One method would be to sample different age groups, because opinions can vary greatly depending on age. Further differentiation may be to split the age groups into male and female and sample the gender groups separately. Typical age groups may be 18-29, 30-41, 42-53, 54-65, and, perhaps, 66 and over. The method would be to use a questionnaire which includes gender and age questions. We can expect the individual standard deviations for each group to be smaller than the standard deviation of the population of indiscriminate ages and genders. The opinion survey must contain a sufficient number of returned valid and completed questionnaires to produce reasonable sized samples (for the example of groups there would be 10 groups with an average sample size of about 150) so the pollster may need to conduct the survey over a number of days at the airport. The use of an airport must be independent of the subject of the survey. Clearly, an airport would be an unsuitable place to conduct an opinion poll on modes of transport! After statistical analysis of the sample groups it will be clear whether the standard deviations of the groups are significantly different from the standard deviation of the population of 1500. This will determine whether or not the stratified method was effective.
