Candidate model list.

The candidate models encode possible answers to the four questions as follows. Q I: “”, haptic information is never integrated vs. “”, haptic information is integrated when available. Q II (which is only applicable to the answer for Q I): “”, internal knowledge of haptic reliability is incorrect vs. “”, internal knowledge of haptic reliability is correct. Q III: “”, internal knowledge of image-size reliability is incorrect vs. “”, internal knowledge of image-size reliability is correct. Q IV: “”, posterior samples are averaged to form distance judgments vs. “MAP”, MAP estimates are used to form distance judgments.