A taxometric investigation of unipolar depression in a large community sample

The question of whether unipolar clinical depression differs categorically from limited depressive complaints has important implications for the disorder's assessment, treatment and research. This crucial issue has proven difficult to resolve, in part because many studies to date have relied on self-report measures or on clinically homogeneous samples. We therefore applied Meehl's taxometric method to a large, clinically heterogeneous sample, and examined the latent structure of depressive episodes using both self-report and structured clinical interview data.

Data were derived from the Oregon Adolescent Depression Project, a large longitudinal community study. All analyses involved more than 1400 participants. MAXEIG (MAXimum EIGenvalue) and base rate estimation were performed separately for Beck Depression Inventory (BDI) items and for DSM-IV-based major depressive episode (MDE) symptoms.

MAXEIG analyses of the BDI and MDE indicator sets appeared to converge on a taxonic structure for unipolar depression. Base rate estimates overall implied a latent depressive episode class that occurs more frequently than diagnosable MDEs but less frequently than persistent depressed or anhedonic mood.

These findings provide tentative support for a categorical conceptualization and make it very clear that the continuity controversy regarding unipolar depression has not yet been decided in favor of dimensionality. To reconcile the conflicting reports to date, several data analytic and sampling issues need to be explored systematically.