Friday, April 21, 2017

Using Big Data Analytics to Measure Learning


A trend in higher education is an increased focus on measuring learning (Adams Becker et al., 2017).  The basic idea is to apply Big Data Analytics techniques that have been in other industries to education.  The goal is to enrich the education experience while identifying students in need of intervention.

Learning analytics (LA) is used to describe the capture, analysis, and use of data in an education related environment.  LA takes a broad view of the education process, to include the environment where it takes place. By capturing large and rich datasets, institutions and learners can generate customized feedback to improve progress.   Just as marketing organizations attempt to gather a holistic view of their target audience, educational institutions are using similar approaches.  For example, social media content is combined with data captured using video cameras and other tracking technologies. Figure 1 presents a simple visual representation of the approach.  Information about what educational resources that are being utilized (such as a Library) are captured and used with the other data to gain insights into the students’ behaviors.

Predictive analysis is used to identify at-risk students prompting intervention.  For example, if a pattern associated with student drop out is identified the educational intuition can intervein and attempt to modify the student's trajectory. Likewise, analytics are performed on overall student activities to gain insight into course and program effectiveness.  This information can be used by curriculum designers to make improvements.

It is not unimaginable to envision the equivalent of shopping basket analysis (associative rules) to make suggestions to students about course or even degree programs.  For example, a student might receive an email saying “based upon your previous academic activities; we recommend you take the following courses and consider the following majors.”


Figure 1: Modern Education Approach

This approach to education has several issues to overcome.  There are serious legal and ethical questions regarding this degree of observation of students.  The previous example of using predictive analytics to identify students at risk of dropping out could also be used by an institution to artificially improve their retention rate. Likewise, without proper disclosure and consent, tracking a student’s movements and utilization of course materials is likely an invasion of privacy. 

The best approach to addressing these issues is with transparency, disclosure, and required consent.  Just as a user of a social network or search engine should be able to see and download all data that is being stored related to them, a student should have the same rights.  Controls should be in place to ensure that the collected data is accessed in an appropriate way.  For example, a student should grant permission for their parents to view their activity.

As with most Big Data applications, LA has both promise and challenges.  Moving to a data-driven educational system has the potential to enable better and previously unimaginable learning.  It also has the potential to take “teaching to the test” to a new level.



References

Adams Becker, S., Cummins, M., Davis, A., Freeman, A., Hall Giesinger, C., & Ananthanarayanan, V. (2017). NMC Horizon Report: 2017 Higher Education Edition.   Retrieved from http://www.nmc.org/publication/nmc-horizon-report-2017-higher-education-edition/


No comments: