Statistical analysis of a combination of objective and subjective environmental noise data using factor analysis and multinomial logistic regression

摘要

The purpose of this study was to demonstrate the efficiency of factor analysis and multinomial logistic regression in analyzing a set of objective (noise measurements) and subjective data (results of interviews about noise perception in an urban setting) in a noise pollution study conducted in the city of Curitiba, in southern Brazil. In this study, noise was monitored at 23 points along three parallel streets covering a distance of 5 km, and 397 interviews were conducted with residents living in the vicinity of these points. The interview script consisted of 21 questions (variables). To reduce these variables, the main factors (seven factors) were extracted and then added to the objective part (noise measurement data). This was followed by a multinomial logistic regression. The dependent variable in this regression was called "Interviewee symptoms and reactions to environmental noise." The outcome of these two statistical procedures led to the conclusion that 85.2 % of the symptoms and reactions could be attributed to the combination of these seven factors with the noise measurement data.