Still no evidence that late-sighted individuals rely more on color for object recognition: Reply to Vogelsang et al.

We thank Vogelsang et al. for their reply [1] to our critique [2]. However, we remain unconvinced that the experimental data presented in their paper supports their claim. First, applying a non-parametric test to the differences in percent correct, as in their reply, does not solve the ceiling problem. Second, there is indeed enough data to apply more appropriate models that assume binomial rather than Gaussian errors; these models provide evidence for the nullhypothesis. Third, we note a design issue that undermines the causal inferences that can be drawn from the experimental data. Based on these statistical and design considerations, we reiterate that in our view, the empirical data do not support the claim that object recognition in Prakash patients is more impaired by color removal than controls.