Essay About 2017Statistical Misuses Statistics And Data Shows

Bus310 Ethics Paper
Garrett O’ConnorBus. 310Chris RachalApril 8, 2017Statistical Misuses Statistics are supposed to make something easier to understand but when used in a misleading fashion can trick the casual observer into believing something other than what the data shows. That is, a misuse of statistics occurs when a statistical argument asserts a falsehood. In some cases, the misuse may be accidental. In others, it is purposeful and for the gain of the perpetrator. When the statistical reason involved is false or misapplied, this constitutes a statistical fallacy. The false statistics trap can be quite damaging to the quest for knowledge. The case study “The Use and Misuse of Statistics” is broken up into 5 sections and it offers some guidelines for using statistics effectively, derived from Frei’s seminar and other various sources, ultimately aiming to help the reader decide what to ask of the analysts whose numbers you rely on. The first section is labeled “know what you know— and what your only asserting” and in this section a professor at Dartmouth College, Victor McGee talks about managers and how they are idea crunchers, spending most of their time trying to persuade people with their assertions. Additionally McGee recommends color- coding knowledge so people know what needs to be tested. The purpose of color coding is to assist people in not asserting assumptions and to only take things serious if they are supported by real knowledge. “Be clear about what you want to discover” is section number two. Some management reports rely heavily on the arithmetic mean or average of a group of numbers and the main point of this section focuses on the possibility that the mean would not be the most helpful metric for a statistical search depending on your desired outcomes. Figure one shows a good representation of this, showing that the mean is four while in fact four was never a variable used so for this instance the mean would not be helpful for your statistical search. “Don’t take causality for granted” is section three and Frei brings up an interesting point about establishing genuine causation by asking yourself three simple questions, “Is there an association between the two variables? Is the time sequence accurate? Is there any other explanation that could account for the correlation?” These questions ultimately help to establish a true cause and effect relationship. Frei also emphasizes the importance to look at the actual raw data, and not just the apparent correlation. Figure two shows a scatter diagram plotting all the individual data points derived from a study of the influence of training on performance. The case study explains how removing a single data point causes the slope of the line to change significantly, so depending on what results are being pursued, one can consider to potentially remove that one point that is influencing the end result. Additionally checking the data closely to make sure that the reverse isn’t true, and making sure to uncover any hidden variable that may be underlying in your information. Only by eliminating other factors can you truly establish the link between correlation and causation. The fourth section asks the question whether statistics can prove things with 100% certainty or not. The answer to that question is nothing is 100 percent certain because with statistics you would have to record all the impressions of all customers who have had the particular experience with a certain product is when you can establish certainty about customer satisfaction but the cost and time consumed would be too inconvenient. So instead a random sample is taken and all sampling relies on the normal distribution and the central limit theorem and this enables us to calculate a confidence interval for an entire population based on a sample. Furthermore, the fewer defects, incidentally, the larger your sample must be to establish a 95% confidence interval. Frei emphasizes in the need to spend more on quality assurance sampling. The fifth and final section is about how the managerial differences in numbers may be useless. For instance the case study mentions how a manager may look at two different numbers representing customer satisfaction ratings, one number will be better then the other but they will be close in range. A manager might look at this data and confer that the person with the highest customer satisfaction number is doing significantly better then the other person with the lower customer satisfaction number. When in fact that both rating numbers may be inside the 95% confidence interval and not statistically, significantly, different.