Reference no: EM132253505
Holistic Assessment - Lying Statistics
Three kinds of lies are possiblem accoriding to Benjamin Disraeli, a British prime minister in the nineteenth century - lies, damned lies, and statistics. A related nationexists that "you can prove anything with statistics." Such statements bolster the distrust that many people ahve for statistical analysis. On the other hand, many nonmathematicians hold quantitative data in awe, believing that numbers are, or at least should be, unquestionably correct. Consequently, it comes as a shock that various research studies can produce very different, often contradictory, results. To solve this paradox, many naive observers conclude that statistics must not really provide reliable indicators of reality after all, and if statistics aren'ts "right," they must be "wrong." It is easy to see how even intelligent, well-educated people can become cynical if they don't understand the concepts of statistical reasoning and analysis.
Consider, for instance, the frequent reporting of a "scientific discovery" in the fields of health and nutrition. The United States has become a nation of nervous people, ready to give up eating pleasures at the drop of a medical report. Today's "bad-for-you" food was probably once good for you, and vice versa. Twenty years ago, many consumers were turned away from consuming real butter to oily margarine, only to learn that the synthetically solidified oils of margine, trans-faty acids, are worse for our arteries than any fat founded in nature. In the year following the publication of this finidng, margarine sales dropped 8.2 percent and butter sales rose 1.4 percent.
Distrust also arises concerning studies that link exercise to health. Numerous studies have estabilished statistically that people who exercise live longer. But the conclusion that exercise is good for you many put the cart before the horse. Are people healthy becase they exercise? OR do they exervise because they are ehalthy? Corelation, once again, does not estabilsh causation.
1. Is the study sample representative of the population involved?
2. Were the statistical procedures used appropriate to the data?
3. Has the research involved a sample of significant size and a sufficient time period of study?
4. Were adequate controls applied to assure that outcomes are actually the result of the studied variable?
5. Has the margin for error been taken into account in interpreting the results?
6. Has any claim of causation been carefully examined using appropriate approaches?