Improved Statistical Methods for Reporting
A more appropriate statistical analytical metric (not using "p values" alone) to quantify the "strength-of-evidence" in data has been identified. This has been incorporated into a refined hypothesis-testing/confidence-interval framework (McBride et al., 2013) and the Time Trends and Equivalence freeware, which is widely used by regional councils.
For the next phase of the research we are developing new statistical approaches to make use of ‘null values’ (or values below the detectable level) in monitoring data.