Statistical Primer

Statistical Testing

These pages present a simple introduction to a variety of statistical methods that are commonly used in the field of biological monitoring to relate changes in the biological assemblage to measures of human disturbance or ecological condition. These pages use examples from the Mid-Atlantic Integrated Assessment project (MAIA) which includes EPA’s Environmental Monitoring and Assessment Program (EMAP) for streams is Region III.

When large amounts of data are collected, statistics represent a way to summarize and manage the data. For example, we might select a response variable such as the average number of species collected at site rather than the entire list of species found. In this case, the statistic might be the mean of taxa richness. Based on an understanding of ecological processes, we might hypothesize that the number of species goes down as human disturbance goes up. Correlation would be an appropriate statistical method to test this idea. When we apply statistical methods or their associated tests to real data, we must make some assumptions as we translate the complexities of the real world into a simpler statistical representation, or model. Simplifying assumptions relate to the distribution of the variables (e.g., normal), the relationships between variables (e.g., liner), and the variables selected to build the model. For correlation, we can select a parametric or a nonparametric model, depending on our assumptions about how measures of taxa richness are distributed across sites.

Sampling Approaches

The only sampling design that can satisfactorily answer questions about both the overall regional condition and the condition of all the individual water bodies is a census sampling design in which all water bodies are sampled. Because census sampling is too expensive, decisions must be made regarding which locations to visit each year. Status, trend, and targeted sampling designs represent different approaches to site selection and answer different types of questions.

Statistical Power Analysis

Statistical power is the probability of detecting a difference when a difference truly exists. These pages describe the measures that are needed to calculate the power of a given sampling design and statistical test to detect change over time.

Data Browser for Western EMAP

The data browser connects data sets from EMAP-West to the Statistical calculation routines needed for the probabilistic survey design used to collect the data. From the interactive web interface, users can dynamically map measures of habitat and benthic condition, calculate summary statistics, and plot the results.

Site Map

The site map includes an outline of the web pages on this site and an index for the statistical tests described in the Statistical Primer.