Term 3 Week 6 Revision Exercises - Statistics

Suggested solutions
1 There are 9 distinct letters: C, O (2), R (2), E, L, A, T, I, N
(i) Treat all the letters as distinct then arrange them according to the restriction. This will be
followed by taking the repeated letters into account.
Number of arrangements
5 9! 6
2721600
2!2!

= =
(ii) Case 1: All distinct
No. of ways =
9
6
60480 P =
Case 2: 1 pair of repeated letters
No. of ways =
2 8
1 4
6!
50400
2!
C C =
Case 3: 2 pairs of repeated letters
No. of ways =
7
2
6!
3780
2!2!
C =
Total no. of ways 60480 50400 3780 114660 = + + =
2 (i) Let X be the weight of a randomly chosen packet of candy.
Since 60 is large, n = by Central Limit Theorem,
2
20
~ 100, approximately
60
X N
| |
|
\
.
P( 98) 0.781 X = (3sf)
(ii) The sample is biased, as only students are surveyed. Not everyone in the population has an
equal chance of being surveyed.
It will be difficult to get an exhaustive list of people of all age groups to do a proper
stratification. (no sampling frame)
Use Quota Sampling
3 (i)


(ii) From GC, the product moment correlation coefficient = 0.940 (3 sig fig)

(iii) From the scatter plot, as t increases, x increases at an increasing rate, which fits the curve
x = Ae
bt
. Transforming
Bt
x Ae = into a linear model: ln ln x A Bt = + . The magnitude of the
product moment correlation coefficient between ln x and t is close to 1, indicating a strong
linear correlation between ln x and t. Thus this model is suitable.

From GC, ln 1.71086 5.53, 0.0476 A A B = = =
(iv) If t is increased by 5, the increase in x is as follows:
( ) ( )
( 5) 5 5(0.0476)
1
(1.269)
Bt
o
B t B
o o o
x Ae
x Ae e x e x x
+
=
= = = =

The rate of chirps is estimated to increase by 26.9% from the rate before the
temperature is increased.

The estimate is not reliable because we are extrapolating beyond the region where the
data is collected and analyzed.

(v) The value of r remains the same since it is not affected by any translation or scaling.
4
2
~ (45000, 2000 ) A N
2
~ (30000,1850 ) B N
(i) ( )
1 2
2 ~ (30000, 22845000) A B B N +
( ) ( )
1 2
2 25000 0.85224 0.852 P A B B + > = = (3 sig fig)

Assumption: The distributions of the lifespans of all televisions are independent of each
other.

(ii) Let W denotes the number of plasma televisions out of 50 with a life span of more than
30000 hours.
~ (50, 0.5) W B
(14 22) (15 21) P W P W < < =
( 21) ( 14) P W P W =
0.16112 0.0013011 =
0.15982 =
0.160 =

(iii)
2
2000
~ 45000, X N
n
| |
|
\

( 46500) 0.99 P X
46500 45000
0.99
2000
P Z
n
| |
|

|
|
|
\


3
0.99
4
n
P Z
| |

|
|
\


3
2.32635
4
n

9.6212 n
Least 10 n =

(iv) (I) will be greater as A > 25000 and B > 25000 is a subset of A+B > 50000.


5
~ ( )
o
X P
(i)
2 3
10
( 2) 10 ( 3) 0
2! 3!
e e
P X P X



= = = =

2
10
1 0
2! 3
e

| |
=
|
\


3
E( ) 0.3
10
X = = = (shown)
(ii) Let U be the total number of errors in a randomly selected textbook.
~ (15)
o
U P
( 10) 0.0486 (3 s.f.) P U = =
(iii) Let A and B be the total number of errors in the pages 1 to 10 plus 41 to 50 and 11 to 40
respectively.
~ (6)
o
A P and ~ (9)
o
B P ,
Probability
( 1) ( 9) ( 0) ( 10)
( 10)
P A P B P A P B
P U
= = + = =
=
=
(0.01487251)(0.13175564) (0.0024788)(0.11858007)
0.04861075
+
=

= 0.046357=0.0464 (3 sig fig)
(iv) ( 18) 1 ( 18) P U P U > = 0.18053 =
Let x be the minimum selling price, in order to make a profit,

1.15(12 12 ( 18))
16.291
x P U
x
+ >

Minimum selling price is $16.30.
(v) Let V be the number of textbooks that has more than 18 errors out of 50 books.
~ (50, 0.18053) V B
Since n = 50 is large and np = 9.0265 > 5 and n(1-p)= 40.9735 > 5,
~ (9.0265, 7.3969) V N approximately
( ) P 6 P( 5.5) by continuity correction
0.0974
V V < = <
=

6 (i) For the die, let x = P(1) = P(3) = P(5).
2 2 2 1
1 1
P(1) (shown)
9 9
x x x x x x
x
+ + + + + =
= =


P(1) = P(3) = P(5) =
1
9
and P(2) = P(4) = P(6) =
2
9

(ii) For the coin, P(tail) =
1
3
and P(head) =
2
3

P(the coin to show tail and the die to show a number that is at most 4)
= P(coin shows tail) P(die shows 1 4) since the events are independent
1 1 2 1 2
.
3 9 9 9 9
| |
= + + +
|
\

2
9
=


(iii)
P(coin shows a tail or the die shows a number that is at most 4, or both)
= P(coin shows a tail)+ P(die shows a number that is at most 4) P(tail and at most 4)
1 6 2
3 9 9
= +
7
9
=
(iv) P(coin shows a head if the score is at most 4)
P(coin is 'tail' score is at most 4)
P(coin is 'tail' and score is at most 4)
=
score is at most 4
=


2
9
P('tail' and 1-4 or 'head' and 1-2)
2
9
2 2 1 2 2
. .
9 3 9 3 9
1
2
=
=
+ +
=



(v) For each throw, Mary is expected to gain $
4 5
1. 0.5
9 9
( | | | |
+
| | (
\ \
1
$
6
=
Therefore, Mary is not expected to make a profit as she is expected to lose $10.00 in the game.
7 (i)
Unbiased estimate of the population mean
44.8
3.7333333
12
x = = = =3.73 (3sf)
Unbiased estimate of the population variance
( )
2
1.9467
0.1769727273
11 11
x x
= = =

= 0.177
(3sf)
(ii) Let X be the time, in hours, that a randomly chosen customer spends shopping at Takayama
Shopping Complex, and be the mean time.
Test : 4
o
H = (managers claim)
vs
1
: 4 H (superiors suspicion)

Assume that X is normally distributed.
Perform a two-tailed test at 6% level of significance.
Under H
o
,
4
~ ( 1)
X
T t n
s n

=
From the sample,
56
15
x = , 0.1769727273 s = , n=12
Using a t-test, - value 0.0504 0.06 p = <
We reject
o
H . There is sufficient evidence at 6% significance level to confirm the company
superiors suspicion.
(iii) Since the p-value in part (ii) is less than 6%, the new p-value would be half of the value in
(ii), therefore it would still be less than 6%. Thus, the conclusion would remain the same.


(iv) Test
0
:
o
H = (managers claim)
vs
1 0
: H < (manager overstated the average time)
Perform a one-tailed test at 6% level of significance.
Under H
o
,
2
0
~ N , X
n

| |
|
\
, where 0.466 =
From the sample, 4.2375 x = , n = 8
Using a z-test, do not reject
o
H if - value 0.06 p > .

0
0
P( 4.2375) 0.06
4.2375
P 0.06
0.466 / 8
4.2375
1.5547
0.466 / 8
X
Z

< >
| |
< >
|
\

>


0
0.466
4.2375 1.5547736
8

| |
>
|
\


0
4.493658 < hrs

0
4 < hrs 29.62 mins
Largest possible value of
0
is 4 hours 30 mins (to nearest min)