Over the last few years we have received requests to review our approach to the way in which we manage populations on scheduled (or charter) air services that have not been sampled in a given period.
This page sets out how quarterly populations will be shared across survey interviews from the start of 2019.
Step 1
In period P, determine if population associated with (LHR-Scheduled-BA1-JFK) can be matched against a survey interview, continue to Step 2 if not.
Step 2
In period P, determine if population associated with (LHR-Scheduled-1-JFK) can be matched against a survey interview, continue to Step 3 if not.
Step 3
In period P, determine if there are any (LHR-Scheduled-BA-JFK) sampled interviews. If there are, share the population from (LHR-Scheduled-BA1-JFK) across all sampled flight (LHR-Scheduled-BA-JFK). Note that the share of BA1 populations will be dictated by the proportion of passengers on each (LHR-Scheduled-BA-JFK) service.
If there are no (LHR-Scheduled-BA-JFK) sampled interviews then continue to Step 4.
Step 4
A dummy scheduled record will can created to represent the population (LHR-Scheduled-BA1-JFK) that cannot be allocated. It should be noted that the creation of dummy records will mean such entries in the survey outputs will not have responses to many core questions. This means that in calculating a metric it will be necessary to exclude any NULL records.
Example:
| Purpose | Status | W1 |
|---|---|---|
| Business | Interview | 300 |
| Leisure | Interview | 200 |
| NULL | Dummy | 100 |
In this example the sum of the weights totals 600 (300+200+100). The proportion of passengers recorded as travelling for business purposes would be given by 300/500 = 60%. In this case the denominator will exclude and Dummy records. To determine the number of passengers that travel on business, working on the assumption that the proportion of business to leisure can be applied to dummy records, can then be calculated as 60% x 600 = 360 passengers.
This new approach of creating dummy records will replace the earlier approach used where unmatched population in period P would have been assigned into a neighbouring period ( e.g. P+1). It is intended that this will mean that the aggregation of weights in Period P, which will include dummy recorded, will now very closely match actual traffic levels reported in period P.
Provide page feedback
Please enter your comments below, or use our usual service contacts if a specific matter requires an answer.
Fields marked with an asterisk (*) are required.