This fellowship would provide me the opportunity to improve my skills in and apply advanced statistical methodology--longitudinal multivariate behavior genetic models. These skills will allow me to investigate environmental and genetic influences in late life depression. I plan to continue using these skills to investigate depression and suicide in older adults following completion of the fellowship. Proficiency in these advanced statistical procedures, particularly longitudinal methodology, is likely to be highly valuable in a career in geropsychological research. My career objective is a faculty position in a research-focused academic setting. The program of research I am currently interested in pursuing is the study of predictors and correlates of depression and suicide in late life. An additional area of research interest is dementia. The proposed study will be a longitudinal examination of the genetic and environmental influences on depressive symptoms from mid-life to old age. It seeks to test a developmental diathesis-stress model of late life depressive symptoms, in which a genetic predisposition to depression interacts with life events over time to predict depressive symptoms. Previous research has shown a relationship between health-related events and depressive symptoms in old age, although the relationship between other life events and depressive symptoms in this age group has been equivocal. In younger populations, on the contrary, non-health related events are significantly associated with depressive symptoms. No studies have been found which test changes in these relationships with age longitudinally. The influence of genetic factors on depressive symptoms has been well established in both older and younger samples, but changes in genetic influences with age have not been tested longitudinally. The proposed study will be based on data from a large longitudinal study of twins, reared either together or apart. Participants range in age from 26 to 94. Life events and depressive symptoms were measured on several occasions over nine years. Analyses will include intraclass correlations, linear regression and multivariate behavior genetic models.
