Causal inference in environmental science: Applying the scientific method to building evidence from the literature

Two widely accepted tenets of science are that observations arising from one study can be repeated in another and the application of the hypothetico-deductive method.  Observations commonly involve causes and effects that are immediately apparent such as flicking a switch to turn on a light.  The flaw in assuming direct cause and effect between the switch and light becomes immediately apparent when there is a power outage, or the bulb fails, indicating several possible component causes that produce the observed effect.  Repeated observations meet the first tenet and by using inductive reasoning may produce a conceptual model of causes and effects, but the proposed relationships need to be tested.    Epidemiologists have addressed testing where experiments are often not possible with human subjects by accumulating evidence from multiple published studies against a set of criteria.  Likewise, rigorous tests are often not possible in freshwaters because an effect has already occurred, or lack of treatment replication.  Thus, applying these tenets to build confidence in causal inference, using evidence from the published literature has much potential in freshwater science.  Fundamental to the application is evaluating the quality of individual studies and subsequently combining the evidence to reach a conclusion.