The final exam used the same model as the mid-term and its rubric suggested it should take around three hours. However, this may have set up a false expectation. In fact, there were ten questions and each one was designed to take you beyond the simple rote application of formulae. You had to think all the way through.
The final question was even prefaced with:
Note: Although this problem can be solved using material from this class, it is intended to stretch you beyond the material that we explicitly taught in this class. Do not work on it until you have finished with the other problems.
Yes, it did go beyond the course and involved solving a non-linear equation iteratively to find the operating point of a triode tube. Tough - very tough.
And three hours?
Don't be silly. Most students reported spending nine or more hours of the 24 available and working as fast and as hard as possible. I certainly dreamed about one of the questions as I split my attempt over an afternoon and a morning.
Finally - all over.
Course complete, all homework done, final marked and a result instantly on display on the student's profile page.
I was in withdrawal.
Where is my next fix of electronics coming from!?
Even though I was mentally exhausted I was already missing the pain.
As mentioned earlier, the testing methods used converted the experience into a challenge, a computer game if you like, but more importantly into an amazing learning experience that went well beyond the material taught in the lecture videos. The instant feedback on questions is a powerful teaching tool in its own right. If it could be augmented with some hints as to what you might be doing wrong then it might be even better.
One of the problems that online courses have is that they have no control over how students tackle the testing components. Put bluntly they can cheat by looking things up on the Internet or "phoning a friend". There is also the problem that the new crop of online courses have little track record and it could be difficult or easy for the students to get a high score.
At the end of the day what matters is, would you give a student who had completed the course a job?
For some of the online courses the I Programmer team has taken over the past few months, the answer is difficult to arrive at because they didn't stretch the above-average student and the testing procedures were weak. Students coming out of the course with the same high mark could have completely different knowledge and abilities.
For this first MITx course this isn't the case. To use the jargon of the educational psychologists, the course has face validity. It looks as if it covers a difficult subject and anyone who gets a good score on the testing components must know and understand their stuff. This is one situation where Wikipedia and the likes just don't increase your performance IQ.
If I was hiring people to do electronics then a good score on the MIT course would be something I would take into account.
Can this approach be generalized to other subjects?
I think that it can for STEM subjects and I'm not qualified to comment on subjects out of this area. I can see that you could teach programming say in the same way and then set questions that pushed the envelope of what had been learned. The general principle, is the same but you need to keep in mind the difficulty of finding such exercises and the need to keep on finding them as the Internet makes each one more widely known.
Would I sign up for another MITx course?
Yes certainly, I can't wait for the next electronics course to start.
Sue Gee adds
I don't think the next MITx course will happen as MITx is now part of edX, the joint venture between Harvard and MIT.
The new organisation is, however, under the directorship of Anant Agarwal who not only pioneered MITx but also taught 6002.x so we can assume that lessons learned from this experimental course will be used in devising the edX curriculum.
No details have yet been released of forthcoming courses, but you can sign up for details here. And of course, we'll bring you details of all those relevant to computer scientists, developers, software architects and hardware enthusiasts like Harry.
Go is an important game for AI in that, until very recently, humans were much better at it than computers. It is a complex and subtle game that has only just yielded one of the most fundamental parame [ ... ]