A deleted cluster jackknifed sample is created by filtering out (deleting) clusters of data, where d is the percent of clusters removed at random.
This problem is about cluster-based jackknife sampling. The data is grouped into clusters, and a random percentage of clusters is removed to form the deleted sample. The key is to delete entire clusters, not individual rows, so the implementation typically relies on grouping and random selection of cluster IDs before filtering the remaining records.
END
