A study was conducted to determine the dependency of sound attenuation capabilities of earmuffs and earcaps on various application instruction techniques including: (1) no instruction, (2) no instruction with 70 dBA auditory feedback, (3) manufacturers' package instruction, (4) detailed instruction, and (5) modeled instruction. The hearing protection devices (HPDs) studied were: two earcaps (Willson #20 Sound-Ban, Flents #055 Peace & Quiet Headband), and four earmuffs (E-A-R model 1000, Siebe Norton Industrial model 4540, Peltor H6A/v, and Willson 365A Sound Barrier). HPD comfort and wearer preferences were also assessed. Furthermore, typical excessive industrial machinery noises were sampled and analyzed across the 1/3 octave frequencies. Finally, an example is provided as to how these machinery noise spectra can be matched with HPD attenuation spectra to ascertain the protector which would afford optimal protection for a given noise situation. Fifty subjects (twenty-five males and twenty-five females) participated in the experiment. Attenuation characteristics of the HPDs were evaluated utilizing the real-ear attenuation at threshold (REAT) method. A three-way mixed factorial design was used for data collection and analysis of attenuation results. Bipolar scales were used to assess individual HPD comfort, and the HPDs examined were ranked to obtain user preferences. In the analysis of attenuation results, analysis of variance CANOVA) and pairwise comparisons were utilized to detect statistical significance. The comfort scales and ranking scores were evaluated using the Friedman one-way block design. Attenuation results for the earmuffs and earcaps tested showed that they were much less susceptible than earplugs (from a previous study by Epps, 1984) to changes in user insertion/donning instruction technique and also not as dependent on user gender. The main effect of gender was not significant, and in general, any instruction was better than no instruction at all but the effects of those instructions did not differ significantly among each other. As expected, there were main effect differences among the HPDs as to their attenuation capabilities, rated discomfort, and user preference. Because the main objective was on assessing donning instruction effects on HPD attenuation, the comfort/preference assessment was based on only a short (25 minutes) wearing time during the attenuation tests. Therefore, the comfort/preference ratings could likely vary given longer wearing periods and different work environments. All results found are discussed on the basis of the sample data obtained and conclusions drawn from these results should be limited to these experimental conditions and subsequent analyses, as actual attenuation achieved in practice may differ. The example of HPD-machinery noise matching illustrates that the attenuation/spectral matching procedure may indeed be a feasible way of selecting optimal protection for workers. / Master of Science