Statistics: Statistical Distributions (Binomial/Normal)
Statistical distributions describe how the probability of an outcome is spread across all its possible values. A Level Statistics focuses on the binomial distribution B(n,p) for independent trials and the normal distribution N(mu, sigma^2) for continuous data, including standardising values, finding unknown parameters, and approximating a binomial by a normal distribution with a continuity correction.
Method
- Check the binomial conditions apply: a fixed number of trials n, each with only two outcomes, a constant probability of success p, and independence between trials.
- Calculate a binomial probability using P(X = r) = nCr x p^r x (1-p)^(n-r), or sum several such terms for P(X <= r) type questions.
- For a normal distribution X ~ N(mu, sigma^2), standardise using z = (x - mu)/sigma, then use the standard normal distribution (tables or a calculator) to find the required probability.
- To find an unknown mean or standard deviation from a given probability, find the z-value matching that probability (using tables or an inverse normal function), then rearrange z = (x - mu)/sigma to solve for the unknown.
- For a normal approximation to a binomial X ~ B(n,p), check that np > 5 and n(1-p) > 5, then use Y ~ N(np, np(1-p)).
- When using a normal approximation to a binomial (a discrete distribution) with a continuous model, apply a continuity correction, e.g. P(X >= r) becomes P(Y > r - 0.5).
Common mistakes
MistakeSwapping the powers in the binomial formula, writing p^(n - r) (1 - p)^r instead of p^r (1 - p)^(n - r).
InsteadThe power of p is the number of successes r, and the power of (1 - p) is the number of failures n - r. Check that the two powers add up to n before evaluating.
MistakeFinding P(X <= r) as P(X = r), or adding only some of the terms for an 'at most' probability.
InsteadP(X <= r) is the sum of P(X = 0) up to P(X = r), or one cumulative binomial function. For 'at least' use P(X >= r) = 1 - P(X <= r - 1) instead of adding many terms.
MistakeStandardising a normal variable by dividing by the variance instead of the standard deviation.
InsteadFor X ~ N(mu, sigma^2) the second number is the variance. Take its square root before dividing: z = (x - mu)/sigma.
MistakeWriting P(X > r) as P(Y > r - 0.5) when applying a continuity correction.
InsteadFor a whole-number variable, X > r means X >= r + 1, so the continuous boundary is r + 0.5: P(X > r) becomes P(Y > r + 0.5). It is P(X >= r) that uses r - 0.5.
Worked examples
Example 1
The random variable X ~ B(12, 0.25). (a) Calculate P(X = 3), giving your answer to 3 significant figures. (b) Calculate P(X <= 1).
- Use P(X = r) = nCr x p^r x (1-p)^(n-r) with n=12, p=0.25, r=3: P(X=3) = 12C3 x 0.25^3 x 0.75^9.
- 12C3 = 220, 0.25^3 = 0.015625 and 0.75^9 = 0.075085 (6dp), so P(X=3) = 220 x 0.015625 x 0.075085 = 0.258 (3sf).
- For P(X <= 1), find P(X=0) = 0.75^12 = 0.031676 and P(X=1) = 12C1 x 0.25 x 0.75^11 = 0.126705.
- P(X <= 1) = P(X=0) + P(X=1) = 0.031676 + 0.126705 = 0.158 (3sf).
Example 2
A survey finds that 35% of the residents of a large town cycle to work. A random sample of 180 residents is taken, and X is the number who cycle to work, so X ~ B(180, 0.35). Use a normal approximation with a continuity correction to estimate P(56 <= X < 70), giving your answer to 3 significant figures.
- Check the conditions: np = 180 x 0.35 = 63 > 5 and n(1 - p) = 117 > 5, so a normal approximation is suitable.
- The approximating variable is Y ~ N(63, 40.95), because np = 63 and np(1 - p) = 180 x 0.35 x 0.65 = 40.95. The standard deviation is sqrt(40.95) = 6.399.
- Apply the continuity correction to each end separately. X >= 56 means X > 55.5 for the continuous variable. X < 70 means X <= 69, so Y < 69.5. Therefore P(56 <= X < 70) becomes P(55.5 < Y < 69.5).
- Standardise: z = (55.5 - 63)/6.399 = -1.172 and z = (69.5 - 63)/6.399 = 1.016.
- P(-1.172 < Z < 1.016) = 0.8451 - 0.1206 = 0.7245.
- Final answer: P(56 <= X < 70) is approximately 0.725 (3 s.f.).
Example 3
Three proposed models have the following exact mean and variance pairs: A: E(X) = 5.4, Var(X) = 2.16; B: E(X) = 5.4, Var(X) = 2.7; C: E(X) = 5.4, Var(X) = 6.48. For each pair, determine whether X ~ B(n,p) is possible, where n is a positive integer and 0 <= p <= 1. Recover the parameters where possible and explain why checking p alone is insufficient.
- For X ~ B(n,p), E(X) = np and Var(X) = np(1-p). The mean is positive in all three cases, so np > 0. Dividing the variance by the mean gives Var(X)/E(X) = 1-p; this determines p first, then n = E(X)/p.
- For A, 1-p = 2.16/5.4 = 0.4, so p = 0.6. Then n = 5.4/0.6 = 9.
- Check both parameters: 0 <= 0.6 <= 1 and 9 is a positive integer. Substituting back gives E(X) = 9 x 0.6 = 5.4 and Var(X) = 9 x 0.6 x 0.4 = 2.16, so A is admissible.
- For B, 1-p = 2.7/5.4 = 0.5, so p = 0.5 and n = 5.4/0.5 = 10.8. Although p is a valid probability, a fixed number of trials must be an integer. B is impossible; rounding n would change the exact supplied moments.
- For C, 1-p = 6.48/5.4 = 1.2, so p = -0.2. A negative probability is inadmissible, so C is impossible. Its variance exceeds its positive mean, whereas np(1-p) <= np whenever 0 <= p <= 1.
- Final answer: A gives X ~ B(9, 0.6). B and C cannot be binomial models with the exact supplied moments: B needs a non-integer trial count and C needs a negative probability.
Practice questions
Q1[2 marks]State two conditions required for a random variable to be modelled by a binomial distribution.Show answer
Answer: Fixed number of independent trials, each with only two outcomes and a constant probability of success (any two of these).
Q2[2 marks]X ~ B(8, 0.3). Write down E(X) and Var(X).Show answer
Answer: E(X) = 2.4, Var(X) = 1.68 (np and np(1-p)).
Q3[3 marks]X ~ N(50, 8^2). Find P(X > 58).Show answer
Answer: awrt 0.159 (z = (58-50)/8 = 1, P(Z>1) = 0.1587).
Q4[2 marks]X ~ B(15, 0.4). Calculate P(X = 5), giving your answer to 3 significant figures.Show answer
Answer: 0.186 (15C5 x 0.4^5 x 0.6^10).
Q5[2 marks]X ~ N(20, 5^2). Find the value of a such that P(X < a) = 0.9, using z = 1.2816.Show answer
Answer: a = 26.4 (3sf) (20 + 1.2816 x 5 = 26.408).
Q6[2 marks]The random variable X ~ B(n, 0.2) has Var(X) = 6.4. Find the value of n.Show answer
Answer: n = 40 (Var = np(1-p) = 0.16n = 6.4, so n = 40).
Exam-style questions
Written in the style of a A Level Maths exam paper, with a full mark scheme.
Exam tips
- Before choosing a binomial model, match each condition (fixed number of trials, two outcomes, constant probability, independence) to a phrase in the question.
- Sketch a normal curve, shade the area you want and mark the mean. The shading tells you whether the answer should be above or below 0.5 before you calculate.
- When using a calculator's cumulative function, write the probability statement you evaluated, such as P(X <= 4), so the tail you found is clear.
Q1[4 marks]Zoe is a quality inspector at a factory producing glass bottles. From long-term records, 6% of bottles produced have a visible flaw. Zoe selects a random sample of 15 bottles. Let X be the number of flawed bottles in the sample. (a) State two conditions needed for X to be modelled by a binomial distribution. (2) (b) Given that X ~ B(15, 0.06), calculate P(X = 1). (2)Show mark scheme
Tick each line you got.
Nothing ticked yet - 4 available
Q2[5 marks]The heights of adult men in a country, H cm, are modelled by H ~ N(178, 7^2). (a) Calculate P(H > 185). (3) (b) Calculate P(170 < H < 185). (2)Show mark scheme
Tick each line you got.
Nothing ticked yet - 5 available
Q3[6 marks]A polling company surveys a random sample of 250 voters. National figures suggest 48% of voters support a proposed policy. Let X be the number, out of 250, who support the policy, so X ~ B(250, 0.48). (a) Give a reason why a normal approximation to X is appropriate here. (1) (b) Using a suitable approximation with a continuity correction, calculate P(X > 130). (5)Show mark scheme
Tick each line you got.
Nothing ticked yet - 6 available
This topic in real past papers
We have tagged some boards' past papers question by question (not every board or year yet). This topic comes up in:
- AQA A Level Maths past papers: 13 questions across 6 papers, latest tagged June 2023
- WJEC A Level Maths past papers: 7 questions across 5 papers, latest tagged Summer 2023
- OCR A Level Maths past papers: 3 questions across 2 papers, latest tagged June 2024
- OCR AS Level Maths past papers: 3 questions across 3 papers, latest tagged June 2024
Free printable worksheet
Want more practice on paper? Download the Statistics: Statistical Distributions (Binomial/Normal) worksheet pack - 7 pages of exam-style questions with a full mark scheme. Free to download, no card or account. Print it for personal and classroom use.
This topic is chapter 4 of A Level Maths Statistics Essentials, the whole course as one free printable PDF. Its worksheet is also chapter 4 of A Level Maths Statistics Topic Workbook and chapter 4 of A Level Maths 10-Minute Tests 2, free printable practice books.
Other versions of this worksheet: