Gas station without pumps

2012 September 10

AngryMath reviewed Udacity Statistics 101

Filed under: Uncategorized — gasstationwithoutpumps @ 19:33
Tags: , , , ,

A part-time college math teacher, who has taught intro statistics several times, took the entire Udacity Statistics 101 course and reviewed it on his blog: AngryMath: Udacity Statistics 101.

As several commenters point out, his review is of a particular instantiation of a particular course, not a general review of all Massively Overhyped Online Courses (MOOCs).  He found that Thrun was a terrible lecturer and that the Udacity Statistics 101 course was badly structured and poorly taught—nowhere near the quality of a standard community college offering of the similar courses.  (Some of the commenters who had taken Thrun’s AI course last year had similar critiques of Thrun as a teacher.)

AngryMath does bring up some good points about MOOCs that generalize beyond just this example of a bad course:

  • The Udacity format provides little or no feedback to the instructor to fix the problems with the lectures or syllabus.
  • Videos are often touted as better than live lectures, because they are rewindable and can be watched many times.  But they are much less useful than text in this regard, because they are extremely difficult to index and access at random, and they provide no quick reference tables.  They need to be supplemented with indexable (or searchable) text.  (Or perhaps the texts need to be supplemented with occasional videos instead.)
  • The Udacity testing mechanism is a joke, as students can keep guessing until the automatic grader accepts it.  None of the MOOCs have yet produced a good way to avoid Massively Orchestrated Online Cheating, and until they do the idea of credentials from MOOCs is ludicrous.

I’ve not had the patience to sit through a MOOC myself (I have little patience for video lectures, which seems to be the only format MOOCs come in), but I’m trying to keep up with what is happening with them, as I fear that they may be the future of college education (following a Gresham’s Law of education, where bad teaching drives out good). See my other posts under the MOOC tag.



  1. I thought that the new Pearson-proctored final for several of the MOOC providers would help provide some outside certification that the student actually did the work. So, the challenge then becomes creating a final that can be offered at Pearson that is sufficient to say that the student actually did learn what was needed from the course, whether earlier coursework was the student’s or not.

    Note that the downside to using Pearson for proctoring is that it costs, of course. So, the MOOC is free, but the final costs ~$89.

    Comment by Jo in OKC — 2012 September 11 @ 06:53 | Reply

    • Another downside is finding a Pearson exam site. The closest one to us that offers the Udacity tests is about 30 miles away (2-3 hours by public transit) and there are several colleges and universities that are closer.

      Writing a good exam is difficult in many courses. Most of my courses have moved to other means of assessment (writing programs and papers, mainly), because the skills I want to test are very difficult to determine with tasks limited to less than 3 hours.

      Comment by gasstationwithoutpumps — 2012 September 11 @ 09:04 | Reply

      • I’m surprised at the distance for you to a Pearson exam site. It’s not nearly as far for me and I’m in flyover country. (Of course, I drive 30 miles to work every day, so 30 miles doesn’t seem very far for me.)

        For essays and computer programs, the difficulty for MOOCs then becomes finding some way to grade a massive number of papers.

        I’m leery of computer-graded essays/papers. I know some MOOCs do peer-grading, but my daughter had experiences (not in a MOOC) that left me thinking that peer grading is only useful if your peers are at your level, which would certainly be hard to guarantee in a MOOC. There are several companies that do plagiarism checking on essays/papers, but it costs so no MOOC has stepped up yet.

        I know that you could test programs online (via automated testing and various test sets, companies do it all the time), but that really validates output not the way the program’s written. I guess you could also collect metrics on the program, too, as a gauge of maintainability, but those metrics would have to be explained and used consistently throughout the course. I think some universities have created programs that check programs for plagiarism. I would think that’s really hard when everyone’s using the same language to do the same thing, though.

        I’m not sure that a reasonable final exam would be a paper or program generated inside 3 hours. :-(

        In most courses, you have more than the final project/paper/exam to determine whether the student knows the material and is competent. It seems like with the MOOC + Pearson model you’re asking the final to be more like a certification test — one shot to show you know it or you don’t.

        Comment by Jo in OKC — 2012 September 11 @ 09:27 | Reply

        • We’re “over the hill” from a large urban area, so a lot of commercial companies feel no need to serve us.

          I’ve actually taken my son to one of the testing centers in San Jose (for the appropriately named SCAT test). It took us most of the day to get there and back, and he was not at his best during the test (despite leaving an extra hour for him to recover from motion sickness—buses on mountain roads are not ideal pre-test conditions). There are people who do the commute daily by bus (the Highway 17 express bus is usually full during peak hours), but I would not be able to do that, as I can’t read on a bus.

          I’m very dubious about computer grading also of anything other than multiple-choice or integer-valued answers. It is very difficult to craft questions with that form of answer that test the skills I’m interested in.

          Comment by gasstationwithoutpumps — 2012 September 11 @ 10:54 | Reply

  2. I’m doing a MOOC right now, partially to review some material that I haven’t seen since college, and partially to see how they work. I hate, hate, hate the video lectures. They are so SLOOOW and boring. If I could just read the same material, I could skim the parts that I already know, and then slow down on the parts where I need to concentrate. Why is there no textbook???? I can’t learn anything without a textbook,

    Comment by Bonnie — 2012 September 11 @ 13:12 | Reply

  3. I’ll just comment on one point. The criticism that: “The Udacity format provides little or no feedback to the instructor to fix the problems with the lectures or syllabus.” This seems to indicate that neither you nor Angrymath spent time in the forum for the course. There is TONS of feedback, far more than I ever got when I taught at a community college. It’s also clear that it’s read by Udacity staff, at least (don’t know how much each main instructor reads), as they post responses to some posts.

    I DO agree with much, although not all, of the other criticism of this particular course, I’ve written a more complete set of thoughts on both this course and MOOCs in general at

    Comment by viennamike — 2012 September 12 @ 12:21 | Reply

    • I have not tried any of the MOOCs.

      I think that AngryMath and I were not clear. What we meant was that the MOOC provides no immediate feedback for fixing presentations on the fly. As you say, the forums provide plenty of delayed feedback. Whether the instructors look at that feedback is unclear, but it is certainly available.

      Comment by gasstationwithoutpumps — 2012 September 12 @ 12:54 | Reply

  4. […] near the quality of a standard community college offering of the similar courses”.  Gas station without pumps Share this:TwitterFacebookTumblrEmailPinterestRedditStumbleUponLike this:LikeBe the first to like […]

    Pingback by A critical look at educational technology – are MOOCs losers? | Cost of College — 2012 September 20 @ 02:25 | Reply

  5. […] AngryMath reviewed Udacity Statistics 101 […]

    Pingback by MathTwitterBlogosphere, mission 1 | Gas station without pumps — 2013 October 6 @ 19:19 | Reply

RSS feed for comments on this post. TrackBack URI

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: