This study evaluated the accuracy of acoustic response tests in predicting conductive hearing loss (CHL) in 161 ears of subjects aged 2 to 10 years old. Clinical decision theory was applied to the prediction of CHL within this population. Acoustic tests included 1) tympanometric peak-compensated static admittance magnitude (SA) 2) tympanometric gradient at 226Hz 3) admittance-reflectance measurement (YR). Multivariate statistical techniques were discussed and utilised to analyse the data collected. The multivariate data analysis identified that admittance-reflectance responses yielded an output that predicted the presence of CHL with a calculated sensitivity of 90% with a specificity of 94% and is well suited to the prediction of CHL in clinical populations, and the inclusion of a 226-Hz tympanogram may slightly improve overall test performance.