The ability of aural acoustic immittance measures and validated pneumatic otoscopy to identify middle ear effusion (MEE) was determined for a group of children with chronic or recurrent otitis media. The measures were made immediately prior to surgery for placement of tympanostomy tubes, with the validating diagnosis of MEE made by the surgeons. Aural acoustic admittance measures were made by a certified and licensed clinical audiologist using an instrument that meets current standards and otoscopic examinations were made by a nurse practitioner validated for use of otoscopy in the identification of MEE. Logistic regression analyses were done to determine the ability of admittance variables, otoscopy, and their combination to discriminate between ears with and ears without MEE. Of the individual admittance variables, tympanometric width had the best performance. Otoscopy alone had good sensitivity but only fair specificity. Combining acoustic admittance data with the otoscopist's findings did not improve sensitivity greatly, but improved the specificity relative to that of the otoscopist. The criterion tympanometric width > 150 daPa or Ytm < 0.2, a criterion commonly used based on interim norms published in an appendix of ASHA guidelines for screening, had good sensitivity but very poor specificity in the group of children scheduled for surgery. Also, a comparison was made for tympanometric width measures from the children undergoing surgery with those from a group of children more representative of the general population. Three distributions were found; one for ears with MEE, one for ears with no MEE of children scheduled for otologic surgery and one for ears with no MEE from the group of children from the general population. The data illustrate the importance of the population characteristics on the test measure used and have implications for choosing test criteria. In addition, group estimates of performance of acoustic admittance variables were approximately the same when otoscopy was used as the gold standard for identification of MEE as when findings at surgery were used. However, when comparing the two diagnostic methods on a case-by-case basis, it was determined that the otoscopist's diagnosis disagreed with the diagnosis of the surgeons in over 21% of the ears.