Keeping Up With... Learning Analytics

This edition of Keeping Up With… was written by Steven J. Bell.

Steven J. Bell is Associate University Librarian for Research and Instructional Services at Temple University, email:

Helpful or Harmful

In the current higher education landscape there is at least one significant development generating enthusiasm and suspicion in equal parts. Learning analytics refers to technologies, usually software tools, that enable the analysis of student data in order to identify learning weaknesses so that faculty, advisers and even librarians could intervene with corrective action. It has its origins in business and industry where statistical analysis of big data is used to analyze market trends and consumer spending in order to predict purchasing behaviors. While this technology has the potential to increase the odds that academically at-risk students will persist to graduation, it also presents ethical issues related to the gathering and analysis of student data.

Academic librarians are already crunching data to discover relationships between the use of library resources and student academic performance. From an assessment perspective this can help justify library expenditures by demonstrating how academic libraries contribute to students’ retention and persistence to graduation. The promise of learning analytics is in applying the analysis to deliver more personalized learning or to create adaptive pedagogical practices that will enable college students to learn more effectively and persist to graduation. Given higher education’s dismal graduation rates, particularly at the community college level, policymakers and administrators believe that learning analytics can lead to dramatic improvement.

Shift From Post- to Pre-

Using student data for performance analysis is hardly a new idea. Education systems use grades, test scores, essay analysis and more to assess our skill levels for placement and certification of learning. With technology improving our capacity to gather large volumes of data and connect it into enterprise resource and learning management systems, higher education is raising the stakes on analytics. The shift is from post-learning analysis to preemptive warning and corrective intervention.

  • Predicative analytics is finding its way into all walks of life. Employers use it to improve employee performance. Governments use it to make better decisions. Health care uses it to control costs. Universities will now use it to identify at-risk students.
  • Course Signals is Purdue University’s predicative analytics system designed to help faculty intervene when student data points to potential failure.
  • The vast stores of student data that higher education institutions can tap into offers potential for programmatic change, improved outcomes assessment and cost containment.
  • Learning management systems are being examined as platforms to provide faculty with analytics dashboards that offer real-time data and analysis on student performance and engagement with course content.
  • Supported by a Gates Foundation grant, the Western Interstate Commission for Higher Education has created the Predictive Analytics Reporting Framework to compile 1.7 million student records and 8.1 million course level records to analyze student performance.

As learning analytics technology becomes accessible to even greater numbers of colleges and universities, there will be growing pressure to adopt it owing to federal and state government actions to connect student success to funding. Changing demographics point to fewer college-age individuals which means any loss of students and their tuition dollars could lead to deficits and layoffs. The predictive power of learning analytics could become the go to solution for what ails higher education if it can truly keep students retained and on the path to academic success.

Red Flags Rising

Despite the rich potential of learning analytics it raises serious concerns about the security and privacy of student data. Mining data to identify students for signs of academic weakness sounds too much like surveillance to privacy proponents. Why the rush to submit to big data enthusiasts when faculty, using traditional methods, have the capacity to identify and correct student learning problems? Some critics believe the threat or experience of failing offers valuable life lessons, and that predictive analytics could endanger an important life skill if we always monitor performance and swoop in to prevent failure. Some fear students may come to depend more on analytics to drive them to success than their own intrinsic motivation. Even when learning analytics can identify at risk students, the research evidence has yet to support it’s accuracy. In the classroom, faculty may still struggle to determine the appropriate solution without fully understanding what learning analytics are telling them about underperforming students. While the systems may hold promise, today there are many questions being asked.

Do the Pros Outweigh the Cons?

Though few academic libraries are encountering it just yet, it is only a matter of time before higher education institutions integrate learning analytics at every level of the organization. According to the 2014 edition of the NMC Horizon Report, the time to adoption is just one year. Within our professional community mixed reactions would come as no surprise. Academic librarians are passionate about helping students succeed and would gladly leverage technology in the pursuit of that goal. Learning analytics could identify students who need additional research support, enabling librarians to target those to whom they provide personalized assistance. Academic librarians are also particularly sensitive to protecting the privacy of student data, and would no doubt prefer to avoid situations and technologies designed to exploit that information.

Thoughtful Decisions Needed

This unique dualism stands to present an ethical dilemma. Does the potential value of learning analytics outweigh the concerns about how student data is being used?  Academic librarians will need to balance learning analytics’ capacity for personalizing library services in support of students who struggle academically with their professional commitment to protect patron privacy and use personal data ethically. One solution may be give students an opt in or out option for allowing learning analytics to track their academic performance. There may be other ways to work with our academic support or institutional assessment colleagues to apply learning analytics so as to offer maximum protection to students’ private data. We might discover that students, owing to exposure to learning analytics through their K-12 school experience, are well adjusted to predictative analysis and might anticipate it. Academic librarians are no strangers to adapting learning technology to connect with students and faculty. Learning analytics may just be the next technology that advances efforts to integrate the library into the learning process. We will need to proceed, but should do so cautiously and thoughtfully with our students’ best interest in mind.

Resources and Reading

Bell, Steven. “Taylorism Comes to Campus.” From the Bell Tower (Library Journal). July 23, 2014. Accessed October 3, 2014.

Brown, Malcolm. “Learning Analytics: Moving From Concept to Practice”. EDUCAUSE Learning Initiative. July 2012. Accessed October 7, 2014.

New Media Consortium. NMC Horizon Report: 2014 Higher Education Edition. See “Learning Analytics” pp. 38-39. Retrieved at

Sclater, Niall. “Snooping Professor or Friendly Don? The Ethics of University Learning Analytics.” The Conversation. February 26, 2014. Accessed October 6, 2014.

Sclater, Niall. “Taking Learning Analytics to the Next Stage.” JISC Effective Learning Analytics. September 16, 2014. Accessed October 1, 2014.

Warner, John. “The Costs of Big Data.” Just Visiting (Inside Higher Ed). July 6, 2014. Accessed September 22, 2014.

“7 Things You Should Know About Analytics” EDUCAUSE. Accessed September 25, 2014.

Steven Bell’s DIIGO page for “Learning Analytics.”