Table 1 Data collection: sampling frame, success criteria and invitation strategy for sampling.

Data collection.

For each measurement criterion we compared the individual baseline scores with subsequent years.

Success criteria.

Criteria measured: 'relative to the individual baseline score, maintained for two or more years.'

The activity diaries covered 24 hours for three days and activity was scored in half hour divisions on a MET score scale.

Increased activity: (n = 20, Control = 7, Intervention = 13) 'More than 10% increase in activity.'

The food diaries covered three days and were analysed using Microdiet software to determine total calories consumed.

Calorie reduction: (n = 13, Control = 5, Intervention = 8) 'More than 20% calorie reduction.'

Calorie intake attributable to fat consumption was calculated from the food diary analysis.

Fat reduction: (n = 18, Control = 9, Intervention = 9) 'More than 5% fat reduction (calorie equivalent).'

Invitation strategy

A 'pool' of potential participants was selected on the basis of the diary data success criteria (n = 38). We excluded seven for specific reasons (e.g. 1. her husband had just died, 2. she had just had extensive surgery, 3. he had already completed five years in the trial and had requested no further study invitations) and we added three, randomly selected from the trial participants who had not achieved success as defined to broaden the range of experience considered. We then invited participants in groups of five. After each group acceptance we looked at the demographic data to maintain an overall sample balance. For example: if 3 females accepted in the first group and 2 males declined, we would invite proportionately more males for the next group so that the demographics (age, gender, trial group, success measure and trial years completed) of our interviewed sample reflected the demographics of the sample pool.