Data Torturing and the Misuse of Statistical Tools
Statistical concepts, methods, and tools are often used in the implementation of statistical thinking. Unfortunately, statistical tools are all too often misused by not applying them in the context of statistical thinking that focuses on processes, variation, and data. The consequences of this misuse may be ''data torturing'' or going beyond reasonable interpretation of the facts due to a misunderstanding of the processes creating the data or the misinterpretation of variability in the data. In the hope of averting future misuse and data torturing, examples are provided where the application of common statistical tools, in the absence of statistical thinking, provides deceptive results by not adequately representing the underlying process and variability. For each of the examples, a discussion is provided on how applying the concepts of statistical thinking may have prevented the data torturing. The lessons learned from these examples will provide an increased awareness of the potential for many statistical methods to mislead and a better understanding of how statistical thinking broadens and increases the effectiveness of statistical tools.
- Research Organization:
- Sandia National Lab. (SNL-NM), Albuquerque, NM (United States); Sandia National Lab. (SNL-CA), Livermore, CA (United States)
- Sponsoring Organization:
- US Department of Energy (US)
- DOE Contract Number:
- AC04-94AL85000
- OSTI ID:
- 10185
- Report Number(s):
- SAND99-2178C; TRN: AH200125%%301
- Resource Relation:
- Conference: Annual Quality Congress, Indianapolis, IN (US), 05/08/2000--05/10/2000; Other Information: PBD: 16 Aug 1999
- Country of Publication:
- United States
- Language:
- English
Similar Records
Sorry, wrong number: The use and misuse of numerical facts in analysis and media reporting of energy issues
New look at energy and GDP (gross domestic product) correlation. [New statistical approach to computer analysis]