Statistical Primer

Sections of this document:

• Background Material
• Invertebrates as Indicators
• Studying Correlations
• Finding Regional Streams
• Working with Raw Data

The National WSA databrowser brings data from the Wadeable Streams Assessment into a web-based, interactive interface that provides easy access to a large number of variables found in this dataset. The goal of this tool is to make it easy to work with this large, sophisticated dataset. Under the hood, the databrowser uses vetted statistical techniques to subset and categorize data and generates completely documented, standardized statistical plots. Addtional links in the user interface provide accesss to subsets of the data as CSV data files, sortable HTML tables or KML files suitable for upload into Google Maps/ Google Earth.

This 'Use Case' page describes how the WSA databrowser might be used by educators at the high school or college in a class that covers stream ecology. This page is not intended to serve as educational material by itself. Rather it is intended to give those designing curriculum an understanding of what the WSA databrowser can do and how those capabilities might be put to use in a classroom setting. Additional material on watershed ecology and conservation can be obtained from the following EPA sites:

• Water > Educationl Resources
• Office of Water > Watersheds

Background material describing watershed ecology can be found on the EPA's Office of Water Watershed Academy site. In particular, the Introduction to Watershed Ecology module is an excellent educational resources that introduces the major concepts needed to use the WSA databrowser. Additional information from the EPA's Biocriteria page will help students better understand the difference between biological and chemical assessments of stream health.

The EPA's Adopt Your Watershed site has great information for showing students how they can become directly engaged in their own watershed. The Volunteer Monitoring section may be of particular interest as it describes how citizens can volunteer to sample exactly the same sort of variables that are found in the WSA databrowser.

A first classroom activity could be focused on learning more about the stream bugs (benthic macroinvertebrates) that are a major focus of the Wadeable Streams Assessment. Using the WSA databrowser, students will be able to investigate an Index of Biotic Integrity (IBI) that is based on populations of benthic macroinvertebrates sampled at WSA sites.

The following pages provide in-depth information describing concepts related to stream invertebrates and IBI indices:

• Bioassessment and Biocriteria
• Biological Indicators of Watershed Health
• About Biological Indicators
• Invertebrates as Indicators
• Intro to the Index of Biotic Integrity

In order to make use of the WSA databrowser students must first become familiar with the specialized terms and variables used in the databrowser. The WSA glossary of terms will be a valuable resource as will the WSA Databrowser Documentation page.

When students first come to the WSA databrowser it will be initialized with the variable titled '6 METRIC NATIONWIDE IBI'. The plot will display a map above a "Cumulative Distribution Function" (CDF) representation of the IBI variable. The CDF plot will probably be unfamiliar to students. A more familiar representation, the box plot, can be obtained by choosing 'Map above box plots (no weights)' from the 'Display statistics as' selector. A description of all plot types is displayed when the user hovers over the question mark associated with this selector:

The task for students in the 'Invertebrates as Indicators' lesson would be to explore the IBI and other bug indexes in different ecoregions with the 'Data subset' and 'Create categories' selectors. The following questions are examples of how students might be engaged in exploring the databrowser:

• Which of the 9 aggregated ecoregions has the highest mean IBI score?
• Subsetting for that ecoregion, which state has the highest/lowest mean IBI score?
• Do other 'Benthic Macroinvertebrate' 'Metric' variables give similar results?
• Which EPA region contains your home state?
• In this EPA region, which aggregated ecoregion has the highest/lowest IBI score?
• In this EPA region, which state has the highest/lowest IBI score?

Having students combine the written documentation on EPA pages with the statistical information available in plots from the WSA databrowser will help them appreciate how ecology, biology and statistics are all used together to inform policy decisions regarding the health of our streams

Having explored the IBI values, essentially the 'results' of the Wadeable Streams Assessment, students can begin applying what they know about ecology to these results by exploring correlations. This second lesson will encourage students to work with scatter plots to explore correlations between various habitat variables and the IBI assessment of stream condition.

If needed, students can restore the WSA databrowser to its initial state by holding down the CTRL-SHIFT keys and then clicking on the browser refresh button. From the initial state, students can beging exploring correlations by choosing 'Scatter plot' from the 'Display statistics as' selector. The initial plot will show 'IBI' vs. 'Mean Embeddedness' which have a strongly negative correlation -- as Embeddedness goes up, IBI goes down. This correlation is also statistically significant as show by the '**' in the legend box on the plot. (See the description of plot types for more information.)

This is a case where a causal relationship can be argued -- a healthy assemblage of stream bugs needs a variety of stream habitats including loose substrate in the stream bed to crawl under.

Again, a set of assignments can motivate students to explore potential correlations visible in the databrowser plots:

• Find two other 'Habitat' variables (drivers) that show a statistically significant negative correlation with IBI score. Suggest how they might be 'causal'.
• Find two other 'Habitat' variables (dirvers) that show a statistically significant positive correlation with IBI score. Suggest how they might be 'causal'.
• For each correlation plot above, use the 'Create categories' selector to determine if the correlation is significant across ecoregions.
• Use the 'Data subset' selector to select the EPA region where you live and then categorize by state. Can still find significant correlations?

Up to this point, the databrowser has been used to explore the the graphical WSA dataset. A more hands-on, quantitative assignment might have the students create their own scatter plots with subsets of data available from the databrowser. To do this, students will need to access the data used in databrowser plots.

The WSA databrowser always creates data files containing the data used in creating each plot. Clicking on the 'HTML table' link underneath a plot will launch a new window or tab that contains the data used in a plot as a sortable table for easy visual inspection. Data for use in spreadsheets or with statistical software can be obtained by clicking on the 'CSV data' link underneath each plot. Depending on how your browser is configured, clicking on 'CSV data' will either display the CSV file in the same window/tab as the HTML table or launch spreadsheet software on your computer, preloaded with the data.

Having students work with scatter plots and especially raw data will give them experience doing quantitative work with ecological data. This lesson can also provide a starting point for discussions about statistical significance and the difference between causal relationships and merely correlated ones.

One way to make the WSA data relevant to students is to ask them to find those WSA sampling sites that are closest to where they live and then visit them either virtually or in real life. Variables with names like 'Mean Bank Canopy Density (%)' are much more meaningful when students can look at aerial photos of the sample site.

The WSA databrowser makes this easy by provding KML files for the sites used in each plot. In the classroom setting students would begin by using the 'Data subset' selector only use the EPA Region containing their state. Clicking on the 'KML file' link below the plot will download the file to their local machine. (If the KML file appears in a browser window/tab just use the browser 'Save File' functionality.)

Many applications can display KML, including Google Earth, Google Maps, Google Maps for mobile, NASA WorldWind, ESRI ArcGIS Explorer, Adobe PhotoShop, AutoCAD, and Yahoo! Pipes. If students have a Google account they can use the WSA KML file with Google Maps right in their browser with the following steps:

1. Direct a browser window to Google Maps
2. log in to their Google account
3. click on 'My Maps' in the blue bar on the left
4. click on 'Create new map'
5. click on 'Import'
6. use the 'Import KML' window to browse to and then upload the WSA KML file downloaded from the databrowser
7. wait for a moment until the sites appear on Google Maps

Named streams and associated MMI (Multi-Metric Index) scores appear in the left panel with site locations on the map. Clicking on a site location on the map will return the name and MMI score while clicking on a named stream in the left panel will pan the map to include that location.

Students can zoom in to a particular site to investigate the level of urbanization or canopy cover near WSA sampling sites.

A set of assignments involving Google Maps might include questions like the following:

• Zoom in on the site nearest your home. Report the site, its MMI score and explain which habitat features visible in the satellite/aerial photos might cause a low/hi score.
• Use the 'HTML table' product to sort sites based on MMI score and repeat the above for three low scoring and three high scoring sites in your EPA region.