Question 1
1 markA car hire company wants to estimate the mean distance driven on the 12 000 hires it had last year. For each hire, its records show the customer's age and gender, the number of days the vehicle was hired for, and the type of vehicle (small car, family car or van).
The company will take a sample of the hires, stratified by one variable other than age or gender.
Suggest which variable it should use, and give a reason.
Model answer
The number of days of the hire (in groups, such as 1 day, 2 to 6 days and 7 days or more). A vehicle hired for two weeks is likely to be driven much further than one hired for a day, so stratifying by the length of the hire makes sure that short and long hires are each represented in proportion to how common they are.
Explanation
A good stratification variable is one that is likely to affect what is being measured, and one that is known for every member of the population before sampling. Here the population is the 12 000 hires, not people; the length of each hire is in the records and is likely to have a big effect on the distance driven, so stratifying by it stops short or long hires being over- or under-represented by chance.
Mark scheme
-
1 mark
B1 for the number of days of the hire (length of hire, in groups such as 1 day, 2 to 6 days, 7 days or more), with a reason, eg vehicles hired for longer are likely to be driven much further, so short and long hires must each be represented in proportion
oe. Accept type of vehicle with a reason linked to the distance driven, eg vans are often hired for short local moves, while family cars may be hired for long holiday journeys. The variable and a reason linked to the distance driven are both needed. Do not accept a variable that is not in the records (eg the purpose of the hire or the customer's home town), or a reason such as 'to make it fair' or 'to make it representative' with no link to the distance driven.