Mean, variance and Poisson approximation to the binomial
Studying this course? Keep your progress in one place.
Key facts
- Binomial : and , so the variance is smaller than the mean.
- Poisson : .
- From a binomial mean and variance: , then .
- When is large and is small, .
- The approximation overestimates the spread, so it is unsuitable when is not small.
Binomial mean and variance
For , the mean is and the variance is .
If , with independent trials each with success probability , then and . Derivations are not required.
The variance is always smaller than the mean, because it is the mean multiplied by . The standard deviation is the square root of the variance.
Do not use for a binomial: only the Poisson variance equals the mean.
- Mean
- Variance
Worked example
. Find the mean, variance and standard deviation.
- 1
.
- 2
.
- 3
The standard deviation is .
. What is ?
Poisson mean and variance
Both equal .
If then and . The mean and variance are equal. Derivations are not required.
If the variance is and the standard deviation is .
This equality is a useful check: if a data set has a mean and variance that are very different, a Poisson model is probably unsuitable. A binomial has variance less than its mean. Compare the sample mean and variance as a quick test, but also consider the context.
- Mean
- Variance
| Binomial B(n, p) | Poisson Po(λ) | |
|---|---|---|
| Mean | ||
| Variance | ||
| Variance compared with mean | Smaller | Equal |
Binomial B(n, p)
- Mean:
- Variance:
- Variance compared with mean:
- Smaller
Poisson Po(λ)
- Mean:
- Variance:
- Variance compared with mean:
- Equal
. What is the standard deviation of ?
Finding n and p
Divide variance by mean to get , then find .
If you know and for a binomial variable, divide to find : .
Then find from .
Dividing mean by variance and using the result as is a common mistake: the ratio is . Always check that lies between 0 and 1 and that is a positive integer.
- 1
Divide variance by mean
9 ÷ 12 = 0.75 = 1 − p
- 2
Find p
p = 1 − 0.75 = 0.25
- 3
Find n
n = mean ÷ p = 12 ÷ 0.25 = 48
Worked example
A binomial variable has mean and variance . Find and .
- 1
, so .
- 2
.
- 3
Check: and .
A binomial variable has mean and variance . Find and .
Poisson approximation
When is large and is small, .
When is large and is small, can be approximated by (the Poisson approximation). Derivations are not required.
It works because when is small, so the binomial mean and variance are nearly equal, as for a Poisson.
As a guide, should be at least about 50 and no more than about . State both conditions and give the new parameter .
- B(100, 0.03)
- Po(3)
- large, small
Worked example
. Use a Poisson approximation to find .
- 1
is large and is small, so approximate with .
- 2
.
- 3
For , .
is approximated by a Poisson distribution. Which one?
Assessing the approximation
It keeps the mean but overestimates the spread, so it is poor when is not small.
The approximation keeps the mean but replaces the variance by , so it overestimates the spread, by a factor of . For the effect is tiny.
For the mean is 6 and the variance 4.2, while has variance 6. Also is exactly but under the approximation, about 41% too large.
Conclude the approximation is unsuitable when is not small, using evidence from the numbers.
- B(20, 0.3)
- Po(6)
Why is a poor approximation to ?
Try an exam question
The random variable . Use a suitable approximation to find , stating why the approximation is appropriate.
[4 marks]
- [1] is large and is small, so the Poisson approximation is appropriate.
- [1], so .
- [1].
- [1].
That's the notes covered.
Carry on to the next subtopic.
Try a question
Find .
1 mark, marked against the mark scheme
Find .
1 mark, marked against the mark scheme