To be effective, environmental monitoring programs must be able to detect a significant difference in environmental conditions whenever one actually occurs. Poorly designed programs may not be able to do this with the result that environmental impacts are left unmanaged. The ability to detect statistical differences when they actually exist is termed statistical power.
Statistical power is dependent upon four factors as outlined in Table 1.
Table 1: Factors affecting statistical power.
How do we optimise our monitoring effort and balance cost and effectiveness? Some of the more obvious ways are outlined in Table 2.
Table 2: Means of improving statistical power.
Bioequivalence – reversing the burden of proof
The current approach to monitoring tends to “reward” poor monitoring design with results that never show, and are unlikely to ever show, a significant impact. Bioequivalence is an approach where good monitoring designs are rewarded (for example, Rogers et al. 1993).
Bioequivalence, which is often used to test the efficacy of generically branded medicines, reverses this burden of proof, so that the null hypothesis is that an impact has occurred and the onus is on rejecting the null hypothesis, by demonstrating that environmental conditions are significantly higher than a minimum threshold. Rejection of the null hypothesis is more likely to occur if the monitoring has sufficient power. This approach has not often been used in environmental monitoring, but deserves further consideration, especially if a precautionary approach is preferred.
Astron’s capability in monitoring design
Astron has a broad range of experience in designing and executing monitoring programs, and has the tools and expertise to examine and optimise the statistical power of monitoring designs. For further information on the preparation of monitoring plans generally, please contact our Biodiversity Team via our online form or call us on (08) 9421 9600.
Reference
Rogers JL, Howard KI, Vessey JT. 1993, ‘Using significance tests to evaluate equivalence between two experimental groups’. Psychological Bulletin. Vol. 113, pp. 553-565.
Field, S, O'Conner, P, Tyre, A & Possingham, H 2007, 'Making monitoring meaningful', Austral Ecology, vol. 32, pp. 485-91.