5 manieren waarop de Open Onderwijs API kan worden ingezet voor learning analytics

Excerpt blogpost Tom Kuipers op de innovatieblog van SURF

“Kan de Open Onderwijs API (OOAPI) worden uitgebreid met gegevens voor learning analytics? Die vraag stelde SURFnet aan Tom Kuipers, ontwikkelaar afdeling Onderwijs en Onderzoek, ICT Services van de UvA. Hij ontdekte 5 manieren waarop de OOAPI van dienst kan zijn voor learning analytics.

De Open Onderwijs API is een standaard voor het delen van onderwijsdata. Met behulp van de OOAPI kunnen gegevens zoals cijfers en roosters makkelijk worden ontsloten in bijvoorbeeld een studentenapp. Vanuit het project Learning Analytics, onderdeel van het innovatieprogrammaOnderwijs op maat, kwam het verzoek om te bekijken hoe de Open Onderwijs API kan worden uitgebreid met learning analytics gegevens. Gegevens over studievoortgang zouden daarmee onderdeel kunnen worden van apps.”





EU commission report on ‘New modes of learning and teaching in higher education’

The following report is relevant for large scale deployment of Learning Analytics within Europe:


Recommendation 14
Member States should ensure that legal frameworks allow higher education institutions to collect and analyse learning data. The full and informed consent of students must be a requirement and the data should only be used for educational purposes.

Recommendation 15

Online platforms should inform users about their privacy and data
protection policy in a clear and understandable way. Individuals should
always have the choice to anonymise their data

Reporting back on OWD 2014

This is a report back about at the Onderwijsdagen in specific about Learning Analytics.

I should introduce myself. Alan Berg Co-Chair of the Special Interest Group at SURF for Learning Analytics and member of the Innovation Work Group.

OWD are three days of events organized with the support of SURF (http://www.deonderwijsdagen.nl/), reaching a wide national audience. SURF actively busy in a number of themes and their intersections. Learning Analytics being one theme. The example intersection being the relationship between LA and Open Online Education.

Before discussing one of my domains of interest, let’s point at a few excellent presentations captured on video. In no particular order, these are:

Jason Ohler, has a clear perspective that you almost plan innovative new services around. Sharon Klinkenberg works at UvA and is an expert in his Field. His knowledge and evangelism radiate outwards. Pero de Bruyckere is well researched, humorous and a sanity check on our own perspectives.

With my SIG LA hat and SURF supplied T-shirt on I presented twice:

The first subject was ‘Grand challenges’ voor learning analytics en open en online onderwijs. The presentation description was: In deze sessie vertellen Alan Berg en Robert Schuwer over (een inventarisatie van) de ‘grand challenges’ die er zijn op het snijvlak van learning analytics en open en online onderwijs. In het kader van het door OCW ondersteunde programma Open en online onderwijs gaat SURF de komende maanden een aantal nieuwe activiteiten op dit gebied oppakken. Een van de eerste onderwerpen is de verbinding tussen learning analytics en open en online onderwijs. 

De eerste stap is het in kaart brengen van uitdagingen voor open en online onderwijs waarvoor learning analytics mogelijk oplossingen kan bieden, en van uitdagingen voor learning analytics waarbij vormen van open en online onderwijs voor antwoorden kunnen zorgen. In deze sessie presenteren Alan Berg en Robert Schuwer de eerste resultaten van een brede inventarisatie van deze ‘grand challenges’. Er is gelegenheid voor eerste reacties en aanvullende ideeën. 

I enjoyed the interaction with Robert Schuwer, it is relaxing to share a podium with someone who can pick up questions outside my main area’s of expertise. The room was filled and the audience was active. From the presentation, it was clear that the LA community is more research orientated than the OOO community. Within the LA community the trend is towards operationalization through services, but clearly we are in general behind in the service lifecycle race. It is also clear that the LA community has clearer short term challenges. For example, sorting out the ethics and privacy policies at the national level and how do we benchmark and validate services so that we can be sure of the products relative value to the Educational market.

A plug here for the need for a national decision tree (as suggested by Stefan Mol from UvA), where we can consistently find replies for any ethical, privacy and legal issues as we explore LA services.

The second presentation was Ethics & Privacy issues in the context of Learning Analytics with Maren Scheffel. The presentation description was:

During this session, participants will gain awareness of legal and ethical requirements for Learning Analytics tools and potential solutions. Participants will learn to properly address those requirements when implementing Learning Analytics in education.

The massive production, collection and processing of information from various learning platforms and online environments used have led to ethical and privacy concerns regarding individual and societal harms. Previously these types of concerns have impacted on areas as diverse as computer science, legal studies and surveillance studies. Within the SURF SIG Learning Analytics and the FP7 EU LACE project we conducted a series of workshops about issues and research challenges on ethical and privacy aspects in the emerging field of Learning Analytics. We  had asked for short descriptions of issues faced by Learning Analytics stakehoders. We organized an Ethics & Privacy Expert session with international experts in the field to find proper answers. 

Within this session we will present the initial insights gained and open the floor to the audience to contribute additional aspects from their organisations. For more details on the LACE community visit: http://www.laceproject.eu/lace/. For a report back from one of the workshops review: http://www.laceproject.eu/blog/learning-analytics-dilemmas-ethics-trusting-google-apps-teaching.

Maren, took the main load of presenting and fearlessly stepped into the philosophical shadows and grey area’s of ethics and privacy. +1 for the success of her mission.

The first and foremost lesson from the presentation was that the audience/ crowd (http://en.wikipedia.org/wiki/The_Wisdom_of_Crowds) is wise and should contribute to forming policy. The question therefore is how to make contact and pull further into the feedback process. For the ethics and privacy workshop the results were stored in an Open Book. We should moderate the modification of the open book by the crowd. The book acting as an oracle for policy debate.

The second interrelated theme was the bugger is in the detail. Lot’s of gaps and complexity. No one person can cover all the potential for inconsistency or incorrectness. We need to shine a light in this area with all the stakeholders and balance their needs.

The third theme was that Privacy by design should be a driving principle for any large scale deployments of LA (http://www.privacybydesign.ca/index.php/privacy-policies-enough-need-software-transparency/). One can even argue that this is a conversation that we should widen to our cloud services and the moral responsibilities we have as suggested by the themes behind Privacy laws.

Sadly, I did not have time to view other LA presentations, such as:

  • Blended learning als manier om studiesucces en studierendement te verhogen
  • Aspecten van learning analytics bij een digitale leeromgeving

If you read this and were in the presentation I hope I can motivate you to blog here…

Looking forward to OWD 2015 and seeing how far we LA progressed.

Data Driven Education (Quantified Education)

William Gibson maakte de opmerking, “the future is already here, it’s just not evenly distributed.”  Deze quote is de essentie van het huidige debat over de impact van de digitale wereld in het (hoger) onderwijs. Verongelijkte (onderwijs) wetenschappers die niet-erkend worden door de self-made onderwijsvernieuwers en entrepreneurs die de wereld wel even zouden veranderen. Alle ingrediënten voor een daadwerkelijke innovatie zijn aanwezig om de toekomst die zich om ons heen ontvouwt in het onderwijs te trekken.

In een presentatie uit 2007 sprak ik over de kansen, uitdagingen en bedreigingen voor de belofte van “jawel” herbruikbare leerobjecten. De beschreven conclusies n.a.v. een onderzoek van de Digitale Universiteit  zijn nog steeds herbruikbaar en herkenbaar.  Open educational resources en in hun kielzog MOOC’s hebben de plek in de spotlight overgenomen soms op hetzelfde platgetreden pad zonder nieuwe oplossingen en inzichten te bieden.

Er lijkt niets verandert en toch is alles veranderd. 

Mijn stelling is dat de werkelijke uitdaging voor Open Onderwijs niet de inpassing van Open Onderwijs in het huidige onderwijsmodel is. Durf en visie zijn nodig om een oude belofte in te lossen. Dat is de belofte van een sterk gepersonaliseerde leerervaring, in een maatschappelijk zinvolle context, uitgaande van persoonlijke leerpaden en een relevant gepersonificeerd curriculum Het is mijn betoog dat enkele essentiële voorwaarden zijn ingevuld om deze belofte daadwerkelijk gestand te doen.

De twee belangrijkste maatschappelijk technologische ontwikkelingen in de wereld om deze visie vorm te geven zijn enerzijds, Ubiquitous  Information & Technology. Dit zit ondertussen in de haarvaten van de wereld om ons heen en in alle apparaten die wij dagelijks gebruiken om informatie tot ons te nemen, te maken en te verspreiden. Anderzijds de Quantified Society. Niet meer weg te denken is deze drang tot meten die steeds meer aspecten van het persoonlijke leven stuurt. Of het nu Big Data discussies  zijn of de Nike armband die de quantified Self in de persoonlijke levenssfeer brengt. Ik betoog dat dit moet leiden tot een vorm van Quantified Education, onderwijs op maat en in de juiste fase van de persoonlijke ontwikkeling.

Het onderwijs is de ultieme kraamkamer  om deze nieuwe technologieën uit te proberen en de werkelijke uitdaging is welk onderwijs genoten kan worden en waarom. Is het slechts substitutie, dwz de technologie wordt ingezet om de bestaande praktijk te optimaliseren of leggen we de lat daadwerkelijk hoger en gaan we de uitdaging aan om het onderwijs te transformeren en te onderzoeken welke nieuwe vormen van persoonlijk online open onderwijs, Data Driven Education, mogelijk zijn.

Towards a uniform code of ethics and practices for Learning Analytics

Originally posted @ the SURF Special Interest Group Learning Analytics’s site by Alan Berg


In this article, I argue that we need a clearly defined, uniformly enforced sector wide code of ethics. Without the code of ethics and the associated practices we are vulnerable to external pressures. We may well react ad-hoc per institution to external pressures undermining consistency of practice, risking decreasing overall fairness.

Luckily, the Learning Analytics field is not fully matured. In general, we in Holland are starting to explore the possibilities. Fully mature campus wide deployments are going to slowly emerge over a number of years. We have time to discuss, evolve and generally tinker with a code of ethics and the derived practices.

The law tells us what we are allowed to do without legal consequences. Ethics is the true measure of which actions are morally allowable. It may be legal to read most everyone’s emails, Google docs, listen in on their Skype conversations and lie to congress. It may well be legal to filter information to support agenda’s, obscuring truth.  It may well be legal for the BND or other national secret services to share massive amounts of data with the NSA and lets not forget about the dangers of large-scale Internet surveillance and its negative impact on cloud economics. The Guardian Newspaper is doing a great job at publicizing many aspects of this evolving scandal.

A well-formed ethical code leaves little room for ambiguity and that is what the Learning Analytics field needs. When asking tough questions about our methodologies and comparing costs of practices, the first question to answer is, is this going to be ethical?

If we don’t have a unified code of ethics then the answer per University may well be different. As individual organizations we might react differently to external pressures. Higher-level management might reasonably have a different viewpoint than the practitioner in the field. Who is the final arbitrator of action? I would argue we should use a properly debated code of ethics and associated practices as that arbitrator. Not so important for Proof Of Concepts. However, an ethical code is much more important as we start to aggregate data sources. These resources can potentially be used as part of mass surveillance or influencing employment opportunities.

Where are we with our code of ethics?  Two recent papers worth reading are:

Learning Analytics: Ethical Issues and Dilemmas by Sharon Slade and Paul Prinsloo and an evaluation of policy frameworksfor addressing ethical consideration in learning analytics by the same authors.

The ethical issues paper states: “This research indicates that some higher education institutions’ policy frameworks may no longer be sufficient to address the ethical issues in realising the potential of learning analytics.”  This implies a potential of uneven action between different institutions if asked the same practical questions.

The second paper outlines a number of possible principles including transparency. This paper is a great point of departure for a code of ethics.

There are already laws that can help guide our code of practices. For example, laws reinforcing cloud ethics for the legal profession. We can use these types of already predigested practices to strengthen and clearly define our uniformly applied code of practices.

Spurred forward by the NSA scandal, the EU has plans to implement stronger data protection act with significantly greater enforcement around May in 2014. However,  the deadline feels ambitious to this author because there are so many interested parties and pressures involved.

Potentially, the best method for defining a relevant ethical code of practice is to measure against emerging contemporary questions. Here are some examples of questions that our sector collectively needs to ask and then measure:

Scope: If we are asked by external party to provide historic data about student activity.  What level of granularity is acceptable: At the user level, all the peope that might know the user level  or all the data? In other words, have we an ethical urge to ignore bulk surveillance requests.

Secrecy: If we are asked or expect that information be passed on to a third party. Should we make this clear to our students even if laws may forbid us to do so?

Practicalities: What are the conditions under which we are allowed to use cloud services, services that may well be part of a mass surveillance program? What are the implications for the use of social media

Terms and conditions: Should our emerging understanding of mass surveillance change the terms and conditions of our LA projects? Do we have to communicate to our audience the practices that we use and their inherent weakness, given the current state of technology?

Changing awareness: As new knowledge emerges, with potentially awkward implications for specific deployments. Are we going to be forced to react during the projects and at what cost?

The value of large datasets: If we discover we can do greater good by merging large datasets, for example help improving study success for migrant populations.  Are we ethically forced to work across institutions or are the increased privacy risks a greater ethical issue?

No doubt, I  have missed relevant questions. Please add to the list, comment, discuss and react with your own blogs. Do you think a sector wide ethical code of practice is needed or do you think institutional codes of practices, a bottom up approach is more viable? Let your voices be heard.