A concept on the forefront this week was action analytics. It refers to the application of academic or learning analytics with clear action component in mind. Most universities collect a lot of institutional stats, but struggle to provide more than reports. In an earlier blog post I briefly described the Signals Project at Purdue University.
This project provides an “early warning system” for students at risk of dropping out. Socio-economic data, educational data mining techniques and (increasingly) social network analysis.
The background for this trend is an increasing demand for higher education, combined with a drive towards increasing cost efficiency and quality. Action analytics can (among others) help to increase the retention rate (early drop-outs are costly), improve admission and provide a more personalized learning experience (for example through the use of recommender systems, cfr. Amazon). Action analytics are currently most found in for-profit educational institutions.
Linda Baer from the Gates Foundation recalls in her LAK11 talk the use of action analytics (coupled with measurable targets) as central in the organization’s functioning. Analytics is not limited to the research department, but requires buy-in from all departments (and the leadership) within an institution. Furthermore, she argues that the use of analytics is not limited within the institution, but that the availability of cross-institutional performance data will grow dramatically. A driving force is the growing success of “open architecture” approaches, instead of earlier, closed “data silos”. For example, Education Watch Online! provides for the USA “state by state” summaries of achievement and best practice examples of schools (Norris et al, 2008).
Finally, the skills, competences and habits of mind, needed to be employable in today’s flattening world, cannot be assessed by today’s measures of grades and resumes. The
use of e-portfolios will rise. Individuals will be responsible for maintaining a lifelong personal learning record in order to demonstrate their capabilities (Oblinger, 2007).
The 2011 Horizon Report (shortlist) predicts an adoption time for learning analytics of 4-5 years.
Here is a link to the recording of the guest lecture by Linda Baer.