CHAPTER 76 CHI-SQUARE AND DISTRIBUTION-FREE TESTS

EXERCISE 313 Page 823

1. A dice is rolled 240 times and the observed and expected frequencies are as shown.

Face 1 2 3 4 5 6
Observed frequency 49 35 32 46 49 29
Expected frequency 40 40 40 40 40 40

Determine the 2-value for this distribution.

Face Observed Expected o-e

frequency, o frequency, e

1 49 40 9 81 2.025
2 35 40 -5 25 0.625
3 32 40 -8 64 1.6
4 46 40 6 36 0.9
5 49 40 9 81 2.025
6 29 40 -11 121 3.025

Hence, the Chi-square value,

2. The numbers of telephone calls received by the switchboard of a company in 200 five-minute

intervals are shown in the distribution below.

Number of calls 0 1 2 3 4 5 6 7
Observed frequency 11 44 53 46 24 12 7 3
Expected frequency 16 42 52 42 26 14 6 2

Calculate the 2-value for this data.

1137
Number Observed Expected o-e
of calls frequency, o frequency, e

0 11 16 -5 25 1.5625
1 44 42 2 4 0.0952
2 53 52 1 1 0.0192
3 46 42 4 16 0.3810
4 24 26 -2 4 0.1538
5 12 14 -2 4 0.2857
6 7 6 1 1 0.1667
7 3 2 1 1 0.5000

Hence, the Chi-square value,

EXERCISE 314 Page 829

1138
1. Test the null hypothesis that the observed data given below fits a binomial distribution of the

form 250(0.6 + 0.4) at a level of significance of 0.05

Observed frequency 8 27 62 79 45 24 5 0

Is the fit of the data 'too good' at a level of confidence of 90%?

= 250[0.02799 + 0.13064 + 0.26127 + 0.29030 + 0.19354 + 0.07741 + 0.01720

+ 0.00164]
= 7 + 33 + 65 + 73 + 48 + 19 + 4 + 0 correct to the nearest whole number

Observed Expected o-e

frequency, o frequency, e

8 7 1 1 0.14286
27 33 -6 36 1.09091
62 65 -3 9 0.13846
79 73 6 36 0.49315
45 48 -3 9 0.18750
24 19 5 25 1.31579
5 4 1 1 0.25000
0 0 0 0 0

Degrees of freedom,  = N – 1 = 8 – 1 = 7

For and  = 7 from Table 76.1, page 825 is 14.1

Hence, the null hypothesis is accepted, i.e. the observed data fits.

= 2.83, hence the data is not ‘too good’.

2. The data given below refers to the number of people injured in a city by accidents for weekly

1139
 periods throughout a year. It is believed that the data fits a Poisson distribution. Test the

goodness of fit at a level of significance of 0.05

Number of people injured in the week 0 1 2 3 4 5 6

Number of weeks 5 12 13 9 7 4 2

To determine the expected frequency

The expectation or average occurrence is given by:

 = = = 2.404

The expected frequencies are calculated using a Poisson distribution, where the probabilities of
there being 0, 1, 2,..., 6 people injured are given by the successive terms of

e taken from left to right, i.e. e , e , ,

Calculating these terms for  = 2.404 gives:

Number of people injured 0 1 2 3 4 5 6

Probability 0.0904 0.2172 0.2611 0.2092 0.1257 0.0605 0.0242
Expected frequency 5 11 14 11 7 3 1

To determine the 2-value

The 2-value is calculated using a tabular method as shown below.

Observed Expected o-e

frequency, o frequency, e

5 5 0 0 0
12 11 1 1 0.09090
13 14 -1 1 0.07143
9 11 -2 4 0.36364
7 7 0 0 0
4 3 1 1 0.33333
2 1 1 1 1.00000

To test the significance of the 2-value

1140
The number of degrees of freedom is  = N - 1, where N is the number of rows in the table above,

giving  = 7 - 1 = 6. The percentile value of 2 is determined from Table 76.1, for ( ,  = 6), and

is 12.6. Since the calculated value of 2 , i.e. 1.86 < 12.6, the data fits a Poisson distribution at a

level of significance of 0.05

3. The resistances of a sample of carbon resistors are as shown below.

Resistance (M) 1.28 1.29 1.30 1.31 1.32 1.33 1.34 1.35 1.36
Frequency 7 19 41 50 73 52 28 17 9
Test the null hypothesis that this data corresponds to a normal distribution at a level of

significance of 0.05

MΩ

Class Class z-value for class Area from Area for class Expected frequency
mid-point boundaries, x 0 to z

boundary = from Table 71.1,

page XX
1.275 -2.50 0.4938
1.28 0.02 6
1.285 -1.94 0.4738
1.29 0.0561 17
1.295 -1.39 0.4177
1.30 0.121 36
1.305 -0.83 0.2967
1.31 0.1864 55
1.315 -0.28 0.1103
1.32 0.2206 65
1.325 0.28 0.1103
1.33 0.1864 55
1.335 0.83 0.2967
1.34 0.121 36
1.345 1.39 0.4177
1.35 0.0561 17
1.355 1.94 0.4738
1.36 0.02 6
1.365 2.5 0.4938

1141
 Resistance Observed Expected o-e
frequency, o frequency, e

1.28 7 6 1 1 0.1667
1.29 19 17 2 4 0.2353
1.30 41 36 5 25 0.6944
1.31 50 55 -5 25 0.4545
1.32 73 65 8 64 0.9846
1.33 52 55 -3 9 0.1636
1.34 28 36 -8 64 1.7778
1.35 17 17 0 0 0
1.36 9 6 3 9 1.5000

Degrees of freedom,  = N – 1 – M = 9 – 1 – 2 = 6

For and  = 6 from Table 76.1 is 12.6

Hence, the null hypothesis is accepted, i.e. the data does correspond to a normal distribution.

4. The quality assurance department of a firm selects 250 capacitors at random from a large

quantity of them and carries out various tests on them.

The results obtained are as follows:

Number of tests failed 0 1 2 3 4 5 6 and over

Number of capacitors 113 77 39 16 4 1 0

Test the goodness of fit of this distribution to a Poisson distribution at a level of significance of

0.05

To determine the expected frequency

The expectation or average occurrence is given by:

 = = = 0.896

1142
The expected frequencies are calculated using a Poisson distribution, where the probabilities of
there being 0, 1, 2,..., 6 tests failed are given by the successive terms of

e taken from left to right, i.e. e , e , ,

Calculating these terms for  = 0.896 gives:

Number of people injured 0 1 2 3 4 5 6

Probability 0.4082 0.3657 0.1639 0.0489 0.0110 0.00196 0.0003
Expected frequency 102 91 41 12 3 0 0

To determine the 2-value

The 2-value is calculated using a tabular method as shown below.

Observed Expected o-e

frequency, o frequency, e

113 102 11 121 1.18627

77 91 - 14 196 2.15385
39 41 -2 4 0.09756
16 12 4 16 1.33333
4 3 1 1 0.33333
1 0 1 1 0
0 0 0 0 0

To test the significance of the 2-value

The number of degrees of freedom is  = N - 1, where N is the number of rows in the table above,

giving  = 7 - 1 = 6. The percentile value of 2 is determined from Table 76.1, for ( ,  = 6), and

is 12.6. Since the calculated value of 2 (i.e. 5.10) is smaller than the percentile value, the

hypothesis according to a Poisson distribution is accepted, the data fits a Poisson distribution

at a level of significance of 0.05

5. Test the null hypothesis that the maximum load before breaking supported by certain cables

produced by a company follows a normal distribution at a level of significance of 0.05, based on

the experimental data given below. Also test to see if the data is 'too good' at a level of

1143
 confidence of 95%

Maximum load (MN) 8.5 9.0 9.5 10.0 10.5 11.0 11.5 12.0
Number of cables 2 5 12 17 14 6 3 1

Class Class z-value for class Area from Area for class Expected frequency
mid-point boundaries, x boundary = 0 to z

8.25 -2.51 0.4940

8.5 0.0276 2
8.75 -1.83 0.4664
9.0 0.0915 5
9.25 -1.15 0.3749
9.5 0.1977 12
9.75 -0.46 0.1772
10.0 0.2643 16
10.25 0.22 0.0871
10.5 0.2288 14
10.75 0.90 0.3159
11.0 0.1271 8
11.25 1.58 0.4430
11.5 0.0451 3
11.75 2.26 0.4881
12.0 0.0103 1
12.25 2.95 0.4984

Load Observed Expected o-e

frequency, o frequency, e

8.5 2 2 0 0 0
9.0 5 5 0 0 0
9.5 12 12 0 0 0
10.0 17 16 1 1 0.0625
10.5 14 14 0 0 0
11.0 6 8 -2 4 0.5000
11.5 3 3 0 0 0
12.0 1 1 0 0 0

1144
Degrees of freedom,  = N – 1 – M = 8 – 1 – 2 = 5

For and  = 5 from Table 76.1 is 11.1

Hence, the null hypothesis is accepted.

, hence the results are ‘too good to be true’.

EXERCISE 315 Page 833

1. The following data represent the number of hours of flight training received by 16 trainee pilots

prior to their first solo flight:

11.5h 20h 9h 12.5h 15h 19h 11h 10.5h 13h 22h 14.5h
16.5h 17h 18h 14h 12h

Use the sign test at a significance level of 2% to test the claim that, on average, the trainees solo

after 15 hours of flight training.

Using the procedure for the sign test:

1145
(i) Null hypothesis,

Alternative hypothesis,

(ii) Significance level,

(iii) With + sign  15 and - sign  15:

- + - - + + - - - + - + + + - -

(iv) There are 7 + signs and 9 – signs, hence, S = 7

(v) From Table 76.3, with n = 16 and S2

hence, the null hypothesis is accepted.

2. In a laboratory experiment, 18 measurements of the coefficient of friction, , between metal and

leather gave the following results:

0.60 0.57 0.51 0.55 0.66 0.56 0.52 0.59 0.58 0.48 0.59 0.63
0.61 0.69 0.57 0.51 0.58 0.54
Use the sign test at a level of significance of 5% to test the null hypothesis  = 0.56 against an

alternative hypothesis   0.56

Using the procedure for the sign test:

(i) Null hypothesis,

Alternative hypothesis,

(ii) Significance level,

(iii) With + sign  0.56 and - sign  0.56:

+ + - - + + - + + - + + + + + - + -

(iv) There are 12 + signs and 6 – signs, hence, S = 6

(v) From Table 76.3, with n = 18 and S4

hence, the null hypothesis is accepted.

1146
3. 18 random samples of two types of 9 V batteries are taken and the mean lifetimes (in hours) of

each are:

Type A 8.2 7.0 11.3 13.9 9.0 13.8 16.2 8.6 9.4
3.6 7.5 6.5 18.0 11.5 13.4 6.9 14.2 12.4
Type B 15.3 15.4 11.2 16.1 18.1 17.1 17.7 8.4 13.5
7.8 9.8 10.6 16.4 12.7 16.8 9.9 12.9 14.7

Use the sign test, at a level of significance of 5%, to test the null hypothesis that the two

samples come from the same population

Using the procedure for the sign test:

(i) Null hypothesis,

Alternative hypothesis,

(ii) Significance level,

(iii) A - B -7.1 -8.4 +0.1 -2.2 -9.1 -3.3 -1.5 +0.2 -4.1

-4.2 -2.3 -4.1 +1.6 -1.2 -3.4 -3.0 +1.3 -2.3

(iv) There are 4 + signs and 14 – signs, hence, S = 4

(v) From Table 76.3, with n = 18 and S4

Since from (iv) S is equal to 4, then the result is significant at hence, the alternative

hypothesis is accepted.
EXERCISE 316 Page 837

1. The time to repair an electronic instrument is a random variable. The repair times (in hours) for

16 instruments are as follows:

218 275 264 210 161 374 178 265 150 360 185 171 215 100 474 248

Use the Wilcoxon signed-rank test, at a 5% level of significance, to test the null hypothesis that

the mean repair time is 220 hours.

1147
Using the procedure for the Wilcoxon signed-rank test:

(i) t = 220 h

t  220 h

(ii)

(iii) Taking the time difference between the time taken for repair and 220 h gives:

-2 +55 +44 -10 -59 +154 -42 +45

-70 +140 -35 -49 -5 -120 +254 + 28

(iv) Ranking gives:

Rank 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Difference -2 -5 -10 +28 -35 -42 +44 +45 -49 +55 -59 -70 -120 +140 +154 +254

(v) T = 4 + 7 + 8 + 10 + 14 + 15 + 16 = 74

(vi) From Table 76.4, page 834, for n = 16, , T  29

Hence, since 74  29, the null hypothesis is accepted, i.e. the mean repair time is 220
hours.

2. 18 samples of serum are analysed for their sodium content. The results, expressed as ppm are as
follows:
169 151 166 155 149 154 164 151 147 142 168 152
149 129 153 154 149 143
At a level of significance of 5%, use the Wilcoxon signed-rank test to test the null hypothesis

that the average value for the method of analysis used is 150 ppm

Using the procedure for the Wilcoxon signed-rank test:

(i) s = 150 ppm

s  150 ppm

(ii)

(iii) Taking the difference between the analysed ppm serum and 150 ppm gives:

1148
 + 19 + 1 + 16 + 5 - 1 + 4 + 14 + 1 - 3

- 8 + 18 + 2 - 1 - 21 + 3 + 4 - 1 - 7

(iv) Ranking gives:

Rank 3 3 3 3 3 6 7.5 7.5 9.5 9.5 11 12 13 14 15 16 17 18

Difference -1 -1 -1 1 1 2 -3 +3 4 4 5 -7 -8 14 16 18 19 -21

(v) T = 3 + 3 + 3 + 7.5 + 12 + 13 + 18 = 59.5

(vi) From Table 76.4, page 834, for n = 18, , T  40

Hence, since 59.5  40, the null hypothesis is accepted, i.e. the average value for the

method of analysis used is 150 ppm

3. A paint supplier claims that a new additive will reduce the drying time of their acrylic paint. To

test his claim, 12 pieces of wood are painted, one half of each piece with paint containing the

regular additive and the other half with paint containing the new additive. The drying time (in

hours) were measured as follows:

New additive 4.5 5.5 3.9 3.6 4.1 6.3 5.9 6.7 5.1 3.6 4.0 3.0
Regular additive 4.7 5.9 3.9 3.8 4.4 6.5 6.9 6.5 5.3 3.6 3.9 3.9

Use the Wilcoxon signed-rank test at a significance level of 5% to test the null hypothesis that

there is no difference, on average, in the drying times of the new and regular additive paints.

Using the procedure for the Wilcoxon signed-rank test:

(i) N=R

NR

(ii)

(iii) Taking the time difference between N and R gives:

(N – R) -0.2 -0.4 0 -0.2 -0.3 -0.2 -1.0 +0.2 -0.2 0 +0.1 -0.9

(iv) Ranking and ignoring the zero’s gives:

1149
 Rank 1 4 4 4 4 4 7 8 9 10

Difference +0.1 -0.2 -0.2 -0.2 -0.2 +0.2 -0.3 -0.4 -0.9 -1.0

(v) T = 1 + 4 = 5

(vi) From Table 76.4, for n = 10, , T8

Since from (v) T is less than 8, there is a significant difference in the drying times.

EXERCISE 317 Page 843

1. The tar content of two brands of cigarettes (in mg) was measured as follows:

Brand P 22.6 4.1 3.9 0.7 3.2 6.1 1.7 2.3 5.6 2.0
Brand Q 3.4 6.2 3.5 4.7 6.3 5.5 3.8 2.1
Use the Mann-Whitney test at a 0.05 level of significance to determine if the tar contents of the

1150
 two brands are equal.

Using the procedure for the Mann-Whitney test:

(i)

(ii)

(iii) Brand P 0.7 1.7 2.0 2.3 3.2 3.9 4.1 5.6 6.1 22.6

Brand Q 2.1 3.4 3.5 3.8 4.7 5.5 6.2 6.3

(iv) P P P Q P P Q Q Q P P Q Q P P Q Q P

0 0 0 1 1 4 4 6 6 8 writing the Q’s that precede the P’s

(v) U = 1 + 1 + 4 + 4 + 6 + 6 + 8 = 30

(vi) From Table 76.5, page 838, for a sample size of 10 and 8 at , U  17

Hence, is accepted, i.e. there is no difference between brands P and Q

2. A component is manufactured by two processes. Some components from each process are

selected at random and tested for breaking strength to determine if there is a difference between

the processes. The results are:

Process A 9.7 10.5 10.1 11.6 9.8 8.9 11.2 12.0 9.2
Process B 11.3 8.6 9.6 10.2 10.9 9.4 10.8

At a level of significance of 10%, use the Mann-Whitney test to determine if there is a difference

between the mean breaking strengths of the components manufactured by the two processes.

Using the procedure for the Mann-Whitney test:

(i)

(ii)

1151
(iii) Process A 8.9 9.2 9.7 9.8 10.1 10.5 11.2 11.6 12.0

Process B 8.6 9.4 9.6 10.2 10.8 10.9 11.3

(iv) B A A B B A A A B A B B A B A A

0 2 2 5 6 6 7 writing the A’s that precede the B’s

(v) U = 2 + 2 + 5 + 6 + 6 + 7 = 28

(vi) From Table 76.5, page 838, for a sample size of 9 and 7 at , U  15

Hence, is accepted, i.e. there is no difference between the processes.

3. An experiment, designed to compare two preventive methods against corrosion gave the

following results for the maximum depths of pits (in mm) in metal strands:

Method A 143 106 135 147 139 132 153 140

Method B 98 105 137 94 112 103

Use the Mann-Whitney test, at a level of significance of 0.05, to determine whether the two tests

are equally effective.

Using the procedure for the Mann-Whitney test:

(i)

(ii)

(iii) A 106 132 135 139 140 143 147 153

B 94 98 103 105 112 137

(iv) B B B B A B A A B A A A A A

1 3 writing the A’s that precede the B’s

(v) U = 1 + 3 = 4

(vi) From Table 76.5, for a sample size of 8 and 6 at ,U8

Hence, the null hypothesis is rejected, i.e. the two methods are not equally effective.

1152
4. 18 random samples of two types of 9 V batteries are taken and the mean lifetimes (in hours) of

each are:

Type A 8.2 7.0 11.3 13.9 9.0 13.8 16.2 8.6 9.4
3.6 7.5 6.5 18.0 11.5 13.4 6.9 14.2 12.4
Type B 15.3 15.4 11.2 16.1 18.1 17.1 17.7 8.4 13.5
7.8 9.8 10.6 16.4 12.7 16.8 9.9 12.9 14.7

Use the Mann-Whitney test, at a level of significance of 5%, to test the null hypothesis that the

two samples come from the same population

Using the procedure for the Mann-Whitney test:

(i) Null hypothesis,

Alternative hypothesis,

(ii)

(iii) A 3.6 6.5 6.9 7.0 7.5 8.2 8.6 9.0 9.4 11.3 11.5 12.4 13.4 13.8 13.9 14.2 16.2 18.0

B 7.8 8.4 9.8 9.9 10.6 11.2 12.7 12.9 13.5 14.7 15.3 15.4 16.1 16.4 16.8 17.1 17.7 18.1

(iv) A A A A A B A B A A A B B B B A A A B B A B A A A B B B B A B B B B A B

1 2 2 2 6 6 6 8 9 9 9 13 17

(v) U = 1 + 2 + 2 + 2 + 6 + 6 + 6 + 8 + 9 + 9 + 9 + 13 + 17 = 90

(vi) From Table 76.5, for a sample size of 18 and 18 at , U  99

Hence, the null hypothesis is rejected and is accepted, i.e. the two means are not

equal.

1153