|Conducting Environmental Impact Assessment in Developing Countries (United Nations University, 1999, 375 p.)|
|7. EIA communication|
Communication of EIA findings to policy and decision makers is difficult because they are often not technically trained. The task of communication is one of translation and interpretation from the language of the scientist into a clear and concise summary that matches the client's constraints and timetables. Another problem in using research results is the gap between the expectations of decision makers for certainty and the probabilistic realities of science. If a scientist reveals this uncertainty, the client may reject the findings as unhelpful, whereas disguising the uncertainty may cause the scientist to lose credibility when unpredictable results do occur.
Prediction should be straightforward, logical, and systematic regardless of the completeness or accuracy of the data available. All assumptions must be explicitly stated. The users of the assessment can follow the predictive method and, if they wish, substitute alternative assumptions where factual information is lacking.
A four-part format in reporting predictions is helpful in avoiding misunderstanding by the users about the uncertainty that inevitably accompanies EIA results.
First, the prediction should state what is known and with what confidence (a narrative statement of the statistical reliability). For example, "With the chance of being wrong one time out of ten, we believe that the increase in turbidity will result in a 25 to 50% lower catch of fish.''
Second, state what is not known and why. For example, "Tests have been run only on species A and B so we cannot be sure of the response of other organisms.''
Third, explain what could be learned from further investigation if more time and money were available.
Fourth, indicate what should be known in order to proceed in a prudent manner, that is the risk of going ahead based on present knowledge as opposed to the risk of delaying the project.
To summarize, the EIA predictions should state what we do know, do not know, could know, and should know.