CURRENT KNOWLEDGE ABOUT THE LIFE-HISTORY OF THE HEROIN USER IS DISCUSSED, AND IT IS INDICATED THAT MOST INDIVIDUALS WHO EXPERIMENT WITH HEROIN AND OTHER NARCOTICS DO NOT PROCEED TO COMPULSIVE USE. FURTHER, IT HAS BEEN FOUND THAT MANY INDIVIDUALS WHO DO CONSUME NARCOTICS ON A REGULAR BASIS ARE ABLE TO REDUCE OR ELIMINATE THEIR DEPENDENCE ON DRUGS WITHOUT TREATMENT INTERVENTION. IT IS POINTED OUT, IN THE LIGHT OF THESE FINDINGS, THAT DESIGN AND ANALYSIS STRATEGIES THAT CAN COPE WITH THE MORE COMPLEX SEQUENCE AND STAGES THAT ACCOMPANY DRUG USE MUST BE DEVISED. THE WEAKNESSES OF PRESENT EVALUATION DESIGNS ARE REVEALED, AND SUGGESTIONS FOR IMPROVEMENT ARE PROPOSED. IT IS SUGGESTED THAT TYPE OF ANTI-SOCIAL BEHAVIOR, RATHER THAN A GROSS OVERALL MEASURE, SHOULD BE REPORTED. IT IS ALSO CONSIDERED IMPERATIVE TO ESCHEW THE MAN/YEARS MEASURE, AND REPLACE IT WITH REPORTAGE OF AN INDEX OF CRIME, OR OF PROPORTIONS CHARGED WITH SPECIFIC CRIMES, WITH CONTROL ON LENGTH OF TREATMENT. ORGANIZATION OF THE ANALYSIS SO THAT HOMOGENEOUSLY GROUPED INDIVIDUALS ARE COMPARED IS CONSIDERED ESSENTIAL. IT IS BELIEVED POSSIBLE TO ESTABLISH FIRMLY WHETHER THERE HAVE BEEN SIGNIFICANT CHANGES IN THE PARTICULAR GROUPS EXAMINED...RCB