| Index | Last Week | Contact DQ News | Latest Issue |
| Featured Articles | Data & Information | Science & Technology | Education |
| Economics | Business | Law | Medicine |
Industrial Quality Methods Have Data Quality Applications
A lengthy article in The New York Times on December 7th discussed the "Six Sigma" industrial quality methodology that has promising applications in data and information quality. "Six Sigma" refers to a statistical quality control method that has been successfully used by Motorola, General Electric and other industrial corporations to greatly reduce the production of defective products. The term "sigma" refers to the the Greek letter statisticians use to define standard deviation. Two standard deviations encompasses about 95% of the area under a "normal" curve, assuming the quantity being measured is randomly distributed. Three sigma is about 99% and six sigma encompasses so small an area that industrial statisticians commonly refer to products produced using six sigma methodology by the term "parts-per-million." (In industrial quality, the term "parts-per-million" almost always means parts-per-million non-conforming or defective. A well-designed and maintained six sigma manufacturing system will produce products at the rate of about three-per-million non-conforming).
The big advantage of "Six Sigma" is that it gets people away from thinking that 95% is good to thinking that 50,000 failures per million is bad. Such thinking could have profound and immediate benefits for data and information quality, where a 95% effort is all too often considered "good enough." As one industrial statistician recently commented: "Data and information quality are at the same place today as industrial and product quality were 20 years ago."
Unfortunately, the Times article concentrated on a narrative about how the bugs were designed out of a high-tech GE product - a fast X-Ray CT scanner - rather than on state-of-the-art industrial quality - how to design excellent products and processes without bugs. The article was written by Claudia Deutsch and appears on page C1.
Oil Firms Spy for Competitors' Data
According to a front page report in the December 7th issue of The Wall Street Journal, almost all of the world's oil companies rely on people who spy on other oil firms - hoping that a steady stream of data, ranging from rumor to confidential production statistics, will give them an edge over their industry rivals.
Spies (who are usually referred to as "scouts" in the oil industry) obtain data from a wide variety of sources - roughnecks to oil company geologists - and feed their "data" to competing corporations or anyone else they deem to be an "interested party."
Hundreds of "scouts" operate around the world, from Azerbaijan to Alberta, Canada. Scouts often use high-tech devices - like night vision scopes - to determine what is happening (or not happening) on drilling platforms. Around the Gulf of Mexico there are regular weekly sessions where "data" about oil corporations are traded among scouts. According to the Journal, the ultimate prize is acquiring a "well log" - detailed geological data that documents a drilling crew's progress. The article was written by Journal staff reporter Christopher Cooper.
Placebo Surgery Used in Clinical Trials
A front page report in the December 11th issue of The Wall Street Journal discussed the increasing use of sham or "placebo" surgery as a method of evaluating the efficacy of surgical operations and techniques. Unlike placebo drug trials, surgeons who employ placebo surgery as part of a clinical trials process make incisions in soft tissue, drill holes in patients' skulls, and perform other procedures that may cause patients immediate discomfort and pain.
Unlike placebo-controlled clinical trials for drugs - where neither patients nor doctors know who is taking a drug and who is receiving a placebo - surgeons performing placebo surgery often know who is receiving the placebo or the experimental treatment. According to the Journal, both the risk of harm to patients and the bias to the experimental data are often outweighed by the desire of surgeons developing new techniques, corporations funding surgical R & D, HMOs, and federal agencies (like the National Institutes of Health) to quickly obtain data about the efficacy of surgical procedures. Many individual surgeons and organizations involved in surgical research feel that control-groups and placebos are preferable to the anecdotal evidence about how well surgical techniques "work." The article was written by Journal staff reporter Laura Johannes.
Go to: Data Quality Home Page
Comments: dqemail@aol.com (1998-12-6)