This README.txt file was generated on <20220727> by Joseph Kadane. ------------------- GENERAL INFORMATION ------------------- 1. Title of Dataset: Datasets for "The reliability of sampling to measure the weight of seized powder containing heroin" 2. Author Information First and Corresponding Author Contact Information Joseph Kadane Department of Statistics & Data Science Carnegie Mellon University, Pittsburgh, PA 15213, USA kadane@stat.cmu.edu Emily Wilkinson Allegheny County Office of the Medical Examiner 1520 Penn Ave., Pittsburgh, PA 15222, USA Josh Yohannan Allegheny County Office of the Medical Examiner 1520 Penn Ave., Pittsburgh, PA 15222, USA --------------------- DATA & FILE OVERVIEW --------------------- Directory of Files: A. Filename: 20220727_RLDAT_KWY2.csv Short description: 200 bags from a drug seizure were weighed. First the gross (filled) weight of each bag was measured. Then each bag was emptied, and the powder contents were weighed. The stamp bags are authentic street samples without any prior laboratory analyst contact. The bags are from an adjudicated case and were retained by the laboratory for training and research purposes. This is real data of the kind our laboratory sees regularly. This file contains 200 rows and four columns indicating ID (bag ID), Gross (total weight of filled bag in grams), Bag (weight of bag in grams), and Powder (weight of powder in grams). B. Filename: 20220727_SCI80_KWY2.csv Short description: For training purposes, the laboratory filled 998 bags with a visually consistent amount of lactose powder. Eight scientists were each given ten randomly selected bags and asked to record the gross and net weight, resulting in a complete dataset of size 80. This file contains 80 rows and four columns indicating ID (bag ID), Gross (total weight of filled bag in grams), Bag (weight of bag in grams), and Powder (weight of powder in grams). C. Filename: 20220727_INT_KWY2.csv Short description: An intern, who had no familiarity with what routine contents of stamp bags look like, was shown what 20 mg of powder looks like, and was asked to fill 200 pristine stamp bags with 5.031 g of a mixture of benzocaine, caffeine, diphenhydramine, lactose and quinine. Although the target was 200 bags, only 179 were filled because the bags were filled through visual estimation. The gross and net weight were then measured by laboratory personnel. This file contains 179 rows and four columns indicating ID (bag ID), Gross (total weight of filled bag in grams), Bag (weight of bag in grams), and Powder (weight of powder in grams). D. Filename: 20220727_SEED_KWY2.csv Short description: This data set is different in content and purpose from the preceding three. In each of the former, the vast majority of the weight (at least 85%) is in the packaging. In SEED, roughly 85% of the weight is in the contents. About 900 g of wildflower seeds were distributed, roughly evenly, among 98 envelopes. Use of this data set is essentially a check on the programming, to see if it behaves properly with such different input. This file contains 98 rows and four columns indicating ID (packet ID), Gross (gross weight of filled packet in grams), Packet (weight of packaging in grams) and Seed (weight of contents in grams).