Question 1
1 markEach day, the manager of a small shop chooses 5 of her customers at random and records the number of items each of them buys. She uses the mean of these 5 numbers as her estimate of the mean number of items bought by a customer.
Her estimates for four days are 3.2, 7.8, 4.4 and 9.0 items.
Suggest how the manager could get a more reliable estimate of the mean number of items bought by a customer. Give a reason for your answer.
Model answer
She should use a much larger random sample, for example 50 customers each day, or combine the samples from several days. With only 5 customers, one or two who buy a lot (or very little) change the mean a lot, which is why her estimates vary so much. A larger sample generally gives a better estimate of the mean for all customers.
Explanation
Just as more trials give a better estimate of a probability, a larger random sample generally gives a better estimate of a population mean. The estimates from samples of 5 vary a lot (from 3.2 to 9.0) because a few unusual customers have a big effect on a small sample. The reason must be about the size of the sample.
Mark scheme
-
1 mark
B1 for use a larger (random) sample, e.g. 50 customers, or combine the samples from several days, because a larger sample generally gives a better estimate of the (population) mean
The suggestion and the reason are both needed. Accept 'ask many more customers, so that one or two unusual customers have less effect on the mean' oe. B0 for a larger sample with no reason, or with a reason not about sample size (e.g. 'so it is fair'). B0 for suggestions that are not about sample size, e.g. 'use the median' or 'choose customers who buy a lot'. Ignore extraneous non-contradictory comments.