Elements Of Deceptive Data Presentation Insurance Protection For Decision Making

Efficiency of industrial research and development depends on the ability of engineers and scientists to collect and present data to top management for decisions. In many cases, the data presented are so deceptive that management makes costly wrong decisions. Since management personnel often are selected on the basis of higher administrative quality than technical ability, it is the responsibility of the industrial scientist or research engineer to make certain that the data presented tell a completely ac- curate story of each industrial problem. It is rare that misleading data are deliberately presented, but when this is done, it is? of course, dishonest. The rise of deceptive data usually stems from the lack of knowledge of experimental design and statistical methods of analysis. Most of the better chemical engineering schools have recognized the need for courses in industrial experimentation and many of them offer complete courses at both graduate and un- dergrarluate levels. By contrast, many of the mining and metallurgical schools have not recognized the importance of industrial experimentation courses and some of them, in fact, do not allow credit toward a mining or metallurgical engineering degree for statistical courses even when taken as an elective. The sole purpose of this paper is to make a plea for mining and metallurgical schools to present required industrial. statistical courses for their graduates. It is true that fine industrial research has been accomplished in the past without the assistance of any statistical techniques. Industrial problems have also been solved without the use of digital or analog computers, electron microscopes or other modern conveniences. The fact remains, however, that the correct use of these modern tools tends to reduce the cost of research and to eliminate costly errors. A working knowledge of experimental design and statistical data analysis must simply be viewed as a valuable tool whose proper use will greatly reduce the cost of experimentation and minimize the cost of false conclusions. These statistical methods, however, simply cannot replace fundamental knowledge or, for that matter, "good horse sense". Nevertheless. an experimenter armed with fundamental knowledge and good horse sense but completely ignorant of elementary statistical techniques, may easily present his data in such a deceptive form to cause management to make incorrect decisions.