Correlations between transportation noise exposure and communityreaction indicate that on average only 18 to 22% (depending on the type of transportation noise) of the variation in reaction is attributable to noise exposure. Given the importance of these studies in determining not only the acceptable levels of noise but also how the noise is measured (equal energy units or not; with or without time of day weightings, etc.), The impact of reliability of measurement ofcommunity reaction and how reliability might be improved are considered. It is often suggested that error in measurement of reaction isa major reason for the relatively low noise-reaction correlations observed. However, few data exist on the internal consistency of composite scales of reaction. Data are presented on the internal consistency of various reaction scales, indicating that the typical generalreaction scale and annoyance scales are reliable (in terms of internal consistency) and superior to the typical disturbance and complaint disposition scales. Further, reliability is increased by the use of several questions in a scale, rather than a single question. Thus, the data suggest that the best measure of community reaction is a composite general reaction scale (based on questions such as, "how much are you personally affected..."; And "how disturbed are you by...") With a number of contributing questions. Noise-reaction correlations would be increased moderately with the use of an extremely reliable measure of reaction. However, when the limited reliability of the measurement of reaction and noise exposure, and modifying variables (especially attitude and sensitivity) are taken into account, theproportion of variation in reaction left unaccounted for is only around 20%. This paper appears in transportation research record no. 1312, Energy and environmental issues 1991.
Abstract