The Mortimer-Filkins test has been used widely as an instrument for detecting problem drinkers among drink-driving offenders. While extensive psychometric testing has been undertaken by the developers of the test, few independent validation studies have been conducted, and few studies have used the Mortimer-Filkins test with general populations. The present study investigated the test-retest and internal-consistency reliability of the instrument, and the stability of problem drinking, as measured by the instrument. The test was administered to moderate and heavy drinkers at an industrial workplace on three occasions. The results indicated that the Mortimer-Filkins test has high test-retest and internal-consistency reliability, and problem drinking, as measured by the test, appears to be a stable characteristic across time.