Adjusting for the effect of environmental variability in outdoor engineering control studies.
Shulman-SA; Mickelson-RL; Mead-K
Joint Statistical Meetings, August 3-7, 2003, San Francisco, California. American Statistical Association: Alexandria, Virginia, 2000 Aug; :300684
In outdoor engineering control studies, measured analyte concentration levels change as environmental conditions change. Since the effectiveness of an engineering control varies with environmental conditions, comparisons of measurements taken with engineering controls operating versus those taken in an uncontrolled environment should adjust for these changes. However, environmental parameters are difficult to estimate. In this work, models based on factor analysis (Fuller 1987) are used to account for the effect of environmental variables. These models also describe the phenomenon that greater control efficiency tends to occur at the highest levels of the uncontrolled environment (Shulman, Mead, and Mickelsen 2002). The approach is combined with the randomized pair (uncontrolled environment determination, engineering control determination) approach that is often used. Also considered are the benefits of sampling different locations and analytes. Asphalt paving results of factor analysis models are compared with those from regressions of log ratios (controlled/uncontrolled ) on uncontrolled determinations. Implications for statistical design are discussed.
Engineering-controls; Outdoors; Environmental-engineering; Environmental-factors; Models; Sampling; Control-methods; Control-technology; Environmental-control; Environmental-engineering; Environmental-factors; Analytical-models; Sampling-methods; Statistical-analysis; Mathematical-models; Measurement-equipment
National Institute for Occupational Safety and Health, 4676 Columbia Parkway, Cincinnati, OH, 45226-1922
Abstract; Conference/Symposia Proceedings
Joint Statistical Meetings, August 3-7, 2003, San Francisco, California