To examine those persons engaging in multiple health-risk behaviors, we first computed odds ratios between each of the four health-risk behaviors. The statistically significant associations in Table 2 point to the clustering of different health-risk behaviors into a "health-risk syndrome" in the same persons.Persons who engaged in one health-risk behavior were more likely to engage in other health-risk behaviors. To identify those persons engaging in multiple health-risk behaviors, we computed a cumulative health-risk index, assigning a point to each sample member for their involvement in each of the four healthrisk behaviors. The majority (78.4%, n = 728) of young adults in the sample did not engage in health-risk behaviors; 16.3% (n = 151) engaged in only one type of health-risk behavior, and smaller proportions of the sample (summing to 5.3%, N = 49) engaged in a syndrome of multiple health-risk behaviors.Next, we compared the personality profiles of three groups of young adults: Those who did not engage in any health-risk behaviors, those who engaged in one health-risk behavior, and those who engaged in two or more health-risk behaviors. Figure 2 shows the mean (z) scores for the three groups on each of the MPQ scales. A MANOVA revealed significant omnibus differences between the groups, F(20, 1742) = 5.20,p < .01, even after controlling for gender differences. Univariate ANOVAs (df — 2, 882 and a = .01) showed that six personality traits discriminated between the three age 21 groups. Persons who at age 18 scored lower on Traditionalism (d = .77), Harm Avoidance (d = .67), Control (d = .79), and Social Closeness (d = .69), and higher on Aggression (d = 1.36) and Alienation (d = .70) were significantly more likely to become involved in a healthrisk behavior at age 21. These significant effects were further analyzed with the Scheffe post hoc test (a = .01). Each of the age 18 personality traits discriminated between persons who engaged in no health-risk behaviors and persons who engaged in either one or in multiple health-risk behaviors. Although persons who engaged in multiple health-risk behaviors had more extreme personality scores than persons who engaged in a single, isolated health-risk behavior, these differences—with the exception of MPQ Aggression—did not reach statistical significance at conventional levels. To summarize at the superfactor level, persons who at age 18 scored low on Constraint (d = 1.01) and high on Negative Emotionality (d = .94) were significantly more likely to be involved either in one or in multiple health-risk behaviors at age 21.