A Confirmatory Factor Analysis of Subjective Air Quality

Abstract

A complete assessment of air quality requires a better understanding of subjective evaluative measures. An example of the types of measurement issues that should be addressed and an important way to proceed in the study of subjective air quality is provided. The measurement error in four indicators of subjective air quality is examined by means of a confirmatory factor analysis. The indicators of overall quality, clarity, color, and odor of the air are from a survey of Shenandoah National Park visitors in Virginia's Blue Ridge Mountains in July 1980. Analysis reveals both random and nonrandom measurement error in all four indicators. The measure of overall air quality has the highest reliability with nearly 90% of its variance explained by the unobserved subjective air quality variable. The three other indicators have reliability coefficients of 0.5 to 0.6. Given this degree of measurement error it is important that future analyses explicitly consider measurement error when examining the relationship between subjec tive and objective air quality. Otherwise, a highly inaccurate assessment of this relationship will be obtained.