Sunday, March 3, 2013

LAK13 Assignment #1 - Analytics: Logic and Structure Postscript

In Learning Analytics and Knowledge 2013, the first assignment, Analytics: Logic and Structure (link is live only if you are registered for the course), has the following description:
For this assignment, develop an analytics model to gain insight into a complex topic using both qualitative and quantitative methods. Select a particular topic or subject area that interests you (current events, historical activities, a learning challenge) and detail how you will "interrogate" this subject using various analytics tools or techniques. Your project can be in the form of a presentation, a blog post, a video, a simulation, or other digital artifact. The important aspect of this assignment is to walk through the processes and considerations that pre-date tool selection.
There are additional details, but this is the core charge of the assignment. I commented on the assignment in the forums:
Our first assignment seems to allow for a wide variety of projects. This is understandable, given the variety of backgrounds of students, varying understandings of statistics, and the convergence of a number of fields that contribute to Learning Analytics.
I then went on to ask some clarifying questions, and made some suggestions as to the structure of the project, specifically getting a data set to work from and then developing an analytics model. Looking back I realize my suggestions were running counter to the intent of the assignment. Mr. Siemens is looking to replicate the situation that people in analytics are dealt with; hodgepodges of data silos, inconsistent objectives from the different institutions (or within a single institution), regulatory barriers, and a legion of other issues. By having us develop a relevant question we would be put in a position to deal with these barriers, and share our experiences in getting around them.

I appreciate his commitment to using authentic contexts, but feel a bit more direction would have been helpful. The description doesn't specify that the context is necessarily learning analytics, and if it were, the issues above would prevent real data from being gathered. In a later post I mention:
It is starting to look like getting a usable dataset is going to be the primary issue for our projects for this course. The sort of data sets that we are looking to use usually contain sensitive information, and in the case of our US colleagues (myself included), using them in such an open setting would run afoul of FERPA. I have a dataset I am working with, but do not feel my institution is in a place to intelligibly create a data policy, let alone a data openess policy.
All this points to a disconnect between my goals for this course, and Mr. Siemens' goals for this assignment. This course being part of my personal and professional development, I want some clear tools and techniques to analyze student-generated data with when I leave this course. I have serious reservations about using my real data, and thus want an available data set that will help me develop those tools and techniques. In this assignment Mr. Siemens is more concerned about the way we frame our analytics questions, and our plans for how to answer them.

To satisfy both of our goals, I've decided to look at student and teacher performance. There is a wide range of open datasets available, and there are powerful questions about learning analytics that can be approached. LA focuses on related, but different datasets, however the tools and techniques to analyze this data set should transfer to analyzing student generated data in an LMS.

No comments:

Post a Comment