PUNISHMENT IS DEFINED AS AN EDUCATIONAL MEASURE WHICH--CONTRARY TO POSITIVE REINFORCEMENT--REDUCES THE PROBABILITY OF THE FUTURE REOCCURRENCE OF PARTICULAR BEHAVIOR. TWO TYPES OF PREVENTIVE EFFECTS ARE DISTINGUISHED: INDIVIDUAL PREVENTION (I.E., THE EFFECT ON THE INDIVIDUAL EXPERIENCING PUNISHMENT) AND GENERAL PREVENTION (I.E., THE EFFECT ON OTHERS NOT DIRECTLY EXPOSED TO PUNISHMENT). IN TERMS OF INDIVIDUAL PREVENTION, LEARNING EXPERIMENTS SHOW THAT PUNISHMENT IS MOST EFFECTIVE IF THE OFFENDER (1) UNDERSTANDS THE DIRECT RELATION BETWEEN UNDESIRABLE BEHAVIOR AND PUNISHMENT, (2) EXPERIENCES THE PUNISHMENT WHENEVER THE OFFENSE OCCURS, AND (3) CONSIDERS THE PUNISHMENT DISAGREEABLE. IN TERMS OF TRAFFIC VIOLATIONS, THIS MEANS THAT A TICKET WILL HAVE THE GREATEST PREVENTIVE EFFECT IF IT FOLLOWS IMMEDIATELY UPON THE VIOLATION, IT OCCURS MORE THAN ONCE, AND THE OFFENDER CONSIDERS THE LOSS OF TIME OR MONEY AS UNPLEASANT. EMPIRICAL RESEARCH INDICATES THAT PUNISHING TRAFFIC OFFENDERS ALSO POSSESSES A GENERAL PREVENTIVE EFFECT: AFTER SEVERAL DRIVERS HAD BEEN TICKETED IN A PARTICULAR AREA, OTHER COMMUTERS WHO HAD NOT RECEIVED TICKETS SIGNIFICANTLY IMPROVED THEIR DRIVING. SINCE PUNISHMENT CAN ONLY IMPROVE DELIBERATE BEHAVIOR, EXPERTS MUST DETERMINE WHICH VIOLATIONS ARE COMMITTED KNOWINGLY (PROBABLY SPEEDING) AND WHICH RESULT FROM OVERSIGHT (PROBABLY MOST CASES OF GOING THROUGH A RED LIGHT). IN THE SPECIAL CASE OF DRUNKEN DRIVING, ONLY THE SO-CALLED 'SOCIAL DRINKERS' ARE LIKELY TO PROFIT FROM PUNISHMENT, WHILE ALCOHOLICS WHO HABITUALLY DRIVE UNDER THE INFLUENCE OF ALCOHOL TEND TO BE UNAWARE OF THEIR ILLEGAL BEHAVIOR AND CONSIDER PUNISHMENT A STROKE OF BAD LUCK. NO REFERENCES ARE INCLUDED. --IN GERMAN. (SAJ)