A central problem for health education is the lack of baseline data concerning relevant issues such as smoking, alcohol consumption or nutrition. In the absence of this information there is little detailed knowledge of the public's habits or willingness to change those habits. Furthermore, the social context and spatial variation of these factors is unknown. Often the only way in which suitable information can be acquired is through social surveys. This paper uses a case study of a nutrition survey to outline a general framework for the collection of baseline data. The approach centres around postcode based data linkage. The effectiveness of the approach is considered and its shortcomings noted.