Before embarking on the collection of a new set of data, it is prudent to ascertain just how much is already known on the subject in question. Whilst this would appear to be a commonsense step, it is often overlooked, with the result that many surveys either obtain data which simply duplicates that which is already available, or else obtain data which is not completely appropriate to the problem at hand. The review of existing information can provide assistance in two major ways.
First, it may be possible to unearth existing relevant data sources. If one is extremely fortunate, it may be possible to find data which are exactly those required, and hence the need for a further survey is completely obviated. In many cases, however, even though appropriate data is found it may still be necessary to re-analyse the data in the light of the problem at hand. Wigan (1985) gives an excellent review of some of the opportunities and pitfalls involved in such secondary use of transport survey data.
Often not all the required data is found, and hence the need for a further survey is only partially removed. Whilst not removing the need for further data collection, other data sources may be of great assistance in the design of that survey. For example, estimates of population variance for relevant parameters from a previous survey may assist in sample size calculations, while knowledge of the composition of the population (e.g. what percentage of people own cars) may assist in stratified sample design. Finally, data may be found in a different form but one which nevertheless will be useful as a cross-check on the accuracy of the data which is subsequently collected. Many sources of existing data are available for transport planning purposes and a comprehensive list of such data sources would be extremely long - and would vary from country to country. An example of such data sources, compiled for Australian conditions by Bowyer (1980), is shown in Table 2.1. Axhausen (1994) has also prepared a very comprehensive list of major travel surveys which have been carried out throughout the world since the 1950s.
The second way in which an information review might provide assistance is in the revelation of methodological procedures which may be appropriate in the survey. The art of survey design is based on the acquisition of knowledge through experience, and there is no substitute to learning from one's mistakes in the field. Unfortunately, many survey designers appear to take this generalisation to extremes, with the result that their surveys contain the same basic mistakes that have been made in hundreds of previous surveys. Whilst there is no substitute for experience, there are many pitfalls which have been reasonably well documented in the literature. A cursory review of this literature will help avoid some of the common faults in survey design. There is a problem in
transport surveys in that the literature is widely dispersed. Often the survey method is only briefly described as an adjunct to the main topic of the paper which is usually the analysis of the survey data to assist in policy formulation. Nonetheless, there is a growing body of literature which will assist in suggesting survey techniques which are most appropriate for particular problems. For example, the 1981 Sydney Travel Survey adopted the use of a verbal activity recall framework on the basis of work done by Jones and colleagues in the late 70s (reported in Jones et al., 1983) and in response to dissatisfaction with other trip recall techniques used in the 1971 SATS survey.
Table 2.1 Sources of Existing Transport Data
SOURCE VALUE LIMITATIONS
National Bureau of Statistics: Population Census
Provide a broad outline of population groups and characteristics
Limited transport-related items. Individual records not available. Urban Transport Studies;
Home Interview Surveys
Detailed person and household travel data
Little data on non-travel activities nor on likely attitudes to future transport options. Family Expenditure Surveys Household expenditure
behaviour.
Small nationwide sample; not yet suitable for longitudinal studies.
Community Services; Patient and attendee records at hospitals, welfare centres, schools, play groups.
Identifies "disadvantaged" groups; design and operation of small paratransit services.
Does not cover
"disadvantaged" persons who are not on the records. Some records not easily accessed. Employers and Unions;
Employee and Membership data
Basis for identifying potential users of employer-based transport schemes.
Possible problems with privacy and availability.
Operator Data; Transit and Taxi Systems.
Identifying and monitoring users of these services
May not reflect characteristics and attitudes of non-users. Political "Network"; Petitions
by individuals or groups to government.
Enables local "grass roots" needs to surface.
Danger of demands
snowballing or wrong solution being fitted to an expressed need.
Monitoring Schemes; Tracking Patronage and Productivity.
Can aid in "fine-tuning" a particular scheme and provides research data for other schemes.
Requires commitment to monitor over adequate time period, and a statement of success criterion.
Universities; Institutes; Special Purpose Studies.
Exploratory "research value" insights into population attitudes and activity/travel behaviour.
Location specific.
Government Departments and Authorities (e.g. Electric Utility Companies, Motor Vehicle Registrations)
Survey sample frames Access to files
Source: Bowyer (1980)
In addition to identifying survey techniques and practical pitfalls, a preliminary literature review should also be conducted with respect to the subject of the survey. This may enable the survey objectives to be specified more precisely, in the light of both previous survey results and also of theoretical considerations with respect to the subject matter.