Forty-nine subjects (ages 8–55) were tested with the circles and animals of the Titmus Stereo test while wearing Polaroid(tm) filters at axis 135 deg before both eyes so that the right eye's target was presented to both eyes. This procedure removed retinal image disparity as a cue to depth and left lateral displacement as the primary cue to obtain correct responses on the test. The subjects' scores on the circle test were much higher than expected by chance guessing. Different instructions to the subjects did not influence this result. Scores on the animal test were similar to those expected by chance. The subjects used lateral displacement as a response cue on the circle test but not on the animal test. Procedures for improving the validity of the Titmus Stereo test are described.