The results of video head impulse tests (video-HITs) may be confounded by data artifacts of various origins, including pupil size and eyelid obstruction of the pupil. This study aimed to determine the effect of these factors on the results of video-HITs. We simulated ptosis by adopting pharmacological dilatation of the pupil in 21 healthy participants (11 women; age 24-58 years). Each participant underwent video-HITs before and after pupillary dilatation using 0.5% tropicamide. We assessed the changes in the vestibulo-ocular reflex (VOR) gain, corrective saccade amplitude, and frequency of eyelid flicks. After pupillary dilatation, the VOR gain decreased for both right (RAC; 1.12 [Formula: see text] 0.12 vs. 1.01 [Formula: see text] 0.16, p = 0.011) and left anterior canals (LACs; 1.15 [Formula: see text] 0.13 vs. 0.96 [Formula: see text] 0.14, p < 0.001), and right posterior canal (RPC, 1.10 [Formula: see text] 0.13 vs. 0.98 [Formula: see text] 0.09, p = 0.001). The corrective saccade amplitudes also decreased significantly for all four vertical canals. The frequency of eyelid flicks, however, did not change. The changes of VOR gain were positively correlated with the lid excursion in RPC (r = 0.629, p = 0.002) and LPC (r = 0.549, p = 0.010). Our study indicates that eyelid position and pupil size should be considered when interpreting the results of video-HITs, especially for the vertical canals. Pupils should be shrunk in a very well-lit room, and artifacts should be prevented by taping or lifting the eyelids as required during video-HITs.