Do You Need To Show Your Work In The Real World?

I was talking to my son at breakfast about his first set of exams. He was a little contemptuous about the fact that the majority of the marks given were for showing how you arrived at an answer, rather than getting the right answer, even though his teachers have told him that in the real world, all of the marks are for the right answer.

This brought to mind a debate I was having back when I was teaching process design, and an issue I deal with on a high proportion of expert witness engagements. These two are related. I'll start with the teaching issue first. I mostly used to teach process design. This is not maths, which has absolutely precisely right answers. The problem with giving 100% of the marks for the right answer in process design is that there is no right answer if the design exercise is at all realistic. There are however an infinite number of sufficiently right designs. There are a larger infinity of wrong designs.

Part of what I was trying to teach the students, (all of whom had obtained good passes in A level maths for giving precisely right answers and a good show of working) was the idea of an exercise which had no precisely right answer, but instead has an approximate and right enough answer. This was what educators call a threshold concept-one of those things which is very hard to learn, but once you have learned it, you cannot understand how it wasn't obvious. I needed them to show their working for the same reasons all educators do-to demonstrate how they got the answer, and to give me a chance to offer them encouragement for heading in the right direction, and a pointer to where they lost their way.

The difficulty of offering a high mark for a right enough answer is practical. Being certain that each one of the twenty submissions you have received really does have a right enough answer by real world standards takes a lot more time than you have available to do the marking, and it is almost certainly the case that it does not. Real world process design is complicated, and it involves liaison with other disciplines, knowledge of the details of the products of various suppliers, access to inhouse knowhow, and a whole lot of other things which a group of chemical engineering undergrads have no knowledge of.

Many of my students did great designs for undergrads, and they were as well prepared as I think it was possible to get them for real world design practice, but you couldn't have built any of their designs and hoped them to work. I was proud of what we did together, but they weren't professional engineers yet. That said, a lot of "professional engineers" aren't either, which brings me to my next point.

I spend a fair amount of my time looking at real world process design calculations when investigating why plants didn't work as a forensic engineering exercise. Most of these don't show their working, and there's a very good chance under the circumstances that they got the answer wrong too. Very occasionally, I get calculations with a solid methodology (not necessarily the one I'd use, but one I know other experts use), good annotations of assumptions and sources of information, and right enough answers.

More commonly, the methodology is poor, and too many of their answers are plain wrong. Sometimes there is no apparent methodology at all- the cell in the excel spreadsheet doesn't contain a formula, just a number they typed in. Once, the methodology was terrible, but they sort of knew their answer was unreliable, so they upsized the kit by a large factor, and lucked on a good enough answer. If you can't be good, be lucky.

The lawyers don't want me to grade this work as such. They have a pass / fail criterion in mind. They want to know if the designer has demonstrated "the ordinary skill of an ordinary competent man exercising that particular art". That is, mediocrity. Here's where we might think we have moved away from the academic, but now that 75% of engineering graduates get a 2:1 or better degree, mediocrity is the standard we are all working to.

So, I don't exercise the benefit of hindsight, and as with my students, the question is not if their calculations are as good as mine would be now that I have thirty years of experience. I tend instead to compare the quality with my work around the point where I became a chartered engineer, with around four years post graduation experience. I never consider whether they are following best practice, or even good practice. Best practice is exceptional, and good practice is clearly better than mediocrity. The standard here is however "meh". Which makes it all the more surprising that I see so many sets of calculations which fail to reach it.

So, you do need to show your working in the real world, especially since you might have screwed up (on average, you are likely to be average). A company with a good design QA process will require it (or they should-many of these dodgy sets of calcs came out of companies which were supposedly ISO 9000 registered). Show your working, annotate your assumptions and references, get your calcs checked by someone competent. If someone asks you to check their calcs, take it seriously. This is in your own interests. I've been in court watching an engineer get cross-examined on whether work they did twenty years ago was competently executed. Take it from me-you don't want to be that guy.