Cluster analysis was undertaken using IBM SPSS Statistics 20. Variables were first standardized to produce z scores so that each would have equal weighting in the models. Only samples that provided valid data for all three isotope provenancing indices were included in the analysis (n = 122). Eight samples that provided poor δ34S replicates were omitted, and the single sample from Mount Pleasant (MP107) with anomalous δ15N and δ18O values (that can be explained by an unusual diet) was also omitted as this would alter the model for reasons other than origins. The hierarchical cluster analysis was then undertaken to guide the predefined number of clusters, described as K, in K-means cluster analysis. This is a measure of dissimilarity, and one output is a dendrogram, which displays the rescaled distance between samples in terms of their isotope signatures. This exploratory analysis informs the K value by using the number of clusters on the dendrogram at different rescaled distance values. The hierarchical cluster analysis used the “between-groups average linkage” cluster method with a squared Euclidean distance interval measure.

Note: The content above has been extracted from a research article, so it may not display correctly.

Please log in to submit your questions online.
Your question will be posted on the Bio-101 website. We will send your questions to the authors of this protocol and Bio-protocol community members who are experienced with this method. you will be informed using the email address associated with your Bio-protocol account.

We use cookies on this site to enhance your user experience. By using our website, you are agreeing to allow the storage of cookies on your computer.