Abstract
Research indicates a robust association between personality and substance use and misuse. The high prevalence and pervasive detrimental impacts of alcohol use disorder (AUD) and smoking of tobacco necessitate more studies designed to identify factors closely associated with these outcomes in specific populations. The analyses reported in the present paper concern the relative utilities of five measures of personality and personality pathology rated by three sources (self, informant, and interviewer) in predicting AUD and regular smoking in a representative sample of 987 older adults, an understudied and uniquely vulnerable population. All measures and sources contributed to the predictions, with notable parallels as well as some important differences identified across substances and sources of information. In particular, low agreeableness robustly predicted AUD and smoking across self- and informant-reports. High interviewer-rated borderline personality pathology also strongly predicted AUD. Model fit indices suggested that measures of personality and personality pathology have stronger utility in predicting AUD as compared to regular smoking. These findings have important implications for the assessment of older adults in research and clinical settings and for the understanding of enduring risk factors for substance misuse later in life. Multi-source personality information is valuable for generating a complete picture of the relationship between personality and substance misuse.