The attenuation characteristics of each of three commercial audiometric headsets of the noise‐barrier type—(1) Aural Research (AR‐100) “Auraldomes,” (2) Madsen (type ME‐70) “Noise‐Excluding Headset,” and (3) Rudmose (RA‐125) “Otocups”—were measured, as were the attenuation characteristics of a standard audiometric headset (Telephonics TDH‐39 earphone with an MX‐41/AR cushion). The threshold‐shift method was used, employing a pure‐tone sound field in an anechoic room. The resulting attenuation data for each of the three noise‐barrier headsets were compared statistically to those for each of the others and to the data for the standard headset. The data were contrasted with those supplied by the manufacturers. In terms of their attenuation capabilities for the octave bands pertinent to audiometry, the Rudmose Otocup (RA‐125) is ranked first, the Auraldome (AR‐100) second, and the Madsen (ME‐70) headset third. It is suggested that attenuation data alone should not determine the acceptability of any such device. The potential user should, in view of the attenuation data yielded by this study, consider such use only after measurement of octave‐band noise levels in the environment in question.