Document: EUG-3-20-3

How can improving accuracy make a difference in understanding ecology? Some answers from finding animals in observation and experiment.

MEYER, E.R*

Dept of Biology, Loyola College, Baltimore, MD ¹

Abstract:
Two new tools to evaluate accuracy are 1) search experiments for lifelike replicas, and 2) simultaneous searches for living animals. In two search experiments, I placed animal replicas on a forest floor. They consisted of adult turtle shells, turtle hatchlings, small snakes and salamanders. The use of replicas allows the researcher to know exactly the community, populations, and locations. Results: taxa differed strongly in the rates found. Searchers differed in results even when similarly trained. Increasing the number of searchers increased the percent of actual population found. The second tool to measure accuracy, simultaneous search for living animals, was implemented by biologists walking near each other. Some experienced biologists could not detect some species, even when pointed out to them, but made unique finds of other species. Lessons learned from systematic observation: (a) Experienced investigators detected different species at the same place and time. They identified different vertebrate communities. (b) A useful analysis to compare their search techniques is Uniqueness, Completeness, and Peak levels (UCP) of the data. (c) Differences among searchers can be helpful because comparing search results shows the most correct data for the community. The big question: when is evaluating accuracy helpful in understanding the biology of a situation? The answers reflect the ideas being tested. Examples of different accuracies on three levels: (1) For some reproduction and conservation studies, finding all the life stages present. (2) For monitoring a population, the contributions of different observers and techniques in detecting one life stage. (3) For predation and community studies, finding all the key species present, and seeing the comparability of their numbers. Questions about accuracy are thus opportunities to consider which data aspects are most useful.

Keywords: accuracy, observer differences, experiment

This abstract is being presented at: 8:30 AM in session:
Symposium # 15: Measurement Error in Ecological Data.