The current study investigated whether candidate cognitive endophenotypes may be used to validate a schizo-obsessive subtype. Using within-subject random effect regression analyses and cross-trait cross-relative analyses, we evaluated the association between obsessive-compulsive symptoms (OCSs) and cognitive performance in 984 patients with nonaffective psychosis (22.5% with OCSs), 973 unaffected siblings (7.7% with OCSs), 851 parents (4.2% with OCSs), and 573 controls (4.5% with OCSs). No significant within-subject associations between OCSs and cognitive functioning were found for patients and siblings. Severity of OCSs was associated with worse set-shifting ability in parents and worse processing speed in controls, but effect sizes were small (0.10 and 0.05 respectively). Cross-trait cross-relative analyses yielded no significant results. Contrary to our expectations, neither within-subject analyses nor cross-relative analyses yielded a clear association between OCSs and cognitive performance. Results do not support a schizo-obsessive subtype associated with cognitive impairment.