Document: DAV-3-52-30

Design and outcome in ecological experiments: An analysis of tadpole density manipulations.

SKELLY, D.K.* ¹ and J.M.KIESECKER ²

Yale University, New Haven, CT 06250, USA ¹
Pennsylvania State University, State College, PA 16802, USA ²

Abstract:
An ecologist conducting an experiment must make a series of choices regarding its design. Among these, the choice of experimental venue ranks as among the most critical. The decision to experiment in the laboratory, in mesocosms, or in the field has been postulated to constrain other aspects of design and to influence outcome. Relative to field manipulations, manipulations in the laboratory or in mesocosms are thought to allow more complicated designs, greater precision among replicates and a higher degree of replication. However, it has been argued that despite these drawbacks, field experiments may be particularly valuable because of their presumed advantages in extrapolation to understanding natural populations. We evaluated designs and outcomes from 52 studies aimed at quantifying the impact of density on the growth of larval anurans. Contrary to expectations, we found that there was little evidence that experimental venue influenced design complexity or precision among replicates. However, laboratory experiments tended to have moderately greater numbers of replicates. In a meta-analysis of growth responses we determined that experimental outcomes were strongly related to venue. While most manipulations tended to yield effects consistent with competition, those conducted in the field tended to yield more moderate effects. As an example, interspecific manipulations in mesocosms were associated with effect sizes an order of magnitude larger than those from field manipulations. Taken together, our results from tadpole experiments suggest that the tradeoffs presumed to flow from the choice of venue may be more moderate than expected, while the impact of venue on experimental outcome may be quite large. We believe that such analyses will be useful as ecologists work to improve experimental methods and to expand their inferential power.

Keywords: anuran, design, experiment, outcome, venue

This abstract is being presented at: 8:15 AM in session:
Oral Session #38: Amphibian Ecology.