Diver Sampling

Diver Sampling, for on-street market research

In 2001, as part of the London Congestion Charging baseline monitoring, Michelle Diver, then of the TfL Congestion Charging Monitoring scheme, together with Rob Sheldon of Accent Market Research, and Professor Peter Jones, then of the University of Westminster, developed an innovative sampling method that would ensure that the demographics of the sample set would be representative of the people who were walking down the street.

This Diver Sampling was used in various on-street surveys for monitoring the impacts on attitudes of Congestion Charging, and in subsequent surveys across TfL, because of the robustness of its results, and the ease of implementation.

Technical Method

Diver Sampling is a method of dynamically allocating quotas within a continually-adjusted stratified random sample.

Interviewers work in teams: one quota-setter/counter, and a group of interviewers.

The survey period is divided into sampling intervals, typically of 30 minutes or an hour. For a period of time equal to one-third of the sampling interval (10 or 20 minutes respectively) the quota-setter classifies every n^thperson who passes into a stratum specified by the survey protocol (typically based on apparent age, gender and possibly type of clothing). At the end of this period, quota sheets are prepared for the interviewers for the next sampling interval.

The quotas are based on shift targets of for each interviewer and are updated through the shift. They are set by dividing each stratum by the total and multiplying the result by the target number of interviews per interviewer per shift.

The quota setting is based on cumulative data. So, for the second sampling interval, the total observed flows from both sampling intervals are used to set targets for the next interval.

This method ensures that the interviewed sample is representative of pedestrians passing through the study area during the fieldwork period.