Rating level as a method to assess the impact of speech noise on cognitive performance and annoyance in offices

offices. The rating level denotes the energy-equivalent sound pressure level during a measurement period with speech sounds and considers penalties for tonal, informational and impulsive constituents. There is little evidence that the rating level correlates with the performance and perceived annoyance of office workers. This study evaluates 89 different sound conditions under which subjects have to complete a short-term memory task and a questionnaire in laboratory conditions with respect to their relationships with the rating level. The relationships of the penalty for impulsiveness and the penalty for tonality or informational constituents with the rating level are analyzed separately. In addition, the penalty for tonality or informational constituents is substituted by percentile level statistics, namely the difference between the 10th and 90th percentile levels. In contrast to the penalty for tonality or informational constituents, this metric is objectively measurable. Using the rating level to assess the noise at office workplaces could be improved by using percentile level statistics to account for informational constituents. To improve the predictive validity, it is suggested to report the penalties separately.