As a test for detecting middle ear disease among preschool children, tympanometry — as opposed to audiometry — has three advantageous attributes: a high degree of sensitivity, minimal need for subject cooperation, and total objectivity. For these reasons interest has arisen in tympanometry as a method for screening, i.e., identifying children with previously undetected middle ear disease. However, uncertainty persists concerning the importance of detecting apparently asymptomatic middle ear effusions, and concerning optimal methods, or even the advisability, of treating them. Further, the sensitivity and specificity of tympanometry depend on how the pass-fail cutoff point is defined. Defining this cutoff point so as to achieve high sensitivity may result in excessively low specificity, with the production of large numbers of false-positives who then become overreferrals. Data are presented to show how the validity of the test may be increased to some extent by attention to the gradient of “negative-pressure” tympanograms. At the present time, given the various aforementioned uncertainties, and with adequate validation as to the presence or absence of disease often lacking in reported studies of impedance screening in preschool populations, the cumulative results of these studies do not warrant embarking on large-scale screening programs. What is needed instead is additional research to explore the issue further.