Fifty-two optometry students exhibiting various refractive errors determined their own spherical correction by observing, through a phoropter, a defocused laser beam projected on a slowly rotating drum. Their task was to adjust the amount of sphere in the phoropter until the speckle pattern of the laser appeared stationary. Each observer made 20 monocular settings using his dominant eye. The data were analyzed for each refractive type (emmetrope, myope, hyperope) in terms of their variability, and it was found for all refractive groups that the phoropter settings were highly reliable. The data showed that when neutrality was approached from a simulated myopic refractive state (over correction with plus) the reliability for all groups was higher than when the initial simulated refractive state was hyperopic (over correction with minus). The results also indicated for all groups that adjustment of the lenses from an initial over correction with plus resulted in less minus correction than when they were adjusted from an initial over correction with minus.