Statistical analysis of RbSr isotope data by the "bootstrap" method

The "bootstrap" is a method for the evaluation of statistical parameters for data that cannot be assumed to conform to a normal distribution. We have used the bootstrap method to test the results of RbSr isochron calculations for a large number of RbSr isotope data-sets defining both "isochrons" and "errorchrons". The results indicate that in many cases the confidence intervals for ages and intercepts ("initial ratios") are not symmetrically disposed about the calculated age and intercept values. The precisions for ages and intercepts obtained from isochron calculations must therefore be used with caution. This is especially the case for arrays of data points where a few points have very high or very low Rb/Sr and Sr-isotope ratios relative to the majority of the data points. The bootstrap method shows that the MSWD-value obtained for an isochron is a very poor estimator of MSWD-values that could be obtained from other sample suites from the same rock.