Jump to main content or area navigation.

Contact Us

CADDIS Volume 4: Data Analysis

Getting Started

Organizing Data
  • Authors: S.B. Norton, K. Schofield, L. Alexander, S.M. Cormier,
    G.W. Suter II, P. Shaw-Allen

Organizing Data Along Causal Pathways

When beginning a causal analysis, we strongly recommend using a conceptual diagram to systematically classify data according to causal pathways linking sources, stressors (candidate causes), and the biological responses that constitute impairments. Mapping available data onto the hypothesized causal pathways makes apparent what types of analyses are possible and where data gaps exist.

The most compelling analyses demonstrate that a proximate stressor is associated with the biological response at the site. Other associations between steps in the causal pathway can increase confidence in the body of evidence for or against each candidate cause. These include the presence of a source for the stressor or intermediate steps that give rise to the stressor at the site.

Data from other locations and studies also can be related to the conceptual diagram. For example, results from other studies may be used to determine if the mechanisms or modes of action are reasonable, and the intensity of an intermediate step or interacting stressor are sufficient for a credible causal pathway. Finally, data from other locations and studies may be used to quantitatively model relationships between elements in the pathway, so that data from the impaired site can be compared to the model results.

Figure 4 is a conceptual diagram illustrating possible elements and structure of a causal pathway. Conceptual diagrams for common stressors are available for download in Volume 2: Sources, Stressors & Responses of this website. An interactive on-line tool for constructing and modifying conceptual diagrams also is available, in Volume 5: Causal Databases.

causal pathway
Figure 4. Conceptual diagram of elements in a causal pathway.
  • Measures of stressor sources are useful for identifying potential candidate causes, for completing the causal pathway, and, after the causal analysis, for source apportionment. However, source measurements can be difficult to use in a site-specific causal evaluation because they are often large in scale, can distribute stressors broadly and may contribute multiple stressors. Information on sources that produce many proximate stressors cannot be used to distinguish among those stressors. For example, increases in impervious surface area have been linked to proximate stressors such as increased flow extremes, temperature spikes, increased toxic substances, and decreased dissolved oxygen (Walsh et al. 2005, Schueler 2003). Therefore, measuring impervious surface does not help distinguish among these stressors.

  • Measures of the proximate stressor in the environment (e.g., degree of siltation, dissolved oxygen concentrations, or chemical concentrations) establish whether it occurs at elevated levels when compared to regional references or some other standard.

    • Obtaining measurements of the proximate stressor that can be associated with the effect can be challenging. In some cases, the candidate cause is the lack of a required resource, such as nesting habitat. In these cases, measurements can establish that the resource is indeed missing at the place and time it would be required by an organism.

    • When measurements of the proximate stressor are not available, surrogates can be sought. Information on the location and attributes of possible sources can be useful surrogates. This information is particularly important for intermittent stressors (e.g., high flow events) or stressors that degrade quickly (e.g., some pesticides).

  • Measures of factors influencing or interacting with the proximate stressor are important in establishing whether complex causal pathways are complete. For example, one pathway by which excess nutrients affect stream biota is by stimulating periphyton growth, which respires or decays and reduces dissolved oxygen. To evaluate this pathway, data on dissolved oxygen concentrations may be supplemented with data on other steps in the causal pathway (e.g., nutrient concentrations and periphyton biomass). Evidence of an interacting stressor, such as low or fluctuating pH, alerts the analyst to consider possible effects on proximate stressors, such as increased solubility and bioavailability of metals or nutrients.

  • Measures of exposure, physiological mechanisms or other evidence of relevant modes of action may be used to verify that biologically significant exposure has occurred. Measurements might include biomarkers of exposure, tissue residues, or abundances of organisms representing different functional feeding groups (e.g., increase in filter feeding insects).

  • Measures of the biological impairment characterize the biological responses of primary interest (blue oval in Figure 4). If responses are very specific, they may be diagnostic of the cause. Be aware that biological measures also may be proximate stressors (e.g., an introduced predator) or steps in a causal pathway (e.g., carp as part of the causal network for increased suspended solids).

Helpful tips

  • Remember that data (evidence) may strengthen or weaken any given pathway.
  • If data characterizing any portion of a pathway are absent, the role of that pathway in contributing to the impairment becomes less certain.
  • Where there are sufficient data, structural equation models linking several strata within a conceptual diagram could be applied. This approach may be particularly useful for complex pathways (see the above example for nutrients).

Top of page


Jump to main content.