Abstract
The non-obligatory nature of a number plate-interview and the hand-out-interview seems open to objection. If the classic roadside-interview proves to be the only method of obtaining a random sample, then it must be applied also with high traffic intensities. The funnel-theory will in such cases render good service. This paper was presented at the working days on traffic engineering 1981 (see IRRD abstract no 255833) but is not published in the conference papers.