The location of the SSPO study at the margins of two different models of knowledge-production – one dealing with personal and social relations, the other with the quantification of the body in the discourse of the medical and physical sciences – had many implications for the way in which the study was done. The different methodologies embedded in these alternative models are usually described as quantitative and qualitative, and they are distinguished from each other in three main ways. First, with regard to the definition of data, quantitative research isolates and defines variables which constitute data, pre-specifying hypothetical relationships between variables before data collection takes place; the qualitative approach, on the other hand, is concerned with categories rather than variables, and these are expected to change their nature and definition as the research progresses. Secondly, the qualitative tradition obliges the researcher to pay some attention to the assumptions and values of those carrying out the research and collecting the data, as these are viewed as constituting part of the research data themselves; while the quantitative researcher labours in the belief that the world of personal values is separate from the world of scientific work. Thirdly, qualitative research mines the terrain while quantitative work maps it: quantitative research is geared towards the generalizability of research findings, and qualitative research to the exploration of the internal relations of, and links between, categories and processes.

In the SSPO study, these traditional methodological divisions could be summarized as follows:

Quantitative methods used:

• 1. An experimental research design.

• 2. Use of a table of random numbers to allocate subjects to intervention and control groups.

• 3. Pre-specified hypotheses as to relations between key variables.

• 4. Use of pre-coded structured research instruments.

• 5. Quantification of data.

• 6. Effectiveness of intervention assessed by means of pre-specified ‘hard’ outcomes.