Objective: to verify the tax of false-negative result in protocols of auditory selection. Methodology: 1,326 protocols of children evaluated in the auditory selection of a particular doctor's office between 2002 and 2009 had been analyzed. The selection consisted of anamneses and registers of the transient evoked otoacoustic emissions (EOAET). The criterion of "imperfection" was absence of EOAET in one or both the ears. Results: the age of the children varied of one month the six years. Of the 1,326 protocols, it had "imperfection" in both the ears in 41 (3.09%) of them (Group 1); "it fails" only in the right ear (OD) in 28 (2.11%) (Group 2) and "it fails" only in the left ear (OE) in 29 (2.19%) (Group 3). For the retest, they had returned 13 (31.7%) children from Group 1, being 9 (69.2%) with register of "it passes" and 4 (30.8%) of "imperfection"; 9 (32.14%) children of Group 2, being 100% with register of "pass"; 8 (27.59%) children of Group 3, being 5 (62.5%) with register of "pass" and 3 (37.5%) of "imperfection". Considering the results of the 3 Groups, the tax of false-negative result was of 76,6%. Conclusion: in the analyzed protocols was verified that 76.6% of the children who had failed in the selection they had passed in the retest. Modified conditions of external average ear can justify this finding. Salient the raised index of evasion in the retest, limiter factor so that a more necessary estimate of the false-negative tax and the considered objective met was contemplated of trustworthy form.