Using Real Interpretative Differences to Assess Inter-laboratory Isotopic Variability Due to Sample Preparation

Abstract

Here we present a two-part study to assess isotope data comparability between two sample preparation laboratories, based on the concept of Real Interpretative Differences (RID). Inter-laboratory isotopic variability should be evaluated prior to compiling data from different sources. Isotopic variability can result from two main causes: sample preparation and sample analysis. We compared the carbon and nitrogen isotope delta (δ) values of 20 modern human bone collagen sample pairs—representing seven individuals from diverse backgrounds—that
were prepared close in time at two laboratories and then analyzed at one facility. From this study, values for RID of preparation, or RIDprep, were calculated for bone collagen as 0.43‰ and 0.75‰ for δ¹³C values and δ¹⁵N values, respectively. These RIDprep values were tested using a set of 30 bone collagen sample pairs—representing 30 modern individuals—that were first prepared over a span of five years (2014–2019) at one laboratory; they were prepared again at the second laboratory in 2019. While the differences in isotope δ values between the paired samples in the second study were statistically significant, they did not exceed the calculated RIDprep values for either carbon (30 comparisons) or nitrogen (30 comparisons)—with the exception of one carbon difference in a single paired bone sample. The inherent error rate in combining the isotope test results was thus calculated as 1.7%. This study demonstrated that the isotope data generated from human bone collagen prepared at one laboratory can be reliably compared to data generated from samples prepared at the other laboratory.