Just starting a course in Learning Analytics offered by SOLAR.  It looks to be really interesting.  I’ll be posting more as we proceed.  I’m also supposed to be using hashtags for content related to the course, so this is a post to try and figure that out.  🙂


Posted in #lak12 | Leave a comment

A few resources

Next up, some resources regarding Learning Analytics.

Learning analytics is a fairly new field, just a few years old really.  The best resource set I’ve found is at learninganalytics.net, a site that seems to principally be the bailiwick of  George Siemens.  There’s a lot of good information there.  Two things are worth mentioning to those who may be interested in learning more about the field.

There’s a conference on learning analytics coming up this spring in Vancouver, B.C.  Given that my University has but the kibosh on international travel, no matter how close it may be, it seems unlikely that I will be able to attend.

On the other hand, I will be able to attend an online course in learning analytics being offered beginning in January by the Society for Learning Analytics Research (SOLAR).

Posted in Uncategorized | Leave a comment

Marketing Education

Here’s another way to think of Learning Analytics

Many of us realize that marketing has gotten increasingly data driven.  Every time we swipe our “Rewards” card at the grocery store, our purchases are recorded and the data is used for a variety of purposes.  Some of these are general, like what items should the store should stock more of, or what should go on sale next week.  But some of them are more targeted.  For example, by watching what you buy marketers can do a better job of guessing which items you’re most likely to try out if they print a coupon on the back of your receipt.  The more they know about your purchasing habits, the better they can direct their advertising efforts to you.  In a simple scenario, someone who buys diapers may get a coupon for baby food, while someone who buys veggie burgers may get a coupon for soy milk.

Take a look at how online retailers work.  When you go to Amazon (assuming you’re logged in to your account) you get some recommendations.  Amazon tracks what you buy, and what you look at, and compares that information to all of its other customers to find similarities.  And it’s not just Amazon.  Most of the advertising you see as you browse the web is likely to have been targeted at you.  Retailers want their advertising dollars to be spent showing you the ads that you are most likely to respond to.  If they pay to show a diaper advertisement to a bachelor, they’ve just wasted their money.  Marketers thus have a strong motivation to figure out what you will respond to.  50 years ago, the best they could do was to look at buying trends in a specific area, or try to target certain interest groups in newsletters.  If you didn’t belong to the same organization or live in the same area as someone else, there was no way to predict that you might have similar buying habits.  Things are different now.  Between “Rewards” cards, online shopping, and the increased use of credit cards, savy marketers can in principle track nearly everything you buy as well as follow what news you’ve seen and what you do for entertainment.  None of this was imaginable a few decades ago.  First, there simply wasn’t a central information that could be used to gather the information – thank you internet.  Second, the processing power necessary to draw out hidden connections was far too expensive.  It’s easy to guess that a diaper buyer may want baby food, but what do you offer the person who buys oranges, a newspaper, and a pound of bacon?  Figuring out the answer to that question requires analyzing the habits of millions of shoppers and tracking thousands of variables – it ain’t easy.

Now imagine if we took all of the effort and technological power that is employed in advertising and put it to work improving education.  That’s essentially what learning analytics is about.  At the broadest level, this sort of data driven approach could help guide students to the schools at which they were most likely to excel.  Within a school, teacher selection could be done based on empirically driven conclusions about which teacher was likely to be best for a student rather than by random choice or hunch.  At the finest grained level, curricular paths within a course could be individualized and optimized for each student.  This sort of approach would mesh very well with the increasing emphasis in online learning.  In an online environment there is less reason for every student to approach a subject in the same manner or sequence.  A computer based curriculum could be presented in a way that was most likely to be beneficial to a particular student rather than employing a one size fits all approach.

There are of course huge ethical considerations lurking just in the background.  How, for example, do we balance the desire to optimize education with the need to respect a student’s choices that may be deemed non-optimal?  how do we operate such a system without violating reasonable expectations of privacy?  As well, there are substantial metaphysical issues to be addressed.  The scenarios I have suggested seem to presume that there is something like a best path for a student to take.  Yet much of the successes in our lives seem to depend on random encounters.  It’s not clear that it would be possible to possible to fully optimize a learning path, even if we were agreed that it was ethical to do so.  These and similar questions must be addressed before learning analytics can come close to fulfilling its promise.  Being a philosopher, I’ll try to go some way toward addressing them in this blog, though my primary focus will be on the more technical questions.

Posted in Learning Analytics | Leave a comment

Learning Analytics: A brief introduction

So what is Learning Analytics?  In brief, it’s the use of statistical analysis, data mining, and similar methodologies to improve student learning.  This can be done is a wide variety of ways.  For example, a detailed analysis of student demographics might find that certain groups of students are more likely to succeed in online versus traditional versions of the same course.  With this information, a university or department might attempt to guide students to the type of course which maximizes the likelihood that they will succeed.  Or we might conduct a detailed analysis of student performance throughout a course to look for identifiable predictors of difficulty.  Limited resources could then be targeted to specific students to more effectively influence outcomes.

Here’s a good overview.

Posted in Learning Analytics | Leave a comment

Remember when technology was going to save education?

If you’ve been in education as long as I have, you remember the initial promise of technology.  Computers were going to revolutionize teaching and learning.  Not only would they streamline the work of teaching by taking over the drudgery of grading, but they would also help teachers to develop and present more compelling instructional materials.  Students would have nearly instant access to a wide range of materials from which to learn, so so each student could learn in the way that best suited her.  By lightening teacher workloads and increasing student access, we would break the dawn of a new day in learning with technology allowing students and teachers to work cooperatively toward shared goals.

It was a wonderful vision, but I have to say that reality has been a bit of a let down. Sure, we have lots of technology. Most of my courses now depend heavily on computers and the internet.  My classrooms now sport SMARTBoards instead of chalkboards, easels and flip charts have given way to PowerPoint presentations, and most work is turned in via shiny new electrons instead of pieces of dead forest.  But the assignments, presentations, and in class scrawling is all largely as it was when I first stepped into a college classroom as a student back in 1979.  To be sure, there are some definite advantages to the new media.  Now I can put nearly all of my course content online so that students can access it whenever they wish.  Computers have indeed taken over some of my grading, though essays and much of mathematics seems beyond them still.  If I want to offload my grading onto a computer, I have to construct assignments with an eye to what the computer can handle rather than on what will best help the students to learn.  I don’t consider that an acceptable trade-off. The internet has given students access to a treasure of information I couldn’t have even dreamed of at their age.  Unfortunately, that trove makes plagiarism easier than ever, and many students find the allure of YouTube antics far greater than that of the assignment du jour.  It often seems that for every task that computers have made easier, they have introduced a new task or made another harder.  I used to have to create a syllabus for every course.  Now I have to put that syllabus online and keep up with an ever-changing e-learning platform, Moodle, in my case.  Time that I used to spend in developing content for the classroom now gets spent just learning new technologies.  Unless I make up that time elsewhen, my courses end up technology rich but content poor.  In many ways teaching and learning are the same as they ever were, a tortuous pas de deux of teachers struggling to keep up and students struggling to catch up.  Plus ca change, plus c’est la meme chose.

But despite this, I remain optimistic.  Technology can make a difference, a tremendous difference, it just won’t do it easily.  To my mind, the key to teaching and learning is not the latest web gadget, or that newfangled piece of hardware hanging on the classroom wall.  The heart and soul of teaching lie where they always have, in the always human and often fragile interaction between student and teacher jointly feeling their way toward a common goal of learning.  Anyone who has spent much time in front of a classroom knows that the path to this goal is different with each student. The challenge of teaching lies in finding how to connect with this-particular-student-in-front-of-me-now with regard to this-particular-thing-to-be-learned.  The resolution of that challenge depends critically on the particular teacher, the particular subject, and the particular student, and the particular context, it is never the same twice.  This challenge is never resolved by simply adopting a new technology or a new pedagogical framework, it’s far too complex for that.  Teaching today’s student with today’s technology is essentially no different than Socrates teaching the slave boy by drawing figures in the sand – the technology is simply a method for the teacher and student to connect with each other about the subject at hand.  Stick, chalkboard, PowerPoint, Web application, the difference is hardly worth mentioning when compared to the other elements in the learning context: teacher, student, subject, context.

So whence my optimism?  Even though i see little difference between chalkboards and Prezis, computers have made one thing possible: massive computation.  Of course that massive computational power is what makes Prezi possible, it takes a lot of computation to track what’s going on in a Prezi and make sure that everything makes its way from server to student intact.  But that computation doesn’t really do anything different than the basic physics of a chalkboard does.  No, the promise of massive computational power lies elsewhere.  Of course it lies in part in the promise of better modelling of the brain processes involved in learning, but that will be of little direct use to the teacher in the classroom.  What that teacher needs is an answer to the question: how do I reach this student now?  I believe we are now at the threshold at which computers can help us answer that question.  This blog will be an exploration and imagination of how that might work along with an assessment of what progress we have already made in that direction.

Posted in Technology in Education | 2 Comments