The ultimate decision seemed to stem from a skepticism of the existence of latent taxa. The associate editor (who did much of the reviewing) was convinced that we had dichotomized a continuous variable arbitrarily, when in fact we found two clusters using two-step cluster analysis. We provided literature that explained cluster analysis and contrasted taxometric methods against arbitrary splits, but all for naught. A statistical consultant was brought in to review and talked about William Stephenson's Q-technique. After consulting with our quantitative methods faculty, we are still not sure why. We had tested for moderation by cluster membership and the associate editor thought we should have tested for mediation instead, though the theory that drove our study was not consistent with mediation.