You are here:
A NEW METHOD OF LONGITUDINAL DIARY ASSEMBLY FOR EXPOSURE MODELING
GLEN, G., L. SMITH, K. ISAACS, T. R. MCCURDY, AND J. LANGSTAFF. A NEW METHOD OF LONGITUDINAL DIARY ASSEMBLY FOR EXPOSURE MODELING. Presented at International Society of Exposure Analysis Conference, Tucson, AZ, October 30 - November 03, 2005.
The two main objectives of this research are (1) to improve and update and (2) to analyze the CHAD database.
For objective 1, we will
* Reconfigure the CHAD program into a completely modularized Oracle database.
* Redesign User Interface for effcient utilization of the program's capability.
* Obtain dates for those surveys that did not provide them to us, so that we can obtain associated meteorological/climatic inputs for the person-days of information without them.
* Revise the upper and lower bound delimiters in the energy expenditure distributions used for activity-specific estimates.
For objective 2, we will
* Evaluate data quality.
* Evaluate trends and activities for various subgroups.
* Identify temporal patterns for longitudinal data.
* Characterize resolution required for output for exposure and dose models.
Many stochastic human exposure models require the construction of longitudinal time-activity diaries to evaluate the time sequence of concentrations encountered, and hence, the pollutant exposure for the simulated individuals. However, most of the available data on human activities are from cross-sectional surveys (usually one day per day). The variability of the population for long-term exposure is one property (among others) that depends on the method used to assemble activity diaries for long simulation periods (such as a year) from daily diaries.
A new method, based on diary ranks with respect to a user-chosen key variable, has been developed targeting two statistics, D and A. The D statistic reflects the relative importance of within-person variance and between-person variance. The A statistic measures the lag-one (day-to-day) autocorrelation. The targets are met very closely for all simulation periods longer than 30 days, and reasonably well for shorter simulations. Longitudinal diary data from a field study suggest that D and A are stable over time (and perhaps over cohorts as well). The new method can be used with any cohort definitions and any day-type and season combinations, making it easily adaptable to most exposure models.