Statistics and Reality – Part 2 Mini Paper Taken From Winter 2002 Newsletter
Abstract: The purpose of this article is to debunk the myth that statistics can be used to “massage” data and prove anything. Simple time plots of process data are powerful tools, and they often yield far more profound questions about the process than complex statistical analyses. To illustrate this point, the paper uses the example of infection rates from three different hospitals. A sophisticated analysis included bar graphs, histograms, a normality test and an ANOVA. The result was that there is no significant difference between the infection rates from each of the three hospitals. However, a time plot of the data for each hospital suggests quite the contrary. So what went wrong in this analysis? The answer has to do with the fact that nobody asked whether this system was stable on the first place. The paper proposes three important questions: 1. How were these data defined and collected, and were they collected specifically for your current purpose? 2. Were the systems that produced these data stable? 3. Were the analyses appropriate given the way the data were collected and stability state of the system? In conjunction with a process’ stability, one must understand the difference between common and special causes. The most common error in improvement efforts is to treat common causes (inherent) as if they were a special cause (unique). This is known as tampering and can cause undesirable effects. In order to avoid this, don’t treat the monthly data points or the individual “accidents” uniquely in isolation, but rather undertake a process-oriented view of the system over time.
Keywords: Time plots - Analysis of Variance (ANOVA) - Run charts - Common cause - Special cause