Using human activity data in exposure models: Analysis of discriminating factors

• Published in: Journal of exposure analysis and environmental epidemiology Vol. 13; no. 4; pp. 294 - 317
• Main Authors: Mccurdy, Thomas, Graham, Stephen E
• Format: Journal Article
• Language: English
• Published: New York Nature Publishing Group US 01.07.2003; Nature Publishing Group
• Subjects: Cross-Sectional Studies; Education; Environmental Exposure - analysis; Environmental Monitoring - methods; Environmental protection; Epidemiology; Human Activities - statistics & numerical data; Humans; In vehicle; Indoor environments; Longitudinal Studies; Medicine; Medicine & Public Health; Models, Statistical; Motor Vehicles; original-article; Risk Assessment - methods; Seasons; Time Factors; United States; United States Environmental Protection Agency; Weather; United States; exposure modeling; human activity data; indoor time; outdoor time; cross-sectional data; motor vehicle time; longitudinal data
• ISSN: 1559-0631; 1053-4245; 1559-064X
• DOI: 10.1038/sj.jea.7500281

This paper tests factors thought to be important in explaining the choices people make in where they spend time. Three aggregate locations are analyzed: outdoors, indoors, and in-vehicles for two different sample groups: a year-long (longitudinal) sample of one individual and a cross-sectional sample of 169 individuals from the US Environmental Protection Agency's Consolidated Human Activity Database (CHAD). The cross-sectional sample consists of persons similar to the longitudinal subject in terms of age, work status, education, and residential type. The sample groups are remarkably similar in the time spent per day in the tested locations, although there are differences in participation rates: the percentage of days frequenting a particular location. Time spent outdoors exhibits the most relative variability of any location tested, with in-vehicle time being the next. The factors found to be most important in explaining daily time usage in both sample groups are: season of the year, season/temperature combinations, precipitation levels, and day-type (work/nonwork is the most distinct, but weekday/weekend is also significant). Season, season/temperature, and day-type are also important for explaining time spent indoors. None of the variables tested are consistent in explaining in-vehicle time in either the cross-sectional or longitudinal samples. Given these findings, we recommend that exposure modelers subdivide their population activity data into at least season/temperature, precipitation, and day-type “cohorts” as these factors are important discriminating variables affecting where people spend their time.