The DSM-IV criteria recognize the existence of obsessive–compulsive disorder (OCD) with poor insight. However, there is paucity of literature on the clinical correlates and treatment response in poor and good insight OCD. In this study, insight is measured by using the Brown Assessment of Beliefs Scale (BABS) developed specifically to assess insight. One hundred subjects with DSM-IV OCD were ascertained from the OCD clinic of a large psychiatric hospital in India. All subjects were evaluated extensively by using structured instruments and established measures of psychopathology. The subjects were treated with adequate doses of drugs for adequate period. The results showed that 25% of the subjects had poor insight. Poor insight was associated with earlier age-at-onset, longer duration of illness, more number of obsessive–compulsive symptoms, more severe illness and higher comorbidity rate, particularly major depression. Of the subjects who were treated adequately (N = 73), 44 (60%) were treatment responders. Poor insight was associated with poor response to drug treatment. In the step-wise logistic regression analysis, baseline BABS score was highly predictive of poor treatment response. Poor insight appears to be associated with specific clinical correlates and poor response to drug treatment. Further studies are needed in larger samples to replicate our findings.