A review of the statistical principles of geochronometry: towards a more consistent approach for reporting geochronological data

Computational techniques used for determining and reporting isotopic dates must be statistically valid and be consistent between different laboratories in order to permit useful comparison of results. A resume of statistical regression methods presently used is provided, and an attempt is made to illustrate the essential mathematical and conceptual links between different techniques whilst also indicating the pros and cons for different computational models. The model-dependent nature of regression calculations, both isochrons and errorchons, is emphasized throughout, as is the fundamental distinction between calculating a 'date' from a set of isotopic data and assigning real geological 'age' significance to the result. Finally, recommendations are made on the minimum requirements to be met when reporting isotope data. A convention is proposed for reporting data derived by the South African laboratories to facilitate ease of comparison results. 31 refs., 3 figs., 1 tab