A literature survey indicates lack of standardization of a speech‐level measure among published experiments. In addition, experiments reported here show that the measure most commonly used, the VU (volume unit), exhibits high variability, especially among nonexpert observers. The possibility of better measures of speech level is demonstrated by a completely objective measure based on an assumed log uniform distribution of instantaneous voltage levels. It is concluded that there is a need to reassess the general problem of speech‐level measurement and to propose new approaches.