This study aimed to verify the diagnostic value of tympanometry with 226- and 1000-Hz probe tones in infants by comparing tympanometry results with the gold standard of the middle ear state assessed by myringotomy. Furthermore, clinically useful predictors for false peaked tympanograms despite the presence of middle ear fluid should be identified. Pre-operative 226- and 1000-Hz tympanograms were retrospectively compared with intraoperative findings of the tympanic cavity after myringotomy in 111 infants (217 ears) aged ≤ 12 months. In addition to the shape of tympanograms, demographic and clinical characteristics, the results of other audiometric measurements, and the viscosity of middle ear fluid, if present, were evaluated in several subgroups. The sensitivity and specificity of 1000-Hz tympanometry for middle ear effusion (MEE) detection were 97%-98% and 71%-84%, respectively, whereas standard tympanometry with a 226-Hz probe tone achieved a poor sensitivity of 43%-61% and a specificity of 81%-97%. Younger age, low viscosity of the middle ear fluid, and female sex were associated with an increased risk of false peaked 226-Hz tympanograms despite MEE. Furthermore, larger equivalent outer ear canal volumes were linked to false peaked 226-Hz tympanograms. The comparison of myringotomy findings and 226- and 1000-Hz tympanometry confirms the superiority of high-frequency tympanometry in infants younger than 12 months. Although this study identified some vague predictors of potentially false peaked 226-Hz tympanograms, the assessment of the middle ear state by 226-Hz tympanometry remains unreliable in young infants.