As this winter marked my 20th year as an applied mathematician, the most rewarding part of my job has always been explaining math or statistics in such a way that scientists, engineers, and managers could understand and utilize what was being presented. So often, the irony is that statistics is a sore subject to bring up among even the most brilliant technicians, engineers, and scientists. Demystifying the subject to the extent that useful, bottom-line results can be extracted is at times the greatest service I can provide to a customer.
Why do the words “math” and “statistics” cause so many technical people to flinch? I would argue that it has to do with the educational system.
Many math and statistics faculty at universities focus more upon their research careers and less upon teaching. Also, teaching is an art and a talent. People who are technical geniuses do not always make the best teachers. The student usually is left with, at best, a bland, boring, and forgettable presentation, and, at worst, an incomprehensible one that causes the student to feel inadequate.
Also, for many scientists and engineers, it may have been decades since they last took statistics; to boot, many things have changed in the way that we do statistics in the 21st century. Slide rules and simple calculators have been replaced with high-powered software and computer programs. Doing statistics now requires keeping up with the latest paradigms, tools, and software.
Companies who do not keep up with this evolving analytic sophistication or who do not have statistical talent at their disposal are left with massive amounts of data that are like a locked vault: nobody really knows what the backlog of data may be indicating…or hiding. They are left with making key business decisions based upon what boils down to gut feelings and subjective recommendations, and they are left with processes and products that cease improving.
A common complaint that I get from scientists and engineers is that talking with mathematicians and statisticians can be alienating-they don’t “speak English” and tend not to be attentive to whether their points are being understood. Some statistical presentations are full of theory and complex, esoteric explanations that only a Ph.D.-level statistician could understand. The practicality of these complex, esoteric analyses is then also called into question.
In my career experience, scientist, engineers, and decision-making executives want to know the following from their statisticians:
What is the problem that you are solving? An explanation of the problem together with its motivation and application may seem like a moot point, but, if not covered, can set the pace for a disorganized presentation.
What method did you use (try to explain it in plain, simple English), and why did you choose to use this particular method?
What was the outcome of the method? Don’t just tell them what the p-value was, but try to tell them what the results may mean or indicate in a practical sense.
Be inquisitive and see what other information the data can tell you, and present these additional points as is useful to your audience. Often the engineer/scientist/manager will not know what questions should be asked of the data, and they will look to the statistician to tell them what can be deduced from the data.
Most applied statistics can be explained without the use of convoluted statistical jargon. Remember that applied statistics is a tool to help us understand the world around us, not an end in itself. Understand your methods and your results at a fundamental, intuitive level. As you communicate simply yet accurately to others, you will become an indispensable resource-not just because of your ability to do statistics, but because of your ability to communicate it so that others can deeply understand and appreciate the results. Our goal is for our clients to understand all of the information that we give them-because the more they understand, the better they can act upon the information.