THE ANALYSIS IS PART OF A LARGER LONGITUDINAL RESEARCH PROJECT IN WHICH APPROXIMATELY 12,000 TO 14,000 STUDENTS IN GRADES 4 TO 12 WERE STUDIED EACH YEAR FOR 5 YEARS IN SUBURBAN SCHOOL SYSTEMS IN THE BOSTON, MASS., AREA. THE STUDENTS IN THE MARIHUANA USE SAMPLE ARE FROM ONE OF THE SCHOOL SYSTEMS AND WERE 7TH OR 8TH GRADERS IN 1969 AND 11TH OR 12TH GRADERS IN 1973. THE SAMPLE GROUP IS PREDOMINANTLY WHITE AND MIDDLE-CLASS. ALL 651 STUDENTS IN THE SAMPLE REPORTED THAT THEY WERE NONUSERS OF MARIHUANA IN 1969. ONE HUNDRED AND TWENTY-EIGHT STUDENTS INITIATED MARIHUANA USE AS 9TH GRADERS OR EARLIER. THREE HUNDRED AND SEVENTEEN STUDENTS BEGAN USING MARIHUANA IN THE 10TH GRADE OR LATER. TWO HUNDRED AND SIX STUDENTS REMAINED NONUSERS THROUGH THE FINAL YEAR OF THE STUDY. THE ANALYSIS IS BASED ON RESPONSES TO A 400-ITEM PERSONALITY QUESTIONNAIRE DESIGNED TO MEASURE PERSONAL COMPETENCE AND SOCIAL RESPONSIBILITY. OTHER VARIABLES INCLUDE EMOTIONALITY, CIGARETTE SMOKING AND RELATED ATTITUDES, AND ACADEMIC PERFORMANCE. ANALYSIS OF SELF-REPORT AND PEER-RATING DATA SHOWS THAT EARLY USE IS ASSOCIATED WITH LOW SCORES ON PREUSE MEASURES OF PERSONAL COMPETENCE, SOCIAL RESPONSIBILITY, AND ACADEMIC PERFORMANCE, AND WITH HIGH SCORES ON PREUSE MEASURES OF EMOTIONALITY, CIGARETTE SMOKING, AND FAVORABLE ATTITUDES TOWARD CIGARETTE SMOKING. THE FINDINGS INDICATE THAT PRECOCITY OF MARIHUANA USE CAN BE PREDICTED WITH STATISTICALLY SIGNIFICANT ACCURACY FROM PSYCHOSOCIAL VARIABLES SIMILAR TO THOSE THAT PREDICT DEGREE OF SUBSSEQUENT DRUG INVOLVEMENT. SUPPORTING DATA AND A LIST OF REFERENCES ARE INCLUDED. (LKM)