Project Team: Richard Norris, Peter Liston, Sue Nichols,
James Mugodo, Mike Peat, Julie Gallimore
The Application of Multiple Lines and Levels of Evidence (MLLE) to an ecological
question is new, and it is new to all of us.
- Lines of evidence = evidence coming from different groups of animals or
process measures (diatoms, birds, fish, benthic metabolism).
- Levels of evidence = evidence within a particular line (diatoms) that applies
to the different criteria (presence of a biological response, specificity
of association, etc).
 The
D240 team concluded that we needed to conduct a pilot application of MLLE to
clarify our application of the technique in project D240. The University of
Canberra team has led the pilot with the other team members providing input
primarily via email discussion. We needed to clarify our understanding and use
of MLLE, first by making the distinction between MLLE and WOE (weight of evidence):
- WOE involves accumulation of evidence that supports the hypothesis until
one concludes that the hypothesis has been proven.
- MLLE also accumulates evidence but requires that one searches for evidence
that would support alternative hypotheses.
|
MLLE has been used for two quite different purposes; 1) assessment of condition
at a site, and 2) evaluating causality between a potential stressor and a biological
response. In project D240 we are trialing MLLE to assess causality. We are refining
the technique at two levels; 1) ensuring that the different levels of evidence
are meaningful and practical when we apply them and: 2) developing a series
of decision points about the application of MLLE data that will form a framework
for its application. The pilot application currently in progress is identifying
uncertainties, decision points and data needs. Issues under consideration include:
- The definition and quantification of the levels of evidence;
- steps in the overall MLLE framework;
- the type of question to be asked/answerd;
- what are comparable studies;
- what are the rules regarding the literatue search;
- categories used to record information from comparable studies (line, response,
study type, alpha and beta levels, effect size, etc.);
- how do we weight the evidence;
- how do we sub-group the lines of evidence;
- how will we use the various lines and levels of evidence to provide confidence
in our conclusions.
- The pilot is due for completion by 31 March 2004.
|
Professor Richard Norris 