(Better) Learning Analytics - Big data, big problems


Let’s carry on the discussion that couldn’t fit into the limited time we had for the roundtable here.

Here are the questions from the splash page:

  • What are the historical origins of “learning analytics”? Why do these suddenly matter?-
  • What should we be measuring? How does that stack up against what we are
    measuring? (How might “measuring” be a problem in and of itself?)
  • Who does “Learning Analytics” well?
  • Which ideas have shaped your thinking about learning analytics?
  • How do teachers use the output from learning analytics? What are the challenges from informing instruction with data?

And here is the reading that was recommended:

  • Learning Analytics: A Friday Night Rant
  • Big Data Learning Research Breakthrough: Learning Activities Lead to Achievement
  • No More Secret Sauce Analytics


The first recommended reading is pretty funny.


And the third link (discourse has a limit of 2 links early on):

Video is here:


I had a similar post about secret sauce analytics (and why it’s a poor concept):




Thanks for joining us @mike–I really enjoyed your post.

Heres’ the recap from yesterday: http://info.p2pu.org/2014/08/20/better-learning-analytics-lessons-from-edx-hack-education-mozilla-p2pu/

And another take from Laura Park Gogia: http://laurascoloringbook.blogspot.com/


As I reflect on yesterday’s meetup, there’s one topic that I don’t think came up in the conversation. That’s segmentation of students. We kept talking about “students” as if it was a homogenous entity. Is analytics good for students? What do we measure about students? What are the ethics around student data? I’d argue that many of these questions have different answers based on what type of students we are talking about.

For context, here’s my background: B.S. in mathematics at a stereotypical ‘small liberal arts college’ (2,000 students, small classes, collegial campus). MBA at an R1 institution, and then taught as an adjunct for 10 years at a large open admissions school for non-traditional adult learners (U. of Phoenix). The proposition about analytics is different for all three…and DRASTICALLY different for the non-traditional adult segment. I had little to no exposure to the third group until I started teaching there in 2002, but I realize that it’s a significant portion of “students” these days.

My point is that we need to acknowledge that there are different segments and we need to be a little more clear as to which group we’re talking about since a single analytics solution might benefit one segment while simultaneously hindering another.