Sampling methodology

Conducting the Departing Passenger Survey

CAA surveys generally operate for twelve month periods with anywhere from 3,000 to 70,000 interviews being achieved, depending on airport size. The surveys usually follow a stratified sampling design (stratified by carrier, route and quarter), with interviews taking place in the gate rooms. The interviews are then weighted to actual traffic levels. (Less often, a random sample is obtained by interviewing at points in the lounge - in this instance a further level of weighting is required to account for the sampling proportion obtained during the shift).

Surveys usually run on an annual basis, with shifts spread as evenly as practical throughout the year. A record of the flights sampled is regularly monitored to ensure that where possible all routes, and in most cases all flight numbers, are covered regularly.

Departing passengers are selected for interview using a strict sampling system. Interviews are weighted to the two-way passenger flow, under the assumption that over the period, departing and arriving passengers will show the same characteristics. Information is collected from passengers by a skilled interviewing field force.

An interview usually lasts 5 - 7 minutes, and involves around 30 questions being asked. All passengers are eligible, apart from children under two years of age.

It is assumed that passengers who are chosen for interview, but are unwilling or unable to respond (e.g. because of time constraints) have the same characteristics as those who are successfully interviewed. Thus no adjustments are made for non-response, either by item or by questionnaire.

The notes below outline the sampling methodology that is used when interviewing in the gate room. This method can be used irrespective of the number of team members present and ensures that a random selection of passengers are selected for interview.

1. The interviewing area where passengers are sampled should be clearly defined to all team members present.
2. The interviewing area is divided into a number of virtual areas, dependent upon the number of team members available to interview. If there are four team members, the interviewing area should be divided into quarters, where there are only three team members, then the area should be divided into thirds and so on.
3. Each interviewer is assigned an area. It will then be that interviewer's responsibility to sample passengers who move into their area.
4. It is of key importance that interviewers do not avoid certain passengers or show any favouritism to other groups of passengers.
5. On selecting the first passenger (perhaps to the extreme left of the interviewing area) the interviewer will run through the questionnaire in the normal manner. On the completion of this interview, the interviewer will then count three more passengers (in this example, towards the right). The third passenger counted, assuming that they are eligible for interview will then be interviewed. Once this interview has been completed, the interviewer will then count a further three passengers and begin the process again. The same passenger cannot be interviewed twice.
6. The interviewing areas can be redefined during busy or quiet periods in line with the number of team members available to work.