Personalised Evidence Based Practice Framework

Output 3
Personalised Evidence-Based Practice Framework
Version 1.0
The European Commission support for the production of this publication does not constitute an
endorsement of the contents, which reflects the views only of the authors, and the Commission cannot
be held responsible for any use, which may be made of the information contained therein.

SLIDEshow 2 Intellectual Output
Document Change Control
The following is the document control for revisions to this document.
Version
No
Date of Issue Author(s) Brief Description of Change
0.1 27.4.2018 Wolfgang Greller Set-up of document structure
0.2 10.6.2018 Wolfgang Greller Preamble
0.6 7.8.2018 Wolfgang Greller Finish Part 1
0.9 12.9.2018 Wolfgang Greller Finish competences in Part 2
1.0 19.9.2018 Wolfgang Greller Final draft of report and
framework plus feedback from
partners UO, VUB
Authors
Wolfgang Greller, Vienna University of Education (PHW)

Project Information
Project acronym: SLIDEshow
Project title: Showcasing Self-regulated Learning
Implemented through Data-driven Education
Agreement number: 2017-1-BE02-KA201-034791
Sub-programme or KA: Key Action 2 – Strategic Partnership for School Education
(supporting innovation)
Project website: http://www.slideshowproject.eu
Project e-mail address: slideshow.euproject@gmail.com
Twitter: @EU_SLIDEshow
Project period: From 01/11/2017 to 31/08/2020
Project manager: Prof. dr. Koen Lombaerts
Project coordinator organisation: Vrije Universiteit Brussel (VUB)
Consortium partners: University of Oulu (UO)
GO! Education of the Flemish Community (GO!)
Doukas School
Pädagogische Hochschule Wien (PHW)
Copyright: CC BY-NC-SA 4.0

Contents
PREAMBLE.......................................................................................................................................6
1. Aims.............................................................................................................................................6
2. Objectives and Scope ..................................................................................................................6
PART 1: THEORETICAL BACKGROUND...........................................................................................8
3. What is Evidence-Based Learning?..............................................................................................8
4. Learning Analytics for Evidence-Based Learning .........................................................................9
4.1 Distinguishing Factors for Learning Analytics ....................................................................10
4.2 Privacy and the GDPR........................................................................................................12
4.3 Data Evidence for Personalisation.....................................................................................12
5. Competence Requirements for Evidence-Based Learning ........................................................13
5.1 Motivation.........................................................................................................................15
6. European Qualifications Framework .........................................................................................16
PART 2: PRACTICE FRAMEWORK .................................................................................................19
7. Key Competences for Evidence-Based SRL................................................................................19
7.1 Data Literacy......................................................................................................................20
7.2 Self-Observation................................................................................................................23
7.3 Shortfall Awareness...........................................................................................................29
7.4 Data-to-Learning Transfer .................................................................................................32
7.5 Ethical and Legal Knowledge .............................................................................................36
7.6 Social Learning in Teams and Networks ............................................................................38
8 Personalised Evidence-Based Practice Framework ...................................................................40
REFERENCES..................................................................................................................................44

Table of Figures
Figure 1. Learning Analytics Framework according to Greller and Drachsler (2012).............................11
Figure 2. Social network diagram of a Moodle course using the SNAPP tool........................................13
Figure 3. EQF descriptor grid (European Commission, 2005)................................................................17
Figure 4. Process competences.............................................................................................................19
Figure 5. Competence requirements and competence areas ...............................................................20
Figure 6. Example of a social media progression metric .......................................................................25
Figure 7. Heat map of the most difficult questions (Greller et al., 2014) ..............................................26
Figure 8. Areas of self-reaction in SRL ...................................................................................................35
Figure 9. DigCompEdu Progression Model (Redecker, 2017)................................................................40

PREAMBLE
1. Aims
This report aims to outline a practice framework for (self-)assessment of teachers’ abilities to support
self-regulated learning (SRL) and to personalise instruction to the individual needs of students in their
classes. It is particularly looking at the role that learning data plays as evidence for learning, progress,
and achievement. Parallel to the SRL practice framework (IO2)1
, this project output (IO3) will clearly
describe the necessary data competences with the goal of equipping teachers with the skills and
hands-on-experiences with using learner data for the support of their own personal SRL as well as
their students’ SRL.
2. Objectives and Scope
While the IO2 report takes a holistic view of teacher competences for SRL, this current paper
investigates a sub-set of special abilities teachers ought to possess when dealing with evidence in form
of learner data as produced by learning applications or management systems and platforms. We will
outline the necessary competences to interpret and use these data to make informed decisions about
their instructional practice and personalise their support.
To be clear, evidence-based learning can happen without the use of digital data, for example through
direct personal observation or via scientific research (see also the findings of IO1)1
. Here, however, we
shall focus on evidence produced by tracking learner behaviours in digital environments, such as the
tMAIL app, and by processing various sets of logging data or user input data into digested information
using particularly designed Learning Analytics algorithms.
Like in output IO2, we shall refer to the European Qualifications Framework (EQF), which provides a
satisfactory scaffold in terms of levels of independence. With its eight levels, the EQF allows identifying
the readiness of individuals to act by themselves and responsibly.
This report then brings together a number of theoretical themes and merges them into a shared
framework for teaching and learning practice of SRL. The first and core theory involved is Learning
Analytics, i.e. the use of big data in education for the sake of improving learning (cf. Long & Siemens,
2011). Secondly, we shall apply a particular descriptive framework, the EQF, to identify levels of
competences (European Commission, 2005). Thirdly, a perspective based on the research carried out
for output 1 (cf. IO1) will be used, which looks at cognition, motivation, behaviour and emotions, as
impact areas for SRL. Finally, we shall refer to aspects of personalisation and opportunities for fine-
tuning the learning experiences of students and teachers.
1 All project output documents can be found at the SLIDEshow website: https://www.slideshowproject.eu/intellectual-
outputs

There are natural confinements in our investigation coming from the focus and mission to assist
teachers and teacher educators to develop their SRL pedagogies. We will, therefore, not evaluate
individual Learning Analytics products (except for tMAIL as an illustrative example) that are available
on the market or in use by different institutions. We will not analyse or critique how data is gathered
and what indicators or algorithms are applied when processing that data. Instead, we take the
application at face value and treat it as a black box, where users are presented an output that they
work with in their respective educational contexts.
Apart from the generic theory-informed part, later in the project, a case study will be presented
exemplifying what this means for educational practice and how it can be used by individuals to assess
their progress in applying SRL by studying the tMAIL app and monitoring platform. In this way, we are
able to illustrate practical applications of the devised framework.

PART 1: THEORETICAL BACKGROUND
3. What is Evidence-Based Learning?
The underlying question behind the drive for evidence-based learning and assessment is: how do we
know that someone learns? The guiding frame for this is the benchmark that the learner is set to
achieve. This could involve arriving at the specified learning outcomes at a particular point in time
(short-term achievement), or, acquiring a certain mastery or qualification (long-term success).
In a more formal learning environment, involving a teacher and a curriculum, this above question can
be turned into: how do I as a teacher know that my pedagogic strategies and learning designs have a
positive effect on my students’ learning and how can I improve it? As is apparent from the form of the
question, this requires the adoption of a reflective teaching practice, where teachers continuously
evaluate and adapt their pedagogic techniques and learning designs (Lockyer, Heathcote, & Dawson,
2013).
“If an educational strategy is evidence-based, data-based, or research-based, educators compile,
analyze, and use objective evidence to inform the design of an academic program or guide the modifi-
cation of instructional techniques” (taken from: https://www.edglossary.org/evidence-based/).
There is also the more philosophic view that mere ‘being’ consists of learning: “you cannot not learn!“,
but we will exclude this here as it is not a suitable approach to further SRL in schools, where achieve-
ments and qualifications are given by merit of learning evidence with regard to commonly agreed
levels and standards.
The importance of evidence in learning has entered institutional thinking on a bigger scale in the early
2000s, at a time when independent learning and technology enhanced learning became institution-
alised. This followed an earlier expansion in online remote delivery of courses through the internet. It
also mirrored a wider demand for self-governed life-long learning, especially in the workplace, where
this was seen as an economic replacement for traditional CPD courses by hired trainers. Under the
label “learning organisation”, companies were following the developments of a knowledge economy.
In this context, employees were expected to develop their own knowledge and competences “just in
time” and by remote means. To ascertain transparency for the company, evidence of such learning
was required. Similarly, students in work placements or project-based scenarios were held to keep
learning diaries or (e-)portfolios to document their progress and to make it assessable as part of the
evidence provided for marking purposes (cf. Nückles, Schwonke, Berthold, & Renkl, 2004; Prinsloo,
Slade, & Galpin, 2011).
“A widely used adjective in education, evidence-based refers to any concept or strategy that is derived
from or informed by objective evidence — most commonly, educational research or metrics of school,
teacher, and student performance.” (https://www.edglossary.org/evidence-based)

4. Learning Analytics for Evidence-Based Learning
Learning Analytics is a relatively new term for using “Big Data” in education. As far as definitions go, it
is described as “the application of analytic techniques to analyse educational data, including data
about learner and teacher activities, to identify patterns of behaviour and provide actionable infor-
mation to improve learning and learning-related activities” (Van Harmelen & Workman, 2012).
The possibilities for using educational datasets as (objective) evidence are far reaching and reveal
great potential for transforming education and learning in the longer term (New Media Consortium,
2011). Learning Analytics, therefore, offers insights into learning processes as well as delivery and
support processes on various levels of formal learning. For this reason, also institutional management
and education authorities are very much interested in this emergent field of research and
development. In their framework for Learning Analytics, Greller and Drachsler (2012), therefore,
emphasise the stakeholder dimension as one to take into consideration when designing Learning
Analytics set-ups, where decision makers are only one of the groups of beneficiaries. Other
researchers, like Long and Siemens (2011), tend to distinguish between “Learning Analytics” and
“Academic Analytics”, the latter mainly representing the institutional interests in the quest for higher
cost efficiency and effectiveness and for funding and reporting reasons, among other things.
For our purposes, we will disregard the institutional and systemic components of Learning Analytics
(the Academic Analytics part) and focus on the evidence-based learning aspect, i.e. the processes that
support the acquisition of knowledge and competences and the assessment of it.
A recent meta-analysis of scientific articles (Wong, 2017) revealed that common drivers for institu-
tions to adopt Learning Analytics from the learner perspective were to understand students’ learning
behaviours, provide personalised assistance, timely feedback and intervention for students, and to
explore what benefits Learning Analytics can bring to the institution, staff, and students. Similarly, a
recent survey concluded that students appeared to show strong interest in receiving regular updates
on their learning with a preference over a Learning Analytics service that facilitates independent
learning, whereas teachers had an interest in using Learning Analytics to enable personalised support
and to gain an insight into the learners’ progression (Tsai, Scheffel, & Gasevic, 2018).
An earlier survey by Drachsler and Greller (2012) revealed a high level of interest among stakeholders
to more timely information about the learning progress, better insight by institutions on what’s in
each course, and stimulation of reflection by learners on their progress. It also showed the anticipation
of experts that the learner-teacher relationship would be the most affected by applying Learning
Analytics, though they did not analyse in what way this would materialise.
Self-regulated learning (SRL) involves three features: usage of learning strategies, responsiveness to
self-oriented feedback about learning effectiveness, and motivation. It is a cycle of forethought,
performance and self-reflection (Schmitz et al., 2018). In this regard, teachers as well as students
express the need to personalise feedback on learning activities in order to increase the quality of the
learning process. The authors contend that collecting and analysing student data can improve the

quality of the feedback as well as offering the opportunity of using on-demand indicators for evidence-
informed decisions (ibid.).
In a surprisingly early study by Mandinach and colleagues (2006), an interesting observation is that
teachers in their decision making often lack systematicity, from student-to-student, class-to-class, and
year-to-year. Data from learning evidence and the use of Learning Analytics offers a systematic
approach to look for patterns in data at different levels of aggregation, e.g. classroom-wide patterns.
4.1 Distinguishing Factors for Learning Analytics
Not every empirical survey or statistical analysis can be considered Learning Analytics (cf. Greller,
2015; Long & Siemens, 2011). From the above definitions and the scientific literature, we can derive
some characteristic features that encircle the common understanding of Learning Analytics:
SCALE
(1) Learning Analytics uses digital (or digitised) datasets of large proportions (big data). This is a
distinguishing feature from traditional empirical research in the way that it does not use a pre-
selected sample of users as a basis for generalisations. Instead, it is applied across an entire
learner population. The insights, however, can be broken down to individual level and, thus,
provide valuable information for personalisation efforts and SRL.
SPEED
(2) Learning Analytics is automated and runs in real time. Learning Analytics data collection and
algorithms run largely in the background by tracking learner behaviour in digital environ-
ments. On some occasions, nevertheless, direct user input may be required.
(3) Learning Analytics is continuous. Again, this makes it different to traditional empirical or
statistical approaches, which are typically one-off data collections, or, in some cases,
repetitive/cyclical data gathering at specified intervals. In Learning Analytics, continuous data
gathering is coupled with an on-demand and just-in-time approach, where students or other
data clients can spontaneously interrogate the system to find their “position”.
SENSORS
(4) In addition to traditional log data and clickstreams, Learning Analytics can use a large variety
of observational input mechanisms, such as eye tracking, face recognition, GPS or movement
sensors. Its strengths and promises lie in the combinations of different data sources.
Especially point (2) above – the automated real-time feature of Learning Analytics – can be considered
essential for personalised evidence-based learning, as it consists mostly of unbiased data from
clickstream observation (e.g. times a user logged in or accessed a resource) rather than opinionated
inputs to empirical surveys. Objective evidence of learning paths taken by students are recorded from
learner decisions on following certain activities and available options and resources.
In their framework for Learning Analytics, Greller and Drachsler (2012) distinguish six critical
dimensions for Learning Analytics (see figure 1 below). Most notable for this report and for its
application to SRL is the juxtaposition of the “objectives” dimension between (a) reflection and (b)

prediction. In this report, we see Learning Analytics as a reflection amplifier and part of objective
learning evidence that triggers a learner and/or teacher response. This very strongly connects to the
self-evaluation strategy of the Zimmerman model for SRL (Zimmerman, 1989). It is markedly different
to the predictive modelling approaches of Learning Analytics that apply artificial intelligence to
forecast learner situations and positioning in order to anticipate a required action automatically.
Reflection is seen as the critical self-evaluation of a learner as indicated by their own datasets in order
to obtain self-knowledge (Greller & Drachsler, 2012).
Self-observation and reacting to one’s own performance data is sometimes also called the “Quantified
Self” (Wolf, 2009). This implies analysing, benchmarking, interpreting, and acting on your own
performance data as visualised by the respective application. In this SRL scenario, learners take full
charge of their own progress and success. Its usage is not limited to cognitive aspects of learning but
often connected to wanted behavioural changes, supported by a wide variety of dashboards and
mobile apps, e.g. to lose weight or to quit smoking, etc (cf. Didžiokaitė, Saukko, & Greiffenhagen, 2018;
Duval, 2011). Apart from supporting reflection on an individual level, Learning Analytics can offer
personalised information on the progress of the independent learner (Govaerts, Verbert, Klerkx, &
Duval, 2010). However, in social learning environments, reflection can also consist of comparing your
performance to peers and their datasets. From this approach, a range of competitive scenarios are
conceivable that allow for gamification as a learning strategy (e.g. rankings, time to finish, levels, etc.).
Figure 1. Learning Analytics Framework according to Greller and Drachsler (2012)

In classroom settings or guided group learning, a teacher can use the results not only to stimulate and
support student reflection and motivation, but also to reflect themselves on their pedagogic designs.
This may include a utilisation analysis of provided learning materials, learning activities, sequences or
support instruments (e.g. forums, software tools, etc.). In addition, other, non-pedagogic factors could
be highlighted through data, such as the time allowed for certain tasks, or the provision of specific
learning spaces.
4.2 Privacy and the GDPR
It cannot go unmentioned that the utilisation of student data for Learning Analytics has raised much
concern about privacy and data protection that led to an abundance of literature on this topic (e.g.
Drachsler & Greller, 2016; Hoel & Chen, 2016; Ifenthaler & Schumacher, 2016; Prinsloo & Slade, 2016;
Steiner, Kickmeier-Rust, & Albert, 2016). In response to this demand for transparency and the
protection of personal data, the JISC in the UK has developed a “Code of Practice for Learning
Analytics” (Sclater, 2014; Sclater & Bailey, 2015) that sets out how educational datasets can be used
by institutions reassuringly and legally. In 2014, the Open University of the UK laid out their substantial
policy for ethical use of student data and using information to support student learning (Open
University, 2014).
The coming into force of the new European General Data Protection Regulation (GDPR) on 25th
May
2018 has caused intensive debates within institutions and the scientific Learning Analytics community
in the run up (cf. Hildebrandt, 2017; Tsai & Gasevic, 2017). However, the use of learner data to support
learning is relatively straightforward when the provider observes some basic principles. Apart from
general data security (i.e. storage in a save place), informed consent is required for sensitive data and
for direct interventions on the basis of the analytics (JISC, 2017). It should be transparent to the data
subject (the learner) who has access to their data and what it is used for.
For the assisted self-regulated learner, the collection of data and its convergence into analytic infor-
mation requires the learner’s consent. In SLIDEshow and the preceding tMAIL project, this consent is
given at the beginning of using the app. Teachers who pursue the SRL training are able to seek
assistance from a teacher educator (or anyone else they choose) by inviting them to their dataset.
This consent can be withdrawn at any point in time when no longer required or desired.
4.3 Data Evidence for Personalisation
The aspect that feedback to one’s learning can be retrieved at any point in time (cf. 4.1 point (2) above)
allows learners and teachers to react to situations quickly and more effectively than with summative
assessments. Especially in experiential learning situations (e.g. work practice, projects) this is
extremely useful as it enables individual re-adjustments and course corrections during the run-time
of the learning experience, which cannot be done once the placement is over. Learning Analytics,
therefore, provides an opportunity to intervene on a personal level during the run-time of a course
(Schmitz, Van Limbeek, Greller, Sloep, & Drachsler, 2017).

But Learning Analytics holds even more potential for personalising SRL in that students are enabled to
change their learning behaviour based on the feedback from data evidence (Schmitz et al., 2018).
Equally, recommender systems informed by educational data can lead to more personal options for
the learners, allowing for adaptive and personalised learning paths (cf. Clow, Cross, Ferguson, &
Rienties, 2014, with links to more literature).
Zimmerman (1989) describes the three self-regulatory processes for SRL: self-observation, self-
judgement, and self-reactions. Evidence from learning data can support the self-observation phase,
and, in combination with the required skill-set, improve the opportunities and effectiveness of self-
judgement and self-reactions on the micro-level. In an assisted learning setting (such as a course or
classroom), the teacher or mentor can provide personalised learning support using their pedagogic
knowledge and experience.
5. Competence Requirements for Evidence-Based Learning
In their framework for Learning Analytics, Greller and Drachsler (2012) clearly identify the need for a
higher order set of competences for sense-making and utilisation of computer generated evidence in
educational practice. When confronted with the results of an algorithmic analysis of observational
tracking data, the question arises whether learners or teachers possess the competences to interpret
and act upon the output in a pedagogically sensible way.
Figure 2. Social network diagram of a Moodle course using the SNAPP tool
Visualisations can assist the presentation and comprehension of the data, but often appear complex
in terms of the units of measurement or its relationship to the learning design and progression. For
example, a Social Network Analysis diagram (cf. figure 2 above) mirrors the social arrangements and
context in which interactions occur. These may be part of a deliberate design (e.g. role-plays, teacher
centred activity, group work), or they may reflect interactions between individuals independent of the
learning design (self-organisation, random). Schmitz et al. (2017) see understanding learning dash-
boards as a cognitive challenge that also correlates with student satisfaction.
Additional to interpretation skills, Greller and Drachsler (ibid.) mention critical thinking as essential
for understanding learning evidence. They convincingly argue that it is necessary for the data client

and pedagogic agent to know what data is presented in the results and what data is absent. At the
same time, they give the warning that visualisations may obfuscate the full pedagogic reality behind
enticing graphics. In its most recent Digital Education Action Plan, the European Commission (2018b)
goes further on this, by stating:
“Everyday exposure to digital data driven largely by inscrutable algorithms creates clear risks and
requires more than ever critical thinking and the ability to engage positively and competently in the
digital environment.” (European Commission, 2018b)
The following list comprises six skill sets that are critical for reaping benefits from evidence of learning.
In Part 2 of this report, we will go into detail about each individual competence set that we identified
from the literature for SRL as part of the SLIDEshow project.
 Data literacy
General understanding of what data is, how it is generated, which different types of data
exist and are utilised, how it is managed, stored and processed. Ability to evaluate learner
data in connection with the learning objectives and individual situation. Reading and
interpreting visual or textual representations.
 Self-observation
Especially for SRL, the skills to observe and analyse one’s own actions and putting them into
relationship with others is of great importance. What data do I need to observe for a
particular learning goal? Using observational evidence as reflection amplifier.
 Data-to-learning transfer (applying the right action) – intervention design
Deriving the best intervention from a particular learning evidence requires not only
pedagogy, but also transferal skills from learning evidence to learning strategy or alternative
support actions. Finding and articulating relevant indicators and applying appropriate
weighting is an additional requirement to determine progress.
 Legal knowledge
Learning evidence in formal learning is part of a contract between a learner and an
institution. Teachers need to understand what this contract entails, what the ethical and
privacy implications are, such as the obligation to act, and what can be done with the
information gathered from data.
 Shortfall awareness (holistic learner perspective)
This connects to the critical thinking skills mentioned above. It is not to take data
information at surface level but to consider the whole learner and their situation. There is
also a requirement to critically question algorithms and assumptions, as well as the derived
recommendations. Being able to identify bias in data. Recognising when students are trying
to game the system.

 Learning in groups, teams and networks
Evaluating the relative and absolute measurement (compared to peers, resp. compared to
benchmark goals or starting point) of a learning process can provide SRL with efficacy and
confidence, hence stimulating continued effort. Team skills and gaming elements can
provide additional motivation. It includes awareness of roles in collaborative settings.
The above list of competences is not exhaustive. Other key competences such as numeracy or
information literacy can help further in providing a basis for exploiting learning data for SRL. Still, the
six items mentioned above are in many ways unique to evidence-based practices of SRL.
Note, that we do not include mastery of technology in general, or the mastery of particular
applications in the list of meta-cognitive competences above, although this may enhance the
acceptability, affinity and motivation to technology-enhanced SRL. We consider this skill set a precon-
dition for the general adoption of educational data in self-regulated learning.
5.1 Motivation
In addition to the competence items above, learning evidence data can provide some measurable
motivational indicators to the self-regulated learner or their teacher. On a short-term perspective,
feedback from observational data may stimulate engagement and potentially raise satisfaction
levels through bite-size successes and achievements. Measurements on e.g. learning intervals
(how often did a learner participate in a course) or distribution across time (cf. Khalil & Ebner,
2016) indicate the level of motivation and engagement as well as the type of interaction (active
or passive participation). The learning design can, if appropriate, enhance this factor by providing
segmentation of the learning process into “levels” or achievements. Providing relevant feedback,
such as “Congratulations, you have completed level 1, now proceed to level 2” may boost
motivation and improve retention (ibid.).
Aspects of evidence-based motivation that appear relevant to us could be the following:
 Attention and Interest
Information from tracking data that show student engagement and active participation at
regular interaction times. Amount of interactions, karma, strong network links, use of (extra)
learning opportunities.
 Perception of personal success (confidence)
Level of self-efficacy in SRL (tMAIL tests), persistence, vertical advance and progress,
assessment results. Practical application of knowledge and skills in real life (learner diary
documentation and self-assessment).
 Goal orientation
Learning path towards desired or defined outcomes. Managing own expectations. Planning
and applying a learning strategy.

 Social participation and gratification
Helping others, karma. Self-gratification and motivation of or by others.
6. European Qualifications Framework
The European Qualifications Framework for lifelong learning (EQF) was proposed in 2006 and adopted
by the European Parliament and Council in 2007 as an overarching ‘meta’-framework to facilitate
comparison of qualifications and reference levels (European Commission, 2008). Since its inception,
the EQF has risen to become a common tool of reference for European education and training
(CEDEFOP, 2018). It encompasses a set of eight levels of mastery based on knowledge and competence
descriptors (cf. figure 3 below) representing learning outcomes.
There are many competing frameworks in the European member states and beyond, ranging from
defining core competences to specific sets of skills related to individual professions, as well as
interdisciplinary and transferable skills (e.g. interpersonal competences). At a higher, political level,
national qualifications frameworks (NQF) allow the assessment and comparison of learning
achievements within the national education system and/or sectoral labour market. The EQF functions
as a reference instrument to link the different NQFs together and provide a meaningful translation of
qualifications from other countries. What is special for the EQF is that its descriptors cover any and all
learning outcomes from formal, non-formal and informal learning (European Commission, 2008). This
makes it specifically applicable and useful for our personalised evidence-based practice framework.
The descriptors represented in figure 3 (above) aim to distinguish between the level below and the
level above. Higher levels express greater complexity in knowledge, skills and understanding and put
greater demands on the learner (European Commission, 2008). The levels connect to each other in a
continuous way in that each level implicitly includes the levels below, therefore, showing progress in
dimensions of change. Hence, reaching a certain level can be interpreted as a structured progression
in the learning process by being able to demonstrate acquired results and mastery.
Though the EQF is not intended to function as a competences framework but rather as a qualifications
framework, it is still well suited for our purpose of classifying different grades of readiness and
expertise in SRL as its focus is on all forms of learning outcomes. The EQF’s neutrality with respect to
what should be regarded as key competences makes it a versatile and generic tool for all possible
contexts and domains of self-regulated and other learning.

Figure 3. EQF descriptor grid (European Commission, 2005)
The Council of the European Union’s recommendations on the EQF (Council, 2017) contribute to
modernising education and training systems and to increasing the employability, mobility and social
integration of workers and learners. They also aim at better linking formal, non-formal and informal
learning and supporting the validation of learning outcomes acquired in different settings.
Frameworks like the EQF are aiding the validation of learning acquired outside formal education, e.g.
at work and during leisure time. The recent CoEU recommendations encourage further integration of
comprehensive frameworks for all types and levels (ibid.).

Despite some early criticism about the EQF’s neutrality (cf. Bohlinger, 2008; Cort, 2010), the
framework is well accepted by now and its application is suited as the basis for an evidence-based
practice framework in SLIDEshow for two main reasons:
(1) The EQF is considered a qualifications framework for making qualifications and certificates
from different national frameworks transparent and comparable along the lines of the given
descriptors. However, since the EQF is oriented towards learning outcomes, and because the
awarding of qualifications always entails reaching a certain level of knowledge and compet-
ence (Bohlinger, 2008), we are able to turn the perspective around and see the EQF as a table
for mastery of relevant competences to achieve a (hypothetical) SRL qualification.
(2) The EQF descriptors favour autonomy and independence in the acquisition and execution of
competences. By starting from a dependent and supervised instructional setting to becoming
an autonomous self-directed and responsible individual supports the cause of SLIDEshow. It
sets a suitable scaffolding for increased learner self-organisation, including social aspects like
becoming a leader in the field. In our case this implies being a role model for other teachers
as well as a guide to primary and secondary students.
“In the context of the EQF, responsibility and autonomy is described as the ability of the learner to
apply knowledge and skills autonomously and with responsibility.” (European Commission, 2005)
By adopting the EQF, we can integrate empirical evidence from research with an important policy
framework and make it relevant for educational practice on a European scale. Connecting SRL
instruction to the EQF should help teachers, schools, and teacher educators integrate SRL in a more
structured way.

PART 2: PRACTICE FRAMEWORK
7. Key Competences for Evidence-Based SRL
In Part 1 of this report, we briefly introduced the set of six competence areas for evidence-based SRL
against a theoretical background. In Part 2, we want to elaborate and detail the underlying
expectations and descriptors. It is helpful for its applicability in educational practice to embed the
proposed framework in a situated process of Learning Analytics, starting with the technology itself.
This is followed in our case by the pedagogic actions of a human agent. For the conceptualisation of
our personalised evidence-based practice framework (O3), we see the following process flow as
instrumental for organising relevant competence development (see figure 4):
Figure 4. Process competences
Figure 4 above captures the competence requirements relevant to the exploitation of Learning
Analytics for learning support purposes. A different skill set maybe required for the mastery of the
technology and its applications (e.g. Moodle analytics, tMAIL dashboard, mobile apps, wearable tech,
Internet of Things, etc.). The process connected here to the competences for personalised evidence-
based practice follows the challenges posed by understanding the role of the technology and its
output, interpreting it in a pedagogical way, acting on behalf of the learner (or self), and defining or
assessing the learning outcomes: understand – interpret – act – define/validate.
When mapping the above requirements for educational practice from figure 4 to our skill set, the
following picture emerges (cf. figure 5 below):

Figure 5. Competence requirements and competence areas
The diagram in figure 5 groups the competence areas into three pairs for each of the requirements.
The competence areas are bundled by conceptual relevance, but this does not mean they only apply
to this particular field of application. Instead, there is substantial overlap and interrelationship
between them. The define/validate requirement forms the background as this relates strongly to the
motivation, intention and purpose of collecting evidence data, i.e. to the intended learning outcomes
and the need to articulate some form of success or benchmark to reach, which is then validated
against (cf. Matthews, 2016).
Next, we look at the individual skill sets extracted from the relevant scientific literature and policy
documents:
7.1 Data Literacy
Under the heading “A new skills agenda for Europe” (European Commission, 2016), data literacy has
recently become an integral part of the European Commission’s policy and is included in its digital
competence framework DigComp for all citizens (Ferrari, 2013; Vuorikari et al., 2016; Carretero et al.,
2017). It extended the earlier digital competence policy from 2006, recognising that data literacy has
become more and more necessary, thanks to new information visualisation tools and larger amounts
of data available. This acknowledgement originates from the recent massive expansion in digital data
(big data).
“Digital competence involves the confident, critical and responsible use of, and engagement with,
digital technologies for learning, at work, and for participation in society. It includes information and
data literacy, communication and collaboration, digital content creation (including programming),
safety (including digital well-being and competences related to cybersecurity), and problem solving.”
(European Commission, 2018a)

Additionally, a specialised “digital competence for educators” framework, DigCompEdu, has been
released, which includes references to Learning Analytics and data literacy (Redecker, 2017).
“Furthermore, the use of digital technologies in education, whether for assessment, learning,
administrative or other purposes, results in a wide range of data being available on each individual
learner’s learning behaviour. Analysing and interpreting this data and using it to help make decisions
is becoming more and more important – complemented by the analysis of conventional evidence on
learner behaviour.” (Redecker, 2017)
The newest update of the DigComp framework (DigComp 2.1) now introduced eight proficiency levels
for each of the 21 competences, using the eight levels of the EQF (Carretero et al., 2017). The
importance of data literacy has again been reinforced on highest political level in the EC’s “Digital
Education Action Plan” (European Commission, 2018b), recognising that data is vital for education and
training, and by making better data analysis and foresight a special priority (ibid.).
We will break down the individual competence areas, competences and descriptors of DigComp into
items specifically relevant for our goal of a personalised evidence-based SRL framework. This will lead
to some areas of DigComp being deprioritised (e.g. communication competences or content-creation
skills) while in other areas we try to fill existent gaps related to specific learning data knowledge,
competences, and attitudes. In parallel to the SLIDEshow project, another EU project called
Learn2Analyze2
is currently also looking at educational data competences. Though its focus is firmly
on instructional designers and online tutors, there are many congruences in the identified skill areas.
Ridsdale and colleagues (2015) state that a solid foundation of data literacy is nowadays integral to
building domain-specific knowledge. They developed a “data literacy competencies matrix”, oriented
towards post-secondary graduates and organised into five areas: data, collection, management,
evaluation, and application. Particularly useful is the best practices for data literacy education that the
authors investigated and analysed. While their practical approach is very well thought out, not all the
areas they consider seem relevant to everyday practice in SRL. Especially, as data collection and
management are typically not done by users themselves, but via some app or technology, which then
provides feedback to the learners and their teachers. Hence, we do not expect the practitioner in the
classroom to operate at the level of a data scientist or statistician. In our framework further below,
we, therefore, filter and synthesise the approach to suit our objectives of personalised evidence-based
decision making for learners in an SRL environment.
Quoting earlier forerunners, Mandinach et al. (2006) already identified challenges in applying data
evidence in teaching and learning, like the quality and interpretation of data, and the relationship
between data and instructional practices. Although much of their investigation is focused on the
understanding of statistical concepts and test-driven performance, there are many useful and
practically relevant points in their work on a data-driven decision-making framework, which we will
2 http://www.learn2analyze.eu/

incorporate into our proposal below. One such item is the teacher’s ability to see connections
between students’ scores in Learning Analytics and the actual classroom instruction, and then act on
them. The data literacy competences we extrapolate in this report, are based on the principle
understanding “that information is data that is given meaning when connected to a context” (Light,
Wexler, & Heinze, 2004). It is the level of competence of educators that determines this link as well
as the derived action.
Another very useful source of practical information on data literacy is the Data Journalism Handbook
(European Journalism Centre & Open Knowledge Foundation, 2012), published as a freely available
web-book with hundreds of contributors from the media sector. A special chapter has been dedicated
to becoming data literate in three simple steps, which mainly concerns itself with critical questioning
of sources and the understanding of data collection and comparison3
. We can learn some practical
lessons from this resource as they are equally applicable to education.
Matthews (2016) conceptualises data literacy in a way that further suits our context: where data
already exists, skills to source and assess the credibility are required with a need for some under-
standing of the method of production. He also points to the link between analysis and a sub-literacy
of statistics to test out hypotheses and explore relationships and sees “post-analytics competences”
to include sensible archiving of data for reuse and considerations on access and security (ibid.).
Summary
As the amount and importance of data and data-based decision-making continues to rise, data literacy
is a competence area that has been described as a future important core requirement for participation
in society. Approaches to describe data literacy are dating back to even the last century but have just
recently been massively expanded on all levels of education, including governmental policy makers.
Although the wider skill set is naturally very useful to have, for the specific utilisation and adoption of
evidence-based practice in SRL and SRL education, a mere subset and basic understanding is required
in order to reap benefits from learning evidence.
In our approach, we take data as a provided asset. Hence, there is no need for specific competences
in designing data collection methods, data mapping and cleansing, metadata taxonomies, semantic
web sharing protocols, or interlinking and integration of datasets. Nor is there a particular need for
dealing directly with data storage and security concerns. Nevertheless, a superficial understanding
how these things are typically done, at what cost (financial cost or “design sacrifices”), what
information the datasets are therefore able to carry, what possible misinformation or bias might be
contained, and what security risks may be looming is a solid basis for proper exploitation of
educational data. Similarly, awareness of the context in which data is produced and processed is
paramount to the applicability of the information contained.
As we stated earlier in the introduction to Part 2 (cf. figure 5), the skills we propose for SRL are based
on possession or development of an understanding of data and its environment. We, therefore,
3 An updated version of the handbook is envisaged for autumn 2018.

include data literacy competences only on the lower levels of the EQF framework, comprising mainly
of an awareness on what types of data there is, where it comes from, and what can and cannot be
expected of it. Some limited understanding of statistics that goes beyond mere averages is also part
of this competences area, as is the understanding of the connection between datasets and relevant
indicators that extrapolate information on aspects of the learning process, including the motivation
of self-regulated learners.
For SRL, an understanding of how figures derived from data can support particular personal objectives
is required. For example, if you are using a fitness app to improve your health, a step counter isn’t as
useful as a heart rate monitor or a time measurement for distance (pace / speed). Connecting the data
coming from the heart rate monitor with the speed over time can indicate improvements in fitness,
which through reflection on other contextual elements (running before breakfast or after; pauses in
training over a few days, etc.), can yield insights for further advances. So, understanding the
instrument (step counter vs. fitness watch) and the information it produces (units, intervals), analysing
and weighting the output (heart rate should decrease while speed increases) leads to data becoming
a reflection amplifier (why? for what reasons are things the way they look on data?). A feeling for the
information being at the right level unit size (m/h, or s/km?) and measuring interval will also provide
better feedback.
One of the first abilities to develop, therefore, is to be sure you are using the right data and the most
relevant indicators for your goals. For example, some fitness watches tell you how many calories
you’ve burnt while running or on a treadmill. However, this type of data isn’t the best information for
improving fitness or for achieving a time limit for a marathon competition, while it may well help if
the goal was to lose weight. Similarly, looking just at the number of logins in the learning management
system (LMS) might not tell you much about a student’s chances to pass an exam. But with additional
information, e.g. recency or frequency of login, this data may become meaningful.
A basic understanding what range and dispersion (standard deviation) mean in statistics and how
outliers in the data (e.g. “bad days” in our fitness example) can be interpreted and contextualised is a
competence that leads to more self-responsibility for own learning arrangements, i.e. adaptive
behaviours (cf. Winne & Hadwin, 2008).
7.2 Self-Observation
In IO2, we described self-regulated learners as learners who set themselves goals, plan their actions
to pursue these goals, monitor their learning, and, finally, evaluate their learning process (cf.
Zimmerman, 2002). Self-observation, self-monitoring and self-recording are, therefore, essential and
indispensable parts of SRL (Zimmerman, 1989). Because learning data collected from tracking actual
behaviours can be described as authentic and “objective” evidence, observations based on recorded
activity streams lend themselves particularly well for self-observation and self-evaluation, thus
stimulating self-reflection. One of the underlying principles is that the same activities over the same
amount of time will always lead to the same results in the performance data. This algorithmic principle
is only influenced by varying conditions and the variations of possible interpretations (considerations).
To these we will direct our competence descriptors.

IO24
elaborates the generic practitioner competences required for the cyclical triad of the SRL model
proposed by Zimmerman (2002). The model consists of three distinct phases: forethought,
performance and self-reflection (ibid.). As was said in the introduction above, the focus of this report
(IO3) is on the learning evidence from data information. Of the above-mentioned triad, it is the
performance phase that is most affected by evidence-based learning, since this is where most of the
data is produced and documented. The self-reflection phase which influences the later selection or
adaptation of learning strategies (to which cf. further below 7.4) is affected by the performance in the
way it becomes transparent through data analysis and visualisations. Mapped onto Zimmerman’s
terms, our evidence-based practice framework covers the self-observation, self-judgement and self-
reactions areas that define SRL. Intrinsic and extrinsic motivation is also seen as integral for SRL and
data evidence can stimulate emotions and motivation as well as reflection (cf. Derick, et al. 2017). This
leads to better self-efficacy which is generally said to positively influence future learning strategies
(Zimmerman, 1989).
Under the header of self-observation competences, we subsume two self-regulatory processes: self-
observation, and self-judgement (Zimmerman, 1989), as well as interpretation, sensing and sense-
making processes. Self-observation builds on previous planning during the forethought phase (“are
things going according to plan?”). While the process of planning SRL contains a wider variety of
activities, as described in IO2, in our narrower, data-based view, we called it “defining success”,
because we are dealing with measurable outcomes that can be (objectively) assessed and bench-
marked. In this way, self-observation leads to feedback loops from learning evidence that eventually
inform new forethought and planning activities in an iterative process. “Validating success”, therefore,
refers to the adaptation of the “defining success” conditions, based on the results of an analytics
process. It is nothing else then reviewing the learning objectives in the light of the evidence. This may
include an evaluation and reassessment of “shifting goalposts” in a rapidly changing learning
environment (“am I still learning the right things?”). Or it may result in a “mission accomplished” state-
ment once the learning outcomes have been reached satisfactorily.
Zimmerman (2002) says that SRL occurs to the extent a student is able to determine and regulate
behaviour as well as the environment in which it is embedded. Students, therefore, need to be
sensitive to changes in these variables and adapt their behaviour and strategies. Based on this, we
assume that for evidence-based learning applied in a personal and self-determined way, feedback
from data can be an extremely helpful guidance instrument in the learning process but requires the
feeling of “being in control”. Teachers, therefore, are challenged with allowing students the flexibility
to organise their learning strategies, to monitor their achievements, and to develop their beliefs.
Although the scientific literature mentions perceived self-efficacy as being a positive driver for better
learning strategies (e.g. Zimmerman, 1989), we may consider this a bit of a “chicken and egg”
situation, where, on the one end, people receive positive reinforcements while on the other end of
4 All project output documents can be found at the SLIDEshow website: https://www.slideshowproject.eu/intellectual-
outputs

the spectrum, low achievers would find it difficult to get positive feedback, leaving them forever back
in square one. One of the key factors, we would argue, is the belief in self-control. When learners do
not see themselves in charge of their achievements, they tend to “gamble” on success. That is to say,
they will attribute their performance (positive and negative) to the elements of “chance” and
unrelated to the efforts they spent on it. Since students' behavioural performance is assumed to
influence their perceptions of self-efficacy, information from tracking data (i.e. recorded learning
behaviours and status) can illustrate to them the factual investment they have taken and continuous
progression coming from it. Figure 6 (below) exemplifies this, showing that engagement with social
media leads to higher “popularity” and improved visibility – efforts pay off.
Figure 6. Example of a social media progression metric
Among the competences for self-observation we include the ability to set relevant indicators and
weightings, measure against them and connect this to the learning environment (context and
conditions).
7.2.1 Data Visualisation, Interpretation and Analysis
To exercise self-observation, reflection and self-judgement on educational datasets, interpretation
and diagnostic competences are required (cf. Greller & Drachsler, 2012). Visualisations – that is the
communication of information using graphical representations – are the most frequent form of
presentation of data analysis and of emphasising a point. But, as we mentioned in Section 5 above,
they could potentially present cognitive challenges that require a specialised skill set for interpretation
and analysis to facilitate drawing conclusions from data information. Casilli (2010) gave a comprehen-
sive account on what information visualisation means in cognitive terms by comparing it to verbal
language and considering its relationship to memory and cognition.

Visualisation research is a relatively new domain and strongly correlated to the increased importance
data plays in our decision-making processes. It is seen as a valuable tool for understanding and sharing
insights around data. The right visualisation can help to highlight a central point of observation and
open it to further examination – and for consequential (self-)reactions. But, creating effective visual-
isations is difficult, because the problems are often complex and ill-defined. It requires identifying
what attributes in a dataset are important (cf. Fisher & Meyer, 2017). Munzner (2014) gives some
practical advice for utilising graphics. While this may sound to account mostly for activities of a
visualisation designer, very similar interpretation competences are required when being presented as
an ordinary user with a dashboard or infographic, or when pointing learners to the important figures
in a table. Again, consideration of the context is paramount for sensemaking. For example, in a com-
petitive environment some figures may be of lesser importance than, say, in an instruction scenario.
On this basis, more and more learning applications take a visual approach by offering learning dash-
boards. Duval (2011) points out that if they are to have any effect beyond the initial “wow” factor,
they need to be goal-oriented. Greller and Drachsler (2012) point to the danger that enticing graphics
could potentially hide pedagogically relevant insights behind nice-looking but meaningless visualisa-
tions. A critical look at how information is presented, therefore, is indispensable (e.g. table, charts,
graphs, timeline, heatmap, etc.). However, having an overview of student activities in the form of a
dashboard supports self-monitoring and informing the teacher’s learning support and personalisation
efforts.
Furthermore, teachers can benefit from analytics on the provided learning resources or topics. A nice
example is that from a primary school application called Multiplication Trainer to learn the 10*10
tables (Greller, Ebner, & Schön, 2014). It allows teachers to get an overview how their class performs
as a whole or each child. Aside from being able to personalise support for struggling pupils, the class
view identifies particularly difficult items that many children find challenging to solve (figure 7 below).
Based on this insight, teachers are able to react by focussing on exactly these weaknesses in their
lessons.
Figure 7. Heat map of the most difficult questions (Greller et al., 2014)

Analytics results are presented in many different formats. Text presentation, for example in the form
of lists (e.g. rankings), is just one of many but sometimes overlooked as the most informative and
easily comprehensible way to summarise information. Most popular in Learning Analytics applications
are visualisations and dashboards. These aim to show information about learning behaviours, but also
learning paths taken. It is assumed that visualisations of learner activities can motivate students and
reveal participation (Govaerts et al., 2010). In a case study from Engineering, Santos and colleagues
(2012) investigated and evaluated with students a goal-oriented dashboard that supports self-
reflection and comparison with peers. Their goal was to help students learn how they are using self-
tracking applications to achieve concrete goals. In their literature review on the topic of “Visual
Analytics”, Vieira and colleagues (Vieira, Parsons, & Byrd, 2018) state that visualisation tools often
involve and require human participation. This enables an analytic discourse between users and data,
such that users can perform analytical tasks on visual representations, e.g. confirming hypotheses,
exploring hypothetical correlations, or identifying interesting features in data for further examination
(ibid.). As a practical approach, asking high level questions is recommended by some experts.
Interpretation and analysis competences on evidence-based learning go beyond the quantitative
number crunching of statistics, its bar charts and pie charts. For a comprehensive and holistic view of
the learner’s status, effort and progress, hard-to-quantify qualities, like the level of engagement and
enthusiasm of learners, are valued by both educators and learners (Balasooriya, Mor, & Rodríguez,
2018). Furthermore, attention is a core concern in SRL and open educational contexts: as learning
resources become available in more and more abundant ways, attention becomes the scarce factor,
both on the side of learners as well as on the side of teachers (Duval 2011). Attention metadata (cf.
Wolpers, Najjar, Verbert, & Duval, 2007), i.e. clickstream tracking data, can here provide interesting
insights into some more abstract qualities of learning. For example, such data can reveal which re-
sources or learning paths are more popular than others. With this knowledge, teachers can help stud-
ents to manage information overload and the phenomenon of paradox of choice5
.
Summary
Understanding self-observation encompasses increasing levels of complexity when it comes to using
analytics, starting from simple measurement and self-recording. Key for advanced mastery is to know
what you are going to measure (indicators, units, context), when (timespan, interval, frequency) and
how (instruments, sensors). On a higher level of sophistication this involves exercising active control
over these factors, not taking them as given, but changing the settings and tunings. Systematically
experimenting with different variables in controlled environments enhances the possibilities of
exploitation even further but requires more advanced skills.
Some of the greatest value in evidence-based SRL comes from incorporating data into your daily
routine. This begins with finding a few metrics that are simple to monitor and incorporate them into
your regular learning/teaching process (e.g. handraising of children). Setting the best intervals to see
progress is a task that can be fulfilled with experience. For example, when writing an essay, using the
word count feature after every sentence isn’t much good, and bears the danger of diverting attention
5 cf. Wikipedia: https://en.wikipedia.org/wiki/The_Paradox_of_Choice

to other things than learning. However, looking back at the end of your session on how much you’ve
accomplished and reflecting on factors and levels of satisfaction has the potential of evaluating and
finding the right strategy. Establishing a baseline and watching how the metrics evolve over time, is
another good approach. For example, when starting your vocab rehearsals, you might be at a low level
of recollection. From there the expectations are that you bring the percentage of memorised words
up, for example through intensive drills. Hence, being aware of simple metrics will give you a much
deeper understanding of your learning performance as it provides early visibility into positive or
negative changes happening. Metrics can have the effect of gamifying your learning experience and
instilling motivation.
According to Zimmerman (2002), the skills to observe and analyse one’s own actions and putting them
into relationship with learning outcomes and other learners are of great importance. What data do I
need to observe for a particular learning goal? What value do I attach to individual metrics? Using
observational evidence functions as a reflection amplifier through continuous feedback loops. It
begins as a goal-oriented diagnostic exercise and leads to differential evaluation of progression or the
absence of it. Self-observation also means being context aware. Are the environmental circumstances
suitable for optimal learning, e.g. noise levels? Am I as a learner in a good emotional state for learning
or reflection?
For self-evaluation, the influence of progress over time, which can be positive or negative (e.g.
retention of memorised facts) and the learning goal (benchmark) against which to measure are the
main parameters. Because learning is always situated, the state of self and the environment are signi-
ficant influences that need consideration. On a meta level, a critical approach to one’s own learning
strategies, formulation of hypotheses and experiments to optimise them are equally important.
Self-observation is not as trivial as looking into a mirror. It requires reflective and analytic self-
evaluation skills, but also humbleness about one’s strengths and weaknesses. Hence, the formulation
of concrete and measurable learning outcomes and commitment to these learning goals is one of the
competences essential for self-organised independent learning. This is encapsulated in our “de-
fine/validate” criterion above (cf. fig. 5), belonging to the forethought and planning phase. What it
means is that the self-regulated learner should be able to identify what kind of expertise they want to
reach, to what level, and for what purpose. One way to go about this, for example, is a project-based
approach, where the well-known SMART criteria can be applied:6
S – specific;
M – measurable;
A – achievable;
R – relevant;
T – time-bound.
6 cf. Wikipedia: https://en.wikipedia.org/wiki/SMART_criteria

Especially, the component of setting “measurable” goals is of relevance for our evidence-based
approach. It raises the question for the self-regulated learner of what to measure to adequately
determine progress and success, as well as the interval of measuring. For example, when studying a
larger volume of vocabulary in a foreign language, it may be unimaginable to achieve a 100% success
rate. Instead, the learner can determine that they would be satisfied with a lower, but still challenging
achievement (e.g. 90%). They can also refine this goal by saying that they want to reach this level, at
the third re-run through the entire set, by checking their improvement from 50% in the first run to
77% in the second, arriving at 92% in the third rehearsal. Further improvements to this strategy could
be the aim to improve retention of the acquired words by setting a goal of still reaching 80% correct
translations after a one-week break.
Understanding visualisations can be a cognitively demanding task. Familiarisation with rationales of
when certain graph types are most commonly used (e.g. line graphs, bar or pie charts, spider diagrams,
etc.) is part of developing higher levels of competence in sensemaking of data-based information
(ChartMogul.com, 2017). It includes the ability of drawing relevant and plausible conclusions from
data visualisations or dashboards. Putting the right weighting and importance to some figures over
others helps to concentrate learning efforts and enables users to set focal points in learning
behaviours and strategies. Teachers and self-regulated learners should determine which variables
drive change in other variables and diagnose why things are happening by formulating hypotheses.
This can be an extremely powerful method when trying to prove the impact of certain pedagogic
designs.
More advanced skills enable the teacher to combine data information from different sources by
understanding the connections between data, including, for example, the ability to distinguish
causation from correlation. Most common is the use of demographic and background information
(e.g. previous knowledge) together with learning evidence. This is particularly useful for teachers in
care of a classroom full of pupils.
Adaptation is yet another important concept in SRL. Adaptive behaviour is seen as a reflective
component of the self-regulated behaviour (Winne & Hadwin, 2008). It means that students are
monitoring and evaluating their own learning activities: searching reasons for their success and
failures and using this information to plan and support their future learning situations (Clearly &
Zimmerman, 2012).
7.3 Shortfall Awareness
For all types of personalised learning, taking a holistic learner perspective is paramount to achieving
successful outcomes. Learning is situated and embedded in a variety of contexts and factors,
momentary and permanent. When using learning evidence data, it is important to not only evaluate
the feedback received directly, but to be aware of the things the data does not show (cf. Greller &
Drachsler, 2012). The DigCompEdu framework recognises this by adding to the increasing importance
of learning data analysis the formulation “complemented by the analysis of conventional evidence on
learner behaviour” (Redecker, 2017).

There is a clear demand for reflective practice of teachers in that the success or failings of a student
may not be lying in the data, nor in the performance of the learner, but in the learning design. Drawing
a connection between the feedback of the application and the learning design, therefore, is influential
for the successful taking of evidence-based decisions (cf. Mandinach et al., 2006).
Data and analytics are not neutral, neither is the technology that produces them. Michael Feldstein
(2016) points to the fact that analytics are narratives. He emphasises that most people aren’t
especially literate in this kind of narrative and don’t know how to critique it well. He sees this as a
limiting factor to economic growth in the sector of educational data until educators understand the
use and limitations of datasets (ibid.). Feldstein argues that in order for data points to make sense, we
are building a narrative hypothesis on top of them, but even if there is a detectable correlation it may
not be correct due to other student circumstances or pure coincidence. The important point in this
view is that we start out with the feedback figures and then try to find explanations (hypotheses)
instead of asking whether there is anything awkward in the assumptions taken. This leaning on
probabilities leads to weaknesses and uncertainties in the execution and follow-up of the results
presented to us.
The afore-mentioned Data Journalism Handbook (European Journalism Centre & Open Knowledge
Foundation, 2012) gives graphic examples on how important it is to know about the sourcing of data.
The higher the transparency is and the better our understanding is of how the data came about, the
more trust we can put into our own evaluation and situated narrative. Only with a certain amount of
trust in the accuracy and relevance is it possible to critically evaluate and derive conclusions from
datasets of all kinds. Scale matters here too, for in order to test a hypothesis on the feedback, you
need to repeatedly test and compare it, otherwise it could just be an exceptional case. Feldstein
(2016), therefore, argues for the ability to independently verify the stories that the algorithms are
telling. He calls for the teachers to embrace the sciences of teaching and learning. Data should extend
our senses not replace them. Hence, once again, the complementarity of the data feedback and other
(conventional) evidence comes to bear. Scientific literature or even feedback from related learning
networks and peers can enhance our evaluation capacities.
The shortfalls might be hidden in the data and invisible to the end user. Erroneous data or data noise
like test personas that aren’t real people could lead to distortions in the results. This is particularly the
case in aggregated datasets from multiple sources. Algorithm-based recommender systems, for
example, can show inaccurate and unsuitable suggestions coming from “bad” data. Then there is the
designer bias to be considered, where programmers deliberately steer attention to some things at the
neglect of others. For example, an online shopping portal may show recommendations of items partly
calculated from one’s own history and partly randomised to allow users to discover new things or
sales items. In most cases, the user doesn’t have control over the accuracy of these recommendations.
Hence, relying on algorithms to make decisions is problematic as the (economic, political, or other)
motivations driving them are invisible to the users (Currion, 2016).
Wilson and colleagues (2017) mention that the relationships between learning and the digital traces
left by participants in online environments are far from trivial, and that any analytics that relies on

these as proxies for learning tends towards a behaviourist evaluation of learning processes. Without
being aware of this, pedagogic bias may be the unwanted consequence. They also warn of a potential
off-the-shelf approach to all learning induced by the vendors of institution-wide Learning Analytics
solutions. Similarly, personalisation support often involves categorising students according to their
digital behaviour and traces, but warnings have been expressed as to the differences between
categorisation and stereotyping (Mittelmeier et al., 2018). So, this is yet another area to be watchful
of.
In the literature and debates around data and analytics there is much attention on bias contained in
the data and its interpretation. Among those implicit dangers is confirmation bias, i.e. data-based
evidence to confirm preconceptions (cf. Arora, 2016). Observation bias is another form, which is
referring to the attitude that one only sees what they expected to see. Again, this could lead to
uncritical affirmation of presupposed hypotheses, or, in the worst case, reinforcement of prejudices.
It is worth noting that such bias not only comes at the interpretation end of the analytics process but
begins with the programmatic design of the algorithm and setting of indicators and weightings.
Summary
Because of various circumstances, being critical about algorithms and learning evidence data and even
self-critical is a needed attitude for practicing Learning Analytics. It may not be easy to detect and
impossible to avoid bias in analytical processes. Nevertheless, we consider having data at our disposal
as a complementary decision tool to be better than having no data. Still, it is good to be wise about its
limitations and not overvalue the conclusions it may suggest. For this reason, critical thinking skills
are emphasised in much of the literature on Learning Analytics (e.g. Greller & Drachsler, 2012).
The mentioned shortfalls call for certain evaluation competences in users that critically examine the
background, credibility, and focus of data-based information and representations. It entails being able
to describe the population and scale covered in the dataset, identify potential cultural issues buried
in the data, such as economic interests, or detecting weighting bias in recommendations or algorith-
mic results that may not be suited for one’s own learning goals. It also means being aware of the
message that data is able to tell us. For example, if the dataset covers student logins, pageviews and
downloads, there might be a pedagogic bias towards an instructionist value of learning materials
rather than on the learning strategies that students choose. This output may not be relevant in a socio-
constructivist context of learning. Alternatively, data information like this may be a useful comple-
mentary source of feedback to such an approach.
The skill set connected to the challenge of navigating around shortfalls contains the cultural notion of
hypothesis formulation, experimentation and reflection. This is a move towards Feldstein’s demand
for a cultural change from being a teacher to a learning and teaching scientist, and involves document-
ation of the steps being undertaken (Feldstein, 2016). It further involves the ability to explain and
confidently act on probabilities rather than clear facts. A familiar example for educators would be the
use of a plagiarism detection tool. Such machine agents don’t present results of their analysis as true
facts, but as probable cases that could be investigated further with the author. But what does it mean
to have a 60% positive?

In chapter 7.2 above, we mentioned that the concept of defining/validating success (when is a learning
objective reached) is covered by the forethought phase. It sets the rationale and the benchmark to
validate against during the run-time of the learning process, which in our specific practice is data
informed to repeatedly and continuously self-monitor progression towards these goals. On the other
hand, there are non-linear learning theories, where strict iterative sequencing would not work.
Serendipitous learning, heuristic learning, rhizomatic learning, surprise-based learning are examples
of this (e.g. Barto, Mirolli, & Baldassarre, 2013). Such scenarios need not necessarily follow the linear
instructionist pattern of Zimmerman’s model, because they don’t emphasise the element of planning
and forethought. Still, even in these approaches, data may reveal important insights through the
process of educational data mining, bringing to light issues like “rare events”. These are statistically
infrequent or spontaneous events or patterns within a much larger dataset of much more frequent,
generally well-understood contexts (cf. Reimann, 2016).
Shortfall awareness also requires competences to detect potential fraud or manipulation. This could
be students trying to game the system and trying to boost their data performance (e.g. using multiple
logins, automation scripts, etc.). Needless to say, that they are cheating themselves out of a positive
learning experience. Still, to be risk aware is part of the class management skills that teachers are
required to demonstrate in their daily practice. A valuable part of evidence-based learning, therefore,
is the mindset it creates. Data information brings us to be accountable for the teaching and learning
decisions we are taking. Because such figures display a certain status of learners and their progress, it
inspires people to strive for continuous improvement.
7.4 Data-to-Learning Transfer
Retrieving diagnostic information from data and putting it into positive interventions for learning is
arguably the most complicated and complex competence area for evidence-based learning. The cur-
rent state-of-the-art in Learning Analytics research is much focussed on the mutual influences learning
evidence has on learning design and vice versa (cf. the extensive literature review in Mangaroska &
Giannakos, 2018). For SRL, this dimension relates to the self-reactions process, that follows the previ-
ously discussed self-observation and self-judgement phases (see chapter 7.2 above and Zimmerman,
1989). It considers applying the right action after looking at the data-based evidence.
Winne (2017) argues that trace data are observable indicators that support valid inferences about a
learner’s metacognitive monitoring and metacognitive control that constitute SRL. He holds that self-
regulated learners actively research what they do to learn and how well their goals are achieved by
variations in their approaches to learning (ibid. and Winne, 2010). But learners have different ways to
construct knowledge and they have different ways to think about how that construction took place
on the basis of the feedback they receive when learning in a self-regulated way. Learners can act and
react in different ways based on that information. Reflective and planning competences are needed
to think of factors necessary to act adequately on the information provided. But awareness is not the
only quality that influences the process of feedback, reflection and behavioural change, i.e. of self-
efficacy and SRL (Butler & Winne, 1995). A personal culture of scientific thinking is necessary, using
basic methods or principles of scientific inquiry to reasoning or problem-solving. This involves the skills

implied in generating, testing and revising theories, applying predictions, and in the case of fully devel-
oped skills, to reflect on the process of knowledge acquisition and change (C. Zimmerman, 2007).
In evidence-based learning, learners often have the option to compare themselves to some standards
or, alternatively, to peers (Winne, 2017). To choose for oneself, whether it is more important to make
an absolute or a relative advance in measurable progress is one of the possible learning strategies
available. In Winne’s concept of Learning Analytics for SRL, learning evidence leads to adjustments of
conditions, operations, or standards. ‘Operations’ in his terminology refers to adaptive actions, while
‘standards’ refers to goals. In terms of translating data information into further self-regulated learning,
Winne (ibid.) states that if analytics are presented without recommendations for action, there is an
opportunity for the learner to investigate previously available alternative learning path options. This
means a learner can revisit and adapt the learning path. This also allows Learning Analytics to record
motivation and existing individual tactics (ibid.). Tempelaar, Rienties, and Nguyen (2017), in this
regard, advocate using Learning Analytics for student disposition to focus on potential causes of
suboptimal learning, such as applying ineffective learning strategies. This can be supported by self-
reporting mechanisms (Shum & Crick, 2012).
Questions that arise in the context of adapting learning behaviours based on data evidence are: What
should change about how learning is carried out and how to go about changing it; what are Learning
Analytics supposed to help improve? Winne (2017) also thinks that many learners may not know or
have the skills to enact changes and recommendations by themselves (ibid.). What is evident, though,
is that Learning Analytics has a lot to contribute to the validation of learning strategies chosen by the
self-regulated learner. This includes, for example, methods for seeking, organising, and transforming
information.
We previously stressed the importance of learning data and the context in which it arises and aims to
be applied. For teachers, this includes the testing of different variations of learning designs that can
be validated, thus potentially leading to ever-improving methods of delivery. Learning designs are
seen by many researchers as a form of documentation of pedagogic intent that can provide the
context for sense-making of learning evidence (Lockyer et al., 2013). But educators are challenged
with extracting actionable information from large quantities of data and linking it to the learning
processes and learning designs. Hernández-Leo and colleagues (2018) identify three layers of
connections that explicitly address this aspect of supporting teacher reflection on their delivery of the
educational design: the learning experience, the learning design itself, and the community of
educators. With their “Analytics Layers for Learning Design” (AL4LD) framework, the authors aim to
support decision-making in learning design on the basis of analytics (ibid.). Evidence on learning
behaviours from a conceived learning design provides substantial information on its impact and helps
to identify design elements that need to be revised for future designs. For teachers this allows
reflection on issues like the actual task completion time in relation to their expectations, or whether
the provided tutor support is adequate for task fulfilment. More specific design-related datasets serve
teachers with awareness and reflection on decisions made during the creation of a learning design
rather than on the impact on students at the receiving and enacting end (Hernández‐Leo et al., ibid.).
An example of this is whether the student load in a design is balanced across individual vs. group tasks.

Generally, the researchers conclude that Learning Analytics outputs are more meaningful when
aligned with pedagogical intentions. Conversely, learning designs can be strongly influenced by the
data available. They posit that good practices for data-driven scenarios in learning design must include
support for evidence-based reflective re-design of learning activities; scaffolding for the design
process through analysis of the pedagogic intentions embedded in the design; and, inspiration and
awareness of colleagues’ design activity (ibid.). Kitto and colleagues (2017) mention that educators
find it difficult to imagine how students might use educational data and hesitate to translate current
analytics dashboards into their classroom practice. They plead for reflection on learning evidence to
be incorporated into teaching protocols and provide two learning design patterns (do-analyse-change-
reflect and active learning squared) for encouraging skills such as sense-making, metacognition, and
reflection.
A strategy for changing learner behaviour using evidence-based learning is manipulating the weighting
of indicators. For example, if active participation in the classroom is important, but low due to lack of
attendance, the teacher can transparently inform students that attendance will be given higher
importance in their overall assessment, thus, hopefully, motivating higher presence levels in class. The
impact of such a change can be observed and measured, and further adjustments made if needed.
Schmitz and colleagues present a number of case studies in this direction also representing SRL
scenarios (Schmitz, Scheffel, Limbeek, et al., 2018).
Evidence-based learning derived from large and diverse datasets with the properties mentioned
above in chapter 4.3 supports personalisation, because due to its nature, data covers simultaneously
the individual micro, intermediate, and aggregated macro levels. Especially the aspect of near real-
time feedback offers new opportunities of monitoring actions and decisions of learners. Teachers are,
thus, enabled to dynamically apply quantitative and qualitative information on the fly in order to map
learner characteristics, validate learning designs, analyse the effectiveness of learning materials and
tasks, adjust difficulty levels, and measure the impact of their interventions and recommendations (cf.
Ifenthaler, Gibson, & Dobozy, 2018).
An important aspect is that teachers in formal learning contexts (schools) operate under contextual
constraints. For every-day educational practice this means that conclusions and actions based on
learning evidence need to lie within the boundaries of such constraints. They could involve not being
able to repeat field trips with students, or teaching to a mandatory curriculum with resources, learning
outcomes and assessment goals predefined by education authorities. Ethical constraint may also
apply. This has important implications on the flexibility to adaptations of the learning design, even
when and if the pedagogic competences to do so are readily available.
Generally, there is a noticeable gap in research on evidence-based intervention design. Much of the
literature concerns itself with the activities to analyse and improve understanding on learning, but
there is still a lack of research-based guidance as to concrete actions informed by classes and types of
data. Such categories could help, for example, to investigate the implications and mappings of temp-
oral data on proposed actions. As Winne (2017) puts it, much research remains to investigate how
temporal features and coarser grained descriptors may play useful roles in Learning Analytics about

SRL. The same is true for other classes of data like social or communicative data, data sequences, path
analysis, mixed data approaches, and so forth. The essential relationship to learning theories and the
learning sciences has been stressed in several scholarly works (Reimann, 2016; Agasisti & Bowers,
2017; Sedrakyan, Malmberg, Verbert, Järvelä, & Kirschner, 2018).
Summary
We see three areas of analytics-induced change in self-reaction: state, behaviour, and environment
(cf. fig.8 below). Firstly, the learner state can be influenced by feedback from learning evidence.
Motivation is the most relevant factor (e.g. interest, effort, altruism), but also emotional states may
be changed (e.g. feeling in control, feeling enthused, etc.). Secondly, environmental parameters are
becoming more transparent and, therefore, where possible can be improved. Thiscan be simple things
like preventing noise or interruptions; it could also refer to competences in using a particular
technology. Finally, and most importantly, the learning behaviour can be adapted while a more
reflective and adaptive learning culture should evolve. This encompasses things like the adaptation
and evaluation of learning strategies (including e.g. repetition), the choice of learning resources
consulted, or putting a special focus on items of relevance. Changes in these domains, it has to be
emphasised, are hardly effective without direct decision-taking and acceptance by the concerned indi-
vidual but can be induced indirectly and unobtrusively, leading to higher levels of self-responsibility.
This evidence-based decision taking involves a certain amount of self-confidence and perceived self-
efficacy.
Figure 8. Areas of self-reaction in SRL
It is worth noting that figure 8 represents a learner-centric view. Therefore, in a teaching scenario, the
learning design would in this model belong to the environment and may be out of the control of the
learner, e.g. certain learning resources or activities being compulsory, or at what point in time
resources become available (e.g. tests). In a way, the teacher acts as a filter and facilitator on the
process of change. Choice and negotiation possibilities are helpful in this case, for better self-directed
control, and should be embedded in the pedagogic design.

An important skill for transferring data information into learning behaviours is hypothesis formulation,
comparison and prediction: this is similar to an “if … then” statement in computing. In educational
practice, this could involve predictive intention. For example, if data from a learning environment
suggests the learner has a deficit in active participation and engagement with other students, then a
predictive intention may be formulated that more forum posts lead to a different situation.
Furthermore, a deeper qualitative analysis might suggest, that engagement with other students
involves more than statement posts, in terms of engaging in a topical discourse and reacting to other
people’s postings (replies, providing answers, social reactions: likes, shares, forwards, etc.). In turn,
this could lead to a new or improved SRL learning strategy for the social context of learning.
We should be aware of a potential pitfall, where the learner in their iterative reflection loop only sets
actions to improve a single metric, rather than improving their learning. A firm focus on the larger
learning success is needed, i.e. the intended learning outcomes and skills (incl. SRL), and not the micro-
data involved in constant goal-(re)orientation. Hence, preparing to take the right action includes
looking ahead at the goals to reach.
Competences for taking decisions and applying actions include correcting previous misconceptions
about one’s learning. An often-neglected quality is that self-regulated learners may find themselves
at the end of their ability, at least at the end of their independent ability (cf. the zone of proximal
development theory by Lev Vygotsky)7
. Here, seeking help or avoiding continuous frustration by aban-
doning an exercise, potentially looking at alternatives, are sometimes originating from personal states
or environmental factors available from the evidence. This is best to compare to gaming experiences
where a player despite several attempts fails to move on to the next level or a higher score. Before
frustration becomes overwhelming and the game is abandoned, there are often possibilities of seeking
peer support or other help. An example from education is the typical self-study MOOC (massive open
online course), where a great deal of learners drops out due to time constraints from other (external)
influences, such as job or family (Eriksson, Adawi, & Stöhr, 2017). At some point, they realise that they
have fallen too far behind to catch up – as would be noticeable even without supporting data. They
can now take concrete action to either put in extra effort or abandon the course and look for other
possibilities. For a teacher in primary school it is important to observe signs of upcoming frustration
and, if possible, provide guidance or alternatives that help pupils to keep a confident outlook on their
learning activities.
7.5 Ethical and Legal Knowledge
Using data evidence and resulting information in an ethical and supportive way rather than as a
surveillance instrument has long been a cause for hesitation and debate among teachers and learners,
researchers and institutions (cf. Drachsler & Greller, 2016). It has become contentious that machine
algorithms might be used to manipulate behaviours into certain directions. Greller and Drachsler
(2012) additionally mention the dangers of stifling educational innovation by mainstreaming be-
haviours into “averages” and ignoring justifiable outliers without question.
7 https://en.wikipedia.org/wiki/Zone_of_proximal_development

In the common definitions of Learning Analytics, there is an emphasis on the fact that evidence from
data needs to be actionable to be useful. Prinsloo and Slade (2016, 2017), therefore, point to the
ethical dilemma of the so-called “obligation to act” that is directly connected to a teacher’s fiduciary
duties, i.e. the intent of helping and supporting students to the best of their knowledge and ability to
overcome potential vulnerabilities and failure. They rightly argue that knowing more about the
students will offer a greater understanding of their learning strategies and behaviours, which leads to
greater responsibility for the teacher to do more in terms of pedagogic or other support (Prinsloo &
Slade, 2017). However, they also stress that students too need to take greater responsibility for their
own learning. Since the data evidence will not distinguish between the fails of the student and the
failings of the teacher or institution, a critical self-reflection attitude is crucial for a balanced mutually
beneficial approach.
In response to the concerns about ethical use of student data, the JISC have developed a ‘Code of
Practice for Learning Analytics’ (JISC, 2015). Similarly, the Learning Analytics Community Exchange
project (LACE) created a checklist to implement trusted Learning Analytics at institutional level
(Drachsler & Greller, 2016). Furthermore, many institutions, like the University of West London
(Lingard, 2016) or Charles Sturt University in Australia (2015), have established principles, approval
processes, complaint procedures and ombudspersons to promote trust among their learners and
teachers in evidence-based decision support and personalisation.
We already mentioned the legal obligations in the use of student data enshrined in the General Data
Protection Regulation (GDPR) and other national and international basic civil rights codes (cf. chapter
4.2 above).
“Ethical and responsible data use is part of knowing how to use data, and that knowledge focuses on
how to protect student privacy and maintain confidentiality of student data.” (Mandinach, Parton,
Gummer, & Anderson, 2015)
Hildebrandt (2017) states that Learning Analytics processes personal data and concerns interventions
at the level of an identifiable student, whether or not they are aware of this. She stresses that this
includes “being singled out” and sees it as a potential ethical and legal issue even though it may
support timely reconsideration of a student’s capacities or learning strategies (ibid.). According to her
analysis, the use of data not only concerns ethics and privacy but also non-discrimination and
avoidance of reconfirmation of prejudices through data. How ambivalent this topic is shows an article
on data bias by Arora (2016) drawing a comparison of database design from Western industrialised
countries being applied to less developed regions of the world.
An interesting ethical risk management tool for what they call an “Ethical Operating System” has been
launched by the global philanthropic Omidyar Network registered in London and the Institute of the
Future in California (2018). It identifies eight risk zones that may impact and harm the peaceful use of
new technologies, including educational technology. Among other things, it mentions the risk of “bad
actors” and even cyber-criminals, and we want to take this warning on board and incorporate it into

a set of practical properties for responsible data handling by educators. This includes safe storage of
sensitive student data be it digital or analog.
For all these reasons, there is a necessity to develop an awareness and understanding of risks
connected to the ethical and legal application of learning evidence data based on clickstreams and
behaviours, which we aim to include in our proposed framework.
Summary
As we can see from the above, the competences required for evidence-based learning goes well
beyond the simple awareness of and compliance with the law. It poses a substantial ethical and moral
challenge, especially when you are in charge of a group of students. Especially for underage children
and for sensitive personal data, the GDPR has a very strict approach of protection. As a teacher, the
fiduciary duty goes even beyond this, as ethical behaviour can influence a child’s future. In self-
regulated independent learning situations too, when entrusting others with your data, it is important
to know your rights and the avenues open to you should something go wrong.
We consider transparency to be the key to successful application of learning evidence in an SRL
context and otherwise. Learners and their educators should know where their data is going (is it
internally or externally processed), who has access to it, what informed consent means and how to
revoke previously given permissions and have your data deleted. Most of all, there needs to be
transparency with regards to the purpose for which data is collected and processed in order to build
trust in the results and consequently in the pedagogic support. This also requires some risk awareness,
especially with respect to commercial advertisement-driven platforms that are seemingly free to use.
Again, this competence area of ethical and legal knowledge is largely a knowledge domain, and much
would depend on the independence of the learner in question versus the learner in the care of a
competent and knowledgeable teacher. Issues connected to this relate to what methods of control
you have at your disposal for your own learning datasets? How do you document and safeguard data
of yourself and others? This can include technical controls in the settings of an application, it can also
include legal avenues like the possibility of deleting your data from a platform, or administrative
processes for eventual complaint procedures.
7.6 Social Learning in Teams and Networks
SRL is situated learning. This means it happens in context and in a continuous mutual exchange with
the environment, including people and even computer agents. One of the dimensions of this
environment, therefore, is social connectivity. Bandura (1997) acknowledged the importance of the
social environment in which the development of SRL takes place and where meaningful interactions
occur. Applying learning evidence to social learning situations is a specialised field of research,
sometimes called Social Learning Analytics, that focuses on how learners build knowledge together in
their cultural and social settings (Shum & Ferguson, 2012). The basis of it is that new skills and ideas
are not solely individual achievements, but are developed, carried forward, and passed on through
interaction and collaboration (Ferguson & Shum, 2012). In our evidence-based approach, this means
that the competences described hitherto (chapters 7.1-7.5) have to be understood not merely in an

individual but also in a social context, as has been elaborated in IO1 for self-regulated learning, co-
regulated learning, and socially shared regulation of learning.
For the social perspective of evidence-based learning, we can distinguish two distinct layers: (1) the
social aspects of learning; (2) the social positioning of the learner. The former aspect considers social
connectivity as a specific part and component of a learning strategy, for example, help seeking from
someone considered more knowledgeable. It also comprises all kinds of collaborative learning
settings, especially when it is part of a learning design. A significant association between effective
learning strategies and high levels of cognitive presence in social knowledge construction was found
(Fincham, Gasevic, Jovanovic, & Pardo, 2018). The latter aspect of social positioning utilises the power
of personal-level micro data and aggregated higher-level macro data to compare the individual
learning performance with that of others in a collective (class, learning network, age group,
community, etc.). We already mentioned this possibility of relative and absolute measurement
(compared to peers, vs. compared to a standard or baseline) above in chapter 7.4. For teachers, this
second aspect offers substantial opportunities to compare individual students to a group, or different
groups (e.g. previous years) which affords insights into individual progress and personalised support
requirements. Social Network Analysis (SNA) in combination with learning evidence data can be used
to explore social structures promoting or inhibiting learning. A detailed analysis of its potential for all
our setting has been undertaken by Cela Rosero (2015).
Teachers operate in practitioner networks and collaborative spaces to (co)design, tag, share, explore,
reuse and refine learning designs, pedagogies, and design patterns (Hernández‐Leo et al., 2018). An
example of such a community of practice forming around Learning Analytics is the approach by the
US Department of Education’s Connected Educators project, which aims to connect teachers to
enhance their professional effectiveness (Cambridge & Perez-Lopez, 2012). Through the lense of social
presence, teachers are also able to identify patterns in individual and aggregated evidence, or, as we
already mentioned, “rare events” in student behaviours. This enables them to test and verify and even
reject pedagogic theories, but, more importantly, allows for adaptation of personalised intervention
in student support.
Siadaty and colleagues (2012) state that designing systems that unlock the collective knowledge, and
the collective intelligence in higher levels of inference for the purpose of scaffolding learning is not a
straightforward task. Although there is a growing number of tools available for social network analysis
and team effort analysis, this field is still in its infancy.
Summary
In order to exploit learning evidence to its fullest potential, we must consider the social dimension as
part of the learning context. Specifically, the roles of learners in a team or network are of importance.
In groupwork as part of a designed learning activity, students are often given a particular role in a
team, either by the teacher or decided by the group. For example, one person can act in the reporting
role, observing team activities, collecting documentation and summarising the group effort and
results. Naturally, this role will influence the production of learning data, and comparison between

this dataset and that of a student in a coordinating or moderating role will show substantial differen-
ces.
Social scenarios resulting in aggregated data can be stimulating for gamification and motivation en-
hancements, by instilling competition or collaborative components in the learning design. Even self-
competition can be rewarding, such as best-time recordings. Creativity is one of the key competences
for experimenting with and exploiting social data scenarios. In learning networks or groups, embedded
reputation systems to rank users, called “karma”, are sometimes useful for identifying trust relation-
ships between individuals. The relationship between user's karma level and the social network
structure has been investigated by Ganley & Lampe (2009). For an external observer (e.g. tutor or
teacher) karma can validate the strength of social connections and trust relationships in knowledge
construction. In a practical example, it may help a teacher tackling misconceptions, if they consider
social influences. Strong connections can act as multipliers in certain pedagogic surroundings.
8 Personalised Evidence-Based Practice Framework
In this chapter, we will merge and converge the knowledge, skills and practices of the six competence
areas discussed above in Parts 1 and 2 of this report into a proposed framework for personalised
evidence-based SRL. For practical reasons, we don’t want to go into describing specific pedagogic
settings and demands, since the spectrum of potential application is extremely wide, covering the
entire domain of pedagogy in all its facets. Hence, we chose a higher level and more generalist
approach of presentation, which aims to be simple to understand and use, using easy language and
familiar terminology. It is our aim to facilitate every-day practice in schools, rather than scientific
researchers of Learning Analytics.
The DigCompEdu framework, which is one of the underlying models for data and information literacy,
distinguishes six layers of progressive attributes in the professional as a SRL learner (fig. 9 below).
Figure 9. DigCompEdu Progression Model (Redecker, 2017)

While this is a useful illustration of the self-efficacy in increased proficiency, it is difficult to map onto
the eight proficiency levels of the EQF, that we intend to follow. Therefore, we use it here only as an
indicator of self-description in an evaluative readiness model.
The proposed personalised evidence-based practice framework considers the eight proficiency levels
of the EQF (as discussed in chapter 6 above). It is, however, noteworthy that in the SLIDEshow project,
we cover two distinct roles of SRL. We describe the recipient audience as primary and secondary
school teachers, who are themselves self-regulated learners, and who, in their daily practice, aim to
further promote SRL and evidence-based learning as a pedagogic means in class.
The proposed framework was presented, tested and validated through a practitioners’ workshop in
Vienna, 17-19 October 2018. 25 participants from four European countries (Belgium, Austria, Finland
and Greece) were exposed to the framework to check its completeness, applicability, relevance and
usability. The results are summarised in the workshop summary report and evaluation (C1).

Personalised Evidence-Based Practice Framework
EQF Knowledge/Skills (declarative and procedural knowledge) Context/Autonomy/Responsibility (conditional knowledge) Level of Self-
Regulation
Level 1: Basic browsing, searching and filtering data
Basic understanding of learning evidence and presentation methods (lists, visualisations)
Basic awareness of the usage of data for SRL
Can read simple visualisations (graphs) and data representation (text)
Simple knowledge of statistics (means, median)
Limited data comprehension and interpretation
Understand legal policy statements for data collection and use
Navigating data, unfocused understanding of data
Assess the credibility of simple data information sources
Individually and collectively reflect on and evaluate one’s own pedagogical practice
Use data information for simple decision-making
Ask for help from others in applying data evidence
With guidance
and supervised
Level 2: Find goal-oriented data sources through systematic search
Simple interpretation of data evidence (through visualisation) in the context of own learning
Understand feedback loops and procedures to stimulate feedback use by learners
Limited knowledge about embedding learning evidence in teaching designs
Understand documentation, safe storage and data security
Understand rules of copyright and licenses that apply to data and digital information
Understand data protection regulations (GDPR)
Match learning evidence with information needs
Organise evidence data
Comparing data information from different sources
Identify, select and assess evidence indicators for measuring learning
Critically evaluate learning evidence data
Use learning evidence to plan, monitor and reflect on learning
Connect learning evidence with learning progress
Use data information for deciding on SRL learning strategies
Promote student understanding of evidence-based learning
Evaluate the credibility and usefulness of educational datasets and data applications
Limited guidance
increased
autonomy
Level 3: Protect personal data and privacy
Managing one’s digital footprints
Knowhow to support data techniques and conclusions of learners
Formulate narrative hypotheses from data evidence and play with different variables (learning
strategies) to see changes in feedback
Connecting data presentation with goal orientation
Able to choose, combine and apply pedagogic interventions based on evidence
Understand insights and evidence of progress
Understand more advanced statistical approaches (regression, quartiles, variance, outliers)
Awareness of most common risks
Understand measurement errors and discrepancies
Acting with ethical consideration and under contextual constraints
Evaluate the appropriateness of privacy policy statements on how personal data are used
Analysis, comparison and evaluation of sources of data
Plan and implement digital learning tools and data collection in the teaching process
Confident application of data information in learning and teaching
Foster learner engagement with content and peers supported by data feedback
Require learners to articulate information needs and self-evaluate the credibility and relevance of
feedback
Incorporate data feedback into learning activities, assignments and assessments
Allow students to make own data-based decisions to support SRL development
Identify key take-away points from data information
Self-evaluate data-driven interventions
Independent
Level 4: Learning Analytics for SRL and personalisation (data feedback for learning strategies and context)
Master complex and interactive visualisations and extract goal-oriented information
Identify learners’ diverse needs of learning evidence
Identify data competences gaps (self and students)
Detect misconceptions through data
Apply data skills to own teaching practice in support of students’ SRL
Analyse one’s own teaching performance through data reflection
Identify and evaluate diverse risks from data evidence
Awareness and detection of shortfalls and bias
Awareness of appropriate time/unit factor in data-evidenced progress (granularity, interval, etc.)
Apply pedagogic interventions to support SRL reflection of learners (individual and in groups)
Assess information needs and learning evidence
Connect evidence with learning behaviour
Analyse learning evidence to appropriately orchestrate interventions
Consider the specific learning objective, context, pedagogical approach, and learner group, when
selecting relevant data indicators and weightings
Promote data as reflection amplifier to learners
Use learning data analysis to enhance the interaction with learners, individually and collectively
Use data feedback to offer timely and targeted guidance and assistance
Require learners to articulate feedback needs and its impact on planning their own learning strategies
Detect fraud and manipulation of data and attempts to game the system
Independent and
personalised

Draw connections between data and instructional practices
Level 5: Learning Analytics for formative assessment strategies in addition to other (conventional)
evidence
Learning Analytics for Learning Design (data as feedback for teaching strategies)
Wider theoretical knowledge about the Learning Analytics domain(s) (e.g. design frameworks,
ethics, predictive analytics, etc.)
Creating and managing own educational datasets, including visualisation techniques
Demonstrate goal orientation and focus of evidence to students
Demonstrate data access to others
Guide others in managing their digital footprints
Use data feedback to foster and enhance learner collaboration
Support SRL processes by enabling learners to plan, monitor and reﬂect on their own learning
Enable learners and parents to understand the evidence from learning data
Enable learners to use feedback data as part of collaborative assignments
Increase learners’ active involvement in analysing and reflecting on evidence
Support improvement of perceived self-efficacy of learners
Manage the mutual impact of learning evidence and instructional practices
Guiding and
advising others
Level 6: Good grasp of indicators and weightings, integration of different sources of data in a goal
oriented pedagogic way
Formulate learning contexts into data evidence (contextual attention data)
Interpreting more complex data-based information into relevance for learning (e.g. social
network analysis, reputational emotional and motivational data, sensors)
High beliefs of own efficacy for promoting evidence-based SRL
Manage risks from bias and shortfalls
Interpret social data for collaborative learning designs and for social positioning
Adapt (own) learning behaviours according to learning evidence
Draw connections between data information and context, formulate hypotheses
Identify and explain outliers and anomalies
Critically question data inherent assumptions
Enable learners to understand feedback and to change learning behaviours
Cope with integrated data contexts (multiple environments combined)
Use evidence for social aspects of learning (collaboration)
Apply situated feedback loops and evaluate its effectiveness
Create motivating gamification scenarios to stimulate SRL
Stimulate adaptive behaviour and flexibility in students based on learning evidence
Complex
adaptations and
contextualisation
Level 7: Integration of learning evidence with daily routines
Define and assess measurable progress and achievements for individual learners and
communities
Experiment with new SRL data environments and contexts (e.g. mobile learning)
Problem-solving competences to overcome barriers for evidence-based SRL
Create focused visualisations from data
Experiment with data indicators and SRL methods for own benefit and in providing optimised learning
support
Coping with changing contexts and shifting goalposts
Explore relationships between different datasets, indicators and practice
Develop practice-based situated scientific knowledge on SRL with data support (action research)
Integration with
context,
experimentation
Level 8: Propose new ideas for the field
Create new pedagogic applications from data evidence sets
Identify new technologies for the production and exploitation of learning evidence
Combine personalisation for learners and own professional development in new innovative ways
Scientific thinking using principles of inquiry and verification, systematic reasoning, and
predictions
Contribute to professional practices and knowledge
Sharing and participating in a community of practice
Experiment with and develop new formats and pedagogical methods for evidence-based learning
Develop abstract scientific and transferable knowledge for the practitioner community
Advancing the
field

REFERENCES
Agasisti, T., & Bowers, A. J. (2017). Data Analytics and Decision-Making in Education: Towards the
Educational Data Scientist as a Key Actor in Schools and Higher Education Institutions. In
Handbook of Contemporary Education Economics (pp. 184–210). Edward Elgar Publishing.
https://doi.org/10.4337/9781785369070
Arora, P. (2016). Bottom of the Data Pyramid: Big Data and the Global South. International Journal of
Communication, 10(0), 19.
Balasooriya, I., Mor, E., & Rodríguez, M. E. (2018). Understanding User Engagement in Digital
Education. In P. Zaphiris & A. Ioannou (Eds.), Learning and Collaboration Technologies.
Learning and Teaching (pp. 3–15). Springer International Publishing.
Bandura, A. (1997). Self-efficacy: The exercise of control. New York, NY, US: W H Freeman/Times
Books/ Henry Holt & Co.
Barto, A., Mirolli, M., & Baldassarre, G. (2013). Novelty or Surprise? Frontiers in Psychology, 4, 907.
https://doi.org/10.3389/fpsyg.2013.00907
Bohlinger, S. (2008). Competences as the core element of the European Qualifications Framework.
European Journal of Vocational Training, (42/43), 17.
Butler, D. L., & Winne, P. H. (1995). Feedback and Self-Regulated Learning: A Theoretical Synthesis.
Review of Educational Research, 65(3), 245–281.
Cambridge, D., & Perez-Lopez, K. (2012). First steps towards a social learning analytics for online
communities of practice for educators. In Proceedings of the 2nd International Conference on
Learning Analytics and Knowledge - LAK ’12 (p. 69). Vancouver, British Columbia, Canada:
ACM Press. https://doi.org/10.1145/2330601.2330622
Carretero, S., Vuorikari, R., & Punie, Y. (2017). DigComp 2.1 the digital competence framework for
citizens with eight proficiency levels and examples of use.
CEDEFOP. (2018). Qualifications frameworks in Europe 2017 developments, 5.
Cela Rosero, K. L. (2015). Social Network analysis in eLearning environments: a study of learner’s
interactions from several perspectives. Retrieved from
http://repositorio.educacionsuperior.gob.ec/handle/28000/1925
Charles Sturt University. (2015). CSU Learning Analytics Code of Practice. Retrieved from
http://www.csu.edu.au/__data/assets/pdf_file/0007/2160484/2016_CSU_LearningAnalyticsC
odePractice.pdf
ChartMogul.com. (2017, July 3). The Ultimate Data Literacy Cheat Sheet. Retrieved 9 August 2018,
from https://chartmogul.com/resources/data-literacy-cheat-sheet/
Clow, D., Cross, S., Ferguson, R., & Rienties, B. (2014). D2.5 Evidence Hub Review. Retrieved 12 July
2018, from http://www.laceproject.eu/deliverables/d2-5-evidence-hub/
Cort, P. (2010). Stating the Obvious: The European Qualifications Framework is Not a neutral
evidence-based policy tool. European Educational Research Journal, 9(3), 304–316.
https://doi.org/10.2304/eerj.2010.9.3.304

Council of the European Union. (2017). COUNCIL RECOMMENDATION of 22 May 2017 on the
European Qualifications Framework for lifelong learning and repealing the recommendation
of the European Parliament and of the Council of 23 April 2008 on the establishment of the
European Qualifications Framework for lifelong learning. Council of the European Union.
Retrieved from https://publications.europa.eu/en/publication-detail/-/publication/ceead970-
518f-11e7-a5ca-01aa75ed71a1/language-en
Currion, P. (2016, July 11). Slave to the algorithm. Retrieved 28 August 2018, from
http://www.irinnews.org/opinion/2016/07/11/slave-algorithm
Derick, L., Sedrakyan, G., Munoz-Merino, P. J., Delgado Kloos, C., & Verbert, K. (2017). Evaluating
emotion visualizations using AffectVis, an affect-aware dashboard for students. Journal of
Research in Innovative Teaching & Learning, 10(2), 107–125. https://doi.org/10.1108/JRIT-05-
2017-0011
Didžiokaitė, G., Saukko, P., & Greiffenhagen, C. (2018). The mundane experience of everyday calorie
trackers: Beyond the metaphor of Quantified Self. New Media & Society, 20(4), 1470–1487.
Drachsler, H., & Greller, W. (2012). The Pulse of Learning Analytics Understandings and Expectations
from the Stakeholders. In Proceedings of the 2Nd International Conference on Learning
Analytics and Knowledge (pp. 120–129). New York, NY, USA: ACM.
https://doi.org/10.1145/2330601.2330634
Drachsler, H., & Greller, W. (2016). Privacy and Analytics: It’s a DELICATE Issue a Checklist for Trusted
Learning Analytics. In Proceedings of the Sixth International Conference on Learning Analytics
& Knowledge (pp. 89–98). New York, NY, USA: ACM.
Duval, E. (2011). Attention please!: learning analytics for visualization and recommendation. In
Proceedings of the 1st international conference on learning analytics and knowledge (pp. 9–
17). ACM.
Eriksson, T., Adawi, T., & Stöhr, C. (2017). “Time is the bottleneck”: a qualitative study exploring why
learners drop out of MOOCs. Journal of Computing in Higher Education, 29(1), 133–146.
https://doi.org/10.1007/s12528-016-9127-8
European Commission. (2005). Descriptors defining levels in the European Qualifications Framework
(EQF). Retrieved 21 June 2018, from https://ec.europa.eu/ploteus/en/content/descriptors-
page
European Commission. (2008). Explaining the European Qualifcations Framework for Lifelong Learning
(p. 16). Retrieved from https://ec.europa.eu/ploteus/sites/eac-eqf/files/brochexp_en.pdf
European Commission. (2016). A New Skills Agenda for Europe. Retrieved from
https://ec.europa.eu/transparency/regdoc/rep/1/2016/EN/1-2016-381-EN-F1-1.PDF
European Commission. (2018a). Digital competences and technology in education. Retrieved 10
August 2018, from https://ec.europa.eu/education/policy/strategic-framework/education-
technology_en
European Commission. (2018b). Digital Education Action Plan.pdf. Retrieved 10 August 2018, from
https://ec.europa.eu/education/sites/education/files/digital-education-action-plan.pdf
European Journalism Centre, & Open Knowledge Foundation. (2012). The Data Journalism Handbook.
O’Reilly Media. Retrieved from http://datajournalismhandbook.org/1.0/en/

Feldstein, M. (2016). Analytics Literacy is a Major Limiter of Ed Tech Growth. Retrieved 28 August
2018, from https://mfeldstein.com/analytics-literacy-is-a-major-limiter-of-ed-tech-growth/
Ferguson, R., & Shum, S. B. (2012). Social learning analytics: five approaches. In Proceedings of the
2nd International Conference on Learning Analytics and Knowledge - LAK ’12 (p. 23).
Vancouver, British Columbia, Canada: ACM Press. https://doi.org/10.1145/2330601.2330616
Ferrari, A. (2013). DIGCOMP: A Framework for Developing and Understanding Digital Competence in
Europe. Luxembourg: Publications Office of the European Union. Retrieved from
http://publications.jrc.ec.europa.eu/repository/bitstream/JRC83167/lb-na-26035-enn.pdf
Fincham, E., Gasevic, D., Jovanovic, J., & Pardo, A. (2018). From Study Tactics to Learning Strategies:
An Analytical Method for Extracting Interpretable Representations. IEEE Transactions on
Learning Technologies, 1–1. https://doi.org/10.1109/TLT.2018.2823317
Fisher, D., & Meyer, M. (2017). Making Data Visual: A Practical Guide to Using Visualization for
Insight. O’Reilly Media, Inc.
Ganley, D., & Lampe, C. (2009). The ties that bind: Social network principles in online communities.
Decision Support Systems, 47(3), 266–274. https://doi.org/10.1016/j.dss.2009.02.013
Govaerts, S., Verbert, K., Klerkx, J., & Duval, E. (2010). Visualizing activities for self-reflection and
awareness. In International conference on web-based learning (pp. 91–100). Springer.
Greller, W. (2015). Learning Analytics - hype re-labelling of matter. Retrieved 9 July 2018, from
http://www.greller.eu/wordpress/?p=4414
Greller, W., & Drachsler, H. (2012). Translating Learning into Numbers: A Generic Framework for
Learning Analytics. Journal of Educational Technology & Society, 15(3), 42–57.
Greller, W., Ebner, M., & Schön, M. (2014). Learning Analytics: From Theory to Practice – Data
Support for Learning and Teaching. In M. Kalz & E. Ras (Eds.), Computer Assisted Assessment.
Research into E-Assessment (pp. 79–87). Springer International Publishing.
Hernández‐Leo, D., Martinez‐Maldonado, R., Pardo, A., Muñoz‐Cristóbal, J. A., & Rodríguez‐Triana, M.
J. (2018). Analytics for learning design: A layered framework and tools. British Journal of
Educational Technology, 0(0). https://doi.org/10.1111/bjet.12645
Hildebrandt, M. (2017). Learning as a Machine: Crossovers between Humans and Machines. Journal
of Learning Analytics, 4(1), 6–23. https://doi.org/10.18608/jla.2017.41.3
Hoel, T., & Chen, W. (2016). Privacy-driven Design of Learning Analytics Applications – Exploring the
Design Space of Solutions for Data Sharing and Interoperability. Journal of Learning Analytics,
3(1), 139–158. https://doi.org/10.18608/jla.2016.31.9
Ifenthaler, D., Gibson, D., & Dobozy, E. (2018). Informing learning design through analytics: Applying
network graph analysis. Australasian Journal of Educational Technology, 34(2).
https://doi.org/10.14742/ajet.3767
Ifenthaler, D., & Schumacher, C. (2016). Student perceptions of privacy principles for learning
analytics. Educational Technology Research and Development, 64(5), 923–938.
https://doi.org/10.1007/s11423-016-9477-y
JISC. (2015). Code of practice for learning analytics. Retrieved 21 August 2018, from
https://www.jisc.ac.uk/guides/code-of-practice-for-learning-analytics
JISC. (2017). Consent for learning analytics: some practical guidance for institutions | Effective
Learning Analytics. Retrieved 11 July 2018, from

https://analytics.jiscinvolve.org/wp/2017/02/16/consent-for-learning-analytics-some-
practical-guidance-for-institutions/
Khalil, M., & Ebner, M. (2016). Learning Analytics in MOOCs: Can Data Improve Students Retention
and Learning? (pp. 581–588). Presented at the EdMedia + Innovate Learning, Association for
the Advancement of Computing in Education (AACE). Retrieved from
https://www.learntechlib.org/primary/p/173003/
Kitto, K., Lupton, M., Davis, K., & Waters, Z. (2017). Designing for student-facing learning analytics.
Australasian Journal of Educational Technology. https://doi.org/10.14742/ajet.3607
Light, D., Wexler, D., & Heinze, J. (2004). How Practitioners Interpret and Link Data to Instruction:
Research Findings on New York City Schools’ Implementation of the Grow Network, 22.
Lingard, M. (2016). Learning Analytics Policy, 10.
Lockyer, L., Heathcote, E., & Dawson, S. (2013). Informing Pedagogical Action: Aligning Learning
Analytics With Learning Design. American Behavioral Scientist, 57(10), 1439–1459.
Mandinach, E. B., Honey, M., & Light, D. (2006). A Theoretical Framework for Data-Driven Decision
Making. Presented at the AERA, April 9, 2006, San Francisco.
Mandinach, E. B., Parton, B. M., Gummer, E. S., & Anderson, R. (2015). Ethical and appropriate data
use requires data literacy. Phi Delta Kappan, 96(5), 25–28.
Mangaroska, K., & Giannakos, M. N. (2018). Learning analytics for learning design: A systematic
literature review of analytics-driven design to enhance learning. IEEE Transactions on
Learning Technologies, 1–1. https://doi.org/10.1109/TLT.2018.2868673
Matthews, P. (2016). Data Literacy Conceptions, Community Capabilities. The Journal of Community
Informatics, 12/3(Special Issue on Data Literacy), 47–56.
Mittelmeier, J., Edwards, R. L., Davis, S. K., Nguyen, Q., Murphy, V. L., Brummer, L., & Rienties, B.
(2018). ‘A double-edged sword. This is powerful but it could be used destructively’:
Perspectives of early career education researchers on learning analytics. Frontline Learning
Research, 20–38. https://doi.org/10.14786/flr.v6i2.348
Munzner, T. (2014). Visualization Analysis and Design. A K Peters/CRC Press.
https://doi.org/10.1201/b17511
New Media Consortium. (2011). NMC Horizon Report: 2011 Higher Ed Edition. NMC.
Nückles, M., Schwonke, R., Berthold, K., & Renkl, A. (2004). The use of public learning diaries in
blended learning. Journal of Educational Media, 29(1), 49–66.
Omidyar Network, & Institute of the Future. (2018). A Guide to Anticipating the Future Impact of
Today’s Technology. Retrieved from https://ethicalos.org/wp-
content/uploads/2018/08/Ethical-OS-Toolkit.pdf
Open University. (2014). Student Policies and Regulations - Open University. Retrieved 11 July 2018,
from https://help.open.ac.uk/documents/policies/ethical-use-of-student-data
Prinsloo, P., & Slade, S. (2016). Student Vulnerability, Agency and Learning Analytics: An Exploration.
Journal of Learning Analytics, 3(1), 159–182. https://doi.org/10.18608/jla.2016.31.10

Prinsloo, P., & Slade, S. (2017). An elephant in the learning analytics room: the obligation to act. In
LAK ’17 Proceedings of the Seventh International Learning Analytics & Knowledge Conference.
New York, NY: ACM. Retrieved from http://oro.open.ac.uk/48944/
Prinsloo, P., Slade, S., & Galpin, F. (2011). A phenomenographic analysis of student reflections in
online learning diaries. Open Learning: The Journal of Open, Distance and e-Learning, 26(1),
27–38. https://doi.org/10.1080/02680513.2011.538562
Redecker, C. (2017, November 20). European Framework for the Digital Competence of Educators:
DigCompEdu. Retrieved 10 August 2018, from https://ec.europa.eu/jrc/en/publication/eur-
scientific-and-technical-research-reports/european-framework-digital-competence-
educators-digcompedu
Reimann, P. (2016). Connecting learning analytics with learning research: the role of design-based
research. Learning: Research and Practice, 2(2), 130–142.
https://doi.org/10.1080/23735082.2016.1210198
Ridsdale, C., Rothwell, J., Smit, M., Ali-Hassan, H., Bliemel, M., Irvine, D., … Wuetherick, B. (2015).
Strategies and Best Practices for Data Literacy Education (Knowledge Synthesis Report).
Dalhousie University. Retrieved from http://hdl.handle.net/10222/64578
Santos, J. L., Govaerts, S., Verbert, K., & Duval, E. (2012). Goal-oriented visualizations of activity
tracking: a case study with engineering students. In Proceedings of the 2nd international
conference on learning analytics and knowledge (pp. 143–152). Vancouver, Canada: ACM.
Retrieved from https://dl.acm.org/citation.cfm?id=2330639
Schmitz, M., Scheffel, M., Limbeek, E. van, Halem, N. van, Cornelisz, I., Klaveren, C. P. B. J. van, …
Drachsler, H. (2018). “Investigating the relationships between online activity, learning
strategies and grades to create learning analytics-supported learning designs.” Lifelong
Technology Enhanced Learning: Dealing with the Complexity of 21st Century Challenges: ,
Proceedings 13th European Conference on Technology Enhanced Learning, EC-TEL 2018.
Retrieved from https://research.vu.nl/en/publications/investigating-the-relationships-
between-online-activity-learning-
Schmitz, M., Scheffel, M., van Limbeek-Reinaerts, E., Bemelmans, R., & Drachsler, H. (2018). ‘Make it
personal!’ - Gathering input from stakeholders for a Learning Analytics-supported Learning
Design tool. Presented at the ECTEL 2018, Leeds.
Schmitz, M., Van Limbeek, E., Greller, W., Sloep, P., & Drachsler, H. (2017). Opportunities and
challenges in using Learning Analytics in Learning Design. In European Conference on
Technology Enhanced Learning (pp. 209–223). Springer.
Sclater, N. (2014). Code of practice for learning analytics. JISC, 64.
Sclater, N., & Bailey, P. (2015). Code of practice for learning analytics. JISC. Retrieved from
https://www.jisc.ac.uk/guides/code-of-practice-for-learning-analytics
Sedrakyan, G., Malmberg, J., Verbert, K., Järvelä, S., & Kirschner, P. A. (2018). Linking learning
behavior analytics and learning science concepts: Designing a learning analytics dashboard for
feedback to support learning regulation. Computers in Human Behavior.
https://doi.org/10.1016/j.chb.2018.05.004
Shum, S. B., & Crick, R. D. (2012). Learning Dispositions and Transferable Competencies: Pedagogy,
Modelling and Learning Analytics, 11.

Shum, S. B., & Ferguson, R. (2012). Social Learning Analytics. Journal of Educational Technology &
Society, 15(3), 3–26.
Siadaty, M., Gašević, D., Jovanović, J., Milikić, N., Jeremić, Z., Ali, L., … Hatala, M. (2012). Learn-B: a
social analytics-enabled tool for self-regulated workplace learning. In Proceedings of the 2nd
International Conference on Learning Analytics and Knowledge - LAK ’12 (p. 115). Vancouver,
British Columbia, Canada: ACM Press. https://doi.org/10.1145/2330601.2330632
Steiner, C. M., Kickmeier-Rust, M. D., & Albert, D. (2016). LEA in Private: A Privacy and Data Protection
Framework for a Learning Analytics Toolbox. Journal of Learning Analytics, 3(1), 66–90.
https://doi.org/10.18608/jla.2016.31.5
Tempelaar, Rienties, & Nguyen. (2017). Adding dispositions to create pedagogy-based Learning
Analytics. Zeitschrift Für Hochschulentwicklung, 12(1). https://doi.org/10.3217/zfhe-12-01/02
Tsai, Y.-S., & Gasevic, D. (2017). Learning Analytics in Higher Education — Challenges and Policies: A
Review of Eight Learning Analytics Policies. In Proceedings of the Seventh International
Learning Analytics & Knowledge Conference (pp. 233–242). New York, NY, USA: ACM.
Tsai, Y.-S., Scheffel, M., & Gasevic, D. (2018). Enabling Systematic Adoption of LA through a Policy
Framework. Presented at the ECTEL 2018, Leeds.
Van Harmelen, M., & Workman, D. (2012). Analytics for learning and teaching. CETIS Analytics Series,
1(3), 1–40.
Vieira, C., Parsons, P., & Byrd, V. (2018). Visual learning analytics of educational data: A systematic
literature review and research agenda. Computers & Education, 122, 119–135.
https://doi.org/10.1016/j.compedu.2018.03.018
Vuorikari, R., Punie, Y., Carretero, S., & Brande, L. V. den. (2016). DigComp 2.0: the digital competence
framework for citizens. Luxembourg: Publications Office. Retrieved from
http://dx.publications.europa.eu/10.2791/11517
Wilson, A., Watson, C., Thompson, T. L., Drew, V., & Doyle, S. (2017). Learning analytics: challenges
and limitations. Teaching in Higher Education, 22(8), 991–1007.
https://doi.org/10.1080/13562517.2017.1332026
Winne, P. H. (2010). Bootstrapping learner’s self-regulated learning. Psychological Test and
Assessment Modeling, 52, 472–490.
Winne, P. H. (2017). Learning Analytics for Self-Regulated Learning. In Handbook of Learning Analytics
(pp. 241–249). Society for Learning Analytics Research. Retrieved from
https://solaresearch.org/hla-17/hla17-chapter21/
Winne, P. H., & Hadwin, A. F. (2008). The weave of motivation and self-regulated learning. The Weave
of Motivation and Self-Regulated Learning, 2, 297–314.
Wolf, G. (2009). Know thyself: Tracking every facet of life, from sleep to mood to pain 24/7/365.
Wired, 92–95.
Wolpers, M., Najjar, J., Verbert, K., & Duval, E. (2007). Tracking Actual Usage: the Attention Metadata
Approach. Journal of Educational Technology & Society, 10(3), 106–121.
Wong, B. T. M. (2017). Learning analytics in higher education: an analysis of case studies. Asian
Association of Open Universities Journal, 12(1), 21–40. https://doi.org/10.1108/AAOUJ-01-
2017-0009

Zimmerman, B. J. (1989). A social cognitive view of self-regulated academic learning. Journal of
Educational Psychology, 81(3), 329.
Zimmerman, B. J. (2002). Becoming a Self-Regulated Learner: An Overview. Theory Into Practice,
41(2), 64–70. https://doi.org/10.1207/s15430421tip4102_2
Zimmerman, C. (2007). The development of scientific thinking skills in elementary and middle school.
Developmental Review, 27(2), 172–223. https://doi.org/10.1016/j.dr.2006.12.001

Personalised Evidence Based Practice Framework

More Related Content

What's hot (11)

Similar to Personalised Evidence Based Practice Framework (20)

More from slideshowproject (7)

Recently uploaded (20)

Personalised Evidence Based Practice Framework