Criteria for Causation

Criteria similar to those listed by Adams (2003) are used to establish causality and are derivatives of Koch s postulates and Hume s criteria. The list includes (1) strength of association, (2) consistency of association, (3) specificity of association, (4) time order or temporality, (5) biological gradient over space and time, (6) experimental evidence available, and (7) biological plausibility. In many instances, especially at a regional scale and over long periods of time, meeting the requirements for each of these criteria can be difficult. 

Items 4 and 5 are related. Item 4 is a temporal gradient and requires a dataset of sufficient length compared to the dynamics of the effect and the potential causative agents. If a long duration dynamic is involved such as the PDO, then 30 years of data reflects only one cycle. There may also be multiple causative 

Experimental evidence such as toxicity tests or the induction of disease under controlled conditions (Item 6) can also be coupled with field observations to establish cause-effect mechanisms. Experimental evidence is critical for testing specific predictions made by hypotheses designed to predict large-scale relationships and should be included whenever practical. 

Item 7 is a composite of Items 1 to 6, but also is related to the sensitivity of the observer to accept uncertainty in whatever defines plausible. Clearly a mechanism that has been confirmed experimentally and confirmed by field observations is ideal. The paradigms in which the risk assessment group is working is also bound by the expectations of plausibility. In working at a landscape level over a period of decades where multiple causes are likely to be present, this criteria becomes less attainable. 

---

With some possible exceptions (to be discussed in Chapter 9), dose-response relations identical or similar to those shown in Figure 3.1 are observed for all expressions of chemical toxicity. Indeed, the absence of such a dose-response relationship is often used as evidence that a chemical has not caused a particular response. Criteria for causation in... 

---