5478

.02

0.1

.5478

Sta

ndard

ized N

orm

al P

roba

bili

ty

Tab

le (

Port

ion)

0.0

0

= P

(Z <

0.1

2)

P(X

< 8

.6)

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-22

Fin

din

g N

orm

al

Up

pe

r T

ail

Pro

ba

bili

tie

s

�S

uppose X

is

norm

al w

ith m

ean 8

.0 a

nd

sta

ndard

devia

tion 5

.0.

�N

ow

Fin

d P

(X >

8.6

)

X

8.6

8.0

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-23

�N

ow

Fin

d P

(X >

8.6

)…

(co

ntin

ue

d)

Z

0.1

2

0Z

0.1

2

0.5

478

0

1.0

00

1.0

-0.5

478

= 0

.4522

P(X

> 8

.6)

= P

(Z >

0.1

2)

= 1

.0 -

P(Z

≤0.1

2)

= 1

.0 -

0.5

478 =

0.4

52

2

Fin

din

g N

orm

al

Up

pe

r T

ail

Pro

ba

bili

tie

s

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-24

Fin

din

g a

No

rma

l P

rob

abili

ty

Be

twe

en

Tw

o V

alu

es

�S

up

po

se

X

is

no

rma

l w

ith

me

an

8.0

an

d

sta

nd

ard

de

via

tio

n 5

.0.

Fin

d P

(8 <

X <

8.6

)

P(8

< X

< 8

.6)

= P

(0 <

Z <

0.1

2)

Z0.1

20

X8.6

8

05

88

σ

µX

Z=

−=

−=

0.1

25

88

.6

σ

µX

Z=

−=

−=Ca

lcu

late

Z-v

alu

es:

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-25

Z

0.1

2

So

lutio

n: F

ind

ing

P(0

< Z

< 0

.12

)

0.0

478

0.0

0

= P

(0 <

Z <

0.1

2)

P(8

< X

< 8

.6)

= P

(Z <

0.1

2)

–P

(Z ≤

0)

= 0

.5478 -

.500

0 =

0.0

478

0.5

000

Z.0

0.0

1

0.0

.50

00

.50

40

.50

80

.53

98

.54

38

0.2

.57

93

.58

32

.58

71

0.3

.61

79

.62

17

.62

55

.02

0.1

.5478

Sta

ndard

ized N

orm

al P

roba

bili

ty

Tab

le (

Port

ion)

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-26

�S

uppose X

is

norm

al w

ith m

ean 8

.0 a

nd

sta

ndard

devia

tion 5

.0.

�N

ow

Fin

d P

(7.4

< X

< 8

)

X

7.4

8.0

Pro

ba

bili

ties in t

he L

ow

er

Tail

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-27

Pro

ba

bili

ties in t

he L

ow

er

Tail

Now

Fin

d P

(7.4

< X

< 8

)…

X7.4

8.0

P(7

.4 <

X <

8)

= P

(-0.1

2 <

Z <

0)

= P

(Z <

0)

–P

(Z ≤

-0.1

2)

= 0

.5000 -

0.4

522 =

0.0

47

8

(co

ntin

ue

d)

0.0

478

0.4

522

Z-0

.12

0

The N

orm

al dis

trib

utio

n is

sym

metr

ic, so t

his

pro

bab

ility

is

the s

am

e a

s P

(0 <

Z <

0.1

2)

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-28

Em

pir

ical R

ule

s

µ±

en

clo

ses a

bo

ut

68.2

6%

of

X’s

f(X

)

µ+

µ-1σ

Wh

at

ca

n w

e s

ay a

bo

ut

the

dis

trib

utio

n o

f va

lue

s

aro

un

d t

he

me

an

?

Fo

r a

ny n

orm

al d

istr

ibu

tio

n:

σσ 6

8.2

6%

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-29

The E

mpiric

al R

ule

�µ

±2σ

co

ve

rs a

bo

ut

95

%o

f X

’s

�µ

±3σ

co

ve

rs a

bo

ut

99

.7%

of

X’s

95.4

4%

99.7

3%

(co

ntin

ue

d)

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-30

�S

teps to fin

d the X

valu

e for

a k

now

n

pro

babili

ty:

1.

Fin

d th

e Z

va

lue

fo

r th

e k

no

wn

pro

ba

bili

ty

2.

Co

nve

rt t

o X

un

its u

sin

g th

e f

orm

ula

:

Giv

en a

No

rmal P

roba

bili

tyF

ind t

he X

Valu

e

µX

+=

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-31

Fin

din

g t

he X

valu

e f

or

a

Know

n P

roba

bili

ty

Exa

mp

le:

�Let

X r

epre

sent

the tim

e it ta

kes (

in s

econ

ds)

to

do

wn

load a

n im

ag

e f

ile f

rom

the inte

rnet.

�S

up

pose X

is

norm

al w

ith m

ean 8

.0 a

nd s

tand

ard

devia

tion 5

.0

�F

ind X

such t

hat 20%

of

dow

nlo

ad t

imes a

re less t

han

X.

X?

8.0

0.2

000

Z?

0

(co

ntin

ue

d)

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-32

Fin

d t

he Z

valu

e f

or

20

% in

the L

ow

er

Ta

il

�2

0%

are

a in

th

e low

er

tail

is c

onsis

ten

t w

ith

a

Z v

alu

e o

f -0

.84

Z.0

3

-0.9

.17

62

.17

36

.20

33

-0.7

.23

27

.22

96

.04

-0.8

.2005

Sta

ndard

ized N

orm

al P

roba

bili

ty

Tab

le (

Port

ion)

.05

.17

11

.19

77

.22

66

… … … …

X?

8.0

0.2

000

Z-0

.84

0

1.

Fin

d th

e Z

va

lue

fo

r th

e k

no

wn

pro

ba

bili

ty

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-33

2.

Co

nve

rt t

o X

un

its u

sin

g th

e f

orm

ula

:

Fin

din

g t

he X

valu

e

80

.3

0.5)

84

.0

(0.

8

µX

=

−+

=

+=

So 2

0%

of th

e v

alu

es f

rom

a d

istr

ibutio

n

with m

ean 8

.0 a

nd s

tandard

devia

tion

5.0

are

less than 3

.80

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-34

Evalu

ating N

orm

alit

y

�N

ot a

ll con

tin

uo

us d

istr

ibu

tio

ns a

re n

orm

al

�It is im

po

rta

nt to

eva

lua

te h

ow

we

ll th

e d

ata

se

t is

a

pp

roxim

ate

d b

y a

no

rma

l d

istr

ibu

tio

n.

�N

orm

ally

dis

trib

ute

d d

ata

sh

ou

ld a

pp

roxim

ate

th

e

the

ore

tica

l n

orm

al d

istr

ibu

tio

n:

�T

he

norm

al d

istr

ibu

tion

is b

ell

sha

ped

(sym

me

tric

al)

whe

re th

e m

ean

is e

qua

l to

the

med

ian

.

�T

he

em

piric

al ru

le a

pplie

s to

the

norm

al d

istr

ibu

tio

n.

�T

he

in

terq

ua

rtile

ra

nge

of a

no

rmal d

istr

ibu

tion

is 1

.33

sta

ndard

devia

tion

s.

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-35

Evalu

ating N

orm

alit

y

Co

mp

ari

ng

da

ta c

ha

racte

ristics to

th

eore

tica

l

pro

pe

rtie

s

�C

onstr

uct chart

s o

r gra

phs

�F

or

sm

all-

or

mo

dera

te-s

ize

d d

ata

se

ts, co

nstr

uct a s

tem

-an

d-l

ea

f d

isp

lay o

r a

bo

xp

lotto

ch

eck for

sym

metr

y

�F

or

larg

e d

ata

sets

, d

oes th

e h

isto

gra

m o

r p

oly

go

n a

pp

ear

be

ll-sh

ap

ed

?

�C

om

pute

descriptive s

um

mary

measure

s

�D

o th

e m

ea

n, m

ed

ian

an

d m

od

e h

ave

sim

ilar

va

lue

s?

�Is

the

in

terq

ua

rtile

ra

nge

ap

pro

xim

ate

ly 1

.33

σ?

�Is

the

ra

ng

e a

pp

roxim

ate

ly 6

σ?

(co

ntin

ue

d)

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-36

Evalu

ating N

orm

alit

y

Co

mp

ari

ng

da

ta c

ha

racte

ristics to

th

eore

tica

l

pro

pe

rtie

s

�O

bserv

e the d

istr

ibutio

nof th

e d

ata

set

�D

o a

pp

roxim

ate

ly 2

/3 o

f th

e o

bserv

atio

ns lie

with

in m

ea

n ±

1

sta

nd

ard

de

via

tio

n?

�D

o a

pp

roxim

ate

ly 8

0%

of th

e o

bserv

ation

s lie

with

in m

ea

n

±1

.28 s

tan

da

rd d

evia

tio

ns?

�D

o a

pp

roxim

ate

ly 9

5%

of th

e o

bserv

ation

s lie

with

in m

ea

n ±

2

sta

nd

ard

de

via

tio

ns?

�E

valu

ate

norm

al pro

bab

ility

plo

t

�Is

the

no

rma

l p

rob

ab

ility

plo

t a

pp

roxim

ate

ly lin

ea

r (i

.e. a

str

aig

ht

line

) w

ith

positiv

e s

lop

e?

(co

ntin

ue

d)

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-37

Co

nstr

uctin

g

A N

orm

al P

roba

bili

ty P

lot

�N

orm

al p

rob

ab

ility

plo

t

�A

rrange d

ata

into

ord

ere

d a

rray

�F

ind c

orr

espon

din

g s

tand

ard

ize

d n

orm

al qu

antile

valu

es (

Z)

�P

lot th

e p

air

s o

f poin

ts w

ith o

bserv

ed d

ata

valu

es (

X)

on t

he v

ert

ical a

xis

an

d t

he s

tand

ard

ized n

orm

al

qua

ntile

va

lues (

Z)

on the h

orizonta

l a

xis

�E

va

luate

the p

lot

for

evid

ence o

f lin

eari

ty

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-38

A n

orm

al p

rob

ab

ility

plo

t fo

r d

ata

fro

m a

no

rma

l d

istr

ibu

tio

n w

ill b

e

ap

pro

xim

ate

ly lin

ea

r:

30

60

90

-2-1

01

2Z

XTh

e N

orm

al P

rob

abili

ty P

lot

Inte

rpre

tatio

n

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-39

No

rma

l P

rob

ab

ility

Plo

t

Inte

rpre

tatio

n

Le

ft-S

ke

we

dR

igh

t-S

ke

we

d

Re

cta

ng

ula

r

30

60

90

-2-1

01

2Z

X

(co

ntin

ue

d)

30

60

90

-2-1

01

2Z

X

30

60

90

-2-1

01

2Z

XN

onlin

ear

plo

ts

indic

ate

a d

evia

tion

from

norm

alit

y

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-40

Evalu

ating N

orm

alit

yA

n E

xam

ple

: M

utu

al F

unds R

etu

rns

40

30

20

10

0-10

Return 2006

Boxplot of 2006 Returns

Th

e b

oxp

lotap

pe

ars

re

ason

ab

ly s

ym

me

tric

, w

ith

fo

ur

low

er

ou

tlie

rs a

t

-9.0

, -8

.0, -8

.0, -6

.5 a

nd

on

e u

pp

er

outlie

r a

t 35

.0.

(Th

e n

orm

al d

istr

ibu

tio

n is

sym

me

tric

.)

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-41

Evalu

ating N

orm

alit

yA

n E

xam

ple

: M

utu

al F

unds R

etu

rns

Descri

ptive

Sta

tistics

(co

ntin

ue

d)

•T

he m

ean (

12.5

142)

is s

lightly less t

ha

n t

he

media

n (

13.1

).

(In a

norm

al dis

trib

ution t

he

mean a

nd m

ed

ian a

re e

qua

l.)

•T

he inte

rquart

ilera

nge o

f 9.2

is a

ppro

xim

ate

ly

1.4

6 s

tandard

devia

tions.

(In

a n

orm

al

dis

trib

ution t

he inte

rqu

art

ilera

nge is 1

.33

sta

ndard

devia

tions.)

•T

he r

ange o

f 44 is e

qua

l to

6.9

9 s

tandard

devia

tions.

(In

a n

orm

al d

istr

ibution t

he r

ang

e is

6 s

tandard

devia

tions.)

•72.2

% o

f th

e o

bserv

atio

ns a

re w

ithin

1 s

tan

dard

devia

tion o

f th

e m

ean.

(In

a n

orm

al dis

trib

ution

this

perc

enta

ge is 6

8.2

6%

.

•87%

of

the o

bserv

ations a

re w

ithin

1.2

8

sta

ndard

devia

tions o

f th

e m

ean.

(In

a n

orm

al

dis

trib

ution p

erc

enta

ge is 8

0%

.)

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-42

Evalu

ating N

orm

alit

yA

n E

xam

ple

: M

utu

al F

unds R

etu

rns

40

30

20

10

0-10

99.99

99

95

80

50

20 5 1

0.01

Return 2006

Percent

Probability Plot of Return 2006

Normal

(co

ntin

ue

d)

Plo

t is

appro

xim

ate

ly

a s

traig

ht

line e

xcept

for

a few

outlie

rs a

t th

e low

en

d a

nd t

he

hig

h e

nd.

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-43

Evalu

ating N

orm

alit

yA

n E

xam

ple

: M

utu

al F

unds R

etu

rns

�C

on

clu

sio

ns

�T

he r

etu

rns a

re s

lightly left

-skew

ed

�T

he r

etu

rns h

ave m

ore

valu

es c

on

centr

ate

d a

round

the m

ean than e

xp

ecte

d

�T

he r

ange is larg

er

than e

xpecte

d (

caused b

y o

ne

outlie

r at

35.0

)

�N

orm

al pro

bab

ility

plo

t is

reasona

bly

str

aig

ht

line

�O

vera

ll, t

his

data

set does n

ot

gre

atly d

iffe

r fr

om

the

theore

tica

l pro

pert

ies o

f th

e n

orm

al d

istr

ibution(c

on

tin

ue

d)

Busin

ess S

tatistics: A

First C

ours

e, 5e ©

2009 P

rentice-H

all,

Inc..

Chap 6

-44

Cha

pte

r S

um

mary

�P

rese

nte

d n

orm

al d

istr

ibu

tio

n

�F

ou

nd

pro

ba

bili

tie

s f

or

the

no

rma

l d

istr

ibu

tio

n

�A

pp

lied

no

rma

l d

istr

ibu

tio

n t

o p

rob

lem

s