Probabilistic cross-identification has been successfully applied to a number of problems in astronomy from matching simple point sources to associating stars with unknown proper motions and even radio observations with realistic morphology. Here we study the Bayes factor for clustered objects and focus in particular on galaxies to assess the effect of typical angular correlations. Numerical calculations provide the modified relationship, which (as expected) suppresses the evidence for the associations at the shortest separations where the 2-point auto-correlation function is large. Ultimately this means that the matching probability drops at somewhat shorter scales than in previous models.