What are Learning Analytics?

Learning analytics is an educational application of “big data,” a science that was originally used by businesses to analyze commercial activities, identify spending trends, and predict consumer behavior. The rise of the Internet drove research into big data and metrics as well as the proliferation of web tracking tools, enabling companies to build vast reserves of information they could study and apply to their marketing campaigns. Education is embarking on a similar pursuit into data science with the aim of improving student retention and providing a high quality, personalized experience for learners. Learning analytics research uses data analysis to inform decisions made on every tier of the educational system. Whereas analysts in business use consumer data to target potential customers and personalize advertising, learning analytics leverages student data to build better pedagogies, target at-risk student populations, and assess whether programs designed to improve retention have been effective and should be sustained — outcomes for legislators and administrators that have profound impact. For educators and researchers, learning analytics has been crucial to gaining insights about student interaction with online texts and courseware. Students are beginning to experience the benefits of learning analytics as they engage with mobile and online platforms that track data to create responsive, personalized learning experiences.

INSTRUCTIONS: Enter your responses to the questions below. This is most easily done by moving your cursor to the end of the last item and pressing RETURN to create a new bullet point. Please include URLs whenever you can (full URLs will automatically be turned into hyperlinks; please type them out rather than using the linking tools in the toolbar).

Please "sign" your contributions by marking with the code of 4 tildes (~) in a row so that we can follow up with you if we need additional information or leads to examples- this produces a signature when the page is updated, like this: - Larry Larry Feb 8, 2012

(1) How might this technology be relevant to the educational sector you know best?

  • Australian higher education (in common with other commonwealth countries) is under heavy political scrutiny. The accountability for public funds and the political importance of education for society and the economy mean that Government is interested in being seen to be a tough manager of the sector extracting the most value possible. Inevitably this has led to the creation of performance accountability measures. The challenge is that these are typically activity measures or based on factors outside of institutional control (e.g. employment in specific industries). Learning analytics potentially allow for the creation of a better quality evidence base that can reflect and respond to pedaogical designs and disciplinary differences, which can then be used to engage with external stakeholders. There is also the possibility that LA might provide better tools for institutions supporting students to be more successful (both those struggling and those doing well) and in helping staff improve teaching practices. - stephen.marshall stephen.marshall Feb 10, 2015
  • - helen.carter helen.carter Feb 24, 2015I would agree with Stephen's observations here - UK ref. http://analytics.jiscinvolve.org/wp/2015/02/24/uk-learning-analytics-network-notes-from-kick-off-meeting/
  • I also agree with Stephen's observations above. One other aspect of the continued growth in learning analytics is that it can be used to better understand learning and by that I mean in relation to the science of learning as opposed to just a phenomenographic view of student learning. - jason.lodge jason.lodge Feb 26, 2015
  • Learning analytics becomes a powerful application when combined with personalised learning- mike.keppell mike.keppell Feb 26, 2015
  • Learning analytics, at present, reveals only the most granular feedback. The publishing industry now gathers a click stream from all the publications, and in a presentation a key player reveled that the most significant finding was that they could predict who was going to complete the course before the course began with a very high accuracy from a single data point. That a single click (in the case the click was whether a student open their e-pub prior to the course beginning or not) offers the promise of being able to see things we have never been able to know before and the theory goes "as we get more information on what students are doing we will be able to learn more about what works so we can improve their education (in many as yet not clearly defined ways which we call personalisation). I carry on the case in 2 & 3 below to explain why the data will however likely be used as analytics of performance of the employees than learners) - daniel.ingvarson daniel.ingvarson Mar 1, 2015

(2) What themes are missing from the above description that you think are important?

  • I think it is worth acknowledging that Learning Analytics is closely related to the concept of Educational Data Mining, with two research communities in existence. The differences are primarily of emphasis, but both terms will generate useful information in this space. - stephen.marshall stephen.marshall Feb 10, 2015
  • One of the assumptions of LA is that the activities that are being measured are a good representation of student experiences, this assumes that the activities in all courses for all students are able to be measured. In practice, the available measures are much more limited. - stephen.marshall stephen.marshall Feb 10, 2015
  • LA raises issues of student privacy and ethics that need to be carefully considered. Students need to be comfortable that the data being collected is being used to support their learning in positive ways. It is unclear that there is yet any evidence of LA activities resulting in positive changes to student outcomes or rates of success. - stephen.marshall stephen.marshall Feb 10, 2015 - helen.carter helen.carter Feb 24, 2015http://theconversation.com/snooping-professor-or-friendly-don-the-ethics-of-university-learning-analytics-23636
  • Shane Dawson's slides from ASCILITE are worth reviewing http://www.slideshare.net/shaned07/rhetoric-reality-ascilite2014upload - stephen.marshall stephen.marshall Feb 10, 2015
  • - helen.carter helen.carter Feb 24, 2015Agree with Stephen's observations here
  • The current state of LA seems to be coalescing around three categories each with its own particular value: learning analytics that is of value to the individual student or teacher; educational data mining analytics across a larger population which is of value to policy makers (institutional, government, industry and communities); and visual/interactive data analytics which brings the power of visualisation and adaptive modelling to test research questions (of value broadly). Personally I think the fields of analytics is critically important for the future. - philcounty philcounty Feb 25, 2015
  • In a similar way to the way in which epidemiology has raised questions about health, I suspect that learning analytics will begin to contribute to a fuller understanding of how students learn in higher education in the 2010s. Notwithstanding the arguments made above and those of Shane (all of which I tend to agree with), learning analytics needs to grow beyond an end in itself and become part of an evidence ecosystem that will include everything from laboratory studies about learning through to phenomenographic and ecological work. This would be a similar progression to what occurred in the emergence of health informatics and, more recently, psychoinformatics. - jason.lodge jason.lodge Feb 26, 2015
  • What worked in the case above was that there are two clear data points which can be compared to reach a conclusion. The process of learning does not currently have such clear points out side of basic skills and fields which are based on singular answers (eg maths). The thing that is missing is that in discussing the technology we miss that it is actually understanding learning that we need to have in order for the analytics to know the meaning of what they are analysing. There is a hope that huge piles of big data will uncover it, however be our ability to assess what someone has learnt is so course grained out ability to determine what caused that learning will only ever be that course grained. - daniel.ingvarson daniel.ingvarson Mar 1, 2015
  • Additionally the topic needs to state that it is the publishes who are collecting and analysing the data more than our institutions. A consideration of this topic must be who owns this data and who gets to make the analytics. - daniel.ingvarson daniel.ingvarson Mar 1, 2015
  • Lastly the successful players are mostly coupling analytics to content (and that content to a learning model which maps the concepts which are being taught together as a map(and accompanying assessments). This means that current analytics is tied deeply to highly specific content. (rather than as we think of it as analyzing the learner, it is the effect o of the content which is measured)

(3) What do you see as the potential impact of this technology on teaching, learning, or creative inquiry?

  • The pressure for data being imposed by managers and external stakeholders is likely to result in pressure for wider adoption of technology within courses, such as minimum presence policies. Measures of activity online are likely to see at least some institutions imposing responsiveness and activity targets on staff (particularly adjunct or contract teaching staff). - stephen.marshall stephen.marshall Feb 10, 2015
  • - helen.carter helen.carter Feb 24, 2015Hopefully our use of data will be a bit more sophisticated than Stephens analysis above, though no doubt individual institutions and their senior executive might look for the silver bullet solution. I expect there to be significant use of corporate tools in this space, matching what is happening in the consumer sector. Blackboard, OUA, D2L, Moodle all have offerings in this space, making it potentially easier for institutions to implement at an institutional level.
  • Following on from above. I therefore concluded that it is easy to use the analytics for aggregated comparisons through time or across courses and hard to do for content for specific learners. This means that the early data which will come from Learning Analytics will be from course assessments and so we will have performance metrics. Where there is is similar course content and similar conditions it will be tempting to look at what are the correlating factors between these. stephen notes this, however it is now happening in k-12. In New York the Govener has pushed to have 50% of a teachers performance metric based on the learning growth of their students during the course based on two standardised assessments. This is not learning analytics, however I submit it as evidence that if administration feels they can create performance accountability measures from data generated in the process of learning analytics, they will do that as the primary use and it will become analytics not a for learning. - daniel.ingvarson daniel.ingvarson Mar 1, 2015

(4) Do you have or know of a project working in this area?

  • OLT have funded an Australian LA project that is undertaking a stocktake, Shane's slides include some preliminary information, the project site is here: http://www.olt.gov.au/project-student-retention-and-learning-analytics-snapshot-current-australian-practices-and-framework. Reports and other publications are apparently due later this year. - stephen.marshall stephen.marshall Feb 10, 2015
  • - helen.carter helen.carter Feb 24, 2015This paper presented in 2013 reports on 4 Australian projects, which could be worth following up. As the paper indicates, a lot of the work in this area is not made public outside the particular institution. Ref. Atif, Amara, Richards, Deborah, Bilgin, Ayse, Marrone, Mauricio. North Ryde, NSW : Macquarie University; 2013. Learning analytics in higher education : a summary of tools and approaches.
  • - helen.carter helen.carter Feb 24, 2015Deborah Richards and Chris Froissard, Macquarie University have been working on a project looking at validating the effectiveness of the Moodle Engagement Block and piloting a student alert system to optimise retention. The project looks at learning analytics together with student alerts and how they can be used to improve student outcomes. (will try and source a ref.)
  • At the Learning Analytics Research Group (LARG) at the University of Melbourne - https://le.unimelb.edu.au/elearning/larg.html - we're attempting to draw on expertise from across disciplines to situate learning analytics within a broader approach to better understanding student learning in higher education. - jason.lodge jason.lodge Feb 26, 2015
  • Knewton - and the big 4 publishing houses now all have major investments ($100m plus) so its now outside the education institutions. - daniel.ingvarson daniel.ingvarson Mar 1, 2015

Please share information about related projects in our Horizon Project sharing form.