Symposium # 15: Measurement Error in Ecological Data.

Organized by: B.A. Maurer, M. Taper and H. Caswell.
Sponsored by: Statistical Ecology Section.
Tuesday, August 8, 2000
8:00 AM to 11:45 AM
Ballroom III - Cliff Lodge

To acquire their data, ecologists typically rely on measurement processes that are subject to a relatively large amount of uncertainty. This is particularly true of data collected at larger scales of resolution. At the largest scales, remotely sensed data contain many sources of error that can enter into ecologically relevant data such as estimates of vegetation density or habitat classifications. Error also enters into ecological research any time populations are censused. It is rarely possible to obtain an accurate and complete count of the number of individuals in a population, and many sources of measurement error potentially increase the uncertainty in estimates of population size, density, and demographic properties. Uncertainties can be compounded or propagated when inaccurate data are aggregated at larger scales to describe properties of large ecological units. Significant statistical problems arise when measurement error is not properly modeled in the process of data analysis.

The purpose of this symposium is to gather together individuals who have had experience in estimating, modeling, and evaluating the effects of measurement in ecological data. The speakers will address a variety of problems that arise when measurement error contributes a significant component to the overall uncertainty of an ecological data set. A number of solutions will be discussed, including various statistical techniques that can be used to evaluate the role of measurement error in the data. The symposium will give practicing ecologists an introduction to statistical solutions to the problems encountered when there is a substantial amount of measurement error in a data set.