Fitting of binomial distribution

Source : Bio Statistics - A foundation for analysis in the health sciences by

Daniel
In a study designed to determine patient acceptance of a new pain reliever, 100 physicians

each selected a sample of 25 patients in the study. Each patient, after trying the new pain

reliever for a specified period of time, was asked whether it was preferable to the pain

reliever used regularly in the past.

A total of 500 patients out of the 2500 patients participating in the study said they preferred

the new pain reliever, so that our estimate of p is p = 500 / 2500 = 0.2 The expected relative

frequencies can be obtained by evaluating the binomial function.

N o of patients out No of doctors Expected relative Expected

of 25 prefering reporting this number frequency f(x) frequency(E)

new pain reliever (o)

1
5¿} ¿ 11
.0038

.0236
0.38¿} ¿ ¿ 2.74
2 8 .0708 7.08
3 10 .1358 13.58
4 10 .1867 18.67
5 15 .1960 19.60
6 17 .1633 16.33
7 10 .1109 11.09
8 10 .0623 6.23
9 9 .0295 2.95
10 ¿ 0 .0173 1.73
Total 100 100

f ( x )=¿ ( 25 ¿ ) ¿ ¿ ¿
¿ x = 0, 1, 2.

2 (O−E )2
χ =∑ =47 . 624
E

2
I compare the computed χ with the tabulated with (10-1-1) = 8 df and find it is
significant at 0.005 level of significance.
Therefore the data came from a binomial distribution.

Example – 2

Source: Doyle.P (1974) – Statistical Exercises in Medical Research by John F.Osborn

The data below show the number of spontaneous abortions suffered by a sample of 70

women who stated that they wish to have children and who had been pregnant four times. If

the risk of abortion were independent of previous reproductive history and the same for all

women, the number of abortions out of four pregnant should follow a binomial distribution

with n = 4.The total number of abortion is 81 and hence the estimate of the probability of

abortion is 81 / (4 x 70) = 0.2893.

No of spontaneous Observed number of Expected number of

Chi-square
abortions women women
0 24 17.857 2.1133
1 28 29.078 0.0400
2 7 17.941 6.6722
3
4 5¿}¿ ¿ 11 4.816¿}¿¿¿ 5.306 6.1104

Total 70 14.9359

2
I compare the computed χ with the tabulated with (5-1-1) = 3 df and find it is significant
at P < 0.001.

Therefore the data came from a binomial distribution.

Fitting of Poisson distribution

Example - 1

Source:Radhakrishna:‘Statistical methodology and Tuberculosis – A fruitful

association’-Ind J Tub 1992 , 39 , 71

Under the assumption that the leprosy cases are more or less evenly distributed in high endemic

situations, to investigate whether leprosy cases given in the following table follows Poisson

distribution.
If P denotes the probability of observing the characteristic, then the chance that there will be exactly

d individuals with the event in a sample of size n is given by the following expression

¿
d
e−m md
Pr ( d≤d ¿ ) = ∑
d=0 d ! where m = npo

The most frequently employed tool in tuberculosis research has been the controlled clinical
trial. The present study investigates the distribution of contaminated cultures from spot
specimens and collection specimens follows Poisson distribution.

Table 4. Poisson Fit for distribution of contaminated cultures

(Constant risk)

Numberof Spot specimens Collection specimens

contaminated
cultures
Observed Expected Observed Expected

0 1454 1465 334 279

1 590 574 489 513

2 128 129 483 511

3 17 21 344 363

4 205 204
 5 96 95

>6
¿} ¿ ¿ 2 ¿} ¿ ¿ 2 71 57

Total 2191 2191 2022 2022

P pooled >0.2 0.01

Example – 2 Malaria Parasites

Source : Wang C.C (1970) Multiple invasion of erythrocyte by malaria parasites,

Transactions of the Royal Society of Tropical Medicine and Hygeine,64,268-270.

Multiple invasion of erythrocytes is thought to be more likely to occur with Plasmodium

falcipuram than with other malaria parasites. Do these parasites attack certain red blood cells

preferentially, or are multiple infections random? The data are observed frequencies of the

number of parasites per erythrocyte in the blood film of a patient with Plasmodium

falcipuram malaria. If multiple infections are due to chance, the distribution of the number of

parasites per cell will be Poisson.

Number of parasites Observed Expected Chi-square

per erythrocytes frequency frequency

0 40000 39710.8 2.1061
1 8621 9149.37 30.5130
2 1259 1054.01 39.8676
3 99 80.94 4.0296
4

5+ 21¿}¿ ¿ 21 4.67¿}¿¿¿ 4. 53.2488

88
Total 50,000 50,000 129.77