Open Research Online

The Open University’s repository of research publications

and other research outputs

An overview of learning analytics

Journal Item
How to cite:
Clow, Doug (2013). An overview of learning analytics. Teaching in Higher Education, 18(6) pp. 683–695.

For guidance on citations see FAQs.

c 2013 Taylor Francis

Version: Accepted Manuscript

Link(s) to article on publisher’s website:

http://dx.doi.org/doi:10.1080/13562517.2013.827653
http://www.tandfonline.com/doi/full/10.1080/13562517.2013.827653#.UhJALGTdLZs

Copyright and Moral Rights for the articles on this site are retained by the individual authors and/or other copyright
owners. For more information on Open Research Online’s data policy on reuse of materials please consult the policies
page.

oro.open.ac.uk
An Overview of Learning Analytics

Doug Clow

Institute of Educational Technology, The Open University, Walton Hall, Milton Keynes,
MK7 6AA, UK

[email protected]

1
An overview of learning analytics

Learning analytics, the analysis and representation of data about learners in order
to improve learning, is a new lens through which teachers can understand
education. It is rooted in the dramatic increase in the quantity of data about
learners, and linked to management approaches that focus on quantitative
metrics, which are sometimes antithetical to an educational sense of teaching.
However, learning analytics offers new routes for teachers to understand their
students, and hence to make effective use of their limited resources. This paper
explores these issues, and describes a series of examples of learning analytics to
illustrate the potential. It argues that teachers can and should engage with
learning analytics as a way of influencing the metrics agenda towards richer
conceptions of learning, and to improve their teaching.

Keywords: learning analytics; analytics; metrics

Introduction

There is a tension between the framing of education as an economic activity and

conceptions of education and learning that are concerned with the development of

meaning and the transformation of understanding. These difficulties are far from purely

theoretical concerns: they increasingly have very practical, concrete consequences for

teachers and learners, notably around resource constraints, class sizes, and time

pressures. Within this constrained environment, teachers are subject to accountability

processes based on and enabled by the deployment of quantitative metrics of their

practices.

Quantitative metrics are increasingly used not only because of theoretical

framings that support them, but also because of a substantial and dramatic change in

their practicability over the last ten or twenty years. This change is often referred to as

Big Data: the quantity, range and scale of data that can be and is gathered has increased

exponentially (or close to exponentially). Accompanying this explosion of data is a

2
series of rapid advances in computational techniques for managing, processing and

analysing these large volumes of data in ways that are actionable. These developments

are transforming enquiry. The scale of data is greatest in science - for instance, the

Large Hadron Collider at CERN produced 23 petabytes (23 million gigabytes) of

information in 2011 (CERN, 2012). The effect is not restricted to science - for instance,

the ability to manage and integrate textual and geographic data is changing scholarly

practice in the classics (see e.g. Project HESTIA: the Herodotus Encoded Space-Text-

Imaging Archive, http://www.open.ac.uk/Arts/hestia/index.html). New approaches

become possible: for instance, rather than sampling, an entire population can be

captured. The volume and scope of data can be so large that it is possible to start with a

dataset and apply computational methods to produce results, and only subsequently to

seek an interpretation or meaning.

Big data is by no means restricted to the academy. Technology companies such

as Google and Facebook make managing staggeringly large datasets their core business,

but even companies such as grocery retailers are increasingly deploying big data

techniques to capture, understand, model, predict and influence consumer behaviour.

The growing field of Business Intelligence is concerned with the management and

processing of data to support corporate practice, including performance metrics.

The framing of education as an economic activity supports the view of

educational institutions as businesses Business Intelligence is increasingly applied in

higher education, in areas such as outreach and advertising, enrolment, management,

and fund-raising, but also in more academic areas. 'Dashboards' showing performance

metrics against targets are increasingly popular with senior managers, and political

pressures such as the current focus on college completion in the US reinforce this

direction.

3
 These developments are not always welcomed by teachers. Two examples are

illustrative. Texas A&M University introduced a system that calculated dollar amounts

for each individual faculty member, ostensibly accounting for that person’s net

contribution to – or subtraction from – the university’s financial position. This was not

uniformly welcomed by all faculty, who argued that the figures were inaccurate and

unfair (Simon and Banchero, 2010). In the United Kingdom, successive Research

Assessment Exercises and the forthcoming Research Excellence Framework seek to

calculate numeric values for research performance – again, to considerable controversy

about validity and the effect on practice.

But what about the learners?

Learning analytics is the application of these Big Data techniques to improve

learning. Learning analytics is currently a fixture in educational horizon-scanning

reports (see e.g. Johnson et al 2011; Johnson, Adams and Cummins 2012; Sharples et al

2012) and in a raft of other publications aimed at practitioners and aspiring practitioners

from organisations concerned with technology in education, such as Educause

(http://www.educause.edu/library/analytics), JISC (http://jisc.cetis.ac.uk/topic/analytics)

and SURF

(http://www.surf.nl/en/themas/InnovationinEducation/learninganalytics/Pages/default.as

px). Vendors of learning technology are providing analytics packages: for instance,

Blackboard, Desire2Learn, Instructure and Tribal have all released analytics tools, and

there is also activity in the Moodle community. The high-profile providers of Massively

Open Online Courses (MOOCs) - Coursera, Udacity and edX - are all using analytics

tools to inform their practice.

There is also a growing research community around the topic. An annual

conference, Learning Analytics and Knowledge, has been organised (Long et al 2011;

4
Buckingham Shum, Gasevic and Ferguson 2012), a special issue on the topic has been

published recently (Siemens and Gasevic 2012), and an international research network

set up: SoLAR, the Society for Learning Analytics Research

(http://www.solaresearch.org/).

This increasing activity has a range of drivers and facilitators. Firstly, there is

the pressure towards performance management, metrics and quantification. Secondly,

there is an increasing volume of data available about learners and learning, particularly

as more learning takes place online in Learning Management Systems or Virtual

Learning Environments (LMS/VLEs). Every page visited, every interaction, every click

can in theory be recorded and stored. Thirdly, statistical and computational tools to

manage large datasets and to facilitate interpretation have become available as a result

of Big Data activity.

The most commonly-cited definition of learning analytics emerged from an open

online course on learning and knowledge analytics (LAK11,

http://www.learninganalytics.net/?p=28) and was adopted by the associated First

International Conference on Learning Analytics and Knowledge in 2011 (Long et al

2011):

"the measurement, collection, analysis and reporting of data about learners and
their contexts, for purposes of understanding and optimising learning and the
environments in which it occurs"

As with any field of activity, particularly new ones, drawing clear distinctions

between related endeavours is problematic, contested and liable to change; however,

broad outlines can be drawn. Two other emerging areas have significant overlap with

learning analytics. The first is academic analytics, which is the use of business

intelligence in education. This tends to focus more at the institutional and national level,

5
rather than on individual students and courses (Long and Siemens, 2011). The second is

educational data mining (EDM), which seeks to develop methods for analysing

educational data, and tends to focus more on the technical challenges than on the

pedagogical questions (Ferguson, 2012). Learning analytics is first and foremost

concerned with learning.

A key concern in learning analytics is the need to use the insights gathered from

the data to make interventions to improve learning, to generate 'actionable intelligence'

(Campbell, DeBlois and Oblinger, 2007) which informs appropriate interventions. This

is addressed in accounts of the learning analytics process address. Campbell and

Oblinger (2007) set out five steps: Capture, Report, Predict, Act, Refine. Clow (2012)

places this as the central idea in his Learning Analytics Cycle (figure 1). The cycle

starts with learners, who generate data, which is processed in to metrics, which are used

to inform interventions, which in turn affect learners. The learners may be students in a

traditional higher education setting, or in less formal contexts. The data can include

demographic information, online activity, assessment data, and final destination data.

The metrics can be presented in a wide range of ways: from a simple indication of

whether learning progress is being made, to a comparison of outcomes to some

benchmark, or a visual picture of activity in an online forum. The interventions again

range widely, from students taking action in the light of metrics showing their activity

compared to that of their peers, to teachers contacting students identified as requiring

some particular extra help.

[INSERT FIGURE 1 HERE]

Figure 1: The Learning Analytics Cycle, from Clow (2012).

Learning analytics is not so much a solid academic discipline with established

methodological approaches as it is a 'jackdaw' field of enquiry, picking up 'shiny'

6
techniques, tools and methodologies, including web analytics (the analysis of logs of

activity on the web), social network analysis, predictive modelling, natural language

processing, and more (examples and explanations of these are given below). This

eclectic approach is both a strength and a weakness: it facilitates rapid development and

the ability to build on established practice and findings, but it - to date - lacks a

coherent, articulated epistemology of its own.

Having set out learning analytics and its context in broad terms, this paper

presents a set of more concrete examples of learning analytics practice, to provide a

more grounded view of the field. The examples are not intended to be exhaustive, but

were selected to give an overview of the range of possibilities opened up by learning

analytics. They are presented in a rough order of maturity and deployment, starting with

approaches that are widely deployed and validated in use with real students, and ending

with more speculative ideas under active development but not yet proven in practice.

Predictive modelling

The first example of learning analytics - in this paper, and indeed in the field - is

predictive modelling. The basic concept of predictive modelling is fairly

straightforward: a mathematical model is developed, which produces estimates of likely

outcomes, which are then used to inform interventions designed to improve those

outcomes.

Predictive modelling can be applied to education in a wide range of ways. The

best-established application is estimating how likely it is that individual students will

complete a course, and using those estimates to target support to students to improve the

completion rate. Sophisticated mathematical techniques like factor analysis and logistic

regression are applied to a large dataset containing information about previous students

on the course. This information includes things that are known at the start of the course

7
- such as the students' previous educational experience and attainment, demographic

information (such as age, gender, socio-economic status, etc), and things that become

known during the course - data about their use of online course-related tools (how often

they log in, how many postings that make) and formative and summative assessment

data. The final key piece of information is whether the students went on to complete the

course. A model is developed from this data, and then applied to the information

available for current students, to give a quantified prediction of whether each student

will complete the course. These predictions are typically displayed in some way to

teachers, department heads, administrators and so on in a 'dashboard'.

In principle, predictive modelling is not profoundly different from a traditional

teacher noticing which students are struggling in class and giving them extra help;

predictive modelling could be seen as simply extending this ability to the online

learning world. However, there are important practical differences. Firstly, the output of

predictive modelling is a set of estimated probabilities, and it is widely established that

many people struggle to correctly understand probabilities and to make consistent

decisions based on probabilistic information. Secondly, the output is not (typically)

restricted to a student's teacher: the information is readily made available to others

beyond the immediate learning context. Thirdly, the output can be used directly to

trigger actions and interventions without involving a teacher at all.

It is important to stress that the predictive power of these models is far from

perfect. Not only do they produce probabilities, but those probabilities suffer from

significant error: it is not possible to perfectly and accurately predict the chances of a

student completing a course based on the data available. However, they are significantly

more often right than wrong, and it is possible to use them to improve student

completion.

8
Course Signals at Purdue

The Course Signals project at Purdue University

(http://www.itap.purdue.edu/studio/signals/) is the most prominent and arguably the

most successful application of predictive modelling to student completion in higher

education.

The predictive model at the heart of Signals was first developed by Campbell

(2007), and is based on four components: demographic characteristics, previous

academic history, interaction with the LMS/VLE during the course, and performance on

the course to date (Arnold 2010). The predictions from the model are translated in to a

signal: green, denoting a high chance of success; yellow, denoting potential problems;

or red, denoting a high chance of failure.

Teachers run the model and generate signals for the students on their course.

The teacher can then choose what interventions to trigger: sending a personalised email

or text, posting the signal on the LMS (where the student alone can see it), referral to

support services, or arranging a face to face meeting.

The first pilot deployment of Signals was in 2007, and it is not applied to all

courses at Purdue. Results so far are impressive (Arnold and Pistilli 2012).

Overwhelmingly, students' signals tend to improve over a course, rather than worsen.

This in-course improvement is reflected in improved grades: the increases vary between

courses, but all see an improvement on previous semesters when Signals was not used,

with an average of 10 percentage-point increase in grades A and B, and a 6 percentage-

point decrease in grades D, F and withdrawals. There is increased overall retention: of

the 2007 cohort, 69% of students with no exposure to Signals are retained, compared to

87% of students with exposure to at least one course using Signals. Qualitative feedback

is very largely positive too, with students reporting that they perceive the emails as

9
personal contact, and faculty reporting that the tool helps them provide help to students,

and that Signals leads to students becoming more pro-active in seeking support. There is

at least anecdotal evidence that students carry the support-seeking behaviours from one

course to another, even where the subsequent course does not use Signals (John

Campbell, personal communication, May 2012). Importantly, the Course Signals are not

used in a decontextualised environment: the teacher is central to the process, and uses

their judgement to direct students to appropriate existing resources within the

university.

Other implementations

Predictive modelling has been used in many different universities (see e.g. Campbell,

DeBlois and Oblinger 2007), often with powerful results. It is quite possible to transfer

the overall approach between contexts, models themselves cannot be transferred, and

significant work is needed to develop an implement a successful system: there will be

variation not only in what data is available, but in its predictive power. As an example, a

project at the UK Open University (RETAIN, http://retain.open.ac.uk/) found that the

level of activity itself was not predictive of success or failure, but a fall-off in activity

was a clear indicator of trouble (Wolff and Zdrahal 2012): students could be successful

without being active online, but if a previously-active student stopped being so, they

were unlikely to complete.

Far from all modelling efforts are written up and made available to the research

community, particularly where the tools used are part of a proprietary system. One

notable exception is Desire2Learn, which in addition to being a high-tier sponsor of the

first two Learning Analytics and Knowledge conferences (LAK11 and LAK12) has also

published details of the approach to predictive modelling it uses in its products (Essa

and Ayad 2012).

10
Social network analysis
Social network analysis (SNA) is a set of methods for analysing the connections

between people in a social context, using techniques from the computer science field of

network analysis. Individual people (or, more technically, actors) in the social context

are called nodes, and the connections between them are called ties or links. A map (a

social network diagram, or sociogram) can be drawn by treating the nodes as points and

the connections as lines between them as lines. So, for instance, in an online forum, the

nodes might be the individual participants, and the ties might indicate replies by one

participant to another's post. These diagrams can be interpreted simply by eye (for

example, you can see whether a network has lots of links, or whether there are lots of

nodes with few links). Alternatively, they can be interpreted with the aid of

mathematical analysis of the network.

SNAPP
Social Networks Adapting Pedagogical Practice (SNAPP, http://www.snappvis.org/;

Bakharia and Dawson 2011) is a social network analysis tool specifically developed for

online learning contexts (Dawson 2010).

SNAPP allows teachers to track learner activity in the forums of a LMS/VLE

over time, displaying a social network diagram with the individual learners indicated by

a red circle, and the links between them as lines. SNAPP makes it easy for teachers to

identify, for instance, learners who are entirely disconnected from the network (and

hence are not fully participating), or learners who are central to the network (and hence

are key enablers of the conversation). It also helps teachers to identify the pattern of

interaction in the forum - whether, for example, it is largely teacher-centric, or is more

diffuse with stronger peer interaction. Another use is to identify self-contained groups,

or cliques, who interact with each other but not with those beyond the group. SNAPP is

11
designed to be easy to use, but this does mean that SNAPP is not as flexible and

powerful a tool for analysing social networks as the more general tools.

More advanced analysis

SNAPP lies at one end of a scale of complexity of social network analysis: it focuses on

a single forum (at any one time), and the links between the nodes are simply whether a

person has replied to another person's forum posting. It is possible to use SNA in more

complex educational contexts. For instance, Suthers and Chu (2012) used SNA to

explore the Tapped In community for educational professionals (http://tappedin.org).

Their approach, inspired by Actor-Network Theory, was much more detailed and rich,

based on an 'associogram', rather than a simple social network diagram: a complex

multidirectional mapping of the participants, the artifacts they created (e.g. messages in

chatrooms, postings in discussions, shared files), and the actions taken by the

participants on those artifacts (e.g. writing/posting, and reading). Essentially, they were

able to identify real communities purely from their online activity on the site, without

directly using information about their affiliation, geographic location, and so on. This

approach could be applied, for example, to identify communities among student

populations, which could be used to better inform decisions about group work,

placements, project or assignment topics, and so on.

Usage tracking
The data for learning analytics can come from student activity in the LMS/VLE, or in

similar online community environments. It can also come from students' use of any

application on a computer. Many tools exist to capture what a user does on a computer

over time, and these can be used as a source of data about student activity when the

learning task requires them to use something beyond the LMS/VLE.

12
 For example, Santos et al (2012) developed a dashboard for students on a

software development course at the Katholieke Universiteit Leuven. The students use a

range of applications (word processor, programming environment, and web browser) to

carry out a software development assignment, working in groups. Their activity was

logged by time-tracking software (including RescueTime, http://rescuetime.com), and

this data was presented in a dashboard. They could see, for instance, whether they were

spending more or less time on email or writing code or looking things up than their

peers, and how their web browsing compared. Feedback from students was on balance

positive, but not very strongly so.

It is not yet clear how valuable this sort of approach can be for improved student

learning. It also raises questions about what sorts of feedback and information are

helpful, and also ethical concerns around privacy and monitoring, to which this paper

will return in the Discussion.

Content analysis and semantic analysis

The examples discussed so far have concerned essentially quantitative data generated by

students. However, advances in computation, in fields such as natural language

processing and latent semantic analysis, make it possible to analyse qualitative, textual

data - not just in terms of simple frequency counts (how many times particular words

are used), but in richer, more meaningful ways.

For example, Lárusson and White (2012) have developed the Point of

Originality tool, which enables teachers to track how students develop originality in

their use of key concepts over the course of a series of writing assignments. The data in

this context is the students' writing itself, analysed using a sophisticated database of

English (WordNet, http://wordnet.princeton.edu/). The teacher types in the key words

they want to explore, selects which student's work they want to examine, and the tool

13
displays a series of coloured markers for each assignment, with bigger and 'hotter'-

coloured markers indicating more original use of the key words. Clicking on a marker

displays the writing sample in question. A trial on an introductory general course on

computing showed a strong correlation between originality scores in the Point of

Originality tool and the grades achieved for the final assessment, and also between the

originality of their writing and the quantity of their contributions online.

A more speculative example is automated feedback to students about the nature

of their online writing, with the aim of improving the quality of educational dialogue.

Several frameworks for analysing and characterising the nature of educational dialogue

have been developed, including the work of Neil Mercer and colleagues on exploratory

talk in classrooms (see e.g. Mercer and Littleton 2007). This work has been applied to

the analysis of online educational discussion (Ferguson and Buckingham Shum 2011) to

identify places where exploratory talk took place, which could for learners visiting

archived discussions to find the most useful material. These methods could be used to

analyse students' contributions to an online forum, giving them feedback about the

degree to which their online talk is exploratory (or matches other criteria for

constructive educational dialogue), and offering suggestions for ways in which they

might contribute more effectively.

The learning analytics community does not at present encompass the growing

field of automated assessment, but there are many strong parallels, and one could argue

that automated assessment is a particular form of learning analytics. In particular, any of

the many tools under development to support marking of summative written

assignments could be deployed in a formative way as cues for intervention.

14
Recommendation engines
Recommendation engines (or recommenders) are computational tools that provide

suggestions to individuals for items they may be interested in, based on analysis of the

behaviour of many users. The most famous example is Amazon's 'Customers Who

Bought This Item Also Bought' feature; Amazon also uses a recommendation engine to

suggest purchases based on a customer's purchase history, and on the ratings they have

given to other products. The same techniques can be applied in an educational context.

So, for example, a system could suggest learning resources to a student based on what

resources they have previously used or found helpful, and on other students' behaviour

and ratings.

However, it may be problematic to apply this approach in the context of a

conventional university with a set curriculum: students typically are offered relatively

little choice about the direction of their study, and so have less need for an automated

system to suggest learning resources that might be helpful. It may have more

application at higher levels of study, and perhaps the greatest potential benefit lies in

more open-ended and less formal learning contexts.

Discussion

These examples show some of the potential benefits of learning analytics. They raise a

series of implications for teachers in higher education.

The first and perhaps most obvious area is the ethics of personal data. Foucault

(1991) uses Bentham's Panopticon as a symbol of how institutions and power structures

enforce self-surveillance and control through the belief that scrutiny may occur at any

time. The nightmare vision of Big Data for individuals is that the system does not rely

on self-surveillance to enforce a disciplinary regime: all actions are captured, logged,

and analysed, so transgressions will be noted regardless of whether the jailer happens to

15
be looking in the right direction. A more positive vision of widespread disclosure is

Brin's (1999) conception of a transparent society, where surveillance by those in power

is held in check by openness and 'sousveillance' by those not in power, using

increasingly widely-available tools for capturing and analysing data (e.g. political

demonstrators streaming video live online from their phones). These radical visions of

little or no privacy, and of highly-informed and capable sousveillance, are some way

from the current situation.

In practice, surveillance is far from complete, and is circumscribed by a regime

of policies on the ethical capture and use of personal information. In almost all

developed world jurisdictions outside the US, there is comprehensive data protection

legislation that requires that issues of informed consent, data accuracy, appropriateness

of use, and access to information by the individual are addressed. Universities

themselves typically have policies on data governance, and in a research context, any

learning analytics activity will have to pass scrutiny by a body such as an Institutional

Review Board or Ethics Committee. However, learning analytics is often applied

outside an explicit research context; practitioners then have the responsibility to ensure

that their practice meets those ethical standards.

Being open about learning analytics with students can improve their perceptions

of the activity (as with Signals), but openness need not and arguably should not be

complete in learning contexts. The opportunity to learn by making mistakes in a safe

context can be a powerful learning experience, and far from all learners are happy to

have their mistakes kept on record for all time.

Students typically know and care more about their own learning situation than

even the most dedicated teacher. In numerate disciplines many students are quite

capable of making intelligent use of data about their learning. Using learning analytics,

16
they can be encouraged to take personal responsibility for their own situation - making

use of the feedback available about what they're doing, and making appropriate

decisions about support.

Teachers too have responsibilities. Educators have a professional responsibility

to use tools and methods that can improve student learning, and learning analytics offers

potentially powerful ways of doing this. A learning analytics system can reveal

information about students, which leads to new ethical challenges. If you know before

they start that a potential student is extremely unlikely to complete, should you admit

them? Or will that simply reinforce existing power structures that put them in that

position? Feeding back a negative view of a student's prognosis needs to be handled

sensitively and appropriately. Learning analytics offers the possibility of focusing

resources on where they are most needed. However, if resources are directed entirely

towards students who are in danger of failure, there is a risk of short-changing the

experience of stronger students. The experience of Signals at Purdue suggests that this

need not be the case - as described above, there was a greater improvement in high

grades than there was a reduction in fail grades.

As a field, learning analytics is data-driven, and is often atheoretical, or more

precisely, is not explicit about its theoretical basis. Several authors have sought to

ground learning analytics in theory (e.g. Clow 2012; Suthers et al 2008; Dawson 2008;

Atkisson and Wiley 2011), but this is not universal, running the risk of treating the data

that has been gathered as the data that matters. The choice of what is measured - in

learning analytics terms, the selection of metrics - is critical. If an educational system is

designed to optimise metrics that do not encompass learning, it is likely that learning

will be optimised away. For those who care about learning, the choice is to attempt total

17
resistance to the regime of metrics, or to take a more pragmatic course and insist on the

inclusion of appropriate metrics that do reflect learning.

This raises the crucial question of assessment. If assessment does not reflect and

reward those aspects of learning that are valued, a learning analytics system that

improves assessment scores will not improve learning. Concerns about the

appropriateness and reliability of assessment practices are far from new (e.g. Rowntree

1987), but analytics places a new weight and scrutiny on assessment.

Conclusion

The promise of learning analytics is the empowerment of teachers and students to

understand the wealth of data that relates to their learning. Engaging in this process is a

way of taking control of the agenda, so that the economic framing can be at least

supplemented with a concern for learning. It is not a simple or straightforward process,

and a focus on the data alone is not sufficient: to achieve institutional change, learning

analytics data need to be presented and contextualised in ways that can drive

organisational development (Macfadyen and Dawson 2012).

Learning analytics is a new technology, which affords new social actions. The

question of the nature of technology and its relationship to existing power relationships

and structures is well beyond the scope of this article, but it seems clear that educational

data and can and will be used in attempts to reinforce the status quo. Ewing (2011), in a

comprehensive demolition of the use of value-added modelling for evaluating teacher

and school performance, argues that mathematics is often misused:

"as a rhetorical weapon - an intellectual credential to convince the public than an

idea or process is "objective" and hence better than other competing ideas or
processes. This is mathematical intimidation. It is especially persuasive because so

18
 many people are awed by mathematics and yet do not understand it - a dangerous
combination."

This neatly captures the main risks of the use of analytics in difficult times. The

process of data gathering and interpretation is proceeding apace in higher education,

often driven by the demands and worldview of managers and the economic framing of

education. There is value - and not just in the economic sense - for teachers in more

information about their students. The opportunity afforded by learning analytics is for

educators to refuse to be overawed by the process, to understand the tools and

techniques, their strengths and limitations, and to use that understanding to improve

teaching and learning.

References

Arnold, Kimberly. 2010. "Signals: Applying Academic Analytics." Educause Review,

33 (1).
Arnold, Kimberly and Matthew Pistilli. 2012. "Course Signals at Purdue: Using
Learning Analytics to Increase Student Success." In Buckingham Shum,
Gasevic and Ferguson (2012), 267-270.
Atkisson, Michael and David Wiley, 2011. "Learning analytics as interpretive practice:
applying Westerman to educational intervention." In Long et al (2011), 117-121.
Bakharia, Aneesha and Shane Dawson. 2011. "SNAPP: a bird's-eye view of temporal
participant interaction." In Long et al (2011), 168-173.
Brin, David. 1999. The Transparent Society: Will Technology Force Us to Choose
Between Privacy and Freedom? New York: Basic Books.
Buckingham Shum, Simon, Dragan Gasevic and Rebecca Ferguson. 2012. Proceedings
of the 2nd International Conference on Learning Analytics and Knowledge
(LAK12), Vancouver, BC, Canada — April 29 - May 02, 2012. New York:
ACM.
Campbell, John P. 2007. Utilizing student data within the course management system to
determine undergraduate student academic success: An exploratory study. PhD
diss., Purdue University.

19
Campbell, John P., and Diana Oblinger. 2007. "Academic Analytics". Educause
Quarterly, October 2007. http://net.educause.edu/ir/library/pdf/pub6101.pdf
Campbell, John P., Peter B. De Blois, and Diana G. Oblinger. 2007. "Academic
Analytics: A New Tool for a New Era". Educause Review, 42 (4): 40-57.
CERN (European Organization for Nuclear Research), 2012. CERN Annual Report
2011: Fifty-seventh Annual Report of the European Organization for Nuclear
Research. Geneva: CERN.
http://library.web.cern.ch/library/content/ar/yellowrep/varia/annual_reports/2011
/AnnualReport2011-en.html
Clow, Doug. 2012. "The Learning Analytics Cycle: Closing the Loop Effectively". In
Buckingham Shum, Gasevic and Ferguson (2012), 134-138.
Coffield, Frank, David Moseley, Elaine Hall, and Kathryn Ecclestone. 2004. Learning
Styles and Pedagogy in Post-16 Learning: A Systematic and Critical Review.
London: Learning and Skills Research Centre.
Dawson, Shane. 2008. "A study of the relationship between student social networks and
sense of community." Educational Technology and Society, 11 (3): 224–238.
Dawson, Shane. 2010. "‘Seeing’ the learning community: An exploration of the
development of a resource for monitoring online student networking." British
Journal of Educational Technology, 41 (5): 736–752.
Deakin Crick, Ruth and Guoxing Yu. 2008. "Assessing learning dispositions: is the
Effective lifelong learning inventory valid and reliable as a measurement tool?"
Educational Research, 50 (4): 387–402.
Essa, Alfred and Hanan Ayad. 2012. "Student Success System: Risk Analytics and Data
Visualization Using Ensembles of Predictive Models". In Buckingham Shum,
Gasevic and Ferguson (2012), 158-161.
Ewing, John. 2011. "Mathematical Intimidation: Driven by the Data". Notices of the
AMS, 58 (5): 667-673.
Ferguson, Rebecca and Simon Buckingham Shum. 2011. "Learning analytics to identify
exploratory dialogue within synchronous text chat." In Long et al (2011), 99-
103.
Ferguson, Rebecca. 2012. The State of Learning Analytics in 2012: A Review and
Future Challenges. Technical Report KMI-12-01. Milton Keynes: Knowledge
Media Institute, The Open University.
http://kmi.open.ac.uk/publications/techreport/kmi-12-01

20
Foucault, Michel. 1991. Discipline and Punish: The birth of the prison. Translated by
Alan Sheridan. London: Penguin.
Johnson, L., S. Adams, and M. Cummins. 2012. The NMC Horizon Report: 2012
Higher Education Edition. Austin, TX: The New Media Consortium.
http://www.nmc.org/pdf/2012-horizon-report-HE.pdf
Johnson, L., R. Smith, H. Willis, A. Levine, and K. Haywood. 2011. The 2011 Horizon
Report. Austin, TX: The New Media Consortium.
http://net.educause.edu/ir/library/pdf/hr2011.pdf
Lárusson, Jóhann Ari and Brandon White. 2012. "Monitoring student progress through
their written "point of originality"". In Buckingham Shum, Gasevic and
Ferguson (2012), 212-221.
Long, Phil and George Siemens. 2011. "Penetrating the Fog: Analytics in Learning and
Education". Educause Review, 46 (5): 31-40.
Long, Philip, George Siemens, Gráinne Conole and Dragan Gašević (2011).
Proceedings of the 1st International Conference on Learning Analytics and
Knowledge (LAK11), Banff, AB, Canada, Feb 27 - Mar 01, 2011. New York:
ACM.
Macfadyen, Leah and Shane Dawson. 2012. "Numbers Are Not Enough. Why e-
Learning Analytics Failed to Inform an Institutional Strategic Plan." Educational
Technology and Society, 15 (3): 149–163.
Mercer, Neil and Karen Littleton. (2007). Dialogue and the development of children's
thinking: a sociocultural approach. London: Routledge.
Rowntree, Derek. 1987. Assessing Students: How Shall We Know Them? London:
Routledge.
Santos, Jose Luis, Sten Govaerts, Katrien Verbert, and Erik Duval. 2012. "Goal-
oriented visualizations of activity tracking: a case study with engineering
students". In Buckingham Shum, Gasevic and Ferguson (2012), 143-152.
Sharples, Mike, Patrick McAndrew, Martin Weller, Rebecca Ferguson, Elizabeth
FitzGerald, Tony Hirst, Yishay Mor, Mark Gaved, and Denise Whitelock. 2012.
Innovating Pedagogy 2012: Open University Innovation Report 1. Milton
Keynes: The Open University.
http://www.open.ac.uk/personalpages/mike.sharples/Reports/Innovating_Pedago
gy_report_July_2012.pdf

21
Siemens, George and Dragan Gasevic, eds. (2012). "Learning and Knowledge
Analytics." Special issue, Journal of Educational Technology and Society, 15
(3).
Simon, Stephanie and Stephanie Banchero. 2010. "Putting a Price on Professors". Wall
Street Journal, October 22, 2010.
Suthers, Dan and Kar-Hai Chu. 2012. "Multi-mediated community structure in a socio-
technical network". In Buckingham Shum, Gasevic and Ferguson (2012), 43-53.
Suthers, Daniel, Ravi Vatrapu, Richard Medina, Samuel Joseph, and Nathan Dwyer.
2008. "Beyond Threaded Discussion: Representational Guidance in
Asynchronous Collaborative Learning Environments." Computers and
Education, 50, 4, 1103-1127.
Wolff, Annika and Zdenek Zdrahal (2012). "Improving Retention by Identifying and
Supporting "At-Risk" Students." Educause Review Online,
http://www.educause.edu/ero/article/improving-retention-identifying-and-
supporting-risk-students

22