Background. It is not known if a subject's characteristic level of self-rated depression symptoms index their genetic or environmental liability to major depressive disorder when measurement error and other occasion-specific influences are taken into account.

Method. Monozygotic (N = 408) and dizygotic (N = 295) adult female twin pairs from a population-based registry were surveyed twice with an average follow-up interval of 61 months. At each occasion subjects completed a structured clinical interview (SCID) to assess lifetime history of major depression and the subject-rated Symptom Check List (SCL) to assess current level of depressive symptomatology. A bivariate measurement model was used to estimate the genetic and environmental correlations between liability to reliably diagnosed lifetime history of major depression and the characteristic or temporally stable SCL depression score.

Results. The genetic and non-familial environmental correlation between liability to reliably diagnosed major depression and the characteristic level of SCL depression symptoms (and the proportion of variance shared between measures) is +0·70 and +0·24 respectively.

Conclusions. When allowance is made for diagnostic unreliability and temporal fluctuations in the level of subject-rated symptoms, 70% of the variance in genetic risk factors and 24% of the variance in environmental risk factors is shared by a diagnosis of lifetime major depression and total SCL depression symptom score. SCL depression scores may therefore be a useful screening measure for many of the genetic risk factors which influence liability to major depression.