<Minnerva>

Minnerva has used a wide range of ideas, techniques, and data resources to tackle the issue of minimising costs for household travel surveys. A starting point is the application of stratified sampling methods, for which Census data provides a conventional sampling framework. While this approach is suited to collecting general information, it is insufficiently precise when the results are required to support transport policy analysis. For example, a general sample is unlikely to provide a statistically significant number of rail travellers when rail services are only available in a limited part of the study area, yet the sample needs to be representative of the area as a whole.

Minnerva therefore developed a stratification process with two dimensions that were respectively named as ‘Locality’, which predominantly reflected areas where different types of people lived, and ‘Transport Areas’, which were determined by the nature of the transport infrastructure in the area.

An important element of the household survey costs relates to the time enumerators spend getting to the households selected for survey. It is therefore valuable to provide geographically clustered samples that still maintained the general level of representation of the population as a whole. This has been provided through bespoke software that enabled different levels of clustering to tested, with results tables indicating the strengths and weaknesses of different degrees of clustering, so that an optimal balance was readily achieved.

The system involved many different elements, with MapInfo GIS being an important element, but also included Ordnance Survey AddressPoint data, whose three quarters of a million records for the area where held and processed in an MS SQL Server database. The household survey data and other elements were held in an MS Access database. The GIS was connected to both databases, as was the bespoke sampling software that was written.

The system was developed for Essex County Council, in association with Mouchel Essex.

(Contact: Miles Logie)