Tuesday, January 15, 2013

PS - the Honor Code

Before taking each quiz we had to check a box confirming we were who we said we were and that we weren't cheating.  I should have copied it exactly.  As far as I could tell we could use notes on the quizzes (or else I should return my certificate), so I suppose this checkbox was just for identity and that I didn't have someone feed me the answers.

The disclosures on my certificate were extensive, Johns Hopkins (home for Dr. Caffo) wanted to make sure everyone knew I was not a student there and that they could not verify that I was me.  I know other MOOCs are offering testing sites for identity verification.  This is certainly one hurdle that must be overcome.  It also makes MOOCs seem more like SATs or APs to me.  I still believe that a college course should be more than passing the final at the end.

Monday, January 14, 2013

Final Thoughts

I had two goals in taking this class.  One was to learn more about statistics, and two was to reflect on how MOOCs are going to change the face of higher education and their implications for Trinity University.

In a traditional university course, a credit hour means 3 hours of work per week, usually 1 hour in class and 2 hours outside of class.  A semester is ~15 weeks.  For this course there were usually 4 lectures a week, ranging from 6 to 32 minutes.  The average weekly lecture time was 76 minutes, ranging from 53 to 93 minutes.  The course lasted 7 weeks, and I spend around 2 hours outside of class completing homework assignments and quizzes.  So, that is perhaps 0.5 credit hours?  If this course lasted 15 weeks, I could see it being worth 1-2 credit hours if there were some additional lectures each week.

I don't think that this experience, even for 1 credit hour, compares to a class at a brick and mortar institution.  I would say this course gives a pretty shallow understanding of basic statistics.  This isn't really a criticism of the course - I can't imagine a 7 week online course achieving much more.  I do feel that I have learned something, and my 'grade' was an 86% (so B? B+).  However, at least at TU, I would expect my students to have a deeper understanding for a B.  Personally I'd give myself a C.

But I suppose I have to reiterate, I think for what the class was it was pretty good.  I do have a better understanding on how basic statistics are calculated.  Although I would probably have to look a few things up, I could preform any of the calculations again.  I didn't learn R, but it doesn't seem complicated.  It was a good introduction, and if I were taking a university course on the same subject, this would have been a perfect lead in.

What's missing in these courses is feedback and support from your professor and classmates.  There is very little taught in a university that couldn't be learned from reading books.  The obvious exception is science lab courses.  But for the majority of courses, a library card is the cheaper option.

But that is not why students go to college.  The reason to go to college (and the best one you can get into at that) is to be able to interact with experts in their fields and with students with similar educational background as your own.  I was very frustrated during this class when I couldn't ask questions of the professor.  Although Dr. Caffo did in fact answer some questions in the discussion boards, the days that transpired lessened the impact of his answers.  Because the discussion boards were filled with students of varying education, enthusiasm, and level of commitment, I quickly found wading through the boards unsatisfying.  I wish I had lingered there more for my own education on MOOCs, but I just didn't get much reward from them in taking the coures.

I suppose there are courses in college where you do not interact with the professor and are only given multiple choice tests, but in my opinion those courses are not very good ones.  I realize that in larger schools there is a buffer of TA's between the students and professors - but usually only at the intro level and at least the TA's are 'experts in training'.

If MOOCs have the power to eliminate the need to go to college, I wonder why libraries didn't do the same thing 100 years ago.

Instead, I believe MOOCs are more like a very interactive book.  But since the content is unavailable after the course is over, a book you must return.  I think that in the college setting, MOOCs can supplement a course.  Even be used to prepare students, as long as the scheduling works.

As for Trinity, I think the access our students have to professors and their talented classmates is our number one selling point, and MOOCs don't threaten this.  What we should perhaps using online content for is allowing our students an additional tool to learn.

Perhaps the best application for TU would be a Calculus bridge course over the summer.  Many of our student struggle with math, and this often prevents them from completing science and engineering majors. We could make a 6 week boot camp where we assigned online lectures, along with homework assignments and quizzes that would be graded by TU faculty.  Even set up chat rooms with only TU students.  This type of approach would yield, IMHO, better results than just suggesting a few Coursera courses.

Grading format

Each week, we were to complete one homework assignment and one quiz.  The number of questions ranged from around 6 to 10.  The questions were multiple choice, from at least 4 choices to as many as 11.  The questions were about two thirds calculation and one third conceptual.

The first couple of assignments did in fact require calculus - fairly simple definite integrals.  Although we were supposed to use the language R, I found the required calculations simple enough (if not a bit tedious) to do with just pencil, paper, and a calculator.

Passing the course required an average of 70% or better on the quizzes.  The homework assignments were not graded, and feedback was given on the answers.  We are given three attempts at each quiz.  In the first week, it was your last attempt that counted, and you were given no feedback on your answers.  Students complained, somewhat reasonably, that without knowing which problems they got wrong, taking the quiz multiple times did not necessarily raise their score.  The grading policy was changed such that your best quiz of the three counted, and after submitting each quiz you were informed which questions were answered correctly.

I felt the assignments did a good job of reinforcing the material.  They only took about an hour, certainly no more than two, each week.  But, I did not feel that this amount of work was sufficient for me to 'master' the material.  I don't think at this point, about 2 months after finishing the course, I could remember exactly how to calculate probabilities and confidence integrals as I did in class.  However, I'm not sure that this is the point of these short courses.  More valuable to me is that I know how to solve a class of problems, and with a short look through a book I could solve them again.

Small aside, this is something I hope my students learn while they are in college.  Knowing how to access knowledge and remembering how to use it quickly is more important than memorizing things.

Wednesday, January 9, 2013

Discussion Forums

One aspect of MOOCs I was very interested in were the discussion boards.  The idea of discussing material with potentially thousands of people discussing concepts, homework problems, lives, etc I thought was a unique feature of a MOOC.  Like a huge study group you can access without leaving your room.

I am not finding the forum for this class very useful.  For the first few weeks, the number one topic seemed to be people not understanding that calculus was indeed required and whining about it.  Seriously.

The forums got a bit better after those people either got over it or dropped.  I never did find a comment thread that discussed the material as a study group would, however.  Various geographical regions did form 'study groups', and perhaps I should have joined one of those to find the interactions I wanted.  I believe they were meeting on Google plus and other chat type rooms.

One of the most common threads for the course were people explaining American sports and customs to foreigners.  Many of the examples were on the line of 'if someone makes a free throw 80% of the time, what is the probability of that person making the second but not the first free throw?'  If you don't know basketball, this problem has added difficulty!

If people were discussing the homework, posters were more likely to just submit a worked out answer than a discussion.  I stopped looking at the threads on the homework at this point - once I saw a worked out answer, I couldn't 'unsee' it and looking at the answers is no way to learn something.

After I finished the course, I was at a meeting of the Texas Section of APS, and we were discussing students taking physics courses remotely at another college or university in TX.  One university that had some success with this arrangement said their students to interact with each other across campuses. They use Skype or Google Chat plus document cameras.  It works for them, so I do believe it is possible for students to interact online in a meaningful way.  It it clear that it must be done with intent along with some planning.

Lecture Format

After taking a few weeks of my biostatistics course, I have a few thoughts on the lecture format.

Dr. Caffo basically recorded voice to play over PowerPoint slides.  He frequently adds overlays to the slides - pictures of the statisticians he is referring to, links to Wiki pages, and occasionally corrections to slides or the text of his lecture.  Mostly I found the overlays useful, but some times they were distracting.

One complaint I have about this format was that he flew through the mathematical examples and definitions and I frequently found myself pausing and rewinding the lecture.  I have never thought PowerPoint slides were a good idea when you are doing math in front of a class - because you don't have to write down the equations, you tend to lecture faster than students can copy down the equations.  Seeing this from the other side has confirmed my distaste for PP lectures.  I suppose however that this is an advantage of taped lecture, I can rewind as much as I need.

The biggest disadvantage to the video is that I can't really ask questions.  There are several statistics variables that have similar names and (maddeningly) multiple symbols.  Sometimes Dr. Caffo was less than clear.  Although most of my questions could be answered from context and rewinding a lot, it was still was an issue for me at times.

It took me a few weeks before I realized that I could download a PDF of the slides.  I tried printing them out and taking notes on them.  This was a learning disaster for me - I paid less attention and would get through an entire lecture without retaining anything.  I began to print them out, but still taking notes as I watched.  That way I didn't have to worry too much about my handwriting.

Now that I am through the course, I would say talking over PowerPoint slides is fine, and probably the easiest way to produce one of these courses.  The filming of actual writing that the Khan Academy uses I find more interesting to watch, and if the voice over is filmed while the author is writing you eliminate lecturing too fast through math problems.

I think this may be the most useful thing for me as a college professor - sitting in a classroom trying to learn something.  Must do this more often for the benefit of my own students!

Pre-Course thoughts

My motivation to take a MOOC was twofold.  First, I have never taken a formal course on statistics.  In my physics training, I have been taught enough stats to calculate error for experimental data and I know my way around a Gaussian, but I have never taken a stand alone course from a math background.  Second, as MOOCs gain popularity and more and more universities are jumping on the bandwagon, I wanted to experience one of these courses as a student.  I think my choice of classes will give me a pretty good perspective of what students' experience would be in these courses - I am not taking a class in my area of expertise and am coming to the subject with little background.

The first couse I signed up for was Statistics One.  From Coursera:
Statistics One is designed to be a friendly introduction to very simple, very basic, fundamental concepts in statistics. This course is, quite literally, for everyone. If you think you can't learn statistics, this course is for you. If you had a statistics course before but feel like you need a refresher, this course is for you. Statistics One also provides an introduction to the R programming language. All the examples and assignments will involve writing code in R and interpreting R output. R software is free! It is also an open source programming language. What this means is you can download R, take this course, and start programming in R after just a few lectures. Statistics may seem like a foreign language, and in many ways it is. The ultimate goal of Statistics One is to get people all over the world to speak this language. So consider this your first course in a new and exciting universal language!

The instructor was Professor Andrew Conwayis, a Senior Lecturer in the Department of Psychology at Princeton University.

I dropped this course quite quickly - I wanted a more mathematically intense course.  I watched the first few lectures before dropping.  Prof. Conwayis either lectured over slides or we saw him on camera in front of a screen with his slides on it.  He has some sort of tablet that allowed him to go forward on his slide presentation as well as write on the slides.  It looked like PowerPoint.  The reason I dropped was that I wanted a more mathematical statistics course - this course was more qualitative, and seemed to rely a lot on R to make what I felt was simple calculations.

Luckily another course started ~2 weeks later, Mathematical Biostatistics Boot Camp:

Statistics is a thriving discipline that provides the fundamental language of all empirical research. Biostatistics is simply the field of statistics applied in the biomedical sciences. 

This course puts forward key mathematical and statistical topics to help students understand biostatistics at a deeper level. After completing this course, students will have a basic level of understanding of the goals, assumptions, benefits and negatives of probability modeling in the medical sciences. This understanding will be invaluable when approaching new statistical topics and will provide students with a framework and foundation for future self learning. 

Topics include probability, random variables, distributions, expectations, variances, independence, conditional probabilities, likelihood and some basic inferences based on confidence intervals.

The instructor is Dr. Brian Caffo, associate professor in the Department of Biostatistics at the Johns Hopkins University Bloomberg School of Public Health. 

This course seemed to meet my needs better, and calculus was a pre-requisite so I was certain of its rigor.  I look forward to starting!