Analyses of data from the NIOSH Occupational Noise and Hearing Survey (OHNS) have previously focused on 1172 highly ``screened'' workers. This paper examines excess risk estimates for an additional 894 male workers (609 noise exposed and 285 controls) who were excluded for various reasons (i.e., nonoccupational noise exposure, otologic or medical conditions affecting hearing, prior occupational noise exposure). Data are analyzed by age, duration of exposure, and sound level (8-h TWA) for different definitions of noise-induced hearing impairment. Hearing impairment was defined as a binaural average hearing threshold level greater than 25 dB at selected frequencies (1--4 kHz, 1--3 kHz, and 0.5, 1, and 2 kHz). The analysis suggests that excess risk estimates are sensitive to population characteristics as well as definition of impairment. Lifetime excess risk estimates for noise-induced hearing impairment are similar for the unscreened and screened populations for 8-h TWAs below 85 dB. However, compared to the unscreened population, higher excess risk estimates are observed for the screened population at A-weighted levels greater than 85 dB. This difference in excess risk was most marked for hearing-impairment definitions that include 3 and 4 kHz.