6 Things You Should Know About Learning Analytics
1. What is it?
Learning analytics provide statistically based evaluation of data sources to determine patterns that lead to more informed decisions. For teaching and learning, activity information from sources such as web tools (e.g., Carmen), labs and classrooms (e.g., programs uses, resources), and libraries (e.g., learning objects) can help identify successful course strategies or students at academic risk. In sum, it connects data produced from explicit student activity into information that guides meaningful action.
In a world with increased pressure to demonstrate student success, retention, and ROIs, Learning Analytics has been referred to as the “third wave” of transforming the student experience, equal to such game changers as learning management systems (the first wave) and social media, cloud applications, and web 2.0 tools (the second wave). Source (http://net.educause.edu/ir/library/pdf/ELIB1101.pdf).
From the 2011 Horizon Report: http://wp.nmc.org/horizon2011/sections/learning-analytics/
Image: http://www.elearnspace.org/blog/2010/08/25/what-are-learning-analytics/
2. How does it work?
Learning Analytics work by sharing data or establishing a sharing network allowing for individual, course, and program-based performance to be looked at holistically. Ideally, the program would automatically trigger responses to individuals or coordinators with information based on the data, such as linking to support services (e.g., writing center) should a student’s writing assignments consistently receive poor grades, or congratulatory notes upon marked improvement in a discipline or assignment type. At its core, learning analytics can help define pathways to improve performance for a student at an individual level and improve retention, achievement, and graduation rates at an institutional level.
Without a doubt, learning analytics requires great coordination, collaboration, and security to be effective. Tasks like creating a committee of systems supervisors can be challenging, even at small institutions. A committee like this would be charged with determining what data is exportable, how it could connect to other data streams, and what questions could be answered with the data.
3. Who is doing it? And why?
Learning Analytics is just starting to make major gains in the education sector. Analogous to swiping your local grocer’s loyalty card, examining information related to activity (length and sequence of events) can start creating a culture of predictive intervention and support. Just as your grocer will send you coupons based upon your current purchases, an instructor could send students information about successful habits or information about ways to improve performance based on their current activity in a course.
In higher education, Purdue University’s Signals (http://www.itap.purdue.edu/learning/tools/signals/) is an integrated system to provide students a dashboard visualization of course performance. Other institutions that are early adopters and provide real-time data to students and administrators include Drexel University and University of Wollongong (Australia). Each institution uses the dashboard in an effort to provide students more information and knowledge about their academic progress.
4. What are the downsides?
Significant downsides include the ethical and legal considerations of collecting and using system information to understand individual users and their activities. Even with the best intentions, there are concerns regarding student profiling or the ramifications of actions that are or are not taken for an individual (either purposely or by poor algorithm design). With so much data available from various sources, constructing a foolproof analysis is virtually impossible; incorrect or misleading patterns could emerge. Finally, even with very sound analysis, the appropriate intervention techniques are still being developed.
5. Where is it going?
With increasing financial constraints, predictive abilities and the need for proactive intervention to ensure student success will become paramount. Programs and best practices will help institutions better connect admission information, LMS activity, clicker data, library use, lab programs, e-portfolios, and online discussions into algorithms that can predict how successful a student will be and provide personalized next steps to enhance performance.
At another level, the same data could produce a wealth of information about resource allocation, teaching effectiveness, program effectiveness, and so on.
6. What are the implications for teaching and learning?
Learning analytics requires consistent and heavy use of services and programs connected to authentication, such as computer log-ins, learning management systems, and so on. For those courses, programs, or institutions that aren’t operating via computer networks or using electronic applications, capturing activity data would be difficult. The analysis of data patterns and the implementation of interventions should be done carefully, as common pitfalls of data analysis include assumptions about intention and causality.
The potential benefits of optimizing and aligning dwindling resources, providing real-time progress, and facilitating meaningful communication among myriad stakeholders are truly exciting.