As I write this, I can imagine Jeff stating, "So," pausing for dramatic effect, "if I make magic consultant juice and I say that there is somewhere between 0% and 100% magic included, as long as I am within my stated specifications, I have a quality product, right?"

Well, maybe there is more to quality than I first thought. How do the specifications get set? Don't they contribute to overall quality? I apparently know one of the important parameters of quality (adherence to specifications), but I am clearly missing the whole picture. This is a common mistake. People love to create simple models of complex systems in hopes of someday being able to describe the complex with only a single 'statistic'. I wanted to describe the complex using the simple and it didn't fully work for my definition of quality.

This wasn't my first foray into simple models for complex systems. I have been doing this my whole life in one way or another. In my early career, part of my job as a pharmaceutical stability chemist was to trend stability study results over time. I could do wonders with a regression line! When I started studying statistics formally, I learned the science of reducing large data sets (i.e. the world) into a small number of descriptors. I thought that this was the most powerful scientific tool available until I walked to my boss' desk and saw a quote on his desk.

"There are three kinds of lies: lies, damned lies, and statistics" - Benjamin Disraeli

Later in my career, I had the "privilege" of studying accounting around the same time as the WorldCom and Enron situations. Accounting always seemed like a generally accepted method of fitting a square peg in a round hole, and the happenings at the time surely didn't lend credibility to the accounting model. On top of that, I learned in my accounting class that employees are an expense for a company. That helped explain why the company that I worked for at the time was reducing employees (layoffs) due to stock prices falling. Of course, the stock went up after a 10% workforce reduction. Based on my simple understanding of the accounting model, I deduced that by reducing all expenses (i.e. getting rid of all workers), the company could make that much more money. They didn't try that, though. Maybe the accounting model doesn't factor people in correctly.

Finally, from my earliest memories, I have been trying to discuss and analyze the performance of athletes using simple statistics. Hmmm, if a running back gains 4 yards per carry and a team has 4 downs to gain 10 yards, why doesn't the team run the ball every single time? By that statistic of running efficiency, they can't lose, right? No, there is more to the performance of a team than the statistical performance of an individual.

Trying to reduce a complex system down to a single representation sometimes skews the real picture, I guess. But, like Jeff, I have been unable to explain quality. Thank goodness he will set me straight in next week's blog!