In reviewing the abstracts that were submitted, it became clear that several major themes emerged from the papers being proposed. While these topics are being investigated in many different ways a common sense of direction can be seen.

The content of the conference reflects:

• A concern that while we have made enormous progress answering the most pressing technical problems of CAI, we need to investigate the effects of these new methods and technologies on sampling, interviewer practice, interviewee response and ultimately the quality of the data collected. CAI not only allows us to administer questionnaires in new ways, it also allows us to collect a whole range of new data about the conduct of the survey. How can we best use this new wealth of information?
• The progress being made in stitching data together. CAI has opened new opportunities for data collection, but the exploitation of these facilities creates new problems, for example in the effective documentation of survey instruments. Beyond individual CAI applications the second great step that is explored in many of the conference papers is how and why we should link data from different sources together and make it more accessible outside the confines of individual organisations. There are several good examples of practical attempts to tackle the problems that data linkage creates which show that we are fast moving from theory to application.
• Serious attempts to make qualitative information more accessible and subject to rigorous analysis.

There are also excellent papers on data analysis, good practice and a few that we hope will make people think in new directions.