When two related species interbreed, their hybrid offspring frequently suffer from reduced fitness. The genetics of hybrid incompatibility are described by the Bateson–Dobzhansky–Muller (BDM) model, where fitness is reduced by epistatic interactions between alleles of heterospecific origin. Unfortunately, most empirical evidence for the BDM model comes from a few well-studied model organisms, restricting our genetic understanding of hybrid incompatibilities to limited taxa. These systems are predominantly diploid and incompatibility is often complete, which complicates the detection of recessive allelic interactions and excludes the possibility to study viable or intermediate stages. Here, we advocate research into non-model organisms with haploid or haplodiploid reproductive systems and incomplete hybrid incompatibility because (1) dominance is absent in haploids and (2) incomplete incompatibility allows comparing affected with unaffected individuals. We describe a novel two-locus statistic specifying the frequency of individuals for which two alleles co-occur. This approach to studying BDM incompatibilities requires genotypic characterization of hybrid individuals, but not genetic mapping or genome sequencing. To illustrate our approach, we investigated genetic causes for hybrid incompatibility between differentiated lineages of the haplodiploid spider mite Tetranychus evansi, and show that strong, but incomplete, hybrid breakdown occurs. In addition, by comparing the genotypes of viable hybrid males and inviable hybrid male eggs for eight microsatellite loci, we show that nuclear and cytonuclear BDM interactions constitute the basis of hybrid incompatibility in this species. Our approach opens up possibilities to study BDM interactions in non-model taxa, and may give further insight into the genetic mechanisms behind hybrid incompatibility.