Causal Analysis/Diagnosis Decision Information System (CADDIS)

Step 3: Evaluate Data from the Case

• 3.1. Overview
• 3.2. In-Depth Look
• 3.3. Results and Next Steps

3.2.5. Manipulation of Exposure

Concept
Field experiments or management actions that increase or decrease exposure to a cause must increase or decrease the biological effect.

Figure 3-7a. Manipulation of Exposure, Supports. The impairment (dead fish) occurs when the causal agent (the effluent) is present but not when it is removed (the effluent is shut down).
(General explanation of symbols)

Figure 3-7b. Manipulation of Exposure, Refutes. The impairment (dead fish) occurs when the causal agent is present (the effluent) and when it is removed (the effluent is shut down).

Examples
Consider the release of toxic substances from a known, unpermitted point source as a candidate cause of decreased mayfly (Ephemeroptera) taxonomic richness. What findings support or weaken the case for this toxic discharge as the cause, based on manipulation of exposure?

• Supporting evidence - Elimination of the unpermitted discharge results in increased mayfly taxonomic richness, similar to levels seen at unimpaired sites.
• Weakening evidence - Elimination of the unpermitted discharge has no noticeable effect on mayfly taxonomic richness.

How do I analyze the data?
The most compelling manipulations are controlled field experiments that involve eliminating or reducing exposure to a source (e.g., fencing cattle out of a stream), changing the level of an agent (e.g., adding large woody debris to a stream), or artificially inducing exposure to a stressor (e.g., placing caged organisms at sites with varying levels of the suspected stressor). Ideally, changes in both the proximate stressor and the observed biological effect should be measured before and after manipulation. The power of this type of evidence comes from the control of exposure achieved by deliberate manipulation of events, and even the potential for replication. However, uncontrolled experiments (e.g., elimination of effluent during facility shutdown) also can be useful.

Statistical techniques used to evaluate manipulation of exposure include time series analysis; statistical tests such as the t-test, when manipulations are replicated and randomized; or statistical techniques based on rare events, such as the reappearance of sensitive taxa. Uncertainty in the data can be introduced when other events, natural factors, or other causes co-occur with the variable being manipulated; sampling designs such as before-after-control-impact (BACI) can help control for some of these factors. Recovery rates and treatment effectiveness also are sources of uncertainty and should be taken into account when analyzing results.

What evidence would support or weaken the case for a candidate cause?
Supports

• Data showing that the effect decreases when or where exposure to the candidate cause is eliminated or reduced
• Data showing that the effect appears or increases when or where exposure to the candidate cause is initiated or increased

Weakens

• Data showing that the effect persists when exposure to the candidate cause is removed is eliminated
• Data showing that the effect does not occur when exposure to the candidate cause is initiated

Refutes

• Data showing, with a high degree of certainty, that the effect either does not change or does not change in the expected direction with manipulation of exposure.

How do I score the evidence?

Helpful tips

• The power of this type of evidence resides in the deliberate manipulation of exposure, which can be designed to minimize confounding factors.
• Note the distinction between manipulation of exposure and normal variation in exposure. Normal variation is analyzed in terms of co-occurrence and stressor-response relationships from the field.