PARENT SESSION
TP6b Acceptability of risk: parameters determined by Science and Society
4:30 PM to 6:30 PM, Tuesday, 08 May 2001
Session Chair: J. Costa-David
Room 6

(267) Managing the precautionary principle and uncertainty with statistical power analysis.

Sanderson, Hans¹, Petersen, Søren², ^{1 2}

ABSTRACT- The diversity of interpretation and subsequent lack of implementation and enforcement of the precautionary principle has been an important issue in the European environmental discourse in the recent years. On February 2nd 2000, the European Commission published a communication on it's interpretation of the precautionary principle, and opened a debate about the principle. The distinction between precaution and prevention is not obvious in the communication, hampering the relevance of, and the need for, the principle. The aim of this presentation is a) to discuss briefly the difference between precaution and prevention and b) to present a statistical frame for managing the EU Commission's version of the precautionary principle and uncertainty via statistical power analysis. In null hypothesis testing of complex environmental issues, it is not enough simply to assess the risk of committing a Type I error (=.05) - known as the "producers risk". The determination of as the risk of committing a Type II error (the "recipients risk") and the power of the test (1-) should also be tested on demand by the EPA's as a measure of precaution. It is vital for the regulators to have this complete information as the risk of committing a Type II error increases with complexity of the test, which is typically the case in ecological risk assessments due to increasing variation and uncertainty. The size of and thus the required power of a test should a priori be set according to a societal cost benefit analysis associated with the risk. From the power analysis it is possible to refine the design requirements of the investigation in terms of needed effect size () and replicates (n).

Key words: power analysis, type II error, uncertainty, risk management