Study Design and Logistics - Wetland Techniques - page 37

Estimation of effect size following data collection and statistical analyses is a
relatively recent addition to reporting of results from wetland and other natural
resource studies, but estimation of treatment effect provides additional evidence
and weight for conclusions developed during the study. It is a relatively simple
concept that should not be made any more complicated than necessary regarding
the magnitude of any found effect. Any significant statistical test reported in the
literature should also include the magnitude of the effect to assess biological
significance of the results. For example, one can achieve a statistically significant
difference between means of a treatment and control population but, depending on
the variation within each population, a biologically significant effect may not be an
appropriate conclusion. Such an occurrence is more likely in a laboratory setting,
but can also be found in field studies. Effect size can be as simple as reporting the
percent mean change due to an effect (i.e., increases or decreases by X % due to the
application of the treatment). In addition, a number of indices have been developed
to quantify the strength of the difference between groups (e.g., levels of indepen-
dent variables). The most common effect size index is Cohen's d (Cohen 1988 )
or standardized mean difference whereby calculated effect size index values are
categorized as 0.20

large. These indices can be
calculated for a wide variety of study designs (see for example: http://www.
bwgriffin.com/gsu/courses/edur9131/content/Effect_Sizes_pdf5.pdf ; http://www.
campbellcollaboration.org/resources/effect_size_input.php )

¼ small, 0.50 ¼ medium, 0.80 ¼

1.12 Other Logistical Considerations of Wetland Study

Wetlands are complex and diverse ecosystems; therefore, it is quite difficult to
generalize a logistical approach that can be applied to all wetlands under all study
situations. However, there are a number of common information needs to access
and become familiar with prior to conducting a wetland study. It is imperative to
become well-versed in the system being proposed for study beyond the immediate
question being addressed. All biotic and abiotic elements of a wetland ecosystem
are potential variables in a wetland study no matter whether one is investigating
water quality, hydric soils, plant associations, invertebrates, or animal communities
because of the ecological linkages among all elements in the ecosystem. To fully
document the effects found in any study, one must consider the totality of effects on
all elements of the wetland, which can only be accomplished via a thorough
ecological understanding of the system being studied - including the potential
ecological states of the system under the environmental variation potentially affect-
ing the wetland that may differ from the state measured during the study.
To define the study population of wetlands and the potential scope of inference
relative to research results, one needs to have knowledge of the spatial scale of
occurrence of the wetlands of interest. There are a number of sources of wetland
occurrence, but quality of locations and associated information varies greatly.
Nearly all available mapped locations of wetlands are provided as electronic data