12 monthly samples of 1,250 telephone numbers constitute the "incoming" fraction, with each unit surveyed for three consecutive months.

However certain households are not surveyed : impossibility to contact them, refusals... In general about 1,700 questionnaires are completed each month.

A methodological operation was carried out in April 2008 to measure the impact of survey base coverage bias.

The result was that the survey provides data of sufficient quality for short-term analyses.