In their article ‘In defense of incompatibility, objectivism, and veridicality about color’ P. Roberts and K. Schmidtke offer the results of an experiment supposed to show that if selection of colored samples representing unique hues for subjects has a greater inter-subject variability than identification of sample pairs with no perceptual difference between them the result provides support for the philosophical concept of color realism. On examining the results in detail, we find that, according to standard statistical methodology, the relative magnitude (...) of inter-subject variability in the matching experiment is for both tested colors larger than that in the naming experiment, thus invalidating their claims. In addition, we point out several serious shortcomings in the experiment. (shrink)