I have learned to take most news reports with a pound of salt. The VW diesel scandal may deserve at least a pinch of salt. That is my take from the attached article, “VW Dieselgate…”
Most who have taken exams to receive their medical, law, or engineering licenses would likely not pass the exam on any given day as normally administered. I am sure I could solve a sufficient number of problems to get a passing grade on a PE exam but not in the 8-hour window.
In the real world, our clients or employer would like us to work quickly and arrive at a correct solution. Few will expect you to work at the breakneck speed, required in exams, to solve problems where health and safety are involved. Demonstrating your understanding of the concepts is required but not sufficient as it may be under exam conditions. Testing, whether of machines or people, is artificial. Until you have attempted to test real world conditions, you cannot appreciate the difficulty.
Part of the artificial or “cheat mode” associated with testing people is to have specific books at arm’s length and to have recently used these books and specific charts and tables. Some hire tutors, attend special review classes, use study guides that cover the expected topics. Most of this will not be at the professional’s side after the exam. Much of the exam material will not be addressed in our day-to-day work. Most will find their job focuses on a particular product or technology. They may find themselves in administrative roles not addressed on the exams they took to receive the required certification in their field.
We accept and understand the lack of correlation between exams and real world conditions. So what is the point of testing? Some benchmark is required and this is the best we know - no one suggests cheating.
If we wish to test in real world conditions, whose world will we choose? Sticking with just automobiles within the USA, the environment for a car in Florida is very different from a car in Wyoming. As I address in my blog “Engineering From Behind the Bushes” 6/3/2015, http://www.jagengrg.com/blog correlation between test and real world is never 100% and may not exist at all. In that example, there was some real cheating on top of the known lack of correlation.
I cannot help but see a parallel between the inherent flaws in testing methods used for people and those used for machines. We establish a goal and we design to that goal. Is this criminal or a solution to an otherwise impossible set of goals? Before we crucify VW et al., would it not be better to educate the lawmakers about the impossibility of testing to real world conditions? That would be a step towards creating standards, which achieve the desired goals. What do you think?