Tympanometry can indicate middle ear effusion in children referred for tympanostomy tube placement. In outpatient setting, objective adjunctive tools are needed to diagnose the otitis media spectrum. We enrolled and followed 515 children aged 6-35 months at primary care level. We compared tympanometry with pneumatic otoscopy and evaluated the proportions of type A, C1, C2, Cs and B tympanograms in relation to specific otoscopic diagnoses in 2206 and 1006 examinations at symptomatic and asymptomatic visits, respectively. At symptomatic visits, different peaked tympanograms were associated with a healthy middle ear as follows: type A in 78%, type C1 in 62%, type C2 in 54% and type Cs in 18% of examinations. In contrast, any peaked tympanogram was related to healthy middle ear in 67% of examinations. Flat (type B) tympanogram was related to otitis media with effusion in 44% and to acute otitis media in 56% of examinations, respectively. At asymptomatic visits, the peaked tympanograms together were associated with a healthy middle ear in 87% of otoscopic examinations. Flat tympanogram indicated otitis media with effusion as well in 87% of examinations. Tympanometry is not a useful tool in detecting specific otoscopic diagnoses because it cannot distinguish between otitis media with effusion and acute otitis media. However, among outpatients all peaked tympanograms suggest a healthy middle ear and a flat tympanogram is useful in detecting any middle ear effusion. Thus, tympanometry can be used as an adjunctive tool, but accurate diagnosis requires careful pneumatic otoscopy.