Thursday, June 25, 2015

Pass the MOOCs -- Part 2 (Underestimations)

Last update: Thursday 6/25/15
I posted a note on this blog ("Pass the MOOCs") recently that informed readers that I am currently taking a series MOOCs that will enable me to reenter the workforce with newly acquired skills in machine learning, data analytics, data science, etc, etc, etc. The main point of my previous blog was that dabblers drop out of MOOCs, but job hunters stay the course. In my case, I dropped out of a couple MOOCs that I had previously taken to satisfy my intellectual curiosity.

But this time around I'm a serious job hunter, so I'm going all out for completions and maximum knowledge. All of my current MOOCs provide verified digital certificates upon successful completion that can be posted on Linkedin where they can be viewed by prospective employers or by prospective consulting clients. And all of them charge a modest fee for the certificates.

The purpose of this note is to share a couple of insights that I have experienced thus far in my quest for new knowledge delivered by the new online media. Please note that the knowledge I am seeking is not only new to me; but the combinations of tools and techniques covered by these courses were only developed in the last 10 to 15 years, so they are also new to just about everybody. I was surprised by the insights that I'm about to share. Suspecting that my insights may surprise some of my readers, I am publishing them because I haven't seen similar comments in any of the education news publications or blogs that I usually read.


The MOOCs that I'm taking are sophisticated blends of statistics, computer programming, and funky hacking tools. At the time of this writing, I have passed and received verified certificates for three MOOCs, one from M.I.T. (via edX) and two from Johns Hopkins University (via Coursera). From what I've read, most MOOCS offered by the nation's leading universities cover STEM subjects. In this sense my courses are "typical. But most of the MOOCs offered by most top universities still have an academic orientation, i.e., knowledge for its own sake that pays off in the long run. By contrast, the MOOCs that I'm taking are offered in the same spirit in which I am taking them, i.e., they are career-oriented and promise short to mid-term financial gains.

Don't Believe the Hype!!!
I am referring, of course, to the prerequisites that edX and Coursera declare  for their courses and the hours of effort per week they say will be sufficient for successful completion. Consider these excerpts from their Websites:
  • Data Analytics Edge (M.I.T.)
    "Basic mathematical knowledge (at a high school level). You should be familiar with concepts like mean, standard deviation, and scatterplots. Mathematical maturity and prior experience with programming will decrease the estimated effort required for this class, but are not necessary to succeed." ... Hah!!! ... LOL ... :-)

    Effort: 10 to 15 hours per week ... Hoo-Hah!!! ... ROTF ... :-)
As for my reactions in italics, I post them in the context of my personal assessment that this is one of the five best courses that I have ever taken in my life, anywhere, at any level, from K through PhD. Yes, I'm an avid fan and I highly recommend this course to anyone who wants to get a rigorous introduction to what data analytics, machine learning, etc,  are all about. 

Nevertheless IMHO, like the Gates of Hades, the course Web pages should clearly warn those who would enter that if their math skills are only at high school level and if they really have no prior experience in programming, they will also have to possess an IQ that's off the charts to keep up with this course no matter how many hours they study each week. For example, the programming language used by the course -- called "R" -- is powerful, object-oriented, and functional; but it also has a well-deserved reputation for being "quirky" -- which disqualifies it from anyone's list of computer languages for beginners.

Whereas the M.I.T. MOOC is a single course that lasts 12 weeks, each of the 10 courses that I'm taking from Hopkins is a 4-week mini-course. Each mini-course awards a verified certificate upon successful completion, but the big certificate in "Data Science" is only awarded when students complete all 10 mini-courses. The M.I.T. course was offered by a large team headed by a senior professor with the support of platoon of graduate assistants; whereas the Hopkins mini-courses are offered by three young professors, i.e., by hotshots who really know their stuff.  M.I.T's course provides an introduction to the field, while Hopkins offers coverage at an intermediate level.
  • Data Science (Johns Hopkins University)
    "Some programming experience (in any language) ... Working knowledge of mathematics up to algebra (neither calculus or linear algebra are required)" ... What? No prior introduction to inferential statistics? ... LOL ... LOL ... LOL ... :-)

    Effort: Ranges from 1 to 4 hours each week (overview of program) up to 7 to 9 hours each week (R, Statistical Inference, Regression Models) ... ROTF ... LMFAO ... :-)
Full disclosure requires that I make another confession, this time that I have also become an avid fan of the Hopkins courses. Nevertheless my comments in italics convey my judgments that students who only possess prior knowledge at the level of the stated prerequisites and who study for no more than 9 hours each week will have to have off-the-charts IQs to complete these courses successfully ... the first time. Fortunately, each of the ten courses is repeated every month, so failure or dropping out are viable options. Indeed, the hotshots wave their hands in their videos from time to time that students shouldn't worry if they can't keep up with a course the first time because they can always repeat the course next month. 

In summary, my first unexpected insight is that some top-rated universities are deliberately understating their prerequisites and the level of effort required for successful completion of their career-oriented MOOCs. So no one should be surprised by low course completion rates. 

Question: Why all the fibbing? Why do top-rated universities deliberately understate their prerequisites and the levels of effort required for career-oriented MOOCs???

My Tentative Answer: They have to fib in the context of current expectations about MOOCs because they are top-rated universities.

Before I explain my answer and before I also say why I expect most top-rated universities to stop fibbing in the not-too-distant future, I will have to take the reader on a brief side trip to Udacity in Part 3   

____________________
Related notes on this blog