Document: BRI-3-20-1

Measurement error in ecology: Problems and prospects.

MAURER, B.A.*

Department of Fisheries and Wildlife & Department of Geography, Michigan State University, East Lansing, MI USA ¹

Abstract:
Ecologists have typically ignored the problem that the measurements that they take are imprecise. An increasing awareness of the sources of subjectivity in ecological investigations suggests that it is important to begin to recognize that data collected in the field or laboratory need not provide an accurate description of ecological reality. Once such a recognition is made, then it is possible to begin to account for the subjective decisions and imprecision that enters into the measurement process. This means that statistical models that omit terms accounting for measurement errors are likely to produce statements about the data they model that are inaccurate or misleading. In some cases, mis-specified models can be detected and corrected, particularly if additional information is available about the uncertainty associated with the measurement process. Problems with measurement error are especially likely to be compounded or propagated when the system being studied has a complex, multi-scale structure.

Keywords: statistics, data analysis, scale, bias, estimation, statistical modeling

This abstract is being presented at: 8:00 AM in session:
Symposium # 15: Measurement Error in Ecological Data.