Analytical Reliability in the Decision Making Process-The Numbers Game

This paper is not intended to give the pros or cons of conventional sampling methods (generators of numbers). Rather, our intent is to emphasize the importance of the credibility and integrity of the generated number, by what ever methodology selected, and point out some problems that can be encountered when numbers are poorly generated or analyzed. As generators of numbers we are responsible for the end-products of data analysis for the decision maker. In response to a 1977 Presidential directive to the Council on Environmental Quality, CEQ, an Interagency Task Force was developed to review present federal environmental monitoring and data collection programs and to recommend effective improvements. The Federal Government depends upon the analyses that result from such programs to direct sound policy and decisions. These directives affect the public health and welfare and result in large annual expenditures of funds by all levels of government and the private sector. During the past few years, several major federal environmental and data collection programs were found to be inadequate. Although concern about the reliability of analytical results has always existed, this concern has come more into focus because of the growing interest and activity in environmental pollution control, with its heavy reliance on analytical results for enforcement, regulation, and litigation. This concern primarily is due to the inherent limitations in the conduct and analysis of observed measurements. Certainly, since there are no absolutes in analytical results, in terms of a particular protocol, some indication of value or reliability of the results is needed. Consistent with environmental pollution control, range management deals with precious resources and cannot be ignored when implementing monitoring and data collection programs. Vegetation data, as do other natural resource data, relate thecondition of a system at a particular place and time. It is not possible to perform verifying remeasurement because of the inherent special and temporal variability of the system. Additionally, if a sample is collected and analyzed, the nature of the sample is often changed in the measurement process (i.e. clipping, water quality dynamics, etc.). Because of the impossibility of verifying past measurements, quality assurance can only rest on documented application ofproven methodology by qualifiedpersonnel following accepted guidelines