PANELS: ATTRACTIONS AND PITFALLS

In recent years a common feature of discussions about data collection and policy needs has been the recognition of the shortcomings and limitations of cross-sectional data and the valuable additions and attractions of longitudinal data sets. Suddenly everybody is talking about panels and their potential attractions, because they could make such a useful contribution to answering ever more complex policy questions. The costs which have to be committed to such projects has however limited their development in the field of transport. This paper will discuss the attractions of panels and their potential application to transport policy problems. It will then consider the problems of successfully realising such a project by drawing on a variety of experiences. A number of these problems are well known generally, such as the panel mortality and the learning and conditioning effects. There are also some more subtle problems, however, which are less well known, but which can seriously prejudice the value of the results (e.g. The over-representation of more mobile people). Some existing experience will be reviewed. It will be argued that attention to, and awareness of, some of the likely methodological pitfalls could prevent false conclusions from being drawn. (Author/TRRL)