Causal Analysis/Diagnosis Decision Information System (CADDIS)

• Data Analysis Methods
• Fundamentals of Data Analysis
• Get Data Analysis Tools

DA.1. Data Sources

When assembling evidence, the data used may be collected specifically for the causal analysis at hand or you may have data from elsewhere that was collected for other purposes. In addition, data generated by models can be a vital component of a causal analysis when sufficient data to support or refute relationships between stressors and biological impairments is not otherwise available.

DA.1.1. Data from the Case

Data from the case includes data collected at the impaired site and one or more local reference locations. The primary data you will have in hand are the data used to classify the site as impaired. Additional data may be collected when you need more information to support or refute relationships between stressors and biological impairments at a given site. There are a variety of ways to go about collecting the data, but in any case you must consider the goals and objectives of the study to determine the sampling design.

If you are planning data collection, helpful advice may be found in U.S. EPA's Guidance on Choosing a Sampling Design for Collecting Environmental Data (PDF) (178 pp, 1 MB, About PDF). Suggestions for saving and managing the data can be found at Resources for Planning New Data Collections.

DA.1.2. Data from Elsewhere

Data from other sites were collected for purposes other than to be used in your causal analysis. This includes information collected from journal articles, industry and government publications, and regional monitoring studies. These data usually form the bulk of the data used for causal analysis.

Information from journal articles and other publications may be collected from a data base compilation such as the ECOTOX. Databases are useful for identifying which data you will need and conducting initial exploratory comparisons, but the original source must be consulted if the information becomes critical to the causal assessment.

Monitoring programs provide information on regional field conditions which can be helpful to a causal analysis. When using regional data, the sampling approach used in selecting sites determines how data may be analyzed and how results may be used.

• Data collected using a probability design may be weighted such that analysis results represent any site found within the watershed, region or other sampling frame, not just those that were sampled.

• Data collected using a targeted design produces analyses that represent the sites sampled. This means the data are useful for:

  • Step 3: Evaluate data from the case, when the assessment site is part of the dataset.

  • Step 4: Evaluate data from elsewhere, when the assessment site is not part of the dataset. Analysis should not include data from the assessment site. Rather, stressor and response levels at the assessment site can be compared to stressor-response relationships revealed by the analysis to determine whether observations at the assessment site are consistent with relationships observed in the dataset.

• Observations from similar cases can yield insights into the current investigation, particularly if other investigators have implemented management actions or identified diagnostic symptoms. Unfortunately, there is no central repository of documented field investigations, and these studies often are not published. You must rely on your network of colleagues for much of this information or refer to the resources described above for assistance.

DA.1.3. Data Produced from Models

Modeling can be a means for developing surrogate data when insufficient field or laboratory data are available. For example, a watershed model such as BASINS can generate estimates of water quality parameters at specified locations. A detailed process for developing or selecting pre-existing models to generate quality data can be found at Guidance for Quality Assurance Project Plans for Modeling (PDF) (121 pp, 615 Kb, About PDF).