In response to the low standards in short form development, Smith, McCarthy, and Anderson (2000) introduced a set of guidelines for the construction and evaluation of short forms of psychological tests. One of their recommendations requires researches to show that the variance overlap between the short form and its long form is adequate. This article introduces a new method for estimating this variance overlap. Subjects are randomly assigned to four different groups that are tested with different combinations of the long and the short form. This procedure makes it possible to estimate the effects that attenuate the variance overlap between the short and the long form. By controlling for these effects, the disattenuated coefficient VO gives an unbiased estimate of the variance overlap between both instruments. (Contains 1 table.)