This research report presents the results of tire/pavement noise measurements performed on concrete pavements as a part of the California Department of Transportation (Caltrans) Quieter Pavement Research (QPR) study to investigate tire/pavement noise characteristics on concrete pavements. The On-board Sound Intensity (OBSI) method was used to measure tire/pavement noise. In this third year of the study, a total of 83 pavement sections were tested at 35 different sites, which was reduced from the original 120 sections in the experiment for various reasons. Twenty-four of the 35 sites were divided into three test sections each, while the remaining eleven sites had one test section each that were used for analyses shown in this report. The third-year test sections were comprised of five texture types, with the following distribution of sections used in the analyses: 31 burlap drag, 23 diamond ground, 7 diamond grooved, 4 longitudinally broomed, and 18 longitudinally tined. At this time no single concrete pavement texture type can be considered definitively quieter than the others. Each pavement type has a demonstrated range of noise levels that largely overlap. The difference between the lowest and highest OBSI levels for the same nominal texture type is up to 6 dBA, indicating a large variability within a given texture type when sampled over a wide range of ages. The contributions of joint width, faulting, and sealant overbanding to this variance have not been rigorously investigated, and the results to date include these effects, which are part of the total noise generated by concrete pavements in the state. Taking into account all sections, the OBSI level on existing concrete pavements in California varies from about 101.1 dBA to about 107.6 dBA. The ranking of texture types for the sections with new and aged textures evaluated in this research project from quietest to loudest is: longitudinally broomed, burlap drag, diamond ground, diamond grooved, and longitudinally tined, although the differences between the means of each of these textures is not statistically significant, except for the longitudinally broomed sections. The report includes results from the Mojave test sections on Kern 58 which were all constructed at the same time in 2003. Based on these observations and the conclusions to date, recommendations are made for the fourth and final year of large-scale data collection and analysis on concrete pavement noise to change the experiment to focus on the following textures: diamond ground, diamond grooved and longitudinally tined. Recommendations are also made to obtain more detailed data regarding texture characteristics, joint cross-sectional area, faulting, and sealant overbanding to try and better explain differences within each texture type.