The Carnegie Foundation for the Advancement of Teaching: A Study of Engineering Education, written by Charles Riborg Mann in 1918.  Wow, several hundred pages of text written in early 1900's style. I wonder how I can make reading it more fun. I know! I'll translate it into colloquial English shorthand, chuckling all the while about the gender assumptions (endless references to boys and boyish pursuits!) Mann is making about engineering undergrads.

Carnegie Foundation: Hey, we're the Carnegie Foundation. Once in a while, we ask folks to write reports about the status of education for a certain field. However, since engineering education is (it's 1918) pretty standardized across the various US universities, we don't need to write about the "many different ways of teaching engineering" -- there aren't any! -- so instead we've asked Prof. Charles Mann of the University of Chicago to look at whether the current teaching methods are effective. His take?


Not really. Math is important, but by loading students with theory before they get to practice engineering, we're graduating too many students who never find out if they can actually do practical engineering. We're teaching engineering this way "because that's how we've always taught," which ironically means the way we teach this "applied science" is unscientific -- so read on and Prof. Mann will tell you all about what we can do about it.

Professor Mann: Thank you for that excellent introduction, Coalition Of Important People.

American engineering started when the colonists needed to find ingenious ways around the building of a new country without the help of European tech and expertise. We have been slow to actually acquire schools of engineering (until the Morrill Act funding opened the floodgates), but today (1918) we have 126 engineering schools, and they all believe that "The ultimate aim of engineering education has always been and still is more intelligent industrial production." (p. 336)

Similarities end there. What proportion should be spent in lecture, shop, and lab? How many credit-hours should students spend on different subjects (math, chemistry, humanities, etc)? As industry and tech have exploded, we've crammed more credit-hour requirements on these poor kids; "it is obviously absurd to require from the student more hours of intense mental labor than would be permitted him by law at the simplest manual labor." (p.257) And because of the way school governance is set up, a good chunk of these required classes are in departments siloed off from engineering, with each department thinking their topic is most important, so the English profs will cram them with their material, the Chemistry profs with theirs, the Physics profs with theirs... and you end up with an overloaded and scattered head.

Dropout rates are ridiculous now -- at least 60% -- but they're better than before (91%). Most students drop out before 2nd semester sophomore year, when they haven't even started engineering classes yet (they're mostly taking math, chemistry, physics, English, drawing, and shop). Why should failing English or German mean you're not going to be good at engineering? We compared company work records with grades for some GE engineers and found that college grades didn't correlate with satisfactory job performance. (And how can we expect them to get good grades when we're requiring them to take so many classes at once?) And there's no reason to require theory as a prerequisite to practice. Alas, "...teachers generally believe that the students are incapable of working intelligently at practical engineering projects until they have been well drilled in theoretical principles and mathematical processes, in spite of the astonishing manner in which boys of high school age learn without assistance to man age wireless telegraphy or gas engines." (p. 326)

Let's talk about admissions. We took a big leap forward when we started standardizing acceptance criteria -- instead of individual departments making wildly varying decisions, the College Entrance Examination Board gave us a stable metric to gauge against. It still misses a lot, though -- for instance, why not take into account engineering-style hobbies? Selection of professors is also important; in contrast to law and medicine, which came from apprentice-style teaching by active practitioners (medical professors have treated patients themselves), engineering is taught by academics who have usually not worked in industry. Fortunately, some schools are experimenting with different ways of teaching engineering; I'll list a bunch of cool things folks are trying like the University of Cincinnati's introduction of a co-op program. Others have been looking at grading; is there a way we can make a system students won't want to game, a system that also recognizes and encourages things like creativity?

Anyhow. Concrete recommendations. Here's what I think would work:

  • For heavens' sake, give these students a reasonable workload. Don't require more than 18 credit-hours (4-5 classes) of them at one time, so they can actually learn the material.
  • Get "real" engineering experience into the first two years "to make the boy feel that he has actually left high school and entered upon a professional career." (p. 316)
  • Make sure engineers have broad training that includes the humanities, rather than narrow specialization -- otherwise they can't change the world they'll go out into. But they don't need foreign languages (in 1918) -- our graduates say they never use them.

We'll need to do quite a bit of work on our institutions and curricula to get here, since we can't just swap components in and out -- it all needs to be coordinated.

Hop to it, people.

And nearly 100 years later, we're... still trying to hop to much of the same stuff.