Unravelling the evolution of complex social organization in animals is an important aim, not least because it helps to understand the evolutionary roots of human sociality. Recent advances in comparative methods allow to approach this question in a phylogenetic context. The validity of such comparative approaches depends strongly on the quality of information regarding the behaviour, sociality, and reproduction of animals in natural systems, and on the quality of the phylogenetic reconstruction. Applying a novel comparative approach, a recent study of Dey et al. (, Nature Ecology & Evolution, 1, 137) concluded that evolutionary transitions to cooperative breeding in cichlid fishes were not associated with the social mating pattern. Here we argue that this result was adversely affected by equivocal classifications of mating patterns, and inadequate phylogenetic data. In order to illustrate the impact of the mating system misclassifications, we scored mating patterns as reported in the original literature and re‐analysed the dataset based on Dey et al.’s tree topology. The result suggests that the mating system does in fact significantly explain the evolutionary transition to cooperative breeding in lamprologine cichlids, but we submit that a reliable conclusion cannot be reached before improving the behavioural information and the underlying phylogenetic reconstruction. The problems identified in this case study are not unique and we urge caution in the interpretation of results from comparative phylogenetic studies in general. We do agree with Dey et al. () though that the lamprologine cichlids of Lake Tanganyika may constitute a fundamental test case for the theory of social evolution, but better information on their behaviour and phylogenetic relationships is needed to allow meaningful analyses.