An obvious and important question to ask in regard to hearing protection devices (HPDs) is how much hearing protection, commonly called attenuation or noise reduction, can they provide. With respect to the law, at least, this question was answered in 1979 when the U. S. Environmental Protection Agency (EPA) promulgated a labeling regulation for hearing protection devices (HPDs) that specified a descriptor called the Noise Reduction Rating (NRR) measured in decibels (dB). In the intervening 25 years many questions and concerns have arisen over this regulation. Currently the EPA is considering publication of a proposed revised rule. This report examines a number of the relevant issues in order to provide recommendations for a new label, new ratings, and a preferred method of obtaining the test results from which the ratings are computed. A wide variety of ratings are reviewed, from the putative gold standard, an octave-band calculation, to simplified ratings employing fewer numbers that can be applied to more common noise measures such as C-weighted or even A-weighted sound levels or exposures. Additionally, the most simplified method of all, namely a class or grading scheme is examined. The conclusion is that a Noise Reduction Statistic for use with A weighting (NRS A), an A-A' rating computed in a manner that considers both inter-subject and inter-spectrum variation in protection, yields sufficient precision for most situations. Justification for this recommendation stems from consideration of the inter-wearer variation in fitting, the variation in noise spectra, and the accuracy of the basic measurements of hearing protector attenuation and noise-exposure values. Furthermore, it is suggested that to provide additional guidance to the purchaser, two such ratings ought to be specified on the primary package label-the smaller one to indicate the protection that is possible for most users to exceed, and the larger one to indicate the protection that is possible to achieve by individual highly motivated expert users; the range between the two numbers conveys to the user the uncertainty in protection provided. Guidance on how to employ these numbers, and a suggestion for an additional, more precise, graphically oriented rating to be provided on a secondary label (the Noise Reduction Rating, graphical, NRS G) are also included. Another important consideration is the data from which the new rating is computed. Examination of potential types of data from U. S. or international standards reveals that ANSI S12.6-1997 Method-B data appear to provide the best correlation to field performance and hence the most useful ratings; however, concerns about the reproducibility of Method-B based results led us to also offer an alternative Method-A based value. Since insufficient data are available at this time to clearly distinguish between the two recommendations the need for an interlaboratory study is identified along with suggestions for how it might be conducted.