CORRECTION OF LICENSE PLATE SURVEYS FOR SPURIOUS MATCHES

Abstract To economize and to keep errors of recording in check, it is common practice in license plate surveys to record only a part of the license plate code. License plate records collected at two points in time or space are then searched for matching records. Due to the comparison or partial license plate codes some matches occur between records belonging to different vehicles. In this paper the probability distribution of such spurious matches is derived. On this basis, the importance of the phenomenon is explored. A procedure is given for the correction of survey results and the assessment of the confidence limits for corrected results. The procedure is supported by appropriate graphs and tables.