Abstract
The prevalence of a consequential degree of transportation noise-induced annoyance varies greatly among communities with similar cumulative noise exposure levels. The variation in community response is so great that regression-derived predictions of annoyance prevalence rates in nominally "typical" communities are often substantially in error when applied to many actual communities. These errors of prediction can be so great that noise regulation based upon purely descriptive statistical methods often fails to achieve essential goals. These include uniformity of nationwide regulatory effect, and credibility of definition of "significant" noise impacts. The 2014 revision of ISO Standard 1996-1 includes a non-regression based method for improving the accuracy of dosage-response analysis that can support improved efficacy of transportation noise regulation.