I once had a client from engineering. This is pretty rare, as I usually work with social scientists and biologists. And despite the fact that I was an engineering major for my first two semesters in college, I generally don’t understand a thing engineers talk about.
But I digress. In this consultation, we had gotten about 20 minutes into me not understanding a word he was talking about when I realized he was using “Parameters” when he meant “Variables.” As in, “I measured four flexibility parameters on the doohickey.”
In statistics, Variables are things you measure that vary from observation to observation. Height, weight, flexibility, bending strength, % ground cover–these are all Variables if they vary from one observation to another. (They are constants if they don’t).
Parameters are things you measure about the variables. Their means, their variances, the size of their effect on another variable. And parameters specifically refer to the measurements made about the entire population.
I suppose it makes sense that engineers consider variables to be parameters, since to them, parameters are things you measure about doohickeys. In statistics, variables are the doohickeys getting measured.
So it makes it hard to talk with engineers because I have to translate as they speak. But I’ve come to accept that they speak a different language although with the same words.
But lately, I’ve seen other people (like ecologists) calling their variables Parameters. And in the same sentence as using the terms like p-value and adjusted R-squared, so I know they knew statistics well.
What’s going on?
I had to laugh when I read this posting. Very true – they definitely speak a different language. I work a lot with product design engineers and manufacturing engineers, and long-story made short, they are called parameters because every effort is to be placed on minimizing their variation. We’re talking about part dimensional tolerances that are millimeters or fractions of a millimeter. Yes, they do vary, and we ensure there are measurement systems in place to accurately quantify it, but compared to the social/behavioral sciences the variation is negligible. However, that does not diminish the importance of understanding the variation and its impact on overall product quality….it’s just that the “big-picture” variation is very tiny.