This is more likely to be caused by a hardware difference than a software differnce, but both of these explanations seems really unlikely compared to the theory that this person's self report was confused. If in a controlled experiment, he can reliably differentiate between patterns of light polarization, then I will worry about explaining this.