2015 H2 Maths Paper 2, Question 5

Stratified sampling requires a sampling frame — a complete list of all the customers of the supermarket, divided into strata by age, from which a random sample can be drawn within each stratum. The manager has no such list, since customers are not registered members and simply walk in and out of the supermarket, so stratified sampling cannot be carried out.

The manager first divides the customers into age groups (the strata), for example under 20, 20–39, 40–59, 60 and over. Using known information about the proportion of the supermarket’s customers in each age group (e.g. from past records), the manager sets a quota — the number of customers to be interviewed — for each age group. The manager then stands at the supermarket and interviews customers as they are encountered, asking each one’s age group and cola preference, continuing within each age group until its quota is filled.

Note: within each quota, the customers interviewed are not chosen at random — the manager (or interviewer) is free to choose whom to approach.

The interviewer is likely to approach customers who look more approachable or willing to stop and answer, rather than choosing at random from that age group. This introduces bias, so the customers interviewed within each age group need not be representative of all customers of that age group in the supermarket.