Experimental Design in Aquatic Ecology - Freshwater Ecology: Concepts and Environmental Applications - page 478

Appendix: Experimental
Design in Aquatic
Ecology
Natural Experiments
Simulation Modeling
Manipulative Experiments
Summary
Experimental results supplied in a text or course setting are often pre-
sented as facts with little information on the scientific process that was
used to decide if an effect was “real” or not. Obtaining much of the fol-
lowing information from a statistics course or text would be preferable
(Sokal and Rohlf, 1981); at a minimum, some background in statistical
methods is recommended. Because much ecological knowledge is gained
from formal statistical treatment, the following brief introduction to ex-
perimental design is provided for those having little or no statistical expe-
rience. I hope this will convince even the mathphobic students of the ne-
cessity of statistical courses for the study of aquatic ecology.
Generally, when a scientist approaches a problem, she has an idea of
what causes a particular phenomenon to occur. This initial expectation can
be stated formally as a hypothesis (though such formal statement may not
always occur, and initial hypotheses are often discarded as naive). The hy-
pothesis states that some factor (or factors) has an effect. The job of the
scientist then is to test the hypothesis. The construction of hypotheses and
their testing and usefulness to ecologists have been discussed in detail by
numerous authors (Pickett et al., 1994; Quinn and Dunham, 1981). It will
be helpful to know some basic statistical terms before discussion on the
methods (Table A.1).
479