The effects of age, sex, socioeconomic background and noise exposure on hearing were examined in a population sample. Subjects, aged between 17 and 80 years, were selected from the electoral register using a stratified random sampling technique based on age and reported hearing difficulty. Exposure to occupational, social and gunfire noises was estimated retrospectively from a structured interview. Analysis of the audiograms of 2,162 subjects free from any material conductive impairment led to a simple model requiring only first-order factors of age, occupational group (manual or non-manual) and occupational noise exposure, but with parameters different for the two sexes. These models accounted for up to 58% of the variance in hearing threshold level (HTL), depending on frequency. It was not possible to separate out gunfire noise exposure as a factor because of a confounding effect of age. For females there was no significant effect of noise exposure, possibly owing to small numbers of noise-exposed females. The dependence of HTL on age agreed well with published literature, particularly for males. For males, noise exposure was associated with a modest increase in HTL, mainly between 3 and 6 kHz. The model was consistent with reported HTLs for males with equivalent noise exposure.