0% found this document useful (0 votes)
16 views12 pages

Normal Approximation to Binomial Distribution

The document discusses the normal approximation to the binomial distribution, highlighting the conditions for its application and the need for continuity correction. It provides examples and solutions related to binomial probabilities, including confidence limits and various scenarios involving binomial distributions. Key formulas and calculations are presented to illustrate the application of the normal approximation in different contexts.

Uploaded by

kasiitamartin696
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
16 views12 pages

Normal Approximation to Binomial Distribution

The document discusses the normal approximation to the binomial distribution, highlighting the conditions for its application and the need for continuity correction. It provides examples and solutions related to binomial probabilities, including confidence limits and various scenarios involving binomial distributions. Key formulas and calculations are presented to illustrate the application of the normal approximation in different contexts.

Uploaded by

kasiitamartin696
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

THE NORMAL APPROXIMATION TO THE BINOMIAL DISTRIBUTION

Summary:

(i) Use X ∼ N(np , npq) in case of a binomial distribution whose parameters

n and p are such that np > 5 and nq > 5, where q = 1 − p.


(ii) A continuity correction is needed when a discrete binomial distribution is

approximated by the continuous normal distribution.

(iii) A discrete value x is continuity corrected to a normal interval by obtaining

the limits of accuracy of x along a continuous axis.

∴ x transforms into a normal interval with boundaries x− 0⋅5 and x+ 0⋅5.

P(X = x) → P(x− 0⋅5 < X < x+ 0⋅5 ) “ By continuity correction”

(iv) The binomial confidence limits = np ± Z npq , where Z = critical value

for the desired level of confidence

EXAMPLES:

1. Write down the continuity corrections for each of the following when using

the normal approximation to the binomial distribution.

(i) P(X = 40) (ii) P(X < 30) (iii) P(X ≤ 30) (iv) P(X > 40)

(v) P(X ≥ 40) (vi) P(20 ≤ X ≤ 24) (vii) P(20 < X ≤ 24)

(viii) P(20 ≤ X < 24) (ix) P(20 < X < 24)

1
2. (a) State the conditions under which the binomial distribution X ∼ Bin(n , p)

may be approximated by a normal distribution, giving the parameters of this

normal approximation.

(b) A fair coin is tossed 100 times. Find the probability of obtaining:

(i) exactly 40 heads

(ii) atmost 53 heads

(iii) atleast 45 heads

(iv) more than 60 heads

(v) less than 48 heads

(vi) not more than 42 heads

(vii) not less than 57 heads

(viii) between 43 and 55 heads inclusive

(ix) atleast 41 heads but less than 54

(x) between 47 and 56 heads

Solution:

(a) Required conditions are np > 5 and nq > 5, where q = 1 − p.

Required parameters are µ = np , σ 2 = npq, where q = 1 − p.

(b) If X ∼ Bin(100 , 0⋅5),

⇒ µ = 100(0⋅5) = 50 , σ = 50(0 ⋅ 5) = 5

2
2. One in every four workers of a certain company is a graduate.

(a) If 192 workers are selected at random, find the probability that:

(i) exactly 52 are graduates.

(ii) at most 40 are graduates.

(iii) at least 44 but less than 56 are graduates.

(b) Find the least number of workers to be selected in order to be 95% certain
that at least one of them is a graduate.

3. (a) Two fair octahedral dice with faces numbered from 1 to 8 are thrown 240

times. Find the probability that a sum of 5 appears:

(i) exactly 25 times.

(ii) between 18 and 24 inclusive times

(b) How many tosses would be required if the probability of getting a sum of 5

at least once is greater than 0⋅55.

4. 400 students sit a multiple choice test containing 80 questions whose answers

are either “YES” or “NO”. If the students answer all of them by guess work,

find the:

(i) chance that a student selected at random got exactly 45 correct answers.

(ii) number of students expected to pass if the pass mark is 38.

(iii) new pass mark if only 115 students are expected to pass.

[Ans: (i) 0⋅048 (ii) 285 (iii) 43 ]

3
5. (a) Given that X ∼ Bin(48 , 0⋅25) ,

(i) Show that the distribution can be approximated by a normal distribution

(ii) find P(X = 16)

(b) In a class room, 64% of the students are short−sighted. If 100 students

are tested, find the 90% confidence limits for the mean number of

short−sighted students.

(c) In a packet of 100 seeds only 65 can germinate. If four of such packets are
planted, find the probability that at most 120 seeds will not germinate.

(d) One in every five students in a class room gossips. Find the least number

of students to be selected at random from the classroom so that the

probability that none of them gossips is less than 0⋅007.

EER:

1. (a) The chance of germination of a certain type of seed is 0⋅8. If 250 seeds

are planted, Find the chance that more than 54 seeds will not germinate.

(b) A test has 90 multiple choice questions each with 5 suggested answers

of which one is the correct answer. If a student answers the questions


randomly, find the probability of obtaining at least 20 correct answers.

(c) 15% of the eggs sold in the shop are rotten. Find the probability that a

sample of 300 eggs will have more than 52 rotten eggs.

4
(d) On a farm, 20% of the cows are infected by a tick disease. If a random

sample of 50 cows is selected from the farm, find the probability that not

more than 10% of the cows are infected

(e) 10% of the students travel to school by bus. Find the probability

that in a sample of 400 students, atleast 35 but less than 47 students

travel by bus.

2
(f) The chance that a factory produces a defective item is . Find the chance
3
that of the 72 produced items more than 50 are defective.

[Ans: (a) 0⋅2383 (b) 0⋅3464 (c) 0⋅1125 (d) 0⋅0558 (e) 0⋅6807 (f) 0⋅2660 ]

2. A test has 72 multiple choice questions each with 3 suggested answers of which

one is the correct answer. If a student answers the questions randomly, find the

probability of obtaining:

(i) atleast 30 correct answers.

(ii) between 43 and 57 incorrect answers.

[Ans: (i) 0⋅0845 (ii) 0⋅8529 ]

3. A biased dice is thrown 300 times and the number of sixes obtained is 80.
If the dice is further thrown 12 times, determine the:
(i) probability that a six will occur exactly twice
(ii) variance of the number of sixes.
[Ans: (i) 0⋅2111 (ii) 2⋅3467 ]

5
4. (a) Given that X ∼ Bin(400 , 0⋅1), find P(X ≥ 48).

(b) Given that X ∼ Bin(n , 0⋅2), find the least value of n for which

P(X ≥ 1) > 0⋅87.

(c) Given that X ∼ Bin(1000 , 0⋅5), find the largest value of T for which

P(X ≤ T) < 0⋅95.

(d) The chance that a cow recovers from a certain mouth disease when treated is

0⋅72. If 100 cows are treated by the same vaccine, find the 95% confidence

limits for the mean number of cows that recovers.

[Ans: (a) 0⋅1056 (b) 10 (c) 525 (d) Lower limit = 63⋅1996, Upper limit = 80⋅8004 ]

5. An examination has 100 questions. A student has 60% chance of getting each

question correct. A student fails the examination for a mark less than 55.

A student gets a distinction for a mark of 68 or more. Calculate the probability

that a student:

(i) fails the examination

(ii) gets a distinction

[Ans: (i) 0⋅1308 (ii) 0⋅0629 ]

6. Among the spectators watching a football match 80% were the home team

supporters while 20% were the visiting team supporters. If 2500 of the

spectators are selected at random, find the probability that there are more than

540 visitors in this sample.

[Ans: 0⋅0215 ]

6
7. A baised coin is twice as likely to show heads as tails. If it is tossed 120

times, find the probability of obtaining:

(i) between 42 and 51 tails inclusive.

(ii) between 72 and 90 heads.

[Ans: (i) 0⋅3729 (ii) 0⋅8939 ]

8. (a) In an election, 75% of the voters support party Q. Find the probability

that in a sample of:

(i) 10 voters, atleast 8 support party Q.

(ii) 48 voters, between 30 and 40 support party Q.

(b) Find the least number of voters to be selected in order to be 99% confident

that at least one of them supports party Q.

[Ans: (a) (i) 0⋅5256 (ii) 0⋅845 (b) 4 ]

9. (a) A pair of dice is tossed 180 times, determine the probability that a sum

of 7 appears:

(i) exactly 40 times

(ii) between 25 and 35 inclusive times

(b) How many tosses would be required if the probability of getting a sum of 7

at least once is greater than 0⋅85.

[Ans: (a) (i) 0⋅0108 (ii) 0⋅7286 (b) 11 ]

7
10. On a farm, 35% of the cows are infected by a tick disease. If a random

sample of 400 cows is selected from the farm, find the probability that:

(i) less than 30% of the cows are infected

(ii) more than 155 cows are infected

(iii) between 120 and 150 inclusive cows are infected.

[Ans: (i) 0⋅0158 (ii) 0⋅0521 (iii) 0⋅8487 ]

11. (a) A box of 72 harvested apples has 5% that are rotten. Three such boxes are

harvested. Find the probability that the harvest contains:

(i) less than 10 rotten apples.

(ii) between 10 and 15 rotten apples.

(iii) between 9 and 12 inclusive rotten apples.

[Ans: (i) 0⋅3441 (ii) 0⋅4134 (iii) 0⋅4658 ]

12. A research station supplies three varieties of seeds S1 , S 2 and S 3 in the ratio

[Link]. The probabilities of germination of S1 , S 2 and S 3 are 50%, 60% and

80% respectively.

(a) Find the probability that a seed selected at random will germinate.

(b) Given that 150 seeds are selected at random, find the probability that less

than 90 of the seed will germinate. Give your answer to 2 decimal places.

4
[Ans: (a) (b) 0⋅73 ]
7

8
13. Three in every eight workers of a certain company are graduates. Find the:

(i) probability that in a sample of 10 workers are selected at random, exactly

4 are graduates.

(ii) probability that in a sample 960 workers are selected at random, at least

342 workers but less than 396 are graduates.

(iii) minimum number of workers required to be selected in order to be 97%

certain that at least one of them is a graduate.

[Ans: (i) 0⋅2475 (ii) 0⋅8824 (iii) 8 ]

14. A multiple choice test contains 50 questions whose answers are either “YES”

or “NO”. If a student answers all of them by guess work, find the chance that:

(i) only 30 will be correct.

(ii) more than half will be correct.

(iii) atleast 22 but less than 28 will be correct.

[Ans: (i) 0⋅0416 (ii) 0⋅4439 (iii) 0⋅5991 ]

15. In a certain clan the chance that a person has blood group A is 0⋅75. Find the:

(i) most likely number of clan mates with blood group A in a sample of 9 clan

mates selected at random.

(ii) probability that in a sample of 48 clan mates selected at random, at least

29 clan mates but less than 43 have blood group A.

(iii) minimum number of clan mates required to be selected in order to be 99%

certain that at least one of them has blood group A.

[Ans: (i) 7 (ii) 0⋅9787 (iii) 4 ]

9
15. (a) State the conditions under which a binomial probability model may be

approximated by a normal model.

(b) In a potato garden, the average number of rotten potatoes was 48 with a

standard deviation of 4 potatoes. Find the:

(i) number of potatoes in the garden

(ii) probability that the garden contains not less than 45 rotten potatoes

[Ans: (b) (i) 72 (ii) 0⋅8091 ]

16. (a) In a clan of 80 families each with 60 children, the probability of having a

boy at each birth is 0⋅55. Find how many families are expected to have:

(i) exactly 29 girls

(ii) atleast 33 girls


(b) In a group of 500 people each tosses a fair coin 120 times, Find how many

people should be expected to obtain between 48 and 72 heads.

[Ans: (a) (i) 8 (ii) 7 (b) 483 ]

17. The chance of hitting a target with a single shot is 0⋅3.

(i) Find the least number of shots required to be fired in order to be 95%

confident that at least one of them hits the target.

(ii) If 12 shots are fired, find the probability of hitting the target atmost 4 times

(iii) If 84 shots are fired, find the 90% confidence limits for the mean number

of times the target is hit.

[Ans: (i) 9 (ii) 0⋅7237 (iii) Lower limit = 18⋅291, Upper limit = 32⋅109 ]

10
16. (a) A baised coin is thrice as likely to show heads as tails. If it is tossed 48

times, find the probability of obtaining:

(i) atleast 28 but less than 42 heads.

(ii) between 30 and 40 heads.

(b) A dice is tossed 40 times and the chance of obtaining a six on any one toss is

0⋅122. Find the probability of obtaining between 7 and 9 sixes.

[Ans: (a) (i) 0⋅9649 (ii) 0⋅8450 (b) 0⋅0626 ]

17. The marks of 120 students in a test were normally distributed with mean 60

and standard deviation 16. Find the:

(i) interquartile range of the students marks

(ii) lowest mark for a distinction, if 10% of the students got distinctions.

(iii) probability that atleast 70 students got above the mean mark.

[Ans: (i) 21⋅568 (ii) 80⋅512 (iii) 0⋅0415 ]

1 If a
18. In a class, the chance that a student takes Arabic but not French is .
2
2 Find the chance that
student takes Arabic, the chance that he takes French is .
5
in a sample of 72 students chosen at random between 18 and 32 of them take

both Arabic and French.

[Ans: 0⋅8852 ]

11
19. The lifetimes of electric bulbs are normally distributed with mean 1512 hours.

Given that 20% of the bulbs exceed a lifetime of 1600 hours. Find the:

(i) standard deviation of the lifetimes.

(ii) probability that in a sample of 100 such bulbs at least 25 of them have their

lifetimes exceeding 1600 hours.

[Ans: (i) 104⋅5131 (ii) 0⋅1303 ]

20. A machine produces bolts which are 10% defective. Find the probability that

in a sample of 400 bolts produced,

(i) at most 30 are defective

(ii) at least 55 are defective

[Ans: (i) 0⋅0567 (ii) 0⋅0079 ]

21. A student answers 72 questions. The chance of getting each question correct

is 2 . Find the probability of obtaining more than 50 correct answers.


3

[Ans: 0⋅266 ]

22. In a box containing 12 pieces of chalk, 3 are broken. If a random sample of 48

such boxes is taken, find the probability that between 10 and 17 boxes contain

broken pieces of chalk.

[Ans: 0⋅6247 ]

12

Common questions

Powered by AI

The critical value Z is used to determine the confidence limits of a binomial distribution when approximated by a normal distribution. It corresponds to the chosen confidence level and defines how many standard deviations from the mean the limits should be for that level of confidence. The confidence limits are given by np ± Z√(npq), where np is the mean and √(npq) is the standard deviation .

When using normal approximation for a discrete binomial distribution, continuity correction must be applied. For the probability P(X = x), a discrete value x is corrected to a normal interval by transforming it into an interval with boundaries x−0.5 and x+0.5. Thus, P(X = x) is approximated as P(x−0.5 < X < x+0.5) in the continuous context .

Recalibrating a pass mark necessitates analyzing the distribution of test results with respect to the difficulty changes, ensuring that the pass mark reflects desired competencies fairly. If the test's difficulty increases or decreases, the central tendency and spread of scores could shift, potentially warranting a new pass mark to maintain alignment with performance expectations. Modifications could be based on statistical analyses like standard deviation, confidence intervals, and overall distribution, ensuring that it takes into account consistency in evaluation standards across cohorts .

Without continuity correction, the probability of obtaining at most x successes in n trials using normal approximation is calculated directly from P(X ≤ x) on the normal distribution curve. With continuity correction, this probability includes the continuity adjustment by evaluating P(X ≤ x + 0.5). This correction refines the approximation by accounting for the discrete nature of the original binomial distribution, leading to potentially more accurate results in certain scenarios .

Calculating the least number of individuals required for a certain confidence level in low probability events ensures adequate sample size to reflect the true probability of the event occurring. This is critical for statistical power, which affects the reliability of the study’s conclusions. For probabilities near the extremes (close to 0 or 1), smaller sample sizes might yield inadequately powered studies, resulting in higher chances of type II errors (failing to detect a true effect).

A binomial distribution X ∼ Bin(n, p) can be approximated by a normal distribution when both np > 5 and nq > 5, where q = 1 − p. Under these conditions, the binomial distribution becomes sufficiently symmetric and spread over a sufficiently large portion of the real number line to be approximated by a normal distribution with parameters µ = np and σ² = npq .

Using a normal approximation for a binomial distribution is advantageous when dealing with large sample sizes where direct computation becomes computationally intensive. Additionally, when both np and nq are greater than 5, the binomial distribution resembles the normal bell shape, making approximation more accurate and efficient for calculating probabilities and confidence intervals. This is particularly useful in hypothesis testing, quality control processes, and situations involving large population studies or surveys .

A continuity correction is necessary when approximating a discrete distribution, like a binomial, with a continuous distribution, like a normal distribution, especially when calculating exact probabilities (e.g., P(X = x)). However, for probabilities that inherently incorporate intervals (such as P(X ≤ x) or P(X ≥ x)), the continuity correction's impact may be minimal or less critical since the computed range might overlap adequately with the intervals' spill due to the shape of the continuous distribution .

When the sample size increases from 192 to 960 workers, while keeping the percentage of graduates (one in four workers) constant, the probability distribution of finding a specific number of graduates will shift according to the normal approximation of the binomial distribution. As n increases, the distribution becomes narrower due to increased certainty (lower variance relative to mean), thus the calculation will involve reevaluating based on the changed parameters to maintain accuracy in probability calculations, using np = 960 * 0.25 and nq = 960 * 0.75 .

The probability of a biased coin showing heads can be approximated using a binomial distribution, X ∼ Bin(n, p), where n is the number of trials, and p the probability of getting a head on a single trial. If p is different from 0.5, and both np and nq are greater than 5, normal approximation can be used with adjustments for continuity correction if needed. This model accounts for the asymmetry in probabilities and adjusts the approximation accordingly .

You might also like