ENVIRONMETRICS - Beyond the p-value
by Professor David Fox
Abstract: Environmetrics is a relatively new discipline. It is concerned with the development and application of statistical tools for robust, fit-for-purpose methods of environmental monitoring, sampling, and assessment. While 'conventional' modes of statistical analysis such
as t-tests and ANOVA have served the scientific community extremely well
over the past eight decades it is Professor Fox's contention that, by and large, they are flawed approaches for environmental assessment.
Nevertheless, as noted by Cohen (1994) "after four decades of severe
criticism, the ritual of mechanical dichotomous decisions around a sacred 0.05 criterion still persist". Statistics courses for students in the
biological and environmental sciences have only served to perpetuate what
has been referred to as the "p-value culture" (Nelder 1999) that focus
more on "asterisk hunting" (Lenth 2001) than they do on appropriate model choice and scientific interpretation.
In this talk Professor Fox will illustrate some of the pitfalls of
classical modes of statistical inference for environmental studies and
discuss the emergence of the Bayesian paradigm as a preferred alternative
among many environmental scientists (Clark 2005). Examples include air
quality monitoring in Whyalla and Port Pirie, arsenic levels in a PNG
river, and the establishment of uranium 'trigger' levels for the Magella
Creek in the Northern Territory.
Clark, J.S. (2005) "Why Environmental Statisticans are becoming
Bayesians", Ecology Letters, 8, 2-14.
Cohen, J. (1994) "The Earth is round (p<.05)", American Psychologist,
Lenth, R. V. (2001) "Some Practical Guidelines for Effective Sample Size
Determination", The American Statistician, 55, 187-193.
Nelder, J.A. (1999) "From Statistics to Statistical Science", The
Statistician, 48(2), 257-269.
For More Information: Paul Norbury tel. 8344 5534 http://www.ms.unimelb.edu.au/~pnorbury/colloquium/colloquium-2005.html