A procedure is presented for estimating the relative contributions of geochemical survey, local sampling, and analytical variability to total survey variability. Additionally, F-tests for the statistical significance of the variance components are undertaken to assist in determining if the survey data are fit for purpose in mapping exercises. Based on triplicate analyses, where one of the field duplicates is split to provide an analytical duplicate, the procedure is effective, both statistically and in terms of cost. Open Source, R, Analysis of Variance (ANOVA) software is available to undertake the statistical data investigation. Guidance rules are developed to help determine if the data should be transformed in order to satisfy the assumptions, normality and homoscedasticity, of the ANOVA method, and to accommodate the compositional, constant sum, nature of analytical geochemical data.
- Received October 20, 2011.
- Accepted September 5, 2012.
- © 2013 AAG/Geological Society of London