Sixty-four in-patient cases of deliberate non-fatal self-poisoning were compared for psychosocial problems in a case-control study with a similar number of individually matched community controls. A strongly significant association was found between unemployment and self-poisoning. Further analysis revealed no firm evidence to support the hypotheses that unemployment was causally related to self-poisoning in an indirect manner or that it increased the vulnerability of individuals who self-poison to other stressful life events and difficulties. It is concluded that a possible explanation is that some third factor independently increases the risk of both unemployment and self-poisoning, giving rise to a non-causal relationship between these last two variables.