I have been taking Statistics 101 from Coursera. This course is taught by a Princeton professor. I have been interested in stats for a while, but have never taken any classes. As a bonus, a work project I’m focused requires a lot of linear regression. (I’m using a library for the linear regression, of course, but wanted to understand some of the limits of linear regression before applying it for a business purpose.)
It was very easy to sign up for the class. I was a bit early, (the lectures are put up weekly starting on a given date), so I just added my email address to the wait list. When the class started, they emailed me and I registered.
It was free.
There are some changes coming to the university world. I have some friends who are professors and I’ve been sharing articles like this one and this one for years.
So, that was an additional reason to take the course. What would an actual online class be like?
The class is divided up into 12 weeks, 1 midterm and 1 final. Each week there are between 4 and 6 lecture videos to watch (the longest was approximately 20 minutes, the shortest is approximately 5 minutes), a lab video and a homework assignment. The lab typically examines the lecture concepts and puts them into practice using R, an open source stats tool/language. The homework is an untimed quiz that I have 100 tries to finish. Each quiz has 10 questions, some text input, some multiple choice, and typically is due 2 weeks after the initial lecture on the topic. I can complete the quiz later for reduced credit.
I’m over halfway through the course.
Am I learning something? yes. Definitely. I’ve learned basic concepts of statistics. There has been some handwaving on some complicated concepts and single letter concepts are occasionally introduced with little explanation (t, Z, and F values, for example), but this is an intro course, so I am unsurprised. I definitely have become comfortable with basics of R.
Am I learning what I would be in a normal college classroom? Nope. There’s been no collaboration (because of my time constraints, I don’t participate in the forums, which are the only form of collaboration I have seen). All R scripts are provided in the labs, which means you sometimes just cut and paste. Questions on quizzes are constrained by the online test format. Because I’m jamming it into my schedule, I don’t review the material as much as I should. There’s no opportunity to stop the instructor and ask exactly what he meant.
But….did I mention it was free? And that I’m not in college and don’t have the time for a normal college class?
If you are thinking of taking an online course through a MOOC like this one, here are some tips.
- block out time to watch the lecture videos–I spend about 1-2 hours a week doing this. You can double book this with mild exercise (treadmill), sometimes. Sometimes the concepts were complex enough I could not multi task.
- Coursera has a button on the video player to run the videos faster. I just started using this and find running the lectures at 1.5x is doable.
- plan to spend some time on the labs and quizzes–I spend about 1-2 hours every week.
- know what you want. If I wanted a deep understanding of statistics, I would probably need to spend an additional 2-4 hours each week working on understanding all the concepts covered in the lectures, and get an additional statistics book. (This one looks nice.). But I want a conceptual overview that lets me dig into third party libraries and learn the domain jargon enough to search the internet for further resources. This class is letting me do that.
- get a tablet–these devices are perfect for consuming the lecture videos.
- be flexible in your viewing, but try not to view a week’s worth of material in one day–that much academic knowledge transfer is no fun.
- no credit is available. This might be an issue for some students.
- be committed. It is very easy to sign up for these courses and then drop out, because there really are no consequences. But there are a huge variety of courses from a number of sources: udacity, udemy, edx, Khan Academy.
- enjoy the free world class instruction. Did I mention this was free?
All in all, I’m happy with this course and will come out of it more grounded in statistics.
From my experience in this class, I think that the business of teaching, especially introductory material that lends itself to video lectures, is going to undergo a change as radical as what newspapers have been through over the past 20 years. I don’t know if MOOCs will augment or supplant universities, but the scale and cost advantages are going to be hard to beat.
Your experience is similar to the one I had taking Odersky’s Scala course through Coursera. At this point in my career, I care little for credit, seeking to advance my understanding of topics that interest me, so it is a great fit.
It has gotten me thinking about the future of education. I believe one possible path will be to separate learning from certification. The former will likely be much less expensive while the latter will be where the cost will reside.
This obviously won’t be the case in every field, but for many, it could reduce the cost of entry with little or no downside.
Interesting thoughts Joe.
In what fields does learning matter more than certification? It’s when the stakes of failure are low and when the field changes rapidly and when demand for the skills are high. Software is an obvious case (although there are places where the stakes of failure are very high in software, so maybe my criteria are incorrect). I’m struggling a bit to think of others–most certifications fail one of these criteria–mostly low stakes of failure. I’m thinking of professions like medicine, law, civil engineering, even the trades like plumbing.