Study Design and Logistics

Abstract

Reliable knowledge is critical for management and conservation of wetlands. Essential to the scientific method and achieving reliable knowledge is study design. The primary purpose of study design is the collection of data in an unbiased and precise manner for an accurate representation of a population. Proper study design includes formulation of study questions and objectives, hypotheses to explain an observed pattern or process, conceptual models, appropriate methodology, and a data management plan. Inference of study results and conclusions can be explicitly bounded by defining an appropriate target population. Deductive, Inductive, and Retroductive reasoning are used to infer study results to target populations. Development of multiple competing hypotheses capable of being tested is at the core of the hypothetico-deductive approach that maximizes potential knowledge from a study. Selection of independent and dependent variables to test hypotheses should be done with cost, efficiency, and understanding of the wetland system being studied. Study type (e.g., experimental, observational, and assessment) influences the certainty of results. Randomization and replication are the foundation of any study type. In wetlands, impact studies (e.g., BACI [before-after/control-impact] design) are common and usually follow unforeseen events (e.g., hurricanes, wild fire, floods). Sampling design is dictated by study objectives, target population, and defined study area. A robust sampling effort is essential for accurate data. Reduction in statistical and mechanical errors and data management protocols are overlooked features of study design. In addition to statistical tests, estimation of the magnitude (i.e., effect size) of an effect is crucial to interpretation of study results. When judging the merits of results from a study, investigators should independently assess the hypothesis, methodology, study design, statistical approach, and conclusions without regard to how they would have conducted the study. Doing so will facilitate the scientific process.

Student Exercises

1.1.1 Classroom Exercise

In wetland studies, there are usually a number of acceptable study designs to generate knowledge regarding an observed ecological pattern or process, effects of management or anthropogenic impacts, or approximation to a desirable condition or state. The key is use of a defensible study design that allows an investigator to make reliable conclusions and inference from the results of data collection and statistical analysis. Use of critical thought through the study design process prior to data collection will ensure dependable results that can be used to advance understanding of the wetland system being studied and hypotheses being tested.

Many wetland systems are actively managed for certain ecological responses through application of specific environmental conditions; for example, water-level manipulation. These ecological responses are typically production of food resources (e.g., seeds, tubers, invertebrates) for wetland-dependent wildlife. Development of management prescriptions to maximize food production typically requires a set of manipulative experiments to test wetland response to a variety of different environmental conditions. However, measurements of food resources in wetlands can occur without manipulated experiments by relating (e.g., correlated) resource production to observed environmental conditions. Such an approach does provide some evidence of influential variables relative to production of food resources, but lacks rigor to produce a complete understanding of causal relationships. Therefore, it is crucial for investigators to properly design studies of appropriate rigor to generate knowledge of sufficient scientific quality to meet the study objectives.

When managing wetlands for wildlife-forage resources, characteristic environmental conditions that are frequently tested include frequency and timing of wetland drawdowns (dewater to expose soils and sediments) and flooding that affects soil moisture and temperature; oxygen content in soil and water (i.e., aerobic vs. anaerobic conditions); and nutrient availability (e.g., nitrogen, phosphorus). Typically, investigators collect and measure invertebrate and plant response to (1) determine species composition in response to treatments and (2) estimate available biomass of forage resources. In addition, relative composition, distribution, and variation among studied wetlands of source populations (i.e., seed and egg banks) for food resources are characteristically considered influential on results but not a primary interest in a study. Finally, the wildlife species of interest are enumerated in some manner to evaluate the response to available food resources. Much of this volume is devoted to descriptions and recommendations for collecting ecological field and laboratory data for wetlands. The purpose of this exercise is to develop a hypothetical field study of wetlands including development of experimental treatments, objectives, and testable hypotheses.

A public land manager has developed 16, 10-ha wetland units on the floodplain of major river in the southwestern United States. Each unit has been laser-leveled to (1) allow ease in flooding and draining each unit using water-control structures and (2) create a relatively uniform elevation across each unit. Each unit can be manipulated independently, but up to four adjacent units can be manipulated simultaneously. The goal of the land manager is to maximize annual production of natural foods for migratory birds, which use the units for migration and wintering.

The four treatments of interest that coincided with availability of water for flooding include a (1) control, (2) early growing-season drawdown, (3) late growing-season drawdown, and (4) early growing-season drawdown with a late growing-season flood to achieve soil field capacity. All wetland units can be flooded at any time during the migratory and wintering period.

Working in small groups, design a study to test the effect of treatments on forage production and wildlife use of the wetland units. Methodology to measure variables does not necessarily need to be included. In your study design include a description or response to the following questions or statements:

• 1. List 2–4 detailed study objectives

• 2. Provide at least two testable research hypotheses or predictions

• 3. Define and describe a study control

• 4. Provide a minimum of three dependent variables and three independent variables and the units of measurements for each

• 5. Describe a strategy for allocation of treatments among wetland units

• 6. Define the sample frame, study population, and extent of inference from the generated results.

• 7. Describe a potential sampling strategy for each objective

• 8. Include a statement on data management and storage