Background. Resemblance between spouses can be due to phenotypic assortment, social homogamy and/or marital interaction. A significant degree of assortment can have consequences for the genetic architecture of a population. We examined the existence and cause(s) of assortment for obsessive-compulsive (OC), anxious and depressive symptoms in a population-based twin-family sample. Method. OC, anxious and depressive symptoms were measured in around 1400 twin-spouse pairs and >850 parent pairs. Correlations of twins and their spouse, twin and co-twin's spouse, spouses of both twins and parents of twins were obtained to consider phenotypic assortment versus social homogamy as possible causes of marital resemblance. The association of length of relationship with marital resemblance was also investigated. Finally, we examined whether within-trait or cross-trait processes play a primarily role in marital resemblance. Results. Small but significant within-trait correlations of between 0.1 and 0.2 were seen for spouse similarity in OC, anxious and depressive symptoms. Cross-correlations were significant but lower. There was no correlation between length of relationship and marital resemblance. From the pattern of correlations for twin-spouse, co-twin-spouse and spouses of both twins, phenotypic assortment could not be distinguished from social homogamy. Both within- and cross-assortment processes play a role in marital resemblance. Conclusions. Small within- and across-trait correlations exist for OC, anxious and depressive symptoms. No evidence for marital interaction was found. Spouse correlations are small, which makes it difficult to distinguish between social homogamy and phenotypic assortment. It is unlikely that correlations of this size will have a large impact on genetic studies. © 2008 Cambridge University Press.