You are on page 1of 35

O X F O R D I B D I P L O M A P R O G R A M M E

ANSWERS

:
M AT H E M AT I C S HI GH E R LE V E L

S TAT I S T I C S
C O U R S E C O M PA N I O N

Josip Harcet

Lorraine Heinrichs

Palmira Mariz Seiler

Marlene Torres Skoumal


WORKED SOLUTIONS

Exploring further
1 probability distributions
Skills check 4
V Var( X ) ¦x i
2
pi  P 2
1 a Mode (X) = −1, 1 because those two values i 1

share the highest probability of 0.3.


12 u 0.4  22 u 0.3  32 u 0.2  4 2 u 0.1  22
Median, m = 1 since P(X ≤ 0) = 0.4 and
54 1
P(X ≤ 1) = 0.7
6
P E( X ) ¦x p
i 1
i i 1 u 0.3  0 u 0.1  1 u 0.3 1.1 1.2 1.3
B
Skills check
C D
=OneVar(a
 2 u 0.1  3 u 0.05  4 u 0.15 0.95 8 MinX 1.
6
9 Q1X 1.
V Var( X ) ¦x
i 1
i
2
pi  P 2 10
11
MedianX...
Q3X
2.
3.
12 MaxX 4.
6 D10 =2.
¦x
i 1
2
i i P ( 1)2 u 0.3  0 2 u 0.1  12 u 0.3
1.1 1.2 1.3 Skills check
 22 u 0.1  32 u 0.05  4 2 u 0.15 A B C D
=OneVar(a
3.85 2 2 0.3 x 2.
3 3 0.2 Σx 2.
V 3.85  0.9025 1.72 4 4 0.1 Σx2 5.
5 sx : = sn–... #UNDER...
1.1 1.2 Skills check 6 σx : = σn... 1.
B C D D6 =1.
=OneVar(a
8 MinX –1.
9 Q1X –1. 2 a 1.3 1.4 1.5 Skills check

10 MedianX... 1.
4

{
11 Q2X 2. 3 3 2
12 MaxX 4. 3 – ·x , 0 <x<2
f1(x)= 4 16
D10 =1. 2 0, 0 >x>2
1
1.1 1.2 Skills check
A B C D
–3 –2 –1 0 1 2 3 4 5
=OneVar(a –1
2 0 0.1 x 0.95
3 1 0.3 Σx 0.95
4 2 0.1 Σx2 3.85 Mode (X) = 0, the part of parabola is opened
5 3 0.05 sx : = sn–... #UNDEF...
§ 3·
6 4 0.15 σx : = σnx.. 1.71683 downwards and it has its vertex at ¨ 0, ¸.
D6 =1.7168284713389 © 4¹
m
§3 3 · 1
Median, ³ ¨  x 2 ¸ dx Ÿ
xi 1 2 3 4 0©
4 16 ¹ 2
b 3 1 3 1
pi 0.4 0.3 0.2 0.1 m m Ÿ m 0.695
4 16 2
2
Mode (X) = 1 because it has the highest 3 3
P E( X ) ³ §¨ x  x 3 ·¸ dx
probability (0.4) 0©
4 16 ¹
2
Median, m = 2 since P(X ≤ 1) = 0.4 and ª3 2 3 4º 3 3 3
«¬ 8 x  64 x »¼ 
2 4 4
P(X ≤ 2) = 0.7 0
4 2
2
P E( X ) ¦x p
i 1
i i 1 u 0 .4  2 u 0 .3 V Var( X )
§3 2 3 4· §3·
³0 ¨© 4 x  16 x ¸¹ dx  ¨© 4 ¸¹
 3 u 0 .2  4 u 0 . 1 2

© Oxford University Press 2015: this may be reproduced for class use solely for the purchaser’s institute Statistics: Chapter 1 1
WORKED SOLUTIONS

2 1.5 1.6 1.7 Skills check


ª x3 3x 5 º 9 6 9
«  »  2  4
¬4 80 ¼ 0 16 5 16 π 0.11685

19
80
0.487 ∫ 4
–π (x ·cos(2·x))dx
4
2

√ 0.11685027506808 0.341834
1.3 1.4 1.5 Skills check
|
nSolve ( 34 ·m– 161 ·m = 12 ,m )
3 0.694593
6/99
2 0.75
∫( 0
x1·f1(x))dx
c
y
1.6 1.7 1.8 Skils check

2 0.2375 2 ⎧ 6,3<x<6
∫( 0
x2·f1(x))dx–(0.75)2 1.5 f3(x)=⎨ x
2

⎩ 3>x>6
0,
1
3/99
0.5
1.3 1.4 1.5 Skills check
0 x
2 1 2 3 4 5 6 7 8
0.75
∫( 0
x·f1(x))dx

2 0.2375
∫( 0
x2·f1(x))dx–(0.75)2 Mode (X) = 3 because it is a decreasing
function on the interval [3, 6].
√0.2375 0.48734 m
1
Median, ³ f ( x ) dx Ÿm 4
| 0
2
6
4/99
P E( X ) ³ xf ( x ) dx 4.16
3
b 1.4 1.5 1.6 Skills check
6

1
f1(x)=
0, {
cos(2·x), –π < x <
4
–π
4
>x>
π
4
π
4
V

1.7
Var( X )

1.8 1.9
³x
3

Skils check
2
f (x )  P 2 0.839

(0,1)

–1
–1
0 1 5
nSolve
(∫ 6

m
6
x2
1
dx= ,m
2 ) 4.

6 |4.15888

Mode (X ) = 0, it has the maximum point at (0, 1).


m
1
∫( 3
x
6
x• 2 dx
)
Median, ³ f ( x )dx Ÿm 0 3/4
S 2

4

∫( )
S 1.7 1.8 1.9 Skils check
4 x2
P 3
E( X ) ³ xf ( x )dx
S
0
6 18.


V
4
S
4
2
f (x )  P 2
∫( )3
6
2•
x 2 dx
x
Var( X ) ³x S
0.342 √ 18–(4.1588830833593)2 0.838863

4
3/4
1.5 1.6 1.7 Skills check
0. 3 a 1 − 0.5 + 0.25 − 0.125 + ... ⇒ u1 = 1, r = −0.5
nSolve

π
(∫ m

–π
4
1
cos(2·x) dx = ,m
2
) Sum
1
1  ( 0.5)
1
1 .5
2
3
0.


4

(x·cos(2·x))dx 2 1 1
–π b 2 1  ! Ÿ u1 2, r
4 2 2 2
4/5 2 2 21
Sum u 22 2
1 2 1 21
1
2

© Oxford University Press 2015: this may be reproduced for class use solely for the purchaser’s institute Statistics: Chapter 1 2
WORKED SOLUTIONS

1 1
4 a f (x ) = ,x ≠2 3 a P( X 0) F (0 ) ,
2−x 6
−1 × − 1 1 1 1 2
Ÿ f ′( x ) = = ,x ≠2 P( X 1) F (1)  F (0)  ,
(2 − x ) 2
(2 − x )2 2 6 6
dx 1 1
Ÿ ³ f ( x ) dx  ln(2  x )  c , x  2 P( X 2) F (2)  F (1) 1
³ 2x 2 2
­ x 1
b f ( x ) = e3 x +1 Ÿ f ′( x ) = 3e3 x +1 3 ° , x 0, 1, 2
Ÿ P( X x) ® 6
1 3 x +1 6 °¯
3 x +1 0, otherwise
Ÿ ³ f ( x )dx = ³ e dx =
3
e +c
b The modal value of the variable X is 2 since it
3 π − 2x 1
c f ( x ) = sin has the highest probability, P( X 2) .
2 3 2
3 π − 2x § 2 · π − 2x 1
Ÿ f ′( x ) = cos × ¨ − ¸ = − cos 4 a P( X 1) F (1)
2 3 © 3¹ 3 25
3 S  2x 3
Ÿ ³ f ( x ) dx ³ sin dx P( X 3) F (3)  F (1)
2 3 25
3 1 S  2x 9 S  2x 5
u cos c cos c P( X 5) F (5)  F (3)
2 2 3 4 3 25

3 7
P( X 7) F (7)  F (5)
25
d f (x ) ( x 2  2)2
9
Ÿ f c( x ) 2( x 2  2 ) u 2 x 4 x ( x 2  2) P( X 9) F (9)  F (7)
25
Ÿ ³ f ( x )dx 2
 2)2 dx ­x
³ (x P( X x)
° , x 1, 3, 5, 7, 9
® 25
4 2 x5 4x3 °¯ 0, otherwise
³ (x  4 x  4 )dx   4x  c
5 3
b Median, m = 7 since
Exercise 1A 9 1 16 1
F (5)  and F (7) !
k 1 k 2k 25 2 25 2
1 a   1 Ÿ 3  3k 12 Ÿ k 3
12 12 12 S
S
3
ª 1 º3
b 5 a ³ 2 sin bx dx 1 Ÿ 2 «  cos bx » 1
X=x 0 1 2 0 ¬ b ¼0
3 7 12 § bS · b
F (x) Ÿ cos ¨ ¸  cos 0 
12 12 12 © 3 ¹ 2
­ 0, x  0 § bS · b
° 2 Ÿ cos ¨ ¸ 1 Ÿb 1
° x  7x  6 © 3 ¹ 2
⇒ F (x ) ® , x 0, 1, 2
° 24 1.1 Exercise A

°¯ 1, x ! 2 b•π 0.
( ( ) =1– 2b ,b)
nSolve cos
3
a2 a4 a6 a8 1.
2 a    1
20 20 20 20 ( ( b•3π ) =1– 2b ,b,0.5)
nSolve cos

Ÿ 4a  20 20 Ÿ a 10 1.
π
b X=x 1
8
2
14
3
18
4
20
nSolve
(∫ 3

0
(2•sin(b•x))dx=1,b
)
F (x) 1/3
20 20 20 20
x x
0, x  1
­
° 2
b F (x ) ³ f (t )dt Ÿ F ( x ) ³ 2 sin t dt
°9x  x f 0
⇒ F (x ) ® , x 1, 2, 3, 4 [2 cos t ]0x  2 cos x  2
° 20
°¯ 1, x ! 4 ­
° 0, x  0
14 7 ° S
c P( X d 2) F (2) Ÿ F (x ) ®2  2 cos x , 0 d x d
20 10 ° 3
° S
¯ 1, x !
3
© Oxford University Press 2015: this may be reproduced for class use solely for the purchaser’s institute Statistics: Chapter 1 3
WORKED SOLUTIONS

§ S· § S· §S · 2 a X ~ Geo(0.25) Ÿ P( X d 4 ) 0.684
c P¨ X t ¸ 1 P¨ X d ¸ 1 F¨ ¸
© 6¹ © 6¹ ©6¹ b X ~ Geo(0.7 ) Ÿ P( X ! 6)
§S ·
1  2  2 cos ¨ ¸ 3 1 0.732 = 1 – P(X ≤ 6) = 0.000729
©6¹
6 a Since the function is not defined for the values c X ~ Geo(0.3) Ÿ P(5 d X d 7 )
of −2 and 2 we need to use limits. = P(X ≤ 7) – P(X ≤ 4) = 0.158
b
§b dx · ª1 xº d X ~ Geo(0.991) Ÿ P(1  X d 7 )
lim ¨¨ ³ ¸ lim « arcsin »
b o2 2 ¸ b o2 S 2¼
© b S 4  x ¹ ¬ b
= P(X ≤ 7) – P(X ≤ 1) = 0.009
2
1ª xº 1
arcsin » (arcsin 1  arcsin(1))
S «¬ 2 ¼ 2 S 1.1 1.2 CH1 Exercise B

geomCdf(0.25,1,4) 0.683594
1 §S S·
¨  ¸ 1 1–geomCdf(0.7,1,6) 0.000729
S©2 2¹ +∞ geomCdf(0.3,1,7)–geomCdf(0.3,1,4)
The function is well defined if ³ f ( x )dx = 1 0.157746
−∞ geomCdf(0.991,1,7)–geomCdf(0.991,1,1)
b The modal value is the value of the random 0.009
variable X where probability reaches its
maximum. 4/99
3
1 § 1· 
f c( x ) ¨  ¸ ( 4  x 2
) 2
(2 x ) 3 X ~ Geo(0.73) Ÿ P( X d 3) 0.980
S © 2¹
x 4 X ~ Geo(0.92)
3
, f c( x ) 0Ÿx 0
2 2
S (4  x ) a P( X 5) 0.0000377
1.1 1.2 Exercise A
y
b P( X t 4 ) 1  P ( X d 3) 0.000512
9 5 X ~ Geo(0.15)
6
1.1 1.2 1.3 CH1 Exercise B
3
geomCdf(0.73,1,3) 0.980317
x
–12 –9 –6 –3 0 3 6 9 12 geomPdf(0.92,5) 0.000038
–3
1
–6 f1(x)= 1–geomCdf(0.92,1,3) 0.000512
π·√4 – x2
–9 geomPdf(0.15,4) 0.092119
1–geomCdf(0.15,1,6) 0.37715
By looking at the graph of the probability
function we notice that the graph has only one
extreme point and that is the minimum point 5/99
at 0, therefore the modal value doesn’t exist.
a P( X 4) 0.0921
Exercise 1B
b P( X t 7 ) 1  P( X d 6 ) 0.377
1 a X ~ Geo(0.6) Ÿ P( X 2) 0.4 u 0.6 0.24
b X ~ Geo(0.14 ) Ÿ P( X 3) 6 X ~ Geo( p ) Ÿ P( X ! k ) 1  P( X d k )
f k
= 0.862 × 0.14 = 0.104 1 ¦q n 1
p 1 p ¦q n 1

c X ~ Geo(0.5) Ÿ P( X 4) n k 1 n 1

inf. geom. seq.


= 0.53 × 0.5 = 0.0625
d X ~ Geo(0.88) Ÿ P( X 5) 1  qk 1  qk
1 p 1 p 1  (1  q k ) qk
= 0.124 × 0.88 = 0.000182 1 q p
1.1 CH1 Exercise B

geomPdf(0.6,2) 0.24
OR
geomPdf(0.14,3) 0.103544
geomPdf(0.5,4) 0.0625 X ~ Geo( p ) Ÿ P( X ! k )
geomPdf(0.88,5) 0.000182
f k
1 1
¦ q n 1 p qk p u ¦q n 1
qk p
1 q
qk p
p
qk
n k 1 n 1

inf. geom. seq.

4/99

© Oxford University Press 2015: this may be reproduced for class use solely for the purchaser’s institute Statistics: Chapter 1 4
WORKED SOLUTIONS

Investigation 1 c X ~ Geo(0.5) Ÿ E( X )
1
2,
0 .5
X ~ Geo (p)
0 .5
a p 0.4 Ÿ P( X ! 5| X ! 3) Var( X ) 2
0 .5 2
P(( X ! 5) ˆ ( X ! 3)) 1  P ( X d 5) 1
0.36 d X ~ Geo(0.88) Ÿ E( X ) 1.14,
P ( X ! 3) 1  P ( X d 3) 0.88
P( X ! 2 ) 1  P( X d 2 ) 0.36 0.12
Var( X ) 0.155
0.882
b p 0.7 Ÿ P( X ! 6| X ! 2)
P(( X ! 6) ˆ ( X ! 2)) 1  P( X d 6 ) For 1B, Question 2:
0.0081
P( X ! 2 ) 1  P( X d 2 ) 1
a X ~ Geo(0.25) Ÿ E( X ) 4,
P( X ! 4 ) 1  P( X d 4 ) 0.0081 0.25
c p 0.12 Ÿ P( X ! 12| X ! 5) 0.75
Var( X ) 12
P(( X ! 12) ˆ ( X ! 5)) 1  P( X d 12) 0.252
0.409
P ( X ! 5) 1  P ( X d 5) 1
b X ~ Geo(0.7) Ÿ E( X ) 1.43,
P( X ! 7 ) 1  P( X d 7 ) 0.409 0 .7
1.2 1.3 1.4 CH1 Exercise B 0 .3
Var( X ) 0.612
1–geomCdf(0.4,5) 0.36 0 .7 2
1–geomCdf(0.4,3)
1
1–geomCdf(0.4,2) 0.36 c X ~ Geo(0.3) Ÿ E( X ) 3.33,
0 .3
1–geomCdf(0.7,6) 0.0081
1–geomCdf(0.7,2) 0 .7
Var( X ) 7.78
1–geomCdf(0.7,4) 0.0081 0 .3 2
|
1
d X ~ Geo(0.991) Ÿ E( X ) 1.01,
4/99 0.991
0.009
1.2 1.3 1.4
1–geomCdf(0.4,2)
CH1 Exercise B
0.36
Var( X ) 0.00916
0.9912
1–geomCdf(0.7,6) 0.0081
1–geomCdf(0.7,2) 2 X ~ Geo(0.73)
1–geomCdf(0.7,4) 0.0081 1
1–geomCdf(0.12,12) 0.408676 a E( X ) 1.37
1–geomCdf(0.12,5)
0.73
1–geomCdf(0.12,7) 0.408676 0.27
b V Var( X ) 0.712
| 0.732

6/99 1 0.27
x max E( X )  3V 3u 3.51
0.73 0.732
X ~ Geo( p ); m, n  ]  ; m  n Mario must make four shots to destroy
P(( X ! n ) ˆ ( X ! m )) the balloon.
Ÿ P( X ! n | X ! m )
P( X ! m ) 1.1 CH1 Exercise C
n 1
P( X ! n ) q 1.36986
q n m P( X ! n  m ) 0.73
P( X ! m ) qm
0.711802

Exercise 1C √ 0.27
(0.73)2
3.50527
1 For 1B, Question 1:
1
1
0.73
+3∙
0.27

(0.73)2
a X ~ Geo(0.6) Ÿ E( X ) 1.67,
0 .6 16/99
0 .4
Var( X ) 1.11 1
0 .6 2 3 X ~ Geo(0.54 ) E( X ) 1.85
1 0.54
b X ~ Geo(0.14 ) Ÿ E( X ) 7.14,
0.14 0.46
V Var( X ) 1.26
0.86 0.54 2
Var( X ) 43.9
0.14 2 1 0.46
x max E( X )  3V 3u 5.62
0.54 0.54 2

© Oxford University Press 2015: this may be reproduced for class use solely for the purchaser’s institute Statistics: Chapter 1 5
WORKED SOLUTIONS

Therefore, we must select 6 students at random b X a NB(0.5, 2) Ÿ P( X ! 6)


to ensure that one of them will be familiar with 6
7
the election procedure. 1 ¦ P( X
k 2
k)
64
0.109
1.1 CH1 Exercise C

1 1.1 1.2 CH1 Exercise D


1.85185
0.54
k=3
1.25599 0.050781

√ 0.46
(0.54)2 6

5.61981 1– (nCr(k–1, 2–1)∙(0.5)k–2∙(0.5)2)


1
0.54
+3∙
0.46
(0.54)2√ k=2
0.109375
|
19/99
2/99
Exercise 1D
c X a NB(0.82, 4 ) Ÿ P(5 d X d 7 )
1 a X a NB(1, 0.2) Ÿ P( X 2)
7
§ 2  1· = = k ) = 0.525
¨ ¸ 0 .8 u 0 . 2 0.16 ¦ P( X
©1  1¹ k =5

b X a NB(3, 0.5) Ÿ P( X 4) 1.1 1.2 CH1 Exercise D

k=2
§ 4  1· 3 0.109375
¨ ¸ 0 .5 u 0 . 5 0.188
© 3  1 ¹ 7

c X a NB(7, 0.8) Ÿ P( X 9) (nCr(k–1, 4–1)∙(0.18)k–4∙(0.82)4)


k=5
§ 9  1· 2 7 0.524751
¨ ¸ 0 .2 u 0 . 8 0.235
© 7  1 ¹ |

d X a NB(23, 0.77 ) Ÿ P( X 32) 3/99

§ 32  1 · 9 32 d X a NB(0.43, 5) Ÿ P(8  X d 11)


¨ ¸ 0.23 u 0.77 0.00847
© 23  1¹ 11
1.1 CH1 Exercise D = ¦ P( X = k ) = 0.326
nCr(1,0)∙0.8∙0.2 0.16 k =9

nCr(3,2)∙0.5∙(0.5)3 0.1875 1.1 1.2 CH1 Exercise D

nCr(8,6)∙(0.2)2∙(0.8)7 0.234881 k=5


0.524751
nCr(31,22)∙(0.23)9∙(0.77)32 0.008467
11
(nCr(k–1, 5–1)∙(0.57)k–5∙(0.43)5)
k=9
4/99 0.325959
|
2 a X a NB(0.25, 3) Ÿ P( X d 4 ) 4/99
13
P( X 3)  P( X 4) 0.508
256 24 §3  1· 2
3 a P (X 3) Ÿ¨ ¸ (1  p ) u p
1.1 1.2 CH1 Exercise D 125 © 2  1 ¹
4 ­ p1 = − 0.274
°
( nCr(k–1, 3–1)∙(0.75)k–3∙(0.25)3 ) =
24
Ÿ p 2 − p3 =
12 °
Ÿ ® p2 =
2
k=3 125 125 ° 5
0.050781
°̄ p3 = 0.874
| 2
Thus p = = 0.4
5

1/99

© Oxford University Press 2015: this may be reproduced for class use solely for the purchaser’s institute Statistics: Chapter 1 6
WORKED SOLUTIONS

§ 2· We can say that it is almost certain that he will not


b X a NB ¨ 2, ¸ Ÿ P(3 d X d 5)
5 © ¹ need to interview more than a dozen students.
5 1.4 1.5 CH1 Exercise D
1572 1.3
¦ P( X
k 3
k)
3125
0.503 0.047339
nCr(5,4)∙0.08∙(0.92)5 0.263633
1.1 1.2 1.3 CH1 Exercise D 12

polyRoots (
–p3+p2–
12
,p ) (nCr(k–1, 5–1)∙(0.08) – ∙(0.92) ) k 5 5
125
k=5
{–0.274456, 25 ,0.874456} 0.999999

5 1572 4/99
( nCr(k–1,2–1)∙
3
5
( ) ( ))
k –2

2
5
2
3125

k=3 7 a X a NB(0.85, 3) Ÿ P( X 5) 0.0829


|
2/99
b X a NB(0.85, 3) Ÿ P( X t 7 )

§ 1· 1  P( X d 6) 0.589
4 a X a NB ¨ 2, ¸
© 4¹ 1.3 1.4 1.5 CH1 Exercise D
§ 1· 3 0.999999
b X a NB ¨ 2, ¸ Ÿ P( X x)
© 4¹ 32 nCr(4,2)∙(0.15)2∙(0.85)3 0.082907
x 2 2 6
§ x  1· § 3 · §1· 3
Ÿ¨ ¸¨ ¸ u¨ ¸ 1– (nCr(k–1, 3–1)∙(0.15) – ∙(0.85) )
k 3 3

© 2  1¹ © 4 ¹ 4© ¹ 32
k=3
x 2 0.005885
3 3
Ÿ ( x  1) Ÿx 3 |
4x 32 6/99
§ 1·
c X a NB ¨ 2, ¸ Ÿ P ( X d 5)
© 4¹ Exercise 1E
5 x 4x 4
47 §4· § 1· § 1· §4· § 1·
¦ P (X k )
128
1 P( X x) ¨ ¸¨ ¸ ¨ ¸ ¨ ¸¨ ¸
k 2
© x ¹ ©N
2 ¹ ©
2 ¹
©x¹©2¹
heads tails
1.2 1.3 1.4 CH1 Exercise D

(x–1)·3x–2 3 xi 0 1 2 3 4
nSolve ( 4x

32
=0,x ) 3.
1 4 6 4 1
5 47
pi
16 16 16 16 16
( nCr(k–1, 2–1)∙ (34) – ∙(14) )
k 2 2
128
1 1 3 1 1
k=2 G (t )  t  t2  t3  t4
| 16 4 8 4 16

2 xi 1 2 3 ... k ...
2/99
k 1 k 1
1 5 25 §5· 1 5
pi ... ¨ ¸ u ...
5 a X a NB(0.85, 4 ) Ÿ P( X 5) 0.313 © 6 ¹
one
6 6k
6 36 216
not one

b X a NB(0.85, 4 ) Ÿ P( X t 7 ) 1 5 25 3
G (t ) t  t2  t !
1  P( X d 6) 0.0473 6 36 216
k
5k 1 k 1 f §5 ·
1.3 1.4 1.5 CH1 Exercise D  t ! t ¦ ¨ t¸
6k 6 k © 6 ¹

0
nCr(4,3)∙0.15∙(0.85)4 0.313204
inf. geom. sequence
6
1 1 1 6 t
1– (nCr(k–1, 4–1)∙(0.15) – ∙(0.85) )
k 4 4
6
tu
5 6
tu
6  5t 6  5t
k=4 1 t
0.047339 6
5 5 6 6
t  1 Ÿ 1  t  1 Ÿ   t 
6 6 5 5
2/99
2
6 a X a NB(0.92, 5) Ÿ P( X 6) 0.264 2 3 2 2 2 2
3 a G (t )  t  t
3t 1 3 9 27
1 t
b X a NB(0.92, 5) Ÿ P( X d 12) 0.999999 | 1 3
2 3 2 2
 t  !  k 1 t k  ! Ÿ P( X 0)
81 3 3

© Oxford University Press 2015: this may be reproduced for class use solely for the purchaser’s institute Statistics: Chapter 1 7
WORKED SOLUTIONS

2 2 8 5 For question 1:
b P( X d 1) P( X 0 )  P( X 1) 
3 9 9 1 1 3 1 1
G (t )  t  t2  t3  t4,
c P( X t 3) 1  P( X d 2 ) 16 4 8 4 16
1 3 3 1
1¨
§2 2
 
2 ·
1
26 1 G c(t )  t  t 2  t 3 Ÿ G c(1) 2
¸ 4 4 4 4
© 3 9 27 ¹ 27 27
P( X t k ) 1  P ( X d k  1) 3 3 3
d G cc(t )  t  t 2 Ÿ G cc(1) 3
4 2 4
§2 2 2·
=1 ¨  ! k ¸
©3 9 3 ¹ E( X ) G c(1) 2; Var( X ) = G ′′(1)
2§ 1·
¨1  k ¸ + G ′(1)(1 − G ′(1)) = 3 + 2(1 − 2) = 3 − 2 = 1
3 3 ¹ 1 1
1 © 11 k
1
1 3 3k 1.1 1.2 CH1 Exercise E
f1(1) 1
3
d 2
4 a X  B(1, p ) dx (f1(x)) x=1

xi 0 1 d2
(f1(x)) x=1 3
dx2
pi 1–p p
G(t) = 1 – p + pt
3/99
G c(t ) p, G cc(t ) 0

E( X ) G c(1) p, For question 2:


t 6  5t  t u ( 5)
G (t ) , G c(t )
Var( X ) G cc(1)  G c(1)(1  G c(1)) 6  5t (6  5t )2
0  p (1  p ) p (1  p ) [ pq ] 6(6  5t )2 Ÿ G c(1) 6
r
§ pt · G cc(t )  12(6  5t )3 u ( 5)
b X a NB(r , p ), G (t ) ¨ ¸
© 1  qt ¹
r 1 60(6  5t )2 Ÿ G cc(1) 60
§ pt · p (1  qt )  pt u ( q )
G c(t ) r¨ ¸
© 1  qt ¹ (1  qt )2 E( X ) = G ′(1) = 6; Var( X ) = G ′′(1)

§ pt ·
r 1
1 + G ′(1)(1 − G ′(1)) = 60 + 6(1 − 6) = 60 − 30 = 30
rp ¨ ¸ rp r t r 1 (1  qt ) ( r 1)
© 1  qt ¹ (1  qt )2 1.1 CH1 Exercise E
(f1(x1.2)) x=1
r
G ′(1) = rp r p − ( r +1) = dx2
p f2(1) 1
d 6
G cc(t ) rp r (r  1)t r  2 (1  qt ) ( r 1) dx (f2(x)) x=1

d2 60
 rp r t r 1 u ( (r  1))(1  qt ) ( r  2 ) u ( q ) dx2
(f2(x)) x=1

rp r t r  2 (1  qt ) ( r  2 ) ((r  1)(1  qt )  (r  1)qt ) 6/99

G cc(1) rp r u p  r  2 ((r  1) p  (r  1)q ) For question 3:


2
2(3  t )2
r
p2
 p 
r 
rp  rq  q r
p2
( p  r  q )
G (t )
3t
, G c(t )

1
Ÿ G c(1) 2(3  1)2
r 2
E( X ) G c(1) ,
p 1
G cc(t ) 4(3  t )3 Ÿ G cc(1) 4(3  1)3
2
Var( X ) G cc(1)  G c(1)(1  G c(1)) 1
E( X ) G c(1) ,
r r§ r· 2
( p  r  q )  ¨1  ¸
p2 p© p¹ Var( X ) G cc(1)  G c(1)(1  G c(1))
2 2
 rp  r  rq  rp  r rq 1 1§ 1· 1 1 3
 ¨1  ¸ 
p2 p2 2 2© 2¹ 2 4 4

© Oxford University Press 2015: this may be reproduced for class use solely for the purchaser’s institute Statistics: Chapter 1 8
WORKED SOLUTIONS

1.1 CH1 Exercise E CH1 Exercise E


(f2(x1.2)) x=1 1.1
√3(x1.2)) x=1 2
dx2 dx2
f3(1) 1 f4(1) 1
d 1 d 5
dx (f3(x)) x=1 2 dx (f4(x)) x=1 3
1 19
d2 d2
dx2
(f3(x)) x=1 2
dx2
(f4(x)) x=1 9
| |
9/99 12/99

4 3 1 3 2 2 Var( X ) G cc(1)  G c(1)(1  G c(1))


6 a P( X 1)
; P( X 4 ) u u u
7 7 3 7 3 49 19 5 § 5· 19 10
 ¨1  ¸  1
4 3 2 3 1 4 9 3© 3¹ 9 9
b G (t ) = t + × t2 + × × t3
7 7 3 7 3 7 5 14
x max E( X )  3V 3u 1 4.67 .
3 1 3 2 3 1 3 1 4 3 3
+ × × × t4 + × × × × t3 Therefore, the maximum number of shots to be
7 3 7 3 7 3 7 3 7
3 1 3 1 3 2 made before the game is over is 5.
 u u u u u t4 !
7 3 7 3 7 3 Exercise 1F
2
4 § 1 1 4 · § 1  3t ·
= t ¨1 + t 2 + t +!¸ 1 a G X Y (t ) GX (t ) u GY (t ) ¨ ¸
7 © 7 49 ¹ © 4 ¹
2 2
2 § 1 1 4 · §2  t · § 2  7t  3t 2 ·
+ t 2 ¨1 + t 2 + t + !¸ u¨ ¸ ¨ ¸
7 © 7 49 ¹ © 3 ¹ © 12 ¹
2
§4 2 2·§ 1 2 1 4 · § 2  7t  3t 2 · 4 28
¨ t  t ¸ ¨ 1  t  t  !¸ b G X Y (t ) ¨ ¸  t !
©7 7 ¹© 7 49 ¹ © 12 ¹ 144 144
4 t  2t 2 1 4 t  2t 2 7 4 t  2t 2 4 28 32 2
u u Ÿ P( X  Y d 1) 
7 1 7 7  t2 7  t2 144 144 144 9
1  t2
7 1
c G cX Y (t ) u 2(2  7t  3t 2 )(7  6t )
42 6 144
Verifying: G(1) 1 1 13
7 1 6 Ÿ E( X  Y ) GXc Y (1) u 12 u 13
72 6
4 t  2t 2 ( 4  4 t )(7  t 2 )  2t ( 4 t  2t 2 )
c G (t ) , G c(t ) 3
§ t · § t ·
3
7  t2 (7  t 2 )2 G X (t ) u GY (t )
2 a G X Y (t ) ¨ 2  t ¸ u ¨ 3  2t ¸
© ¹ © ¹
4 (7  7  t 2  t 3  2t 2  t 3 ) 4 (7  7t  t 2 ) 3

(7  t 2 )2 (7  t 2 )2
§ t2 ·
¨ 2 ¸
© 6  7 t  2t ¹
4 (7  7  1) 5
E( X ) G c(1) b E( X  Y ) G c(1) 15
(7  1)2 3
c Var( X  Y ) G cc(1)  G c(1)(1  G c(1))
The expected number of shots before the game
is over is 2. 234  15 u ( 14 ) 234  210 24
1.1 CH1 Exercise F
d G c(t ) 4 (7  7t  t 2 )(7  t 2 )2 ,
G cc(t ) 4 (7  2t )(7  t ) 2 2
((
d t 3 t 3
) ( ))
dt 2–t • 3–2•t t=1
15

2 3
 2(7  t ) u ( 2t ) u 4 (7  7t  t ) 2
d2
dt2 (( 2–t ) •( 3–2•t) ) t=1
t 3 t 3 234

4 ((7  2t )(7  t 2 )  4 t (7  7t  t 2 ))
(7  t 2 )3
4 ( 49  14 t  7t 2  2t 3  28t  28t 2  4 t 3 ) 2/99

(7  t 2 )3
3 a G X (t ) = e0.25( t −1) , GY (t ) = e0.15( t −1) , G Z (t ) = e0.05( t −1)
4(49 + 42t + 21t 2 + 2t 3 )
= b G X Y  Z (t ) G X (t ) u GY (t ) u G Z (t )
(7 − t 2 )3
4(49 + 42 + 21 + 2) 19 = e0.25( t −1) × e0.15( t −1) × e0.05( t −1) = e0.45( t −1)
Ÿ G ′′(1) = =
(7 − 1)3 9 G X +Y + Z (t ) = e0.45( t −1) Ÿ X + Y + Z ∼ Po(0.45)
P( X  Y  Z d 2) 0.989

© Oxford University Press 2015: this may be reproduced for class use solely for the purchaser’s institute Statistics: Chapter 1 9
WORKED SOLUTIONS

1.1 1.2 CH1 Exercise F Review exercise


poissCdf(0.45,0,2) 0.989121 4t 4 1
| 1 a G (t ) tu
5t 5 1
1 t
5
4 § 1 1 2 1 3 ·
t u ¨1  t  t  t  "¸
5 © 5 25 125 ¹
4 4 4 3 4 4
1/99 t  t2  t  t "
5 25 125 625
pt 4 4 4 4 624
4 X a Geo( p ), GX (t ) Ÿ P 1 d X d 4   
1  qt 5 25 125 625 625
Y ∼ NB(r , p ), Y = 
X + X
+!+ X
Ÿ GY (t ) §4 4 · 24 1
r terms(sucesses) b P( X t 2 ) 1  P( X d 1) 1¨  ¸ 1
© 5 25 ¹ 25 25
G X

X  X
( t )

! G X ( t ) u G X ( t
 u ! u G X (t
)
)
4t 4 (5  t )  4 t
r terms r factors c G (t ) , G c(t )
r
5t (5  t ) 2
pt pt pt § pt ·
u u !u ¨ ¸ 20 20 5
 qt 1  qt
1 1  qt
© 1  qt ¹ = Ÿ E ( X ) = G ′(1) = =
r factors (5 − t ) 2
16 4

X1 a B(n1 , p ), G X1 (t ) (q  pt )n1 ½° d G c(t ) 20(5  t )2 ,


5 a ¾ Ÿ G X1  X 2 ( t )
X 2 a B(n2 , p ), G X 2 (t ) (q  pt )n2 ¿°
G cc(t )  40(5  t )3 u ( 1) 40(5  t )3
G X1 ( t ) u G X 2 ( t ) (q  pt )n1 u (q  pt )n2
Ÿ Var( X ) G cc(1)  G c(1) (1  G c(1))
n1  n2
(q  pt ) Ÿ X1  X 2 a B(n1  n2 , p ) 5§ 5· 5 5 5
= 40 × 4 −3 + ¨1 − ¸ = − =
4© 4 ¹ 8 16 16
b 1st Base
1.1 CH1 Rewiev ...ise
For n = 1, the statement is trivial: X1 a B(n1 , p )
For n = 2, in part a we proved that ( )
d 4•t It=1
dt 5–t
5
4

X1  X 2 a B(n1  n2 , p ) d2 4•t It=1


( )
dt2 5–t
5
8
2nd Assumption
Let’s assume that for n = k this statement is true:
5 5
( )
+ • 1–
8 4
5
4
5
16

k
§ k · 3/99
¦X
i 1
i a B ¨ ¦ ni , p ¸
©i 1 ¹
2 a G X (t ) = e0.6( t −1) , GY (t ) = e0.12( t −1) , GZ (t ) = e0.28( t −1)
3rd Step
k 1 k b G X Y  Z (t ) G X (t ) u GY (t ) u G Z (t )
For n k  1, ¦ X i ¦X  X k 1 ,
i 1 i 1
i
= e0.6( t −1) × e0.12( t −1) × e0.28( t −1) = et −1
by the assumption the first term has a binomial G X Y  Z (t ) et 1 Ÿ E( X  Y  Z )
distribution and by the given proposition in the G X Y  Z (1) e11 e0 1
question X k 1 ~ B(nk 1 , p ). Now we use the fact The expected number of animals at the meadow
from part a that the sum of two binomial at the given moment is 1.
variables with the same probability p satisfies
the following: c G X +Y + Z (t ) = et −1 Ÿ X + Y + Z ~ Po(1)
k
§ k ·½ P ( X + Y + Z ≥ 2 ) = 1 − P ( X + Y + Z ≤ 1) = 0.264
¦ X i ~ B ¨ ¦ ni , p ¸ ° Ÿ X  X ~
k

i 1 ©i 1 ¹¾ ¦ i k 1 1.1 1.2 CH1 Rewiev ...ise

X k 1 ~ B(nk 1 , p ) °¿
i 1
1–poissCdf(1,0,1) 0.264241

k 1
§ k · § k 1 ·
B ¨ ¦ ni  nk 1 , p¸ Ÿ ¦ X i ~ B ¨ ¦ ni , p¸
©i 1 ¹ i 1 ©i 1 ¹
4th Conclusion
k
§ k ·
Now we can conclude that Y = ¦ X i ~ B ¨ ¦ ni , p¸
i 1 ©i 1 ¹ 1/99
for all the positive integer values of k.
© Oxford University Press 2015: this may be reproduced for class use solely for the purchaser’s institute Statistics: Chapter 1 10
WORKED SOLUTIONS

3 a i X ~ Geo(0.75) Ÿ P(X = 4) = 0.0117 1


5 a G (1) = 1 Ÿ 1Ÿ a  b 1Ÿ b a 1
ab
ii Y ~ NB(2, 0.75) Ÿ P(Y = 4)
a  bt  t u ( b ) a
§ 4  1· 2 2 27 b G c(t )
¸ 0.25 u 0.75 0.105 a  bt 2 2
¨ a  bt
© 2  1¹ 256
a a
Ÿ E( X ) = G ′(1) = = =a
(a − b )
2
iii P ( X t 3) 1  P( X d 2 ) 0.0625 12
iv Z ~ NB(6, 0.75) c G c(t ) a (a  bt )2 ,
§ 6  1· 2ab
Ÿ P Z d 10 ¨ ¸ 0.75
6
G cc(t ) a u (2)(a  bt )3 u ( b )
© 6  1¹ (a  bt )3
§ 7  1· 6 Var( X ) G cc(1)  G c(1) 1  G c(1)
¨ ¸ 0.75 u 0.25  !
© 6  1¹ 2a (a  1)
 a(1  a )
§ 10  1· 6 4
13
¨ ¸ 0.75 u 0.25 0.922
©6  1 ¹ 2a 2  2a  a  a 2 a2  a

6 X1 ~ Po(m ), G X1 (t ) = em ( t −1) ½°
b E( Z ) = = 8, the expected number of 6 a ¾ Ÿ G X1 + X 2 ( t )
0.75 X 2 ~ Po(m ), G X 2 (t ) = em ( t −1) °¿
students needed to be selected if we need six
involved in the programme is eight. = G X1 (t ) × G X 2 (t ) = em ( t −1) × em ( t −1) = em ( t −1) + m ( t −1)

1.1 1.2 1.3 CH1 Rewiev ...ise e2 m ( t 1) Ÿ X1  X 2 ~ Po(2m )


geomPdf(0.75,4) 0.011719 b 1st Base
2·(0.75)2 approxFraction(5.I
nCr(3,1)·(0.25)
For n =1, the statement is trivial: X 1 ~ Po(m )
27
256
For n = 2, in part a we proved that X1 + X 2 ~ Po(2m )
1–geomCdf(0.75,2) 0.0625

2nd Assumption
3/99 Let’s assume that for n = k this statement is true:
k

1.1 1.2 1.3 CH1 Rewiev ...ise


1–geomPdf(0.75,4) 0.0029
¦X
i 1
1 ~ Po(km )

4 3rd Step
(0.75)6· (nCr(5+k,5)·(0.25) ) k
k  1, ¦ X1
k 1 k

k=0 0.921873
For n
i 1
¦X
i 1
i  X k 1, by the
6 8.
0.75
assumption the first term has a Poisson
| distribution with the coefficient km and by the
5/99 given proposition in the question X k 1 ~ Po(m ).
Now we use the fact that the coefficient of the
22 sum of two Poisson variables is the sum of the
4 a F (2) 1Ÿ 1 Ÿ a 2 Ÿ 22 Ÿ a 2 or
a2 coefficients of those two variables.
a  2 since a must be positive. k
½
­ 2x ¦X ~ Po(km ) °
i
k

b F c( x ) f (x ) Ÿ f (x )
° 2,0d x d2
®2
i 1 ¾Ÿ ¦X i  X k 1 ~ Po(km  m )
X k 1 ~ Po(m ) °¿
i 1
°¯ 0, otherwise
k 1
­x
° ,0d x d2 Ÿ ¦X i ~ Po (k  1)m
®2 i 1
°¯ 0, otherwise
4th Conclusion
c Since f (x) is increasing on the whole interval n
[0, 2] the modal value of the variable X is 2. Now we can conclude that Y = ¦ X i ~ B(nm )
k =1

for all the positive integer values of n.

© Oxford University Press 2015: this may be reproduced for class use solely for the purchaser’s institute Statistics: Chapter 1 11
WORKED SOLUTIONS

Expectation algebra and


2 Central Limit Theorem
Skills check Exercise 2A
np 2½ 1 E( X ) 5.3, Var ( X ) 1 .2
° 3 3
1 3 ¾ Ÿ 2q Ÿq ,
npq 2 4 E(3 X ) 3 u 5 .3 15.9,
2 °¿ a
2
Var(3 X ) 3 u 1 .2 10.8
1 § 1·
p ,n 8 Ÿ X ~ B ¨ 8, ¸
4 © 4¹ b E( X  3) 5 .3  3 8.3, Var( X  3) 1 .2
2 6
§8· §1· §3· c E( 4 X  1) 4 u 5.3  1 22.2,
a P( X 2) ¨ ¸¨ ¸ ¨ ¸ 0.311
©2¹ ©4¹ ©4¹ Var( 4 X  1) 2
4 u 1 .2 19.6
3 k 8k
§8·§1· §3· E(2 X  5) 2 u 5 .3  5 5.6,
b P(1 d X d 3) ¦ ¨© k ¸¹ ¨© 4 ¸¹
k 1
¨ ¸
©4¹
0.786 d

Var(2 X  5) 2 2 u 1 .2 4 .8
1.1 CH2 Skills check

1 0.311462 e E(kX  p ) k u 5.3  p,


( )
binomPdf 8, 4 ,2
Var(kX  p ) k 2 u 1.2 = 1.2k 2
1 0.786072
( )
binomCdf 8, 4,1,3

§ 2· 2
2 X ~ B ¨ 10, ¸ Ÿ E( X ) 10 u 4,
© 5 ¹ 5
2/99 2 3 12
Var( X ) 10 u u 2 .4
5 5 5
a E(3 X  2) 3 u E( X )  2 3u4  2 14
2 X ~ Po(m ) Ÿ E( X ) Var ( X )
12 108
2 3 3 2
b Var (3 X  2) 32 u Var ( X ) 9u 21.6
7a 6a  9 a  2 Ÿ 6a  7 a  9 a  2 0 5 5

Ÿ (2a  1)(3a  1)(a  2) 0 §2· 1 3


3 Y ~ Geo ¨ ¸ Ÿ E(Y ) 1.5,
3 © ¹ 2 2
1.1 1.2 CH2 Skills check
3
3 y
1
f1(x)=6·x3–7·x2–9·x–2
3 3
Var(Y ) 2
0.75
1 §2· 4
(2,0)
(–0.5,0) (–0.333,0) x ¨ ¸
©3¹
–3 1 3
3
E(2Y  1) 2 u E(Y )  1 2u 1 2,
2
–3 3
Var (2Y  1) 22 u Var (Y ) 4u 3
4
1 1
Ÿ a1  , a2  , a3 2
2 3 4 Y ~ Po(2) Ÿ E(Y ) 2, Var (Y ) 2
We see that all three values of a give positive
a E(3  2Y ) 3  2 u E(Y ) 32u2 1
values of the parameter m:
b Var (3  2Y ) 22 u Var (Y ) 4u2 8
7 7
m1 , m2 , m3 28
4 9

© Oxford University Press 2015: this may be reproduced for class use solely for the purchaser’s institute Statistics: Chapter 2 12
WORKED SOLUTIONS

5
§
X ~ NB ¨ 8, ¸ Ÿ E( X )
1· 8
24, Exercise 2B
© 3¹ 1
1 a E( X  Y ) 3  ( 5)  2,
3
2 Var( X  Y ) 0 .5  1 . 4 1 .9
8u
Var(Y ) 3 48
2 b E(2Y  Z ) 2 u ( 5)  12  22,
§1·
¨ ¸ Var(2Y  Z ) 2
2 u 1 .4  2 .8 8 .4
©3¹

a E ( 2 X  3) 2 u E( X )  3 2 u 24  3 45 c E(2 Z  7 X ) 2 u 12  7 u 3 3,
2 2
b Var (2 X  11) 22 u Var ( X ) 4 u 48 192 Var(2 Z  7 X ) 2 u 2 .8  7 u 0 .5 35.7

2 d E( X  Y  Z ) 3  ( 5)  12 20,
6 X ~ B(15, p ) Ÿ E( X ) np Ÿ 15 p 6Ÿ p
5 Var( X  Y  Z ) 0 .5  1 . 4  2 . 8 4 .7
3 18
Var( X ) npq Ÿ Var ( X ) 6u 3.6 e E( X  Y  Z ) 3  ( 5)  12  14,
5 5
Var( X  Y  Z ) 0 .5  1 . 4  2 . 8 4 .7
2 18
Var(5X  3) 5 u 90
5 f E(3Z  2 X  4Y ) 3 u 12  2 u 3  4 u ( 5) 10 ,

6 6 6 Var(3Z  2 X  4Y ) 3 2 u 2 .8  2 2 u 0 . 5
1 1 2 1 ª x3 º
7 E(X ) ³ x u x dx ³x dx « »  4 2 u 1 .4 49.6
0
3 3 0
3 ¬ 3 ¼0
1 2 2 Var ( X ) 2 Ÿ X ~ Po(2) Ÿ E( X ) 2 and
6 6 6
9 3
2
E(Y ) 5 Ÿ Y ~ Po(5) Ÿ Var (Y ) 5
6
1 2 §2 ·
Var( X ) ³ x u x dx  ¨ 6 ¸
0
3 © 3 ¹ a E(3 X  5Y ) 3u2  5u5 31
6
1
6
8 1 ªx4 º 8 b Var(11Y  7 X ) 112 u 5  7 2 u 2 703
³ x 3 dx  « » 
3 0
3 3 ¬ 4 ¼0 3
3 E( X ) 9 Ÿ n1 p 9 Ÿ Var ( X ) 9(1  p )
1 8 8 1
u 62  3 E(Y ) 4 Ÿ n2 p 4 Ÿ Var ( X ) 4 (1  p )
12 3 3 3
1.1 CH2 Exercise A Var(2 X  3Y ) 22 u 9(1  p )  32 u 4 (1 – p)
√6 1.63299

∫(0
1
x· ·x dx
3
)
r1
72  72 p

2
·√6 1.63299 4 E( X ) 8Ÿ 8Ÿ
3 p
√6 1 p 8
∫ ( x2·
1
)
·x dx–(1.6329931618555)2
Var( X ) 8 u
p p
8
1/3
r
E(Y ) 12 Ÿ 2 12 Ÿ
p
1.1 CH2 Exercise A
1  p 12
2
·√6 1.63299 Var(Y ) 12 u  12
3 p p
√6 8 12 20
∫ 0
(
x2·
1
3
)
·x dx–(1.6329931618555)2

0.333333| 5
Var( X  Y )
p
8
p
 12

Given the n independent variables and their


p
 20

corresponding parameters
1/3
X n , E( X n ), Var ( X n ), n  ' 
2 1st Base
E(3X  2) 3u 62 2 6  2,
3 n=1
1
Var (3X  2) 32 u 3 X1 , E( X1 ), Var ( X1 ) Ÿ E(aX1 ) aE ( X1 ),
3
2
Var (aX1 ) a Var ( X1 )

© Oxford University Press 2015: this may be reproduced for class use solely for the purchaser’s institute Statistics: Chapter 2 13
WORKED SOLUTIONS

2nd Assumption b Since we have 6 independent flips of the same


coin we add six instances of the variable X.
Let’s assume that for n = k the statement is true so
k
c Find the expected value and variance of the
the parameters of ¦a X
i 1
i i are calculated as follows:
random variable Y.
k k
§ · 1 1 3
E ¨ ¦ ai X i ¸ ¦ a E( X ) and
i i
E(Y ) 6u 3, Var (Y ) 6u
©i 1 ¹ i 1
2 4 2
k k
§ · 3
Var ¨ ¦ ai X i ¸ ¦ a Var ( X ) i
2
i d E(Y ) r 3 u Var (Y ) 3r3u
©i 1 ¹ i 1 2
3 r 3.67 Ÿ y  [0.67, 6.67]
3rd Step
k 1 k
k  1, ¦ X i We notice that all the possible outcomes {0, 1, 2,
For n
i 1
¦X i 1
i  X k 1 ,
3, 4, 5, 6} are covered by the 99.7% empirical rule.
§ k 1 · § k · 2 Let’s roll an unbiased die 4 times. Record the number
E ¨ ¦ ai X i ¸ E ¨ ¦ ai X i  ak 1 X k 1 ¸ of multiples of 3 obtained.
©i 1 ¹ ©i 1 ¹
a
X=x 0 1
§ k ·
E ¨ ¦ ai X i ¸  E(ak 1 X k 1 ) P{X = x}
2 1
©i 1 ¹ 3 3
k k 1 2 1 1
¦a E X  ai i k 1 E X k 1 ¦a E X
i i
E( X ) 0u  1u ,
i 1 i 1
3 3 3
2
2 1 §1· 1 1 2
§ k 1 · § k · Var( X ) 02 u  12 u  ¨ ¸ 
Var ¨ ¦ ai X i ¸ Var ¨ ¦ ai X i  ak 1 X k 1 ¸ 3 3 ©3¹ 3 9 9
©i 1 ¹ ©i 1 ¹
k b Since we have 4 independent rolls of the same
¦ ai2 Var X i
i 1
ak21 Var X k 1 die we add four instances of the variable X.
1 4 2 8
§ k 1 · c E(Y ) 4u , Var (Y ) 4u
= Var ¨ ¦ ai X i ¸ 3 3 9 9
©i 1 ¹
3 a E( X  X  X ) 3 u E( X ) 3u3 9,
4th Conclusion
Var ( X  X  X ) 3 u Var ( X ) 3u4 12
The parameters of a linear combination of n
random variables can be calculated by: b E(3 X ) 3 u E( X ) 3u3 9,
2
§ n · n Var (3 X ) 3 u Var ( X ) 9u4 36
E ¨ ¦ ai X i ¸ ¦ a E( X ), i i
©i 1 ¹ i 1 c Var ( X  X  X  3 X ) 3 u Var ( X )
2
§ n · n  3 u Var ( X ) 12 Var ( X ) 12 u 4 48
Var ¨ ¦ ai X i ¸ ¦ a Var ( X ) i
2
i
©i 1 ¹ i 1 Var (6 X ) 62 Var ( X ) 36 u 4 144 Ÿ
Var ( X  X  X  3 X ) z Var (6 X )
Exercise 2C
4 a E( X  X  X  X  X ) 5 u E( X ) 5u2 10
1 Let’s independently flip 6 unbiased coins and
record the number of heads obtained. Var ( X  X  X  X  X ) 5 u Var ( X ) 5u1 5
a
X=x 0 1 b E(Y  Y  Y ) 3 u E(Y ) 3u5 15
1 1 Var (Y  Y  Y ) 3 u Var (Y ) 3u3 9
P{X = x}
2 2
c E( X  X  X  X  X  Y  Y  Y )
1 1 1
E( X ) 0u  1u ,
2 2 2 E( X  X  X  X  X )  E(Y  Y  Y )
2
1 1 §1· 1 1 1 10  15 25
Var( X ) 02 u  12 u  ¨ ¸ 
2 2 ©2¹ 2 4 4

© Oxford University Press 2015: this may be reproduced for class use solely for the purchaser’s institute Statistics: Chapter 2 14
WORKED SOLUTIONS

Var(X + X + X + X + X + Y + Y + Y ) 6 E( X ) u E(Y ) E( XY ) Ÿ P u ( P  2) P3
Var( X  X  X  X  X )  Var (Y  Y  Y ) Ÿ P 3  P 2  2P 0
59 14 Ÿ P ( P  1)( P  2) 0

5 a Ÿ P 0 or P 1 or P = 2.
X = xi 1 2 3 4
1 1 1 1 1.1 CH2 Exercise C
P{X = xi}
4 4 4 4 3 y

f1(x)=x3–x2–2·x
1 1 5 1
E( X ) u (1  2  3  4 ) u 10 (–1,0) (0,0) (2,0)
4 4 2
–5 1 5
Y = yi 1 2 3
1 1 1 –3
P(Y = yi)
2 3 6

1 1 1 5
E(Y ) u1 u 2  u3 Exercise 2D
2 3 6 3
1 a P 1  ( 2)  3 0, V 2 0.16  0.25  1.21 1.62
1 1 1
b 1: P ^1, 1` u
4 2 8 Y  Z  W ~ N (0, 1.62)
1 1 1 1 5 Ÿ P(Y  Z  W  0 ) 0 .5
2 : P ^1, 2` or ^2, 1` u  u
4 3 4 2 24
b P 0  1  ( 2)  3 2,
1 1 1 1 1
3 : (1, 3) or (3, 1) o u  u
4 6 4 2 6 V2 1  0.16  0.25  1.21 2.62
1 1 1 1 5
4 : P ^2, 2` or ^4, 1` u  u X  Y  Z  W ~ N (2, 2.62)
4 3 4 2 24
1 1 1 1 1 Ÿ P( X  Y  Z  W ! 0 ) 0.892
6 : P ^2, 3` or ^3, 2` u  u
4 6 4 3 8
c P 3 u 0  1  ( 2)  3 0,
1 1 1
8 : P ^4, 2` u
4 3 12 V2 32 u 1  0.16  0.25  1.21 10.62
1 1 1
9 : P ^3, 3` u 3 X  Y  Z  W ~ N (0, 10.62)
4 6 24
1 1 1 Ÿ P(3X  Y ! Z  W )
12 : P ^4, 3` u
4 6 24
P(3 X  Y  Z  W ! 0 ) 0 .5

Z = zi 1 2 3 4 6 8 9 12 d P 0  2 u 1  3 u ( 2)  3  1,
1 5 1 5 1 1 1 1
P(Z = zi) V2 1  22 u 0.16  32 u 0.25  1.21 5 .1
8 24 6 24 8 12 24 24
X  2Y  3Z  W ~ N (1, 5.1)
1 5 1
E( Z ) 1u 2u 3u
8 24 6 Ÿ P( X  3Z d 2Y  W )
5 1 1
4u  6u  8u P( X  2Y  3Z  W d 0 ) 0.329
24 8 12
1 1 25
9u  12 u e P 0  4 u (2)  3 u 0  2( 2) 4,
24 24 6
5 5 25 σ 2 = 1 + 4 2 × 0.25 + 32 × 1 + 22 × 0.25 = 15
E( X ) u E(Y ) u E( Z ) E( XY )
2 3 6 X  =  X  2 Z  N(4, 15)

Ÿ P( X   = d X  2 Z )

P( X  4 Z  3 X  2 Z d 0 ) 0.151

© Oxford University Press 2015: this may be reproduced for class use solely for the purchaser’s institute Statistics: Chapter 2 15
WORKED SOLUTIONS

f P 3 u 1 2 u 3  9, c Y a N (60, 22 ) Ÿ 4Y a N(4 u 60, 4 2 u 22 )


σ2 1.21  0.16  22 u 0.16  32 u 1.21 12.9
= N(240, 82 )
W − Y − 2Y − 3W  N( −9, 12.9)
Ÿ P(W  Y d 2Y  3W ) X  4Y a N(225  240, 122  82 )
= P(W  Y  2Y  3W d 0 ) = 0.994

= N 15, 4 13
2

1.5 CH2 Exercise D
normCdf(–1000,0,0,√1.62) 0.5 Ÿ P X  4Y ! 0 0.149
normCdf(0,1000,2,√2.62) 0.891697
normCdf(–1000,0,–1,√5.1) 0.671047 d Y ~ N 60, 22 Ÿ
normCdf(–1000,0,2,√14.25) 0.298121
Y  Y  Y  Y ~ N 4 u 60, 4 u 22
normCdf(–1000,0,–9,√17.21) 0.984976
N(240, 4 2 )
5/99
X  Y  Y  Y  Y a N 225  240, 122  4 2
2 a X a N (240, 20 2 ), Y a N (730, 50 2 )

N 15, 4 10
2

Ÿ Y  3 X a N(10, 50 2  32 u 20 2 )
Ÿ P X  Y  Y  Y  Y ! 0 0.118
Ÿ P(Y  3 X t 0 ) 0.551
1.2 1.3 1.4 CH2 Exercise D
2 2
b X a N (240, 20 ), Y a N (730, 50 ) normCdf(250,9.E999,225,12) 0.01861
normCdf(1000,9.E999,900,24) 0.000015
Ÿ 2Y  4 X a N(2420, 2 u 50 2 u 4 u 20 2 ) 1.5463523972415E-5 0.000015
normCdf(0,9.E999,–15,4·√13) 0.149155
Ÿ P(2Y  4 X t 2500 ) 0.162 normCdf(0,9.E999,–15,4·√10) 0.11784

3 a L a N (35, 52 ) Ÿ P( L  30 ) 0.159
5/99
2
b V a N( 45, 8 ) Ÿ
Exercise 2E
W  V  V  V  V a N(5 u 45, 5 u 82 )
2
§ ·

N 225, 8 5
2
1 a X a N ¨¨ 1.5, ¨
§ 2 ·
¸
© 10 ¹
¸¸ Ÿ P 1 d X d 3 0.777
© ¹
P (W ! 240 ) 0.201 § 2
·
§ 3 ·
b X a N ¨¨ 5, ¨ ¸
¸¸ Ÿ P X  5 d 3
c 4 L  3V a N( 4 u 35  3 u 45, 4 2 u 52  32 u 82 ) © © 7¹ ¹


N 5, 4 61
2
Ÿ P( 4 L  3 L ! 0 ) 0.564 P 2 d X d 8 0.992
2
§ § 4 · ·
1.1 1.2 1.3 CH2 Exercise D c X a N ¨¨ 0.2, ¨ ¸ ¸¸ Ÿ P X t 0.8
normCdf(–9.E999,30,35,5) 0.158655 © © 22 ¹ ¹
normCdf(240,9.E999,225,8·√5) 0.200868 1  P 0.8 d X d 0.8 0.639
normCdf(0,9.E999,5,4·√61) 0.563578
1.1 CH2 Exercise E

( 2
normCdf 1,3,1.5, √10 ) 0.776549

3/99
( 3
normCdf 2,8,5, √7 ) 0.991849

2
( 4
normCdf –0.8,0.8,–0.2, √22 ) 0.63867
4 a X a N (225, 12 ) Ÿ P( X ! 250 ) 0.0186

b X  X  X  X a N ( 4 u 225, 4 u 122 ) N (900, 24 2 ) 3/99

Ÿ P( X  X  X  X ! 1000 ) 0.0000155 § 2
·
§ 1.2 ·
2 X  N 5.5, 1.22 Ÿ X  N ¨¨ 5.5, ¨ ¸ ¸¸
© © 15 ¹ ¹

Ÿ P X  5 0.0533

© Oxford University Press 2015: this may be reproduced for class use solely for the purchaser’s institute Statistics: Chapter 2 16
WORKED SOLUTIONS

2 2
§ § 6 · · § § 6 · ·
3 X  N 35, 62 Ÿ X  N ¨¨ 35, ¨ ¸ ¸¸ e X  N ¨ 1, ¨¨ ¸¸ ¸ Ÿ P §¨ X  1 ·¸
© © 5¹ ¹ ¨ © 40 ¹ ¸ © 2¹
© ¹
a P( X t 37) = 0.228
§ 1 1·
P¨  X  ¸ 0.0983
b X + X + X + X + X ~ N(5 × 35, 5 × 6 ) 2 © 2 2¹


N 175, 5 u 6
2
f X  N ¨¨ 1.9, ¨
§ § 6
·
2
· 1
¸¸ Ÿ P §¨ X  2 t ·¸
¸
© © 120 ¹ ¹ © 5¹
⇒ P (X + X + X + X + X > 180) = 0.355
P 1.8  X  2.2 0.280
1.1 1.2 CH2 Exercise E

( 1.2
normCdf 5,9.E999,5.5,√15 ) 0.946708
1.1 CH2 Exercise F

( 6
normCdf 37,9.E999,35, 5 √ )
0.228028 ( 3
normCdf 1.5,2.5,2 √30 ) 0.63869

(
normCdf 180,9.E999,175,√5 ∙6 ) 0.354694 ( 0.2
normCdf 1.25,1.35,1.3, √50 ) 0.9229

normCdf(–0.48,9.E999,–0.5,0.1) 0.42074

3/99 |

3/99
2
§ § 18.5 · ·
2
4 a X  N(62.5, 18.5 ) Ÿ X  N ¨¨ 62.5, ¨ ¸ ¸¸
© © 12 ¹ ¹ 1.1 CH2 Exercise F

P X d 70 0.920 ( 234
normCdf –9.E999,397,400, 85 √ )
0.035295
2
b T X
X 
!  X
 N(12 u 62.5, 12 u 18.5 )
12 terms
–1 1
( 6
normCdf 2 , 2 ,1, 40 √ ) 0.098299

0.280493

N 750, 4107
2
( 6
normCdf 1.8,2.2,1.9,√120 )
⇒ P(T ≤ 800) = 0.782 6/99

1.1 1.2 1.3 CH2 Exercise E

( 18.5
normCdf –9.E999,70,62.5, √12 ) 0.919895
2 X  N ¨¨ 30 000, ¨
§ § 8000 ·
¸
2
·
¸¸
© © 15 ¹ ¹
(
normCdf –9.E999,800,12∙62.5,√12.18.5 )
0.782364
P( X d 32 000) 0.834
(
normCdf –9.E999,800,750,√4107 ) 0.782364
1.1 1.2 CH2 Exercise F
|
3/99 ( 8000
normCdf –9.E999,32000,30000, √15 )
0.833539
|

Exercise 2F

§ § 3 ·2 ·
1 a X  N ¨ 2, ¨ ¸ Ÿ P(1.5 d X d 2.5) 0.639
¨ © 30 ¸¹ ¸ 1/99
© ¹
2
§ § 0.2 · · 2
b X  N ¨1.3, ¨ ¸ ¸ Ÿ P(1.25 d X d 1.35) 0.923 § § 2 · ·
¨
© © 50 ¹ ¸
¹ 3 X  N(63, 22 ) Ÿ X  N ¨¨ 63, ¨ ¸ ¸¸
© © 3¹ ¹
2
§ § 1 · ·
c X  N ¨ 0.5, ¨ ¸ ¸ Ÿ P( X t  0.48) 0.421 a P X ! 64 0.193
© © 10 ¹ ¹
2
§ § 234 · · b P( X  60) 0.00469
¨
d X  N 400, ¨¨ ¸ ¸ Ÿ P X  397 0.0353
¨ © 85 ¸¹ ¸
© ¹

© Oxford University Press 2015: this may be reproduced for class use solely for the purchaser’s institute Statistics: Chapter 2 17
WORKED SOLUTIONS

c P(61 d X d 65) 0.917 § § 4 ·2 § 9 ·2 ·


B  S a N ¨¨ 4, ¨ ¸ ¨ ¸ ¸¸
1.1 1.2 CH2 Exercise F
© © 5¹ © 4¹ ¹

( 2
normCdf 64,9.E999,63, √ 3 ) 0.193238
Ÿ P B ! S P B  S ! 0 0.796
0.004687
(
normCdf –9.E999,60,63,
2
√3
) 1.1 CH2 Rewiew ...ise

( 2
normCdf 61,65,63, √ 3
|
) 0.916736

4/99
( 4 2 9 2
normCdf 0,9.E999,4, √5 + 2
√( ) ( ) )
0.795603

§ § 4 ·2 ·
2
4 X  N(6, 4 ) Ÿ X  N ¨¨ 6, ¨ ¸ ¸¸ 1/99
© © n¹ ¹

§ § 4 · · 2 2 a Var(2X ) = (E(X ))2 – 5 ⇒ 22Var(X ) = (E(X ))2 – 5


a X  N ¨¨ 6, ¨ ¸ ¸¸ Ÿ P(5.4 d X d 7.6) 0.579 ⇒ 4 m = m2 – 5 ⇒ m2 – 4m – 5 = 0
© © 10 ¹ ¹
§56 76·
⇒ (m – 5)(m + 1) = 0 ⇒ m = 5 or m 1 .
¨ dZ d
b P(5 d X d 7) 0.95 Ÿ P ¨ 4 4 ¸
¨ ¸¸ Variance is always positive.
© n n ¹
b P(X ≥ 6) = 1 – P(X ≤ 5) = 0.384
§ n·
0.95 Ÿ P ¨ Z d ¸ 0.975 c Var (3Y ) 18 Ÿ 32 Var (Y ) 18
© 4 ¹
Ÿ Var (Y ) 2 Ÿ Y  Po(2)
n 1
Ÿ ) (0.975) Ÿ n 4 u 1.95996 Ÿ X  Y  Po(5  2)
4

Ÿ n 7.839856 Ÿ n 61.4633 | 62 Ÿ X  Y  Po(7 ) Ÿ P( X  Y  5)

P( X  Y d 4 ) 0.173
1.1 1.2 1.3 CH2 Exercise F

( 4
normCdf 5.4,7.6,6, √10 ) 0.57942
1.1 1.2 CH2 Rewiev ...ise

| 1–poissCdf(5,0,5) 0.384039
poissCdf(7,0,4) 0.172992

1/99

2/99
1.2 1.3 1.4 CH2 Exercise F
y
1
d E(Z ) = E(3X – 4Y) = 3 × 5 – 4 × 2 = 7
0.5 4 Var(Z ) = Var(3X – 4Y ) = 32 × 5 + 42 × 2 = 77
f1(x)=normCdf(5,7,6, )–0.95
√x
e The random variable Z has no Poisson
0 10 (61.5,0) 80 x distribution since 7 = E(Z) ≠ Var(Z) = 77.
–0.5
3 a X  N 400, 20 2 Ÿ P( X ! 450 ) 0.00621

b Y  N(350, 152 ) Ÿ Y  Y  N(2 u 350, 2 u 152 )

Review exercise Ÿ P(Y  Y  670 ) 0.0787


2
§ § 4 · · c Z  N(320, 122 )
1 B  N(202, 4 2 ) Ÿ B  N ¨¨ 202, ¨ ¸ ¸¸
© © 5¹ ¹
Ÿ X  Z  N(400  320, 202  122 )
2
§ § 9 · ·
S  N(198, 92 ) Ÿ S  N ¨¨ 198, ¨
© © 4¹
¸ ¸¸
¹

N 720, 4 34
2

© Oxford University Press 2015: this may be reproduced for class use solely for the purchaser’s institute Statistics: Chapter 2 18
WORKED SOLUTIONS

d
X  Z  2Y  N 720  2 u 350, 4 34  22 u 152
2
c
§
X  N ¨¨12000, ¨
§ 3200 ·
¸
2
·
¸¸
© © n ¹ ¹
N 20, 382
Ÿ P(1050  X  12500) 0.85 Ÿ n 44.3 | 45
Ÿ P( X  Z  2Y  0 ) 0.299

1.1 1.2 1.3 CH2 Rewiew ...ise 1.3 1.4 1.5 CH2 Rewiew...ise
y
1
normCdf(450,9.E999,400,20) 0.00621
normCdf(–9.E999,670,700,15∙√2)
0.5
(
f1(x)=normCdf 10500,12500,12000,
3200
√x
–0.85 )
0.07865
normCdf(–9.E999,0,20,38) 0.299334
0 10 (44.3,0) 100 x

–0.5
3/99

4 a X  B(n, p ) Ÿ Var ( X ) 6 Ÿ npq 6


6 a E X  E( X )
2

E X 2  2 X E( X )  E( X )
2

2
n 27 Ÿ 27 u pq 6 Ÿ p (1  p )
E X 2  2E( X ) u E( X )  E( X )
2
9
Ÿ 9 p2  9 p  2 0 Ÿ (3 p  1)(3 p  2) 0
E X 2  2 E( X )  E( X )
2 2
1 2
p or p
3 3
E X 2  E( X ) t E X 2
2

E(3X  7) 3E( X )  7 3np  7 

t0

b X  Geo( p ) and Y  Geo(q ), where p  q 1


­ 1
°° E(3X  7 ) 3 u 27 u  7
3
20
Ÿ® 1 1
2
Ÿ E( X  Y ) E( X )  E(Y ) 
°E(3X  7 ) 3 u 27 u  7 47 p q
°¯ 3 q  p 1
pq pq
b Z a Po(m ) Ÿ ( Var ( Z ))2 E( Z )  12
Ÿ Var ( X  Y ) Var ( X )  Var (Y )
Ÿ m 2  m  12 0 Ÿ (m  4 )(m  3) 0
q p q 3  p3
m = 4 or m = −3 Variance is always positive.  2
p2 q p2q 2
Var(5 + 2Z) = 22 Var(Z ) = 4 × 4 = 16 
1 

(q  p ) (q 2  pq  p 2 )
p2q 2
5 a X  N (12 000, 3200 2 ) Ÿ P( X > 13 000) = 0.377
2

§ 2 q  p  3 pq 1  3 pq
§ 3200 · ·
b X  N ¨ 12 000, ¨
¨ ¸ ¸¸ Ÿ P( X  11500) 0.196 p2q 2 p2q 2
© © 30 ¹ ¹
E( X  Y )(E( X  Y )  3)
1.2 1.3 1.4 CH2 Rewiew ...ise
normCdf(13000,9.E999,12000,3200) 1 § 1 · 1 3
0.37733 ¨  3¸ 2
2

3200 pq © pq ¹ p q pq
(
normCdf –9.E999,11500,12000, √30 )
1  3 pq
0.19605 Var( X  Y ) Q.E.D.
p2q 2

2/99

© Oxford University Press 2015: this may be reproduced for class use solely for the purchaser’s institute Statistics: Chapter 2 19
WORKED SOLUTIONS

Exploring statistical
3 analysis methods
Skills check 3 a Y ~ Po (0.4) Ÿ P (Y 0) = 0.670
b Y ~ Po (7) Ÿ P (3 d X < 8)
1 1.1
A
1.2
B
CH3 Skils check
C D = P ( X d 7)  P ( X d 2) = 0.569
=OneVar(a 1.2 1.3 1.4 CH3 Skils check
1 1 22 Title One–Var... poissPdf(0.4,0) 0.67032
2 3 37 x 3.61818
poissCdf(3,5) 0.916082
3 5 46 Σx 398.
4 7 5 Σx2 1750. 1–poissCdf(4.7,4) 0.505391
5 sx : = sn–... 1.68633 poissCdf(7,7)–poissCdf(7,2) 0.569078
D2 =3.6181818181818

1.1 1.2 CH3 Skils check


B C D E
4/99
=OneVar(a
5 sx : = sn–... 1.68633
6 σx : = σnx... 1.67865 2.81785 Exercise 3A
7 n 110.
8 MinX 1. 1 a {2, 4, 6, 8, 10, 12, 14, 16, 18, 20}Ÿ x 11, s 2 33
9 Q1X 3. b {21, 24, 36, 28, 30, 22, 25, 26, 38, 32, 34, 29, 37,
E6 =d62 33, 32, 34, 29, 37, 33, 31, 30}Ÿ x 29.75, s 2 25.3
x 3.62, V 2 1.678652 2.82 c {1, 4, 7, 10, !, 133} Ÿ x 67, s 2 1518

§ 1· 1.1 1.2 1.3 CH3 Exercise A


2 a X ~ B ¨ 5, ¸ Ÿ P( X 5) mean(seq(2·x,x,1,10)) 11
2 © ¹
varPop(seq(2·x,x,1,10)) 33
0 5
§5· §1· §1· 1
=¨ ¸ ¨ ¸ ¨ ¸ 0.03125 mean({21,24,36,28,30,22,25,26,38,32,34,29
©5¹ ©2¹ ©2¹ 32 119
4
§ 1· varPop({21,24,36,28,30,22,25,26,38,32,34,29
b X ~ B ¨ 10, ¸ Ÿ P(3 ≤ X < 8) 405
5 © ¹ 16
10 − k k
7
§ 10 · § 4 · §1·
= P(3 ≤ X ≤ 7) = ¦ ¨ ¸¨ ¸ ¨ ¸
3/6
k =3 © k ¹©5¹ ©5¹
0.322 1.1 1.2 1.3 CH3 Exercise A
119
1.1 1.2 1.3 CH3 Skils check 4
1 varPop({21,24,36,28,30,22,25,26,38,32,34,2
binomPdf 5, ,5
2 ( ) 0.03125
405
1 16
0.980338
binomCdf 8, ,5
3 ( ) mean(seq(x,x,1,133,3)) 67
5 2144 varPop(seq(x,x,1,133,3)) 1518

k=0
( nCr(8,k).
8–k
( 23 ) .( 13 )
k
) 2187
|
6/99

3/99
2 x 0.65, s 0.864
1.1

k=0
(
1.2 1.3
nCr(8,k).
CH3 Skils check
( 3 ) .( 3 ) ) 2187
1.1
B
1.2
C
CH3 Exercise A
E D
=OneVar(a
1 22 Title One–Var...
3 0.640537
1–binomCdf 12, ,4
7 ( ) 2 12 x 0.65
3 4 Σx 26.
1 1 4 2 Σx2 46.
( )
binomCdf 10, ,7 –binomCdf 10, ,2
5 5 ( ) 5 sx : = sn–... 0.863802
0.322123 E5 =0.86380197160799
5/99

© Oxford University Press 2015: this may be reproduced for class use solely for the purchaser’s institute Statistics: Chapter 3 20
WORKED SOLUTIONS

3 a x 33.7 b 99% CI [0.104, 0.167]


b s = 23.8 CH3 Exercise B
1.1 1.2

1.1 1.2 1.3 CH3 Exercise A zInterval 0.04,{0.1,0.12,0.15,0.18,0.13,0.12


B C E D "Title" "z Interval"
=OneVar(a "CLower" 0.104389
"CUpper" 0.16652
1 6 Title One–Var...
"x" 0.135455
2 13 x 33.7143 "ME" 0.031066
3 26 Σx 2360. "sx := sn–1x" 0.035317
4 17 Σx2 118550. "n" 11.
5 8 sx : = sn–... 23.7695 "σ" 0.04
E5 =23.769510891752 3/99

c 95% CI [320, 325]


Exercise 3B
1.1 1.2 CH3 Exercise B

1 a 99% CI [4.19, 5.81] zInterval 4.5,{321,325,330,324,325,326,317


"Title" "z Interval"
1.1 CH3 Exercise B "CLower" 320.5
zInterval 1,5,10,0.99: stat.results "CUpper" 325.214
"Title" "z Interval" "x" 322.857
"CLower" 4.18545 "ME" 2.3572
"CUpper" 5.81455 "sx := sn–1x" 5.5727
"x" 5. "n" 14.
"ME" 0.814549 "σ" 4.5
"n" 10. 3/99
"σ" 1.

1/99
3 x  P d 2 Ÿ P  x d 2 Ÿ 2 d P  x d 2
Ÿx 2dPdx 2
b 95% CI [−12.8, − 9.20] 5 .5
V 5.5 Ÿ 1.960 u 2
1.1 CH3 Exercise B n
zInterval 4.3,–11,22,0.95: stat.results
"Title" "z Interval" Ÿ n 5.39 Ÿ n 29.052
"CLower" –12.7968 We should take at least 30 elements.
"CUpper" –9.20318
"x" –11.
"ME" 1.79682 4 95% CI for the weight if elementines is [72.3 g, 77.2 g]
"n" 22.
1.1 1.2 1.3 CH3 Exercise B
"σ" 4.3
zInterval 3.5,{70,75,77,71,68,80,85,72},1,0
"Title" "z Interval"
2/99 "CLower" 72.3247
"CUpper" 77.1753
c 90% CI [2819, 2889] "x" 74.75
"ME" 2.42533
1.1 CH3 Exercise B "sx := sn–1x" 5.70088
zInterval 327,2854,230,0.9: stat.results "n" 8.
"Title" "z Interval" "σ" 3.5
"CLower" 2818.53 3/99
"CUpper" 2889.47
"x" 2854.
14.2  17.4
"ME" 35.4659 5 a x 15.8
"n" 230. 2
"σ" 327.
3
| b V 3 Ÿ 1.645 u 15.8  14.2
3/99 n

2 a 90% CI [3.90, 6.10] Ÿ n 3.08 Ÿ n 9.51


The sample size should be 10.
1.1 1.2 CH3 Exercise B
zInterval 2,{1,2,3,4,5,6,7,8,9},1,0.9: stat.res
"Title" "z Interval"
"CLower" 3.90343
"CUpper" 6.09657
"x" 5.
"ME" 1.09657
"sx := sn–1x" 2.73861
"n" 9.
"σ" 2.
3/99

© Oxford University Press 2015: this may be reproduced for class use solely for the purchaser’s institute Statistics: Chapter 3 21
WORKED SOLUTIONS

Investigation Exercise 3C
Let’s consider samples of different sizes: all have the
1 a 90% CI [14.45, 15.55]
mean value x 100 and they come from a population
1.1 CH3 Exercise C
with V 2 64. tInterval 15,1.2,15,0.9: stat.results
"Title" "t Interval"
a i n 10 Ÿ 95% CI [95.0, 105.0] "CLower" 14.4543
1.2 1.3 1.4 CH3 Exercise B "CUpper" 15.5457
zInterval 8,100,10,0.95: stat.results "x" 15.
"Title" "z Interval" "ME" 0.545722
"CLower" 95.0416 "df" 14.
"CUpper" 104.958 "sx := sn–1x" 1.2
"x" 100. "n" 15.
"ME" 4.95836 1/99
"n" 10.
"σ" 8. b 99% CI [−25.81, −20.19]
|
1/99 1.1 CH3 Exercise C
tInterval –23,5.8,32,0.99: stat.results
"Title" "t Interval"
ii n 25 Ÿ 95% CI [96.9, 103.1] "CLower" –25.8135
"CUpper" –20.1865
1.2 1.3 1.4 CH3 Exercise B "x" –23.
zInterval 8,100,25,0.95: stat.results "ME" 2.81348
"Title" "z Interval" "df" 31.
"CLower" 96.8641 "sx := sn–1x" 5.8
"CUpper" 103.136 "n" 32.
"x" 100. 2/99
"ME" 3.13594
"n" 25.
"σ" 8. c 95% CI [3430, 3526]
|
1.1 CH3 Exercise C
2/99 tInterval 3478,429,310,0.95: stat.results
"Title" "t Interval"
iii n = 50 Ÿ 95% CI [97.8, 102.2] "CLower" 3430.06
"CUpper" 3525.94
1.2 1.3 1.4 CH3 Exercise B "x" 3478.
zInterval 8,100,50,0.95: stat.results "ME" 47.9434
"Title" "z Interval" "df" 309.
"CLower" 97.7826 "sx := sn–1x" 429.
"CUpper" 102.217 "n" 310.
"x" 100. 3/99
"ME" 2.21745
"n" 50. 2 a 99% CI [1.94, 8.06]
"σ" 8.
| 1.1 1.2 CH3 Exercise C
3/99 tInterval {1,2,3,4,5,6,7,8,9},1,0.99: stat.results
"Title" "t Interval"
iv n 150 Ÿ 95% CI [98.7, 101.3] "CLower" 1.93696
"CUpper" 8.06304
1.2 1.3 1.4 CH3 Exercise B
"x" 5.
zInterval 8,100,150,0.95: stat.results "ME" 3.06304
"Title" "z Interval" "df" 8.
"CLower" 98.7198 "sx := sn–1x" 2.73861
"CUpper" 101.28 "n" 9.
"x" 100. 1/99
"ME" 1.28024
"n" 150. b 95% CI [0.112, 0.159]
"σ" 8.
1.1 1.2 CH3 Exercise C
|
tInterval {0.1,0.12,0.15,0.18,0.13,0.12,0.09
4/99
"Title" "t Interval"
"CLower" 0.111728
b At the same significance level the larger sample "CUpper" 0.159181
"x" 0.135455
size we take the narrower a confidence interval we "ME" 0.023726
ª V V º "df" 10.
get. In the formula « x  u zD , x  u z D » the "sx := sn–1x" 0.035317
¬ n 2 n 2 ¼ "n" 11.
length of the interval is symmetrical with respect 2/99
to the mean of the sample by the same term that
is obtained by division by n therefore the larger n
we take the smaller number we obtain.

© Oxford University Press 2015: this may be reproduced for class use solely for the purchaser’s institute Statistics: Chapter 3 22
WORKED SOLUTIONS

c 90% CI [319.6, 324.9] c For the same set of data, a higher significance
1.1 1.2 CH3 Exercise C level will mean a wider confidence interval.
tInterval {321,325,330,324,325,326,317,318 The result is expected since the confidence
"Title" "t Interval"
"CLower" 319.612 interval is symmetrical about the mean value
"CUpper" 324.921 s
"x" 322.267 x r u tc
"ME" 2.65433
n
"df" 14.
"sx := sn–1x" 5.83667 Exercise 3D
"n" 15.
3/99 1 a 1.1 CH3 Exercise D
A B C D E
3 99% CI for the weight of mandarins is [67.6 g, 82.5 g] =a[]–b[]
1 15 18 –3
1.1 1.2 1.3 CH3 Exercise C
tInterval {66,70,75,84,90,77,71,68,80,85,72, 2 17 15 2
"Title" "t Interval" 3 23 23 0
"CLower" 67.6363 4 17 19 –2
"CUpper" 82.5304 5 18 15 3
"x" 75.0833 C1 =–3
"ME" 7.44705
"df" 11. CH3 Exercise D
1.1 1.2
"sx := sn–1x" 8.30617
"Title" "t Interval"
"n" 12.
"CLower" –2.18091
1/99 "CUpper" 1.98091
"x" –0.1
12.345  14.355 "ME" 2.08091
4 a x 13.35 "df" 9.
2
"sx := sn–1x" 2.02485
1 .5 "n" 10.
b s 1 .5 Ÿ t c u 13.35  12.345
8 |

Ÿ t c 1.895 2/99

Ÿ P( 1.895  t  1.895|v 7) = 0.900 99% CI [−2.18, 1.98]


1.2 1.3 1.4 CH3 Exercise C
b
(13.35–12.345)·√8 1.89505 1.1
C
1.2
D
CH3 Exercise D
E F
1.5
=a[]–b[] =d[]–e[]
tCdf(–1.895046173579,1.895046173579,7) 8 1 93 102 –9
0.900069 9 –2 102 108 –6
10 –1 106 109 –3
|
11 99 111 –12
12 85 103 –18
2/99 F13

The confidence level is 90%. 1.1 1.2 CH3 Exercise D

5 a 95% CI [483.4, 588.6] "Title" "t Interval"


"CLower" –13.7287
1.3 1.4 1.5 CH3 Exercise C "CUpper" –7.43794
tInterval 536,95,15,0.95: stat.results "x" –10.5833
"Title" "t Interval" "ME" 3.14539
"CLower" 483.391 "df" 11.
"CUpper" 588.609 "sx := sn–1x" 6.06717
"x" 536. "n" 12.
"ME" 52.6092 |
"df" 14. 5/99
"sx := sn–1x" 95.
"n" 15.
90% CI [−13.73, −7.44]
1/99
c 1.1 1.2 CH3 Exercise D
b 99% CI [463.0, 609.0] F G H I

1.3 1.4 1.5 CH3 Exercise C


=d[]–e[] =g[]–h[]
"Title" "t Interval" 4 –8 0.58 0.69 –0.11
"CLower" 462.981 5 –14 0.82 0.78 0.04
"CUpper" 609.019 6 –7 0.77 0.65 0.12
"x" 536. 7 –13 0.9 0.8 0.1
"ME" 73.0187
8 –9 1.02 0.95 0.07
"df" 14.
I8 =0.07
"sx := sn–1x" 95.
"n" 15.
|
2/99

© Oxford University Press 2015: this may be reproduced for class use solely for the purchaser’s institute Statistics: Chapter 3 23
WORKED SOLUTIONS

1.1 1.2 CH3 Exercise D b H0 : P P0 , H1 : P z P0, D 0.05


"Title" "t Interval"
1.1 CH3 Exercise E
"CLower" –0.060895
zTest 2,0.3,1.9,25,0: stat.results
"CUpper" 0.085895
"Title" "z Test"
"x" 0.0125
"Alternate Hyp" "μ ≠ μ0"
"ME" 0.073395
"z" –1.66667
"df" 7.
"PVal" 0.095581
"sx := sn–1x" 0.08779
"x" 1.9
"n" 8.
"n" 25.
"σ" 0.3
6/99
|
95% CI [−0.069, 0.859] 2/99

2 a Since the p-value is 0.095581 > 0.05 we have no


Blood sufficient evidence to reject the null hypothesis
1 2 3 4 5 6 7 8 9 10 11 12
sample at the 5% significance level.
Bob 144 153 170 183 125 95 148 177 160 155 170 135 c H0 : P P0 , H1 : P z P0 , D 0.01
Rick 141 161 173 174 119 104 135 175 164 158 167 142 CH3 Exercise E
1.1
Bob– zTest –235,12.8,–238,119,0: stat.results
3 −8 −3 9 6 −9 13 2 −4 −3 3 −7 "Title" "z Test"
Rick
"Alternate Hyp" "μ ≠ μ0"
"z" –2.55673
b 95% CI [−4.27, 4.60] "PVal" 0.010566
1.1 1.2 1.3 CH3 Exercise D "x" –238.
A B C D E "n" 119.
=a[]–b[] "σ" 12.8
8 177 175 2
9 160 164 –4 3/99
10 155 158 –3
11 170 167 3 Since the p-value is 0.010566 > 0.01 we have
12 135 142 –7 no sufficient evidence to reject the null
C12 =–7
hypothesis at the 1% significance level.
1.1 1.2 1.3 CH3 Exercise D
C D E F 2 a H0 : P P0 , H1 : P  P0, D 0 .1
=a[]–b[] =tInterval( 1.1 CH3 Exercise E
1 3 Title t Interval zTest 10,4,9,20,–1: stat.results
2 –8 CLower –4.26715 "Title" "z Test"
3 –3 CUpper 4.60048 "Alternate Hyp" "μ < μ0"
4 9 x 0.166667 "z" –1.11803
"PVal" 0.131776
5 6 ME 4.43381
"x" 9.
E1 ="t Interval" "n" 20.
"σ" 4.
|
Exercise 3E 4/99

1 a H 0 : μ = μ0 , H1 : μ ≠ μ0, D 0 .1 Since the p-value is 0.131776 > 0.1 we have no


1.1 CH3 Exercise E sufficient evidence to reject the null hypothesis
zTest 10,2,12,20,0: stat.results
"Title" "z Test" at the 10% significance level.
"Alternate Hyp" "μ ≠ μ0" b H0 : P P0 , H1 : P  P0, D 0.01
"z" 4.47214
"PVal" 0.000008 1.1 CH3 Exercise E
"x" 12. zTest 21.4,0.75,21.2,50,–1: stat.results
"n" 20. "Title" "z Test"
"σ" 2. "Alternate Hyp" "μ < μ0"
"z" –1.88562
1/99 "PVal" 0.029673
"x" 21.2
Since the p-value is 0.000008 < 0.1 we "n" 50.
"σ" 0.75
reject the null hypothesis at the 10%
significance level. 5/99

© Oxford University Press 2015: this may be reproduced for class use solely for the purchaser’s institute Statistics: Chapter 3 24
WORKED SOLUTIONS

Since the p-value is 0.029673 > 0.01 we have no 4 a H0: “The mean weight is 26 g.” ( P 26)
sufficient evidence to reject the null hypothesis H1: “The mean weight is not 26 g.” ( P z 26)
at the 1% significance level.
b Use the z-test.
c H0 : P P0 , H1 : P  P0, D 0.01
1.1 1.2 CH3 Exercise E
1.1 1.2 1.3 CH3 Exercise E A worms B C D E
zTest –235,12.8,–238,119,–1: stat.results =zTest(26,
"Title" "z Test" 1 22 Title z Test
"Alternate Hyp" "μ < μ0"
2 25 Alternate... μ ≠ μ0
"z" –2.55673
"PVal" 0.005283 3 31 z 4.42627
"x" –238. 4 35 PVal 0.00001
"n" 119. 5 28 x 30.
"σ" 12.8 C4 =9.5958692475566E-6

9/99 Since the p-value is 0.00001 < 0.01 we reject


the null hypothesis at the 1% significance level
Since the p-value is 0.005283 < 0.01 we reject the and conclude that the harvested snails are not
null hypothesis at the 1% significance level. from the population.
5 a H0:“The mean level of fat in the drink is 1.4 g.”
3 a H0 : P P0 , H1 : P ! P0, D 0.05
( P 1 .4 )
1.1 CH3 Exercise E
zTest 10,5,12,20,1: stat.results H1:“The mean level of fat in the drink is more
"Title" "z Test" than 1.4 g.” ( P ! 1.4 )
"Alternate Hyp" "μ > μ0"
"z" 1.78885 b Use the z-test.
"PVal" 0.036819
1.1 1.2 1.3 CH3 Exercise E
"x" 12.
A fat B C D
"n" 20.
"σ" 5. =zTest(1.4
1 1.43 Title z Test
7/99 2 1.52 Alternate Hyp μ > μ0
3 1.35 z 0.424264
Since the p-value is 0.036819 < 0.05 we reject the 4 1.38 PVal 0.335687
null hypothesis at the 5% significance level. 5 1.42 x 1.445
C =zTest(1.4,0.3,a [ ], 1,1): CopyVar
b H0 : P P0 , H1 : P ! P0, D 0 .1
Since the p-value is 0.335687 > 0.05 we have
1.1 CH3 Exercise E
zTest 27.3,3.6,28.,40,1: stat.results
no sufficient evidence to reject the null
"Title" "z Test" hypothesis at the 5% significance level and
"Alternate Hyp" "μ > μ0" conclude that the company’s claim is correct.
"z" 1.22977
"PVal" 0.109391 6 a H0: “The mean volume of juice in the bottle is
"x" 28.
"n" 40.
300 ml.” (μ = 300)
"σ" 3.6 H1 : “The mean volume of juice in the bottle is
less than 300 ml.”(μ < 300)
8/99
b Use the z-test.
Since the p-value is 0.109391 > 0.1 we have no 1.2 1.3 1.4 CH3 Exercise E
A volume B C D
sufficient evidence to reject the null hypothesis
=zTest(30
at the 10% significance level.
1 295 Title z Test
c H0 : P P0 , H1 : P ! P0, D 0.01 2 288 Alternate Hyp... μ < μ0
1.1 CH3 Exercise E 3 293 z –2.60364
zTest –73,3.72,–71.6,92,1: stat.results 4 301 PVal 0.004612
"Title" "z Test" 5 302 x 295.75
"Alternate Hyp" "μ > μ0" C =zTest(300,7.3,a [ ], 1,–1): CopyVar
"z" 3.60977
"PVal" 0.000153 Since the p-value is 0.004612 < 0.1 we reject the
"x" –71.6
"n" 92. null hypothesis at the 10% significance level
"σ" 3.72 and conclude that the bottles contain less
|
volume than stated.
9/99

Since the p-value is 0.000153 < 0.01 we reject


the null hypothesis at the 1% significance level.

© Oxford University Press 2015: this may be reproduced for class use solely for the purchaser’s institute Statistics: Chapter 3 25
WORKED SOLUTIONS

Exercise 3F b H0 : μ = μ0, H1 : μ < μ0, α = 0.01

1 a H0 : μ =μ0 , H1 :μ ≠ μ0, α = 0.05 1.1 CH3 Exercise F


"Title" "t Test"
1.1 CH3 Exercise F "Alternate Hyp" "μ < μ0"
"Title" "t Test" "t" –2.99871
"Alternate Hyp" "μ ≠ μ0" "PVal" 0.007494
"t" –0.486504 "df" 9.
"PVal" 0.638237 "x" 119.8
"df" 9. "sx := sn–1x" 2.32
"x" 4.8 "n" 10.
"sx := sn–1x" 1.3
"n" 10. 5/99
|
1/99 Since the p-value is 0.007494 < 0.01 we reject
Since the p-value 0.6382 > 0.05 we have no the null hypothesis at the 1% significance level.
sufficient evidence to reject the null hypothesis c H0 : μ = μ0, H1 : μ < μ0, α = 0.1
at the 5% significance level. 1.1 CH3 Exercise F

b H0 : μ = μ0, H1 : μ ≠ μ0, α = 0.1 "Title" "t Test"


"Alternate Hyp" "μ < μ0"
1.1 CH3 Exercise F "t" –0.816497
"Title" "t Test" "PVal" 0.225675
"Alternate Hyp" "μ ≠ μ0" "df" 5.
"t" –1.66667 "x" 622.8
"PVal" 0.10858 "sx := sn–1x" 12.6
"df" 24. "n" 6.
"x" 1.9 |
"sx := sn–1x" 0.3 6/99
"n" 25.
| Since the p-value 0.225675 > 0.1 we have no
2/99
sufficient evidence to reject the null hypothesis
Since the p-value 0.10858 > 0.1 we have no at the 10% significance level.
sufficient evidence to reject the null hypothesis 3 a H0 : μ = μ0, H1 : μ > μ0, α = 0.1
at the 10% significance level.
1.1 CH3 Exercise F
c H0 : μ = μ0, H1 : μ ≠ μ0, α = 0.01 "Title" "t Test"
"Alternate Hyp" "μ > μ0"
1.1 CH3 Exercise F "t" –0.667483
"Title" "t Test" "PVal" 0.743755
"Alternate Hyp" "μ ≠ μ0" "df" 19.
"t" 3.48125 "x" 0.95
"PVal" 0.013122 "sx := sn–1x" 0.335
"df" 6. "n" 20.
"x" –35.3
"sx := sn–1x" 0.532 7/99
"n" 7.
Since the p-value 0.743755 > 0.1 we have no
3/99
sufficient evidence to reject the null hypothesis
Since the p-value 0.013122 > 0.01 we have no at the 10% significance level.
sufficient evidence to reject the null hypothesis b H0 : μ = μ0, H1 : μ > μ0, α = 0.05
at the 1% significance level. 1.1 1.2 1.3 CH3 Exercise F

2 a H0 : μ = μ0, H1 : μ < μ0, α = 0.05 "Title" "t Test"


"Alternate Hyp" "μ > μ0"
1.1 CH3 Exercise F "t" 3.53553
"Title" "t Test" "PVal" 0.004763
"Alternate Hyp" "μ < μ0" "df" 7.
"t" –1.99172 "x" 26.4
"PVal" 0.027947 "sx := sn–1x" 1.12
"df" 29. "n" 8.
"x" 14.2
"sx := sn–1x" 2.2 8/99
"n" 30.
Since the p-value 0.004763 < 0.05 we
4/99
reject the null hypothesis at the 5%
Since the p-value is 0.027947 < 0.05 we significance level.
reject the null hypothesis at the 5%
significance level.

© Oxford University Press 2015: this may be reproduced for class use solely for the purchaser’s institute Statistics: Chapter 3 26
WORKED SOLUTIONS

c H0 : μ = μ0, H1 : μ > μ0, α = 0.01 Use the t-test on the difference of times on those
1.1 1.2 1.3 CH3 Exercise F two different cubes.
"Title" "t Test"
1.1 CH3 Exercise G
"Alternate Hyp" "μ > μ0"
A cube1 B cube2 C diff D
"t" 1.25463
"PVal" 0.115078 =a[]–b[]
"df" 14. 2 35 38 –3
"x" 758.6 3 41 40 1
"sx := sn–1x" 14.2 4 30 34 –4
"n" 15.
5 28 30 –2
6 46 44 2
9/99
C diff:=a[ ]–b[ ]

Since the p-value 0.115078 > 0.01 we have no


1.1 CH3 Exercise G
sufficient evidence to reject the null hypothesis ff D E F
at the 1% significance level. ]–b[] =tTest(0, c
4 H0 : “The mean volume is 120 ml.” (μ = 120) 2 –3 Alternate Hyp μ ≠ μ0
3 1 t 0.164399
H1 : “The mean volume is not 120 ml.” (μ ≠ 120) 4 –4 PVal 0.874063
5 –2 df 7.
1.1 1.2 CH3 Exercise F
A volume B C D 6 2 x 0.25
E =tTest(0, c [ ] ,1,0): CopyVar Stat.,
=tTest(120
1 119 Title t Test Since the p-value 0.874063 > 0.05 we have no
2 123 Alternate Hyp μ ≠ μ0
sufficient evidence to reject the null hypothesis at
3 121 t 0.283654
4 120 PVal 0.784882 the 5% significance level and conclude that there is
5 118 df 7. no difference in finishing times on the two Rubik’s
C =tTest(120,a [ ], 1,0): CopyVar Stat
Cubes.
Since the p-value 0.283654 > 0.01 we have no 2 H0 : “There is no difference in the scores.” (μd = 0)
sufficient evidence to reject the null hypothesis at H1 : “There is a difference in the scores.” (μd ≠ 0)
the 1% significance level and conclude that the
Use the t-test on the difference of scores on the
factory advertised a correct volume of a particular
two types of dart.
ice-cream product.
1.1 1.2 CH3 Exercise G
5 H0: “The mean life expectancy is 30 000 hours.” A old B new C dartdiff D
(μ = 30 000) =old–new
H1 : “The mean life expectancy is less than 30 000 1 85 90 –5
2 92 95
hours.” (μ < 30 000) –3
3 100 98 2
1.1 1.2 1.3 CH3 Exercise F 4 97 99 –2
A life B C D 5 89 93 –4
=tTest(300 C1 =–5
1 29500 Title t Test
2 28350 Alternate Hyp... μ < μ0 1.1 1.2 CH3 Exercise G
3 30300 t –1.51876 C dartdiff D E F
4 30250 PVal 0.094643 =old–new =tTest(0, c
5 29350 df 5. 1 –5 Title t Test
C =tTest(30000,a [ ], 1,–1): CopyVar S 2 –3 Alternate... μ ≠ μ0
3 2 t –2.13719
Since the p-value 0.094543 < 0.1 we reject the null 4 –2 PVal 0.085622
hypothesis at the 10% significance level and 5 –4 df 5.
E =tTest(0, c [ ] ,1,0): CopyVar Stat.,
conclude that the manufacturer claims a longer life
expectancy of the LED lamps. Since the p-value 0.085622 < 0.1 we have to reject
the null hypothesis at the 10% significance level
Exercise 3G and conclude that players score a better result by
1 H0 : “There is no difference in finishing times.” using the new type of dart.
(μd = 0) 3 H0 : “There is no difference in the weights.” (μd = 0)
H1 : “There is a difference in finishing times.” H1 : “Students who join the programme drop some
(μd ≠ 0) weight.” (μd > 0)

© Oxford University Press 2015: this may be reproduced for class use solely for the purchaser’s institute Statistics: Chapter 3 27
WORKED SOLUTIONS

Use the t-test on the difference of weights before 3 a H 0 : X  Po(45) Ÿ D 1  P(X d 52) = 0.133
and after the programme. b H1 : X  Po(40) Ÿ E P(X d 52) = 0.972
1.1 1.2 1.3 CH3 Exercise G
1.1 1.2 1.3 CH3 Exercise H
A before B after C wdiff D E
1–PoissCdf(45,0,52) 0.132809
=before-a
poissCdf(40,0,52) 0.971943
1 55 52 3
|
2 82 84 –2
3 63 61 2
4 69 69 0
5 65 62 3
C wdiff: =before–after
2/99

1.1 1.2 1.3 CH3 Exercise G


C wdiff D E F E
Review exercise
=before-a =tTest(0,w
1 2 3 Title t Test 1 Use z-interval since the standard deviation is known
2 4 –2 Alternate... μ > μ0 and the value is 114 g.
3 1 2t 1.23335
4 9 0 PVal 0.121576 a 95% CI [602, 702]
5 2 3 df 11. 1.1 CH3 Review...ise
C =tTest(0,wdiff,1,1): CopyVar Stat., zInterval 114,652,20,0.95: stat.results
"Title" "z Interval"
"CLower" 602.038
Since the p-value is 0.121576 > 0.05 we have no "CUpper" 701.962
sufficient evidence to reject the null hypothesis at "x" 652.
"ME" 49.9618
the 5% significance level and conclude that there is "n" 20.
no difference in weight before and after the "σ" 114.
programme. 1/99

Exercise 3H b 99% CI [586, 718]


1.1 CH3 Review...ise
1 a H 0 : X  N(5, 0.4 2 ) zInterval 114,652,20,0.99: stat.results
"Title" "z Interval"
Ÿ D 1  P(4.2 d X d 5.8) 0.0455 "CLower" 586.339
"CUpper" 717.661
2
b H1 : X  N(4.5, 0.4 ) Ÿ "x" 652.
"ME" 65.6609
E P(4.2 d X d 5.8) 0.773 "n" 20.
"σ" 114.
1.1 CH3 Exercise H |
1–normCdf(4.2,5.8,5,0.4) 0.0455 2/99

normCdf(4.2,5.8,4.5,0.4) 0.772796
c 95% < 99% ⇒ [602, 702] ⊂ [586, 718]
|
We notice that a higher significance level means
a wider confidence level.
2 a
Patient A B C D E F G H I J
2/99 Analyzer I 235 352 410 280 341 325 428 388 272 310
Analyzer II 237 343 416 272 336 329 413 396 265 315
§ 1· 1
2 a H 0 : X  B ¨ 50, ¸ Ÿ E(X ) = 50 × = 25 Difference −2 9 −6 8 5 −4 15 −8 7 −5
© 2¹ 2
D 1  P( X d 35) 0.00130 b H0 : “There is no difference in potassium levels.”
(μd = 0)
§ 4·
b H1 : X  B ¨ 50, ¸ŸE P(X d 35) = 0.978 H1 : “There is a difference in potassium levels.”
© 7¹
(μd ≠ 0)
1.1 1.2 CH3 Exercise H
1
(
1–binomCdf 50, 2 ,0,35 ) 0.001301 1.1
A a1
1.2 CH3 Review...ise
B a2 C dl
='a1–'a2
D E

4
(
binomCdf 50, 7 ,0,35 ) 0.977867 1
2
235
352
237
343
–2
9
3 410 416 –6
4 280 272 8
5 341 336 5
2/99 C1 =–2

© Oxford University Press 2015: this may be reproduced for class use solely for the purchaser’s institute Statistics: Chapter 3 28
WORKED SOLUTIONS
10
1.1 1.2 CH3 Review...ise
C dl D F F ¦x
i 1
i
='a1–'a2 =tTest(0,d b x 30.97
1 –2 Title t Test 10
2 9 Alternate... μ ≠ μ0 10 2
§ 3097 ·
3
4
–6 t
8 PVal
0.76657
0.46297
¦ x i2  10 u ¨
© 100 ¹
¸ 2681
s2 i 1
0.298
5 5 df 9. 9 9000
E =tTest(0,dl[ ],1,0): CopyVar Stat.,
1.1 1.2 1.3 CH3 Review...ise
Since the p-value 0.46297 > 0.01 we have no { 31.2,30.8,30.4,30.8,31.3,32.1,30.3,31.4,3
sufficient evidence to reject the null hypothesis 30.97
varSamp({ 31.2,30.8,30.4,30.8,31.3,32.1,30
and we conclude that there is no difference in 0.297889
measurement of the two types of biochemical
analyzers.
15

¦x
i 1
i
80 16 2/99
3 a x 5.33
15 15 3
15
§ 16 ·
2
1280 c H0 : “The average time is 30 s.” (μ = 30)
¦x i
2
 15 u ¨ ¸
© 3 ¹
488 
3 H1 : “The average time is more than 30 s.”
s2 i 1

14 14 (μ > 30)
184 92 We use the t-test since the standard deviation is
4.38
42 21 unknown.
92 1.2 1.3 1.4 CH3 Review...ise
b s = s2 = = 2.09 tTest 30,{31.2,30.8,30.4,30.8,31.3,32.1,30.
21 "Title" "t Test"
1.1 1.2 1.3 CH3 Review...ise "Alternate Hyp" "μ > μ0"
"t" 5.62011
tInterval 16 , 92 ,15,0.99: stat.results
√ 21
3 "PVal" 0.000163
"Title" "t Interval" "df" 9.
"CLower" 3.72456 "x" 30.97
"CUpper" 6.94211 "sx := sn–1x" 0.545792
"x" 5.33333 "n" 10.
"ME" 1.60877 1/3
"df" 14.
"sx := sn–1x" 2.09307 Since the p-value 0.000163 < 0.05 we reject the
1/99
null hypothesis and can conclude that the
99% CI [3.72, 6.94] average time is more than 30 seconds.
c In 99% of the cases the mean value of a sample Therefore, the company sets up the
of 15 observations taken from the population speedometers to show a higher speed.
will fall within the confidence interval
[3.72, 6.94]. 204  216
6 a x 210
2
47.2  55.2 12
4 a x 51.2 b X  N (210, 144) Ÿ σ = 12 Ÿ 1.960 ×
2 n
2 5
b V 25 Ÿ V 5 Ÿ zD u 216  210 Ÿ n 3.92 Ÿ n 15.37
2 6
The sample size should be 16.
4 6
55.2  51.2 Ÿ z D 1.3 1.4 1.5 CH3 Review...ise
2 4
204+326 265
D 2
1.960 Ÿ )(1.960) 0.975
2
The confidence level of the interval is 95%.
(
nSolve invNorm(0.975)·12 =216–210,n
√n )
15.3658

100 |
5 a s = 1000 m, v = 120 km/h = m / s,
3
s 1000 2/99
t = Ÿt = = 30 s
v 100
3

© Oxford University Press 2015: this may be reproduced for class use solely for the purchaser’s institute Statistics: Chapter 3 29
WORKED SOLUTIONS

7 a For the speed values we used the midpoints of 8 H0 = μ = 10, H1 : μ < 10, using the mean of a
the intervals. sample of size 5.
§ 2·
a H 0 : μ = 10 Ÿ X  N ¨ 10, ¸
1.4 1.5 1.6 CH3 Review...ise
A B C D E
© 5¹
=OneVar(a
1 5 9 Title One–Var... x −10
i 10% for N (0, 1) is –1.282 so
2 15 56 x 23.4667 2
3 25 47 Σx 3520
5
4 35 25 Σx2 99150
5 45 13 sx : = sn–... 10.5383 2
D =OneVar(a[ ],b[ ]): CopyVar Stat  1.282 Ÿ x 10  1.282 u 9.19
5
1.5 1.6 1.7 CH3 Review...ise The critical region is the interval ]−∞, 9.19[
Av B f C D E
=OneVar(v 1.6 1.7 1.8 CH3 Review...ise
1 5 9 Title One–Var...
2
3
15
25
56 x
47 Σx
23.4667
3520
( (
nSolve normCdf –9.E999,x,10, √ 25 ) =0.1,x)
9.18948
4 35 25 Σx2 99150
5 45 13 sx : = sn–... 10.5383
D =OneVar(v,f): CopyVar Stat., Stat6
( (
nSolve normCdf –9.E999,x,10, √ 25 ) =0.05,x)
8.9597
i =x 23.5
ii = s = 10.5 2/99

b i = 95% CI [21.8, 25.2] x −10


ii 5% for N (0, 1) is –1.645 so
1.5 1.6 1.7 CH3 Review...ise 2
tInterval ν,ƒ,0.95: stat.results
"Title" "t Interval"
5
"CLower" 21.7664 2
"CUpper" 25.1669  1.645 Ÿ x 10  1.645 u 8.96
"x" 23.4667 5
"ME" 1.70026
"df" 149. The critical region is the interval ]−∞, 8.96[
"sx := sn–1x" 10.5383
§ 2·
"n" 150. b H1 : μ = 9.3 Ÿ X  N ¨ 9.3, ¸
1/99 5 © ¹
i E = P (X > 9.19) = 0.569
ii 90% CI [22.0, 24.9]
1.5 1.6 1.7 CH3 Review...ise ii E = P(X > 8.96) = 0.705
tInterval ν,f ,0.9: stat.results
"Title" "t Interval" 1.6 1.7 1.8 CH3 Review...ise
"CLower" 22.0425 8.9597
"CUpper" 24.8908
"x"
"ME"
23.4667
1.42417
(
normCdf 9.1894753867941,9.E999,9.3, √ 25 )
0.569364
"df" 149.
"sx := sn–1x"
"n"
10.5383
150.
(
normCdf 8.9597033202516,9.E999,9.3, √ 25 )
0.704731
2/99
|
c We notice that [22.0, 24.9] ⊂ [21.8, 25.2], so a 4/99

90% confidence interval is a subset of a 95%


c When the probability of a Type I error
confidence interval.
decreases from 10% to 5%, the probability of a
Type II error increase from 0.569 to 0.705.

© Oxford University Press 2015: this may be reproduced for class use solely for the purchaser’s institute Statistics: Chapter 3 30
WORKED SOLUTIONS

4 Statistical modeling
Skills Check A xcoord B ycoord C D E F G H
= =TwoVar(’ =LinRegM
1 a 2E(Z ) – 3E(Y ) + 2E(X) 1 4 1 Title Two–va... Title Linear R...
=Linear R..
2 1 1 –x 2.8 RegEqn m*x+b
b 4Var(Z ) – 9Var(Y ) + 4Var(X) 3 3 2 Σx 28. m –0.3409..
4 2 4 Σx2 96. b 2.95455
c E(X )E(Y )E(Z) 5 1 3 sx := sn-... 1.39841 r2 0.146104
6 2 4 σx := σn... 1.32665 r –0.3822..
2 a x = 43.5, y = –31.75, 7 5 2 n 10. Resid {–0.5909..
8 4 1 –y 2.
Var(X ) = 46.25, Var(Y ) = 98.69 9 4 1 Σy 20.
10 2 1 Σy2 54.
b d B ycoord C D E F G H
Sy := Sn-...
11 1.24722
= =TwoVar(’
12 σy := σn... 1.18322
1 33 –21 Title Two–va...
13 Σxy 50.
2 49 –44 –x 43.5
3 50 –23 Σx 174. H1 =“Linear Regression (mx+b)”
4 42 –39 Σx2 7754.
5 sx := sn-... 7.85281 Since r = −0.382, there is a weak negative
σx := σn...
6 6.80074
correlation between the two random variables.
7 n 4.
8 –y –31.75
2 1.3 2.1 2.2 Exercise 4A
9 Σy –127.
52.9 y
10 Σy 2 4427.
11 Sy := Sn-... 11.471
12 σy := σn... 9.93416
13 Σxy –5637.

E1 =“Two–Variable Statistics”

3 a 0.382 b 0.951
c 0.938 d 0.732 x
81.6 158.4

1.1 1.2 1.3 CH4 Skills check


A xcoord B ycoord C D E F G H

tCdf(–0.4,0.8,2) 0.382266 = =TwoVar(’ =LinRegM


1 100 33 Title Two–Va... Title Linear R...
tCdf(–∞,1.83,10) 0.951412 2 115 38 –x 126.9 RegEqn m*x+b
3 140 40 Σx 1269. m 0.245495
tCdf(–1.75,∞,7) 0.938203
4 149 51 Σx2 165059. b 9.44674
tCdf(–1.14,1.14,20) 0.732245 5 88 36 sx := sn-... 21.1421 r2 0.630724
6 132 40 σx := σn... 20.0572 r 0.794182
|
7 152 51 n 10. Resid {–0.9961..
8 144 43 –y 40.6
9 121 32 Σy 406.
10 128 42 Σy2 16868
Exercise 4A 11 Sy := Sn-... 6.53537
12 σy := σn... 6.2
1 1.2 1.3 2.1 Exercise 4A
13 Σxy 52509.
0.5 y
B11

Since r = 0.794, there is a strong positive


correlation between the two random variables.

3 1.1 1.2 Exercise 4A


68 y

–0.3 0.2 6.06 x

x
29.4 96.6

© Oxford University Press 2015: this may be reproduced for class use solely for the purchaser’s institute Statistics: Chapter 4 31
WORKED SOLUTIONS

A xcoord B ycoord C D E F G H 7 Var(X + Y) = E(X + Y)2 – [E(X + Y)]2


= =TwoVar(’ =LinRegM
1 55 58 Title Two–Va... Title Linear R... = E(X2 + 2XY + Y 2) – (μx μy)2
2 35 65 –x 66.7
3 66 52 Σx 667.
RegEqn
m
m*x+b
–0.5559...
= E(X 2 + 2XY + Y 2) − μx2 − 2 μx μy – P y2
4 82 35 Σx2 47645. b 86.1833
5 91 36 sx := sn-... 18.7264 r2 0.939034
= (E(X 2) − P y2 )+ (E(Y 2) − P y2 )
6 79 42 σx := σn... 17.7654 r –0.9690... + 2(E(XY ) − μx μy)
7 48 60 n 10. Resid {2.39513..
8 52 55 –y 49.1 = Var(X) + Var(Y) + 2Cov(X, Y)
9 71 50 Σy 491.
10 88 38 Σy2 25147.
Sy := Sn-...
8 Since X and Y are independent variables,
11 10.744
12
13
σy := σn... 10.1926 E(XY ) = E(X )E(Y ) = μx μy.
Σxy 30995.
A11 Hence, Cov ( X , Y ) E( XY )  P x P y Px P y  Px P y 0.
Hence, p = 0.
Since r = −0.970, there is a strong negative
correlation between the two random variables.
Exercise 4C
Exercise 4B 1 A xcoord B ycoord C D E F G
1 Cov(X, Y) = E(XY ) – μx μy = =LinRegtT
1 42 34 Title Linear R...
= E(XY ) – μy μx 2 2 18 Alternate... ß & ρ ≠ ...
= Cov(Y, X ) 3 26 14 RegEqn a+b*x
4 22 20 t 4.05922
2 Cov(X, X) = E(XX ) – μx μy 5 15 41 Pval 0.002289
6 44 44 df 10.
= E(X 2) − P x2 7 50 51 a 9.76713
= Var(X ) 8 38 45 b 0.73773
9 12 17 s 9.69498
3 Cov(aX, Y) = E(aXY ) – E(aX )E(Y ) 10 28 27 SESlope... 0.181742
11 22 28 r2 0.622318
= aE(XY ) – aE(X )E(Y ) 12 1 1 r 0.788871
13
= a[E(XY ) – (μx μy)] 13 Resid {–6.7517...
D1 =“Linear Reg t Test”
= aCov(X, Y)
H0 : p = 0; H1 : p ≠ 0
4 Cov(X, bY) = E(X(bY )) – E(X )E(bY )
Since 0.00229 < 0.01, we reject the null hypothesis.
= bE(XY ) – E(X )E(bY )
There is evidence of significant correlation between
= bE(XY ) – bE(X )E(Y ) the two variables at the 1% level.
= b[E(XY ) – ( μx μy)] 2 A xcoord B ycoord C D E F G
= bCov(X, Y) = =LinRegtT
1 64 69 Title Linear R...
5 Cov( X1  X 2 , Y ) E > ( X1  X 2 )Y @  E( X1  X 2 )E(Y ) 2 66 64 Alternate... ß & ρ ≠ ...
3 68 67 RegEqn a+b*x
= E > X1Y  X 2Y @ 4 69 64 t 0.931872
5 70 62 Pval 0.394177
 > E( X1 )  E( X 2 )@ E(Y ) 6 72 73 df 5.
7 74 71 a 35.6
= > E( X1Y )  E( X1 )E(Y )@ 8 b 0.457143

 > E( X 2Y )  E( X 2 )E(Y )@ 9 s 4.10435


10 SESlope... 0.490564
= C ov( X1 , Y )  Cov( X 2 , Y ) 11 r2 0.147977
12 r 0.384678
13
6 Cov(X, Y1, + Y2) = E[(X)(Y1 + Y2)] – E(X)E(Y1 + Y2) 13 Resid {4.14285...
E1 =“Linear Reg t Test”
= E[XY1 + XY2] – E(X)[E(Y1)
+ E(Y2)] H0 : p = 0; H1 : p = 0
= [E(XY1) – E(X)E(Y1)] + [E(XY2)
Since 0.394 > 0.05, there is not enough evidence to
– E(X)E(Y2)]
reject the null hypothesis. There is not evidence of
= Cov(X, Y1) + Cov(X, Y2) significant correlation between the two variables
at the 5% level.

© Oxford University Press 2015: this may be reproduced for class use solely for the purchaser’s institute Statistics: Chapter 4 32
WORKED SOLUTIONS

3 A xcoord B ycoord C D E F G 2 A body B heart C D E F G


= =LinRegtT = =LinRegM
1 67 104 Title Linear R... 1 30 136 Title Linear R...
2 70 116 Alternate... ß & ρ ≠ ... 2 37 156 RegEqn m*x+b
3 70 121 RegEqn a+b*x 3 38 150 m 1.90885
4 70 126 t 2.62304 4 32 140 b 82.763
5 71 117 Pval 0.034257 5 36 155 r2 0.57174
6 72 113 df 7. 6 32 157 r 0.756135
7 72 124 a –93. 7 33 143 Resid {–4.0286...
8 72 126 b 3. 8 38 160
9 73 127 s 5.78174 9 44 170
10 SESlope... 1.14371 10 38 144
11 r2 0.49569 11
12 r 0.704052 12
13 13
13 Resid {–4.,–0.9... 13
E1 =“Linear Reg t Test” F1

H0 : p = 0; H1 : p = 0 r = 0.756, which indicates a strong correlation between


Since 0.0343 < 0.1, we reject the null hypothesis. the two variables. Hence we find the line of regression,
There is evidence of significant correlation y = 1.91x + 82.8, and when x = 35 g, y = 150 mg.
between the two variables at the 10% level. 3 a H0 : p = 0; H1 : p = 0
b A x B y C D E F G
Exercise 4D = =LinRegM =LinRegT
1 120 80 Title Linear R... Title Linear R...
1 a A score B age C he... D We... E F G H I J 2 125 90 RegEqn m*x+b Alternate... ß & p ≠ ...
= =LinRegM =LinReg =LinRegM 3 130 92 m 0.62303 RegEqn a+b*x
1 62 10 145 44.9 Title... Linear R... Title... Linear .... Title... Linear R... 4 135 98 b 10.8182 t 10.0189
2 94 11 150 42.1 Reg... m*x+b Reg... m*x+b Reg... m*x+b 5 140 100 r2 0.926185 Pval 0.000008
3 81 8 139 30.2 m 0.014779 m 0.1642... m –0.2780... 6 145 103 r 0.962385 df 8.
4 62 9 148 41.4 b 8.48321 b 131.601 b 61.3778 7 150 105 Resid {–5.5818... a 10.8182
2 2 2
5 58 8 130 41.8 r 0.035188 r 0.0426... r 0.298465 8 155 108 b 0.62303
6 86 10 150 38.4 r 0.187584 r 0.2064 r –0.54632 9 160 110 s 2.82414
7 59 11 154 54.1 Res... {0.60049... Res... {3.216... Res... {0.75945... 10 165 110 SESlope... 0.062185
8 72 9 140 38.6 11 r2 0.926185
9 54 10 153 52.4 12 r 0.962385
13
10 60 9 120 38.6 13 Resid {–5.5818...
11 G1
12
13
13
r = 0.962; p = 8 × 10−6
E10

c There is very strong evidence to indicate a positive


Column F shows the regression data for score association between the random variables X and Y.
versus age. d Since r is close to 1, we can find the regression
Column H shows the regression data for score line, y = 0.623x + 10.8, and when x = 138 mm,
versus height. y = 97 mm.
Column J shows the regression data for score
§ 182 ·§ 200 ·
versus weight. 8390  5 ¨ ¸¨ ¸
|r| is largest for the random variables score and 4 m ¦ xy  nx y © 5 ¹© 5 ¹ 0.6
weight, hence we use W. ¦y 2
 ny 2
9850  5 ¨
§ 200 ·
2

¸
© 5 ¹
b From the GDC, S = −0.270W + 61.4
c From the GDC,
x x 0 .6 ( y  y ) Ÿ x 0.6 y  12.4 ; x = 66.4 g
A B C D E F G 5 For the gradient of the Y on X line of regression,
=
§ 30 · § 86 ·
580 − 10 ¨ ¸ ¨ ¸
1 Title... Linear R...
2 Alternate... ß & ρ ≠ ...
3 RegEqn a+b*x m= ¦ xy − nx y = © 10 ¹ © 10 ¹ = 2.48, hence
4 t –1.84488 ¦x 2
− nx 2 § 30 ·
2

5 PVal 0.102277 220 − 10 ¨ ¸


6 df 8. © 10 ¹
7 a 61.3778
8 b –0.2780... yy 2.48( x  x ) Ÿ y 2.48 x  1.17
9 s 6.1981
10 SESlope... 0.150699
For the gradient of the X on Y line of regression,
11 r2 0.298465
12 r -0.54632
13
13 Resid {0.75945.. § 30 · § 86 ·
580  10 ¨ ¸ ¨ ¸
C13
m ¦ xy  nx y © 10 ¹ © 10 ¹ 0.380 , hence
Since p = 0.102 > 0.05 there is insufficient ¦ y  ny 2 2
§ 86 ·
1588  10 ¨ ¸
2

evidence to reject the null hypothesis, i.e. there is © 10 ¹


not evidence of significant correlation between x x 0.380( y  y ) Ÿ x 0.380 y  29.6
the score and the weight at the 5% level.

© Oxford University Press 2015: this may be reproduced for class use solely for the purchaser’s institute Statistics: Chapter 4 33
WORKED SOLUTIONS

Review exercise 5 a A xcord B ycord C D E F G


= =LinRegT
1 a r is the unbiased estimate of ρ, and 1 44.5 41.2 Title Linear R...
2 10.3 11.1 Alternate... ß & ρ ≠ ...

r ¦x y i i  nx y
= 0.975 3 20.1 18.7 RegEqn a+b*x

¦ x i
2
 nx 2
¦ y i
2
 ny 2
4
5 39.6
55 52.3 t
41.2 PVal
20.9954
2.78035...
6 24.1 26.5 df 8.
n2 7 31.2 29.3 a 2.59473
t r = 10.7
1  r2 8 9.5 11.2 b 0.904534
9 22.3 25.1 s 1.9017
Hence, since p =.000039 < 0.05, there is evidence 10 35.1 33.2 SESlope... 0.043083
11 r2
of a strong positive relationship between the two 0.982175
12 r 0.991047
13
random variables. 13 Resid {–1.6465...

D5 =2.7803542327572E–8
§ 440 606 ·
37 000 − 8 ¨ × ¸
b m= ¦ xy − nx y = © 8 8 ¹
= 1.088 From the GDC, r = 0.991; p = 2.78 × 10−8
¦ y 2 − ny 2 § 606 ·
49 278 − 8 ¨
2

b The p-value suggests a strong relationship


¸
© 8 ¹
between the two random variables, hence it
x x 1.088( y  y ) ⇒ x = 1.088 y – 27.4 ;
makes sense to find the equation of the
x 32.4 regression line: y = 0.905 x + 2.59
C ov( X ,Y ) c y = 0.905(19.8) + 2.59 = 20.5
2 Since U , the maximum
Var ( X ) Var (Y ) d p = 0.00620 < 0.01, hence there is a significant
Cov(X, Y ) will occur when ρ = 1. correlation between the random variables.
6 a The gradient of the Y on X line, y = a + bx, is
Hence, Covmax(X, Y) = 17 u 15 10.2
Cov ( X , Y )
, and the gradient of the X on Y line,
3 If p = 0.9 then t = invt (0.9, 18) 1.33039 , and Var ( X )

18 Cov ( X , Y )
solving 1.33039 r , gives r = 0.299 x = c + dy, is . Since r = 0,
1  r2 Var (Y )
Cov(X,Y ) = 0, hence y = a and x = b, and the
Cov (U , V ) Cov (( X  Y ), ( X  Z )) two lines are perpendicular.
4 a U
Var (U ) Var (V ) Var ( X  Y ) Var ( X  Z ) 1
b For y = a + bx, m = b, and for x = c + dy, m = .
d
Cov ( XX )  Cov ( XZ )  Cov (YX )  Cov (YZ )
1
> Var ( X  Y )@ > Var ( X  Z )@ r r1 Ÿ r2 1. Since r 2 bd Ÿ bd 1, b .
d
Var( X )  0 V2 1 Hence, the two lines have the same gradient.
=
2V 2
2V 2 2V 2 2 Since both regression lines must go through
( x , y ), the lines are identical.
Hence it cannot be said that U and V are
independent. c For the regression lines y a  bx and x c  dy ,
Cov (V , W ) Cov (( X  Y ), ( X  Y )) Cov ( X , Y ) Cov ( X , Y )
b U b and d . Hence
Var (V ) Var (W ) Var ( X  Y ) Var ( X  Y ) Var ( X ) Var (Y )
2
Cov ( XX )  Cov ( XY )  Cov (YX )  Cov (YY ) ª Cov ( X , Y ) º
bd « Var ( X ) Var (Y ) » r 2. Since b and d are
> Var ( X  Y )@ > Var ( X  Y )@ ¬ ¼
Var ( X )  Var (Y ) either both positive or both negative, r  bd
0 if both are positive, and r  bd if both are
> Var ( X  Y )@ > Var ( X  Y )@
negative.
We know that if V and W are independent
then ρ = 0, but the converse does not d Since b and d are both positive,
necessarily hold.
r  bd 0.19 u 0.77 0.382

© Oxford University Press 2015: this may be reproduced for class use solely for the purchaser’s institute Statistics: Chapter 4 34