Posts | Prof. Dr. Marco Kalz

Mobiltelefone aus Schulen verbannen

Tue, 18 Mar 2025 09:36:59 +0000

Angesicht einer Interviewanfrage und der andauernden Diskussion zum Smartphone-Verbot an Schulen habe ich mich gefragt, auf welcher wissenschaftlichen Basis diese Diskussion eigentlich stattfindet. Ich verfolge die oft polemische Diskussion schon einige Zeit auf LinkedIn und dort wird wahlweise Stimmung gemacht gegen Tech-Konzerne, Medienpädagogen oder auch die Politik. Vorab frage ich mich zunächst, was man denn überhaupt regulieren möchte und ich will dabei die folgenden Kontexte unterscheiden:

Verbot einer Smartphonenutzung durch Schüler:innen im Unterricht (nicht für Unterrichtszwecke): Dieser Fall ist aus meiner Sicht absolut deutlich und es erklärt sich, dass die private Smartphonenutzung im Unterricht verboten werden sollte. Hier spielen vor allem Ablenkungseffekte und die Störung der Konzentration eine wichtige Rolle. Diese Regulierung wird oft auf Schulebene oder durch Lehrer:innen durchgeführt.
Verbot einer Smartphonenutzung für Unterrichtszwecke: Die privaten Endgeräte können natürlich auch für Unterrichtszwecke zum Einsatz kommen. Dies kann der Fall sein, wenn z.B. zu wenige Endgeräte für Schüler:innen verfügbar sind oder aber wenn z.B. Social Media Kanäle der Schüler:innen als Unterrichtsinhalte thematisiert werden. Eine Regulierung wäre in diesem Fall aus einer pädagogisch-didaktischen Perspektive nicht sinnvoll und würde vor allem Lehrer:innen in ihrem Medieneinsatz im Unterricht einschränken.
Verbot einer Smartphonenutzung außerhalb von Unterrichtszeiten (z.B. in Pausen): Hier gibt es unterschiedliche Ansätze an Schulen. Zum Teil wird hier unterschieden zwischen Altersstufen, zum Teil gibt es ein Smartphoneverbot, welches generell in der Schulzeit gilt. Diese Freiheit sollten Schulen haben, je nach Schulprofil oder Entscheidungen der Schulversammlung hier eigenständige Entscheidungen zu treffen.

Wenn ich es richtig verstehe, will man ein generelles Verbot von Smartphones, die von Schüler:innen genutzt werden, diskutieren und dabei auch die Fälle 2 und 3 aus der Autonomie von Schulen entfernen. Solche ein tiefgreifender Eingriff in die Autonomie der Schulen wird ja sicherlich auf Basis aktueller wissenschaftlicher Erkenntnisse zustande kommen. In meiner Recherche bin ich dabei schnell auf ein Interview von Anfang Januar gestoßen, in dem Klaus Zierer im Magazin “Campus Schulmanagement” interviewt wird zu einer aktuellen Studie, die er zusammen mit Tobias Böttger publiziert hat. Dabei wird erwähnt, dass es positive Auswirkungen auf die Lernleistung und das soziale Wohlbefinden gebe, wobei direkt im nächsten Abschnitt nur von geringen Effekten auf die Lernleistung geredet wird. Daher habe ich mir diese Studie einmal genauer angesehen.

Böttger, T., & Zierer, K. (2024). To Ban or Not to Ban? A Rapid Review on the Impact of Smartphone Bans in Schools on Social Well-Being and Academic Performance. Education Sciences, 14(8), 906. https://doi.org/10.3390/educsci14080906

Zunächst fällt mir hier auf, dass diese Publikation in einem Publikationskanal publiziert wurde, der nicht gerade für seriöse Begutachtung und Qualität bekannt ist. Das Finnische Forum für die Klassifikation von Publikationskanälen (JUFO) hat entschieden, dass ab 2025 alle Journals des Verlags MDPI als “graue Literatur” (Level 0) bewertet werden. Aber das es sich hier ja um ein sog. “rapid Review” handelt, war den Autoren die schnelle Publikation hier evtl. wichtiger als die Qualität des Publikationskanals.

Beim Lesen der Publikation fällt als nächstes auf, dass sich das Review auf gerade einmal 5 Studien stützt und man somit nicht von einer breiten Evidenzbasis reden kann. Bei der Einleitung fällt zudem auf, dass hier einseitig Studien mit negativen Effekten zusammengefasst werden, aber die zahlreichen Publikationen aus der Mobile Learning Community zu positiven Effekten nicht erwähnt werden. Die Autoren erklären weiterhin, dass bei einem “Rapid Review” die klassischen Stufen einer systematischen Literaturanalyse im Zuge der Publikationsgeschwindigkeit verkürzt werden. Dabei wird auf die Cochrane-Richtlinien verwiesen. Etwas ungewöhnlich für ein Review ist aus meiner Sicht die Ableitung von Hypothesen, da ja eigentlich das Ziel eines Reviews sein müsste, keine vorgefertigte Annahme zu bestätigen, sondern ein möglichst objektives Bild der Evidenz für oder gegen in diesem Fall Smartphone-Verboten an Schulen zusammenzufassen und zu synthetisieren. Anhand der Einschlusskriterien wurden aus 324 Studien nur 5 Studien ausgewählt, die für das Review geeignet seien. Dabei wurden nur Studien ausgewählt, aus denen sich Effektstärken und Standardabweichungen ableiten ließen. Verschiedene Einschränkungen der Studien werden diskutiert. Bei der Übersicht der Effekte wird deutlich, dass sich die Berechnungen für Effekte auf Lernleistung und auch soziales Wohlbefinden jeweils auf gerade einmal drei Originalstudien stützen. Zudem fällt auf, dass die Effekte auf die Lernleistung zum einen nicht signifikant sind und zum anderen, wenn man mal Hattie als Orientierung nimmt, als Entwicklungseffekt einstufen kann, der zu vernachlässigen ist. In Bezug auf das soziale Wohlempfinden wird von einem signifikanten mittleren Effekt berichtet. Hier ist die Nutzung des Wortes “moderate” fraglich, da die Effektstärke mit d = 0.22 eher als kleiner und nicht als mittlerer Effekt zu interpretieren ist. Auch diesen Effekt würde ich mit Hattie auf jeden Fall nicht als erwünschte Effektstärke einschätzen. Bei der weiteren Analyse des Papers fällt zudem auf, dass die Aggregation von Effekten in beiden Kategorien als zumindest diskutabel einzustufen sind. In der Kategorie “soziales Wohlbefinden” sind Effekte in Bezug auf Bullying gemischt mit Effekten zu Konkurrenz und Zufriedenheit. In der Kategorie Lernleistung werden Ergebnisse eines nationalen Mathematiktests kombiniert mit einem schulischen Abschlusstest und Schulnoten.

Zusammenfassend würde ich hier sagen, dass das Rapid Review keine ausreichenden Basis für die Einführung von Smartphone-Verboten in Schulen bietet. Daher bin ich noch einmal auf die Suche gegangen, um zu schauen, ob es bessere Wissenschaft zu dem Thema gibt. Dabei bin ich auf ein Scoping Review einer australischen Gruppe von Forschenden gestoßen, welches ich im folgenden vorstellen mmöchte.

Campbell, M., Edwards, E. J., Pennell, D., Poed, S., Lister, V., Gillett-Swan, J., … & Nguyen, T. A. (2024). Evidence for and against banning mobile phones in schools: A scoping review. Journal of Psychologists and Counsellors in Schools, 34(3), 242-265. https://doi.org/10.1177/2055636524127

In diesem prä-registriertem Scoping Review einer Forschungsguppe verschiedener autralischer Universitäten wird die wissenschaftliche Evidenz zu Smartphone-Verboten in Bezug auf die Variablen Lernleistung, mentale Gesundheit und Wohlbefinden und Cyberbullying untersucht. Dabei leiten die Autor:innen das Review zunächst mit Verbreitungsdaten von Smartphones ein und diskutieren, dass Smartphone-Verbote oft als politische Reaktionen auf Sorgen in der Bevölkerung installiert werden. Als theoretische Grundlage nehmen die Autor:innen eine systemtheoretische Perspektive ein, in denen Effekte vom Kontext und Hintergrundvariablen abhängen und auch Wechselwirkungen eine Rolle spielen können. Da ich die Beschreibung so passend für den aktuellen Diskurs in Deutschland finde, zitiere ich hier direkt zum politischen Diskurs:

“Community concerns are amplified by the media, creating moral panics about issues for which little-to-no evidence exists – and to which banning then becomes a seemingly necessary and politically popular response” (Campbell et al. 2024, p. 244).

22 Studien wurden am Ende in das Review einbezogen, die in 12 verschiedenen Ländern durchgeführt wurden. In Bezug auf die methodische Qualität stellen die Autor:innen fest, dass es keine rigorosen Studien gibt wie z.B. randomisierte Kontrollstudien oder Vergleichen zwischen der Situation vor und nach der Einführung eines Smartphone-Verbots. Zudem werden Unterschiede der Samples diskutiert. Dabei haben sich einige Studien hautsächlich auf Schüler:innen mit niedrigerem sozio-öknomischen Status konzentriert, was zu einer Verzerrung der Effekte geführt haben könnte. Aus der Synthese dieser 22 Studien leiten die Autor:innen ab, dass es derzeit keine ausreichende Evidenz gibt für positive Effekte eines generellen Smartphone-Verbots auf Lernleistung, mentale Gesundheit und Wohlbefinden oder die Reduktion von Cyberbullying.

Diskussion und Fazit

Beide hier vorgestellen Publikationen sind als Open-Access-Publikationen ohne Bezahlschranken verfügbar und somit relativ leicht zu finden, wenn man sich über Effekte von Verboten von Smartphones in Schulen informieren möchte. Die erste schnelle Zusammenfassung basiert auf sehr wenigen Primärstudien und identifiziert kleine Effektstärken und nicht signifikante Effekte. Die zweite Studie der australischen Forschungsgruppe ist eine sehr gute Grundlage für die Entwicklung einer Evidenzgrundlage für die Entwicklung von Richtlinien. Ein von den Bildungsministerien vorgeschriebenes Verbot von Smartphones in Schulen ist ein tiefgreifender Eingriff in die Realität der Schulen in Deutschland und sollte nur auf Basis deutlicher positiver Effekte auf Lernergebnisse und/oder Wohlbefinden bzw. Reduktion von (Cyber-)Bullying Anwendung finden. Laut den Autor:innen der zweiten Studie gibt es für diese Effekte auf Basis der aktuellen Studienlage derzeit keine Evidenz.

Mir erscheint ein von den Bildungsministern verordnetes Smartphone-Verbot in Schulen zudem auch als ein unglücklicher Eingriff in die Autonomie von Schulen, die ja man andererseits immer weiter fördern möchte. Im Diskurs wird dabei immer wieder auf die schädliche Wirkung von einerseits übermäßiger Nutzung von Smartphones und andererseits der Nutzung von Social Media Kanälen verwiesen. Hier müsste man aus meiner Sicht eine Differenzierung vornehmen: Die Diskussion zur gesundheitsschädlichen Wirkung der übermäßigen Smartphone-Nutzung geht weit über den Schulkontext hinaus. Schulen können hierzu nur beitragen, wenn sie die Smartphone-Nutzung mit den Schüler:innen kritisch reflektieren. Auch die Beeinflussung von jungen Menschen durch Social Media Kanäle, die zum Teile fragwürdige Inhalte für Jugendliche zugänglich machen, sollte nicht im Kontext von Schulen, sondern im Kontext der Regulierung durch die EU diskutiert werden.

Zudem erscheint mir die Diskussion sehr janus-köpfig, wenn einerseits auch von politischer Seite immer wieder die Bedeutung von neuen Technologie und KI und der notwendigen Medienkompetenz betont wird, andererseits aber Objekte der Lebenswelt einfach wegreguliert werden. An einer weiteren Entfernung von Schule und Lebenswelt hat sicherlich niemand Interesse. Daher wäre meine Schlussfolgerung, dass man die Regulierung den Schulen überlassen sollte und den verantwortungsvollen Umgang mit Smartphones als übergreifende Aufgabe in den Lehrplan aufnehmen sollte mit vielen Anknüpfungspunkten zur Medienbildung und politischen Bildung.

Referenzen

Retraction of a plagiarised edtech book. The final chapter.

Wed, 22 May 2024 02:36:59 +0000

This is hopefully the last chapter of a series of activities I have started nearly one year ago. I had reported a case of AI-supported plagiarism in a book on learning analytics which has been published in German last year (see original post here). One of our papers has been plagiarized by a team of authors who have published a book in German on “Educational Data Mining and Learning Analytics”. After months of lack of reaction by the Springer and authors from Dublin City University I have reported the status in this post.

Today, I am happy to report that the book is finally retracted and it even has the following retraction message on each chapter:

“The publisher has withdrawn this volume in agreement with the editors. The volume was produced as part of a series of volumes by the publisher providing an AI-based summary and translation of the current state of the art in the field, but this process was not properly stated in the volume itself. Furthermore, the authors of the papers presented in the volume were not properly acknowledged. The editors were not made aware of these problems and have followed the publisher’s policies and guidelines in good faith. The publisher and editors apologise for any inconvenience caused.”

How did this happen? I had finally contact to someone with sense for research integrity at DCU. Their research integrity officer took action and finally the authors have initiated the retraction of the book despite the ignorant statement by Springer that this is not a case of plagiarism. For weeks the book completely disappeared but since yesterday the retraction message is visible. I am quite sure that Springer would have not taken any action, given that there are still several books of the same production logic available (1, 2, 3, 4, 5, 6, 7). If you look at the editors you might be able to see a pattern.

What can we learn from this? My only take-home message is that we need to develop more awareness about research integrity and the responsibility of each individual author in this process. In times where “content” can be generated by a click, we will see more of this kind of “publications” which in the end diminishes the value of research and researchers. In the same way that OpenAI has ignored the intellectual property of many individual authors large language models have the potential to automate the results of a process of individual discovery, learning and knowledge production. By employing these approaches in our professional context as researchers, we contribute to the further destabilisation of higher education and research.

KI-Campus. Quo vadis?

Thu, 09 May 2024 09:36:59 +0000

Wir (Mandy Schiefner-Rohs und ich mit Team) haben aus einer kleinen Förderung des BMBF im Jahr 2021/2022 ein Kursangebot zur Datenkompetenz von (angehenden) Lehrer:innen erstellt. Dabei haben wir Konzepten zur Ethik von Daten in Bildungskontexten mit anwendungsorientierten Fragen zur Nutzung von Bildungsdaten kombiniert. Über das Projekt Data2Teach konnten wir im September 2022 den Kurs Data2Teach auf dem KI-Campus anbieten. Der Kurs ist sicherlich nicht der bekannteste und begehrteste auf dem Campus, aber wir beobachten stetigen Zuwachs von Einschreibungen und nutzen den Kurs und/oder Anteile in der eigenen Lehre.

Nun erreichte mich vor einigen Wochen die folgende Mitteilung des Teams des KI-Campus: “Das KI-Campus-Team befindet sich derzeit in einem extensiven Reviewprozess, der eine Schärfung unseres Portfolios vorsieht. Nach sorgfältiger Überprüfung von Kennzahlen und Nutzer:innen-Feedback sowie von Kriterien in den Bereichen Didaktik, Lerninhalte und Assessment des Kurses haben wir uns den entschieden, den Kurs zu archivieren.”

Nach nur 20 Monaten endet also das Hosting unseres Kurses, obwohl die offene Bereitstellung des Kursangebotes Teil der Förderbedingungen des BMBF für unser Projekt war. So viel dann zur Plattform für offene Bildung. Persistenz scheint nicht das Konzept des KI-Campus zu sein. Nach einigen E-Mails finden wir heraus, dass die “sorgfältige Überprüfung der Kennzahlen” ein internes Review von 2 Gutacher:innen im Juli 2023, die insgesamt einige Fehler entdeckt haben und die sog. “KPIs” des Campus schlecht bewerteten (Anzahl der Lernenden, Leistungsnachweise etc.. Hier gibt es Hinweise zum Kriterienraster). Leider hat uns dieses Review nie erreicht. Der Grund dafür blieb leider unbeantwortet.

Zudem wird uns mitgeteilt, dass der sog. Qualitätssicherungsprozess damit verbunden ist, dass das HPI Ende 2024 aus der Produktion des Lernmanagementsystems des KI-Campus aussteigt und damit alle Kurse von der Plattform OpenHPI auf Moodle umgezogen werden müssen. Nur um es noch einmal in Erinnerung zu rufen: Das Projekt KI-Campus 1.0 endete zum Dezember 2022 und es wurden bis dahin laut dieser Auskunft des Bundestages 35,98 Mio EUR verausgabt um eine Lernplattform für KI zu entwickeln und nun ist die technische Basis dieser Entwicklung obsolet. Jemand Fragen? Nein, nur Jubel und fröhliche Gesichter auf LinkedIn. Zudem gibt es ja die Folgeförderung und wenn ich es dem obigen Dokument des Bundestages richtig entnehme, noch viel mehr Geld in den Jahren 2023 - 2025 (Ich verstehe aber nicht, was “gebunden” in diesem Zusammenhang bedeutet).

Bei der Entwicklung des Kurses ist mir neben der recht eingeschränkten technischen Funktionalität im Vergleich zu ähnlichen Plattformen (OpenEdX, Moodle) aufgefallen, dass auch unsere Möglichkeit der forschungsbasierten Weiterentwicklung des Kurses durch den KI-Campus sehr eingeschränkt wurden, da wir nur sehr wenige Fragen zur Standard-Evaluation der Plattform zufügen durften. Daher verwundert es mich auch nicht, dass der KI-Campus in seinem Abschlussbericht so gut wie keine Aussagen darüber macht, wer auf dem Campus eigentlich was gelernt hat. Meine Nachfrage dazu auf LinkedIn blieb leider bis heute unbeantwortet. Ich wiederhole die Fragen einmal hier:

Welche Lernziele sind im übergreifenden Kompetenzrahmenwerk zu Futureskills und Datenkompetenz des KI-Campus erreicht worden sind und welche nicht?
Wie sieht die Abbruchquote in den Kursen aus?
Wie viele Zertifikate wurden über den KI-Campus ausgegeben?
In wie weit werden die Zertifikate in den Hochschulen oder bei Arbeitgebern anerkannt?
Welches Niveau von Transfer in die Praxis konnte aus Lernersicht erreicht werden?

Warum ist das wichtig? Weil man die berichteten Kennzahlen für jede x-beliebige Videoplattform hätte nehmen können. Wenn im Ansatz der Evaluation noch noch nicht einmal ein subjektiver Lernzuwachs erfasst wird und diese Daten für den Zwischenbericht auch nicht notwendig sind, dann verstehe ich es nicht. Interessiert den Beirat nicht, wie erfolgreich die Kurse sind, um die entsprechenden Inhalte zu entwickeln und Lernziele zu erreichen? Falls doch, wieso gibt es kein Wort im Abschlussbericht zu Aspekten, die etwas mit dem Lernerfolg zu tun haben (ein Wort, welches im Abschlussbericht nicht vorkommt)?

Apropos Abschlussbericht: Bei einer so hohen Fördersumme wäre in anderen Ländern eine externe Begutachtung unumgänglich gewesen, aber hier scheint es ja auszureichen, ein paar quantitative Zahlen im Sinne des allseits kritisierten MOOC-Reporting zu nennen (Anzahl Accounts, Abrufe) in Kombination mit einer Auflistung von Aktivitäten und einer sog. SWOT-Analyse auf Post-Its, um dann die technischen Entwicklungsleistungen in den Vordergrund zu stellen, die aber ja Ende 2024 zum Teil obsolet sind, wenn man auf Moodle umzieht.

Ich will hier gar nicht die ursprüngliche Entwicklungsleistung schmälern und auch nicht die Geschwindigkeit, mit der der KI-Campus aufgebaut wurde. Und natürlich wurde über die Förderung zahlreiche Kurse entwickelt und es wurden viele weitere Initiativen gestartet. Trotzdem leisten wir uns eine nationale Lernplattform, die nach Jahren hoher Investitionen auf ein Standard-LMS wechselt und keinerlei Daten berichtet zu den Lerneffekten der Kurse.

In Berlin würde man wohl sagen: Nachtigall ik hört Dir trapsen!

The rise of the scinfluencers and the new narrative emerging from pseudo-scientific policy-making

Sat, 16 Mar 2024 09:36:59 +0000

The topic of science communication has gained a lot of attention in the last years and no higher educaton institution has not been involved in concepts related to “transfer” activities or the so called “third mission” of higher education institutions. While I have initially supported these activities, with the demise of social media platforms I have developed more and more doubts about science communication if it is driven by people who have never been involved in scientific work themselves. In the last time I perceive a trend which worries me. Weingart and Guenther (2016) have formulated this worry in the following way:

“science communication has become an arena in which many different stakeholders battle for attention and the power of definition, because there is money in the game, there are jobs to be captured, and there are professional identities at stake”.

A new profile of science-influencers (scinfluencers) has emerged in the last years especially through the more and more important role of policy-making institutions who want to influence the discourse and mission of higher education institutions. A remarkable trait of these roles seems to be that this job profile does not seem to have as prerequisite that the persons involved in policymaking have themselves experienced the complexity of scientific research or development of study programs. Furthermore, I also have the impression that many of these scinfluencers are also not aware of a scientific approach to policy-development, because often the (international) research about a topic is completely or partially ignored.

This leads to discussion papers, research selection procedures and strategic documents which are often not far away from opinions or crowd-sourced conclusions. Sources are combined relatively randomly, the scientific discourse is replaced with the internal discourse of the policy-making institutions and awareness about emerging paradigms but also the value of the historical discourse is low. Due to the centrality of these institutions the key people who are responsible for developing this discourse further have the power to replace the scientific discourse with a pseudo-scientific discourse. And if scienfluencers talk to scinfluencers there is a perpetuating power-narrative and new funding regimes are developed leading to huge financial investments which are not based on scientific findings but political goals.

My conclusion is that higher education institutions need to recognize the danger of influencers of all kinds and better follow a skeptical approach towards policy-making with an unclear interest. Instead I see a lot of institutions who are going where the money is and I am missing a national discourse on potential dangers of policy-making and influence in higher education.

References

Weingart, P., & Guenther, L. (2016). Science communication and the issue of trust. Journal of Science communication, 15(05), C01.

Retraction activities of EdTech journals

Tue, 12 Mar 2024 04:23:30 +0200

Due to my current side-project of getting a plagiarised, AI-generated book retracted, I was curious to see how selected top-ranked journals in Educational Technology have been actively retracting articles. I have been using the GoogleScholar ranking and have focused on journals I would include in this list. Here is a table of my findings.

Journal	Number of retractions	Reasons
Computers and Education	4	Double publication, plagiarism
Education and Information Technologies	13	Plagiarism, Lack of ethics approval documents, authorship sale
British Journal of Educational Technology	3	Plagiarism
International Journal of Educational Technology in Higher Education	0	-
Educational Technology Research and Development	0	-
Interactive Learning Environments	0	-
Journal of Computer Assisted Learning	1	Plagiarism
Journal of Educational Computing Research	0	-
The Internet and Higher Education	1	(Self-)plagiarism/double publication
Australasian Journal of Educational Technology	0	-
International Review of Research in Open and Distributed Learning	0	-
Educational Technology & Society	1	Data manipulation
Learning, Media and Technology	0	-
TechTrends	0	-
Distance Education	0	-
Language Learning and Technology	0	-

What do these numbers tell us? Assuming a well-design process of peer-reviews, quality-assurance and editorial decision-making (which usually needs to be the basis for being incluced in this list) the amount of retractions is quite low if we take in to account the many years of existence and number of papers published in these journals. Of course, there is probably a percentage of papers which have not been idenfified and potentially, these number will be growing due the availability of AI-based writing support.

A surprising finding is the amount of retracted papers of the journal Education and Information Technologies, but there the lack of ethical approval documents seems to lead automatically to retraction. Furthermore, I was surprised to see that besides plagiarism, some papers have been for sale on marketplaces for academic papers.

A plagiarised edtech book and nobody cares

Fri, 08 Mar 2024 06:36:59 +0000

Some of my regular readers might remember that I have reported a case of AI-supported plagiarism in a book on learning analytics which has been published in German last year see original post here. One of our papers has been plagiarized by a team of authors who have published a book in German on “Educational Data Mining and Learning Analytics”. Here I report about what has happened since I have reported the case.

First of all, I would like to publish the material which I have been using as a basis for my complaint against a. the authors, b. the publisher and c. Dublin City University as affiliated institution of the authors and d. peer researchers whose work has been potentially plagiarized.

The first document is a presentation I have given during an online meeting on 20. July 2023 with Springer Nature representatives. It shows the Springer code of conduct for book authors and uses COPE guidelines to evaluate the plagiarism case and provides examples of text from the book chapter. All parties have received this presentation.

Slides from presentation to Springer Nature on 20. July 2023

In addition, I have provided to all parties a full analysis of the text which has been plagiarized on the book chapter. For this I have retranslated the German text with DeepL and marked to similar text passages.

Full analysis of text similarity from original text and translated text

Next I will summarize how each party has reacted to the case.

Communication with the authors. My initial approach was to contact the editors/authors of the book to see how they are seeing their responsibility and breach of publication ethics. Besides that fact that only one of the authors has seen the need to communicate, the only argument was that it was all a project with Springer and that Springer is responsible for the text that has been produced. There was no awareness that publication ethics are an individual responsibility for researchers. The authors have not taken any action to withdraw or correct any word from the book. Furthermore, some authors have listed the chapters a their original publication in GoogleScholar.
Communication with Springer. I have been talking to Springer Nature and the VP content innovation in an online-meeting. This team has accepted all the things which have gone wrong and they promised to work on that. While they initially offered an erratum, the VP Content Innovation has communicated in an E-Mail to me on 18. September 2023 that a retraction process for this book has been initiated. I had also filed an official complaint via the contact form of Springer Nature about the plagiarism and have never received any reply (not even a confirmation of the request!). Since several months had passed without any action I had contacted the VP again and he informed me in an E-Mail on 13 October 2023 that the case was analysed by “research integrity specialists and I can report that after careful consideration they found that there was not sufficient evidence to support a claim of plagiarism by the editors. There was no misappropriation by the editors of intellectual property used in the machine-generated book and there was no absence of attribution to the originators of intellectual property used in the machine-generated book”. That was the end of the communication and there have not been any changes to the book.
Communication with Dublin City University. On 20 September 2023 I have contact the Vice President for Research at Dublin City University and he promised to address this issue directly. After sending a monthly reminder I have only received excuse E-Mails and my latest request for an update has been ignored. DCU has not documented any action after I have reported the case and I have not received any information about the procedure which they are following.
Communication with peers. After documenting the case I have informed other authors via E-Mail and also marked many on LinkedIn to raise awareness about the case. Nobody of the other authors has reacted or got in contact with me.

What is the conclusion of this? My only conclusion is hat nobody really cares about principles of publication ethics and that all rules related to plagiarism can be streched as it pleases publishers and plagiarists. I am deeply disappointend that also nobody of my colleagues cares for this and I can´t understand why.

Using Nextcloud as project management system

Sat, 03 Feb 2024 09:36:59 +0000

Due to lack of good open source project management systems and deficiencies of federal solutions and lack of a local system, I have explored again Nextcloud as an open source option for project management. While an initial search mostly delivers integrations to Open Project into Nextcloud, I had the goal to model a workflow of a project fully in Nextcloud.

My requirements for project-management besides conformity with national data protection rules are the following:

A group function to assign users to groups (members of a project). Groups should be visible in other apps.
A joint file-sharing option with joint editing support. For the editing I want to be able to start documents in the browser for notetaking during meetings and other purposes. Collaborative editing should also work flawlessly.
An option to maintain joint ToDo lists with deadlines and assignment of tasks to individuals.
A calendar to maintain an overview of events and deadlines for a project.
An overview page which integrates all important project information (ideally with joint editing).

The group function is initially a no-brainer since groups can be easily created in the user management and a user can be assigned to different groups. Later in the implementation, a second group-concept called circle comes into the game. The best description of the difference between those concepts is available here and it boils down to the fact that groups should be used for permission issues (top down) and circles can be used for self-organized groups. This can become later confusing, when apps (e.g. Deck or Collectives) rely on circles and not groups.

For task management there are mainly two options: the Tasks app and the Deck app. I have not found an option to maintain tasks and taks categories collaboratively in the Tasks app, so I have used the Deck app for this purpose. The additional advantage here is that task are ordered in cards and that different ways to manage tasks are supported (e.g. Kanban). In Deck I can assign a task to one ore more users and assign a deadline to a task. Deck uses tasks behind the scene so that it is not possible to deactivate tasks and use Deck only. This is another source of confusion for setting-up projects in Nextcloud.

As an additional nice feature, tasks with deadlines are also entered in a specified calendar of a group. A calendar can be easily configured per group and it can be filled with important events and tasks from the Deck app. In addition, this calendar can be subscribed from a local calendar app but not as a simple subscription but rather with setting up your own CalDav-Account as described here. As a nice additional feature, the tasks in this calendar are also included in the reminders app on Mac (and also in my favourite PopDo app).

What is now missing is a joint overview page for orientation which can be edited collaboratively. For this purpose, the app Collectives is very well suited. It is mainly a joint page collection with WYSIWYG-editing and the options to create sub-pages. In theory, this set can be used for the management of small- to medium-scale projects within one organisation.

What makes this scenario a little difficult is the fact that each app assumes a mixed usage of personal knowledge management and group knowledge management. This means that for example a user choses the Deck app, but the has to select the Deck for the group project. This applies to several apps. While it might be a small issue, to be able to constrain some apps within the group would be an advantage.

AI destroys principles of authorship. A scary case from educational technology publishing.

Fri, 15 Sep 2023 06:36:59 +0000

I have long waited before I share a special case of AI generated publishing in the field of educational technology which needs a public reflection and review. Approximately 3 months ago, I have received a citation alert which made me curious. One of our papers has been cited by a team of authors who have published a book in German on “Educational Data Mining and Learning Analytics”. The book has as subtitle “A machine generated overview”. Especially the fact that authors from Ireland have published a book in German and the use of generative AI for the book has made me even more curious.

After an initial reading I was really confused what I am seeing there. First of all, the introduction and much of the text did not really make a lot of sense and the language was not up to scientific standards. Second, I read the part where our paper should have been cited and I was surprised to see that our publication has been somehow mentioned, but that there is no proper citation in the text (but there is one in the references). After reading the text more intensively, which should be a summary of our paper according to the introduction, I had the impression that the text itself is not a summary, but a direct translation of our original publication. This left me really confused and I translated the so-called summary back into English to see that my initial impression was right. A large proportion of text is not summarized but just translated and put into this chapter without quoting the original text. Puzzled by this case of plagiarism I first checked how Springer as a publisher acts in such kinds of cases.

What I found was the Springer Nature Code of Conduct for book authors of which the following principles where especially important for me:

Principle 1: “The submitted work must not contain any plagiarism and should not have been published elsewhere in any form or language”.
Principle 2: “The work of others should always be properly acknowledged…Clarity should be provided on which text is the Authors’ own and which text has been used from others”.
Principle 3: “A basic rule is that if the Author is not the creator of everything in the manuscript, they must get permission from copyright owners or have a valid license to use their content, unless it is ‘fair use / fair dealing’ or in the ‘public domain’”.

With these principles in mind I contacted the authors regarding the breach of these principles and I was also in touch with Springer regarding this project. The authors argue that they are clearly marked as editors of the book and that the way the book has been produced is even mentioned in the title and that Springer would be responsible for the design of the book. When I look at the individual chapter in which our text has been reused, the editors of the book are also shown as authors. The same applies when I look for this specific chapter in GoogleScholar. Furthermore, if the editors are not the authors of the chapters, who is then the author?

So an original text of me and my co-authors is resued without proper quotes and citation plus the text is marked as intellectual property of the new authors and even the book is published under a commercial license although our original text has a Creativecommons license. This is a really scary approach to scientific publishing and the potential future of authorship and publishing in the age of AI and a breach of publication principles. The publisher is currently still working on an erratum focusing on the part where our text has been reused, but I feel responsible to share this case with other researchers (especially the ones whose publications have been also mentioned or reused). If this is the future of academic publishing, I am very sure we should stay away from it. What do you think?

Exploring BlueSky as social network

Sat, 09 Sep 2023 18:23:30 +0200

I have started using Twitter on the 24th of November of 2007 and have stopped posting any content exactly 15 years after. From my perception, the network which is called “X” now has been converted into a right-wing propaganda machine and I cannot understand my colleagues who are still active there. While I shifted some of my writing to LinkedIn, my writing there is longer and less spontaneus (professional pressure probably ;-). Although I did not really miss to post regularly, I have recently explored alternatives approaches like Nostr and the Iris app, but the approach and technology does not seem to be reliable yet. The more I was curious to explore BlueSky which is currently in a public beta for which invite codes are required. Finally, two days ago I have been invited and I had some time to explore BlueSky.

Vision of BlueSky and first impression

After we have seen so many problems with social networking sites (“If something is free, then you are the product!”), first and foremost it is essential to have a look at the organisation behind it and its technical vision and business model. BlueSky has been initially started under Twitter Inc. but after the takeover by EM the project went parted ways and the former founder of Twitter Jack Dorsey has established with others the Bluesky PBLLC, a Public Benefit LLC (see source). There is a background story on the non-profit status here. The potential to be taken over is low from my perspective due to the business model, the formulated vision and the technical basis (an open-source protocol).

After logging in (No, I don´t have an invite code), the setup was very easy and due to the similarity with Twitter, the adoption process to tweet and seach was really fast. A tweet consists of 300 letters and you can add fotos, links and mention others. In addition, you can chose the language in which you are tweeting. So far, so good. After this initial experience, it is important to configure the service according to your needs, which I will explain next.

Adapting personal preferences

The first important configuration option is the configuration of your home feed. For some reasons, there is a threshold of likes defined before a reply to one of your feeds is visible. Also reposts and quotes can be activated to become visible. The second important configuration are the language options for the home feed. You can decide which languages you want to see in the feed. The third and most important configuration are available under the menu-items “Moderation”. Instead of trusting the owners of the platform regarding moderation like on “X”, on BlueSky the user decides which content should be shown. There are for example settings for explicit images, hate and spam and the user can configure what he wants to be shown. I think this is a very good sign and in general, the atmosphere on BlueSky is like in the old days before fake news, truth.social and alike. After this initial setup, one can search for useres of interest. Since it is not yet a public service, there are not yet as many interesting profiles on BlueSky for my domain of interest, but I expect a steady growth. Next I will explain some more advanced features.

Using your own domain

Instead of authenticating against the “bsky.social”-handle on the service, user who have access to their own domain name and the connected DNS-server can just add a TXT-entry and then the handle will be replaced by the individual domain. Here is an article, which describes the process. BlueSky also supports the buying and configuration of domain names soon.

Configuring feeds

The process to configure your own feeds is at the moment still a little bit complicated. I have used the TweetDeck-lookalike “SkyDeck” for this process (see below). Via some combinations of conditions and regular expressions users can create feeds there which are the made public and are available in the app for the public.

Summary

Overall, my impression of BlueSky is very positive and I will develop my account there for some time to make some more experiences, I hope that there will be an individual subscription option, because for an ad-free network, in which I can configure what I see and in which order, I would definetly be willing to pay.

Don´t believe thy hype. 9 problems with the concept of future skills and 21st century skills

Wed, 05 Jul 2023 09:36:59 +0000

I have written at the end of the last semester a critical review of the concept of future-skills in German (preprint available here, still in review) In Germany, there is a plethora of activism around this concept and many higher education institutions feel the need to react to the hype around these skills. This hype even got stronger when ChatGPT got published and now a similar notion under the umbrella of “AI skills” is pushed into the public discourse. I find it really problematic if we start to assume that with every new methodology, technology or important content-area new models or categories of skills are required and I have summarized 9 problems based on earlier published sythesis of evidence of future skills and 21st-century skills:

Lack of integration and relation between concepts

There are many similar concepts like 21st century skills, transversal skills, future skills and alike and the differences and similarities of these concepts have not been referenced. Furthermore, the differences between knowledge, skills and competences are neglected in many publications leading to a lack of integrative power of these concepts (Halász & Michel, 2011).
Lack of model for specific future skills

The specificity and relation between what is called a future skill is unclear and a never-ending list of these skills has been produced. In a recent scoping review Kotsiou, Fajardo-Tovar, Cowhitt, Major and Wegerif (2022) have identified 99 future skill frameworks with 341 skills in total.
Lack of evidence regarding impact on adult outcomes

According to Pellegrino and Hilton (2012) and Lamb, Maire and Doecke (2017) there is not sufficient evidence that what is coined as a future skill or 21st centurity skills has a positive effect on what authors call “adult outcomes”, namely educational success, job success, job satisfaction or civic engagement.
Lack of measurement approaches for future skills

Without more complex approaches of measurement, future skills can harly be measured. Geisinger (2016) and Greiff & Kyllonen (2016) recommend the measurement via open and unstructured problems and triangulated data.
(Implicit) Downgrading of knowledge

By stressing methods like design-thinking and introducing new skill requirements from external sources, the role of knowledge in higher education curricula is downgraded and future skills are often positioned as a second agenda besides the domain-related knowledge building. This can be harmful since there is so much evidence that transfer of knowlegde happens based on the domain expertise of learners. Furthermore, there is research evidence that the downgrading of knowledge is harmful for curricula (Rata, 2012; Young, 2013; Priestley und Sinnema, 2014).
Unclear relation to transfer of learning

Due to the downgrading of domain knowledge there is an unclear relation to initiatives related to near and far transfer of knowledge (Barnett & Ceci, 2002).
Didactical implications

In the discourse it is not clear if initiatives for future skills are related to setting new learning objectives, new learning activities or new didactucal approaches. This differentiation is important for the development of new teaching approaches.
Missing prioritization of skills

The lack of order in existing lists of skills leads to an unclear empirical research basis. Pellegrino & Hilton (2012) argue that research should build on evidence and predictors.
Lack of attention to learning contexts and learning environments

The discourse often neglects the role of (real-world) learning contexts and new learning environments to work on these complex problems.

For the higher education context, the discourse often neglects that academic education is per se directed at the future since the goal is to educate people to contribute to the creation of knowledge. In the article I further argue that it is harmful for higher education institutions if they act in a “vicious circle” in which new skill models for the future are defined without clear evidence that these skills have an impact on adult outcomes. Furthermore I argue that a focus on learning transfer is a much more promisig and evidence-based course of action.

References

Barnett, S. M., & Ceci, S. J. (2002). When and where do we apply what we learn?: A taxonomy for far transfer. Psychological bulletin, 128(4), 612.

Geisinger, K. F. (2016). 21st century skills: What are they and how do we assess them?. Applied measurement in education, 29(4), 245-249.

Greiff, S., & Kyllonen, P. (2016). Contemporary assessment challenges: The measurement of 21st century skills. Applied Measurement in Education, 29(4), 243-244.

Halász, G., & Michel, A. (2011). Key Competences in Europe: interpretation, policy formulation and implementation. European journal of education, 46(3), 289-306. https://doi.org/10.1111/j.1465-3435.2011.01491.x

Kotsiou, A., Fajardo-Tovar, D. D., Cowhitt, T., Major, L., & Wegerif, R. (2022). A scoping review of Future Skills frameworks. Irish Educational Studies, 41(1), 171-186.

Lamb, S., Maire, Q., & Doecke, E. (2017). Key skills for the 21st century: An evidence-based review. Sydney, New South Wales: NSW Department of Education.

Pelegrino, J. W., & Hilton, M. L. (Eds.) (2012). Education for life and work: Developing transferable knowledge and skills in the 21st century. National Academies Press.

Priestley, M., & Sinnema, C. (2014). Downgraded curriculum? An analysis of knowledge in new curricula in Scotland and New Zealand. Curriculum Journal, 25(1), 50-75.

Rata, E. (2012). The politics of knowledge in education. British Educational Research Journal, 38(1), 103-124.

Young, M. (2013). Overcoming the crisis in curriculum theory: A knowledge-based approach. Journal of curriculum studies, 45(2), 101-118.

Uncontrollable GOLLEM-AI and the total decoding and synthesizing of reality

Sat, 15 Apr 2023 08:05:11 +0100

You have read that I am criticizing the discourse on AI here in the last weeks and I have made a call for regulation of AI in education. But I have really underestimated the potential harmful effects which are coming through the uncontrolled release of Generative Large Language Multi-Modal Models (GOLLEMS).

Tristan Harris and Aza Raskin of the Center for Humane Technology are developing in the video below a very well-grounded summary of the catastrophic risks of the current generation of AI on many fundamental aspects of societies and they are right to demand responsibility and a discourse about potential unwanted effects of large powerful tech-companies striving for market domination through a rat-race for the most powerful GOLLEMS.

If you have one hour today, this is the best video currently available on the dangers of the current generation of AI. Watch it and talk about it!

Einsatz von (generativer) KI im Bildungsbereich regulieren

Wed, 05 Apr 2023 04:23:30 +0200

Die Stellungnahme des Ethikrates “Mensch und Maschine – Herausforderungen durch Künstliche Intelligenz“ ist eine sehr ausführliche Diskussion der Entwicklung und dort werden verschiedenen übergreifende Herausforderungen und Problemfelder von KI in verschiedenen gesellschaftlichen Handlungsfeldern diskutiert. Der Unterabschnitt zu KI in der Schule ist mir aber nicht ausführlich genug und es fehlen aus meiner Sicht einige konkrete Probleme. Zudem ergeben sich aus meiner Sicht einige übergreifende Problemfelder, die auch z.B. für den Hochschulkontext gelten und in verschiedenen Richtliniendokumenten der EU oder der UNESCO ausführlich diskutiert wurden.

Gleichzeitig beginnen Bildungsinstitutionen bereits, auf der Basis von aktuellen KI-Diensten eigene Services zu implementieren und anzubieten, ohne dass der Diskurs zu ethischen Dimensionen und potenziellen nachteiligen Auswirkungen bereits ausreichend geführt wurde. Dabei geht es mir nicht um einen Verbot der Nutzung von KI-basierten Diensten, aber eine ausreichende Analyse der Folgen sowie eine transparente und nachvollziehbare Kommunikation sowie die Möglichkeit des Opt-Out. Zudem bin ich davon überzeugt, dass es für den Bildungsbereich wichtig ist, nicht auf propriertäre und kommerziell orientierte Sprachmodelle zu setzen, sondern auf offene und geteilte Sprachmodelle, die auch nachvollziehbar und veränderbar sind. Aus diesem Grund habe ich eine eigene Stellungnahme zum Einsatz von (generativer) KI im Bildungsbereich gestartet, die hier auf Unterstützende wartet.

Redefining my writing workflow

Sat, 25 Mar 2023 09:36:59 +0000

One of the things I have started recently is a book-project. For this purpose I have been working on a new writing workflow which combines reference management, reading, highlighting and note-taking, writing and publishing. In this short post I will try to describe the setup and workflow.

My book-writing workflow in detail

Sources: Of course, the basis of all writing is reading :-). To identify prior work and to get an overview about the domain I have collected approximately 80 sources which I have started to read. These sources are mainly journal articles, some are books and book-chapters. I mostly start with GoogleScholar and combine it with some meta-databases like EBSCOHost, but I also appreciate article newsletters from journals (and can recommend the VUB Paperboy for our domain). The list of potential sources to identify related work is of course endless and I want to try in the future AI services like Researchrabbit or Elicit (which I both got to know through conversations with Yanay Zaguri: Thanks!).
Reference management: I have switched some years ago from Mendely to Zotero and besides a little outdated interface I really appreciate the reliability and extendability of Zotero. I can recommend for example the extension BetterBibTex. The extension takes for example care for renaming PDF and sorting them into a specific folder. In general, Zotero works great in importing references from the web and it can for example also generate references to books based on ISBN-numbers. You can easily collect a nice list of well-formated references.
Reading and structured notetaking: Logseq has really changed the way I read papers and take notes about them. As described in an earlier post about my PKM the PDF-reader and annotation environment is perfect for taking structured notes. The nice thing is, that I can directly access papers from Zotero. I can start a paper from my database and open it on the left side of the screen while I have a chapter structure for my book-projekt as an outline on the right side. Whenever I think something fits into this structure I can either mark and shift pieces of text from left to rigth, or I can note down ideas for chapters.
Writing: My main target format for the book is a well-formatted PDF. Due to this requirement I am using LyX as a visual editor which can produce nicely formatted documents via LaTeX. Again, the interface looks a little outdated, but the functionaliy is really great. LyX can access directly my reference libary from Zotero and has sufficient styling options to produce a nice looking text.
Publishing: As already mentioned, one important output format is a PDF file. This is the standard format of the editor I am using for writing. At the same time, I would like to produce a web-version of the text. For this purpose I have configured LyX to produce MarkDown files from the same text which I can easily integrate into my website which is running on the basis of Wowchemy. Via this route I can easily give people access who can provide feedback to the chapters during the writing process.

Positionality statements in research articles as biaswashing

Sun, 12 Mar 2023 09:36:59 +0000

During a recent review process I have requested from authors that they should address from which standpoint they have approached their qualitative study. While I was referring in this comment the ontological and theoretical levels as formulated by Twining, Heller, Nussbaum & Tsai (2017) in the guidelines for the journal Computers & Education, the authors have understood my request as a call for a „positionality statement“ in which the authors disclose some of their biographical details (white, cis, middle-age) as potentially influencing their research. I was surprised why the authors were assuming that my judgement of their article could be improved by adding some author-related identity information since this is also to some extend in conflict with an anonymous review process. Confused by this experience I have searched further about this practice and I was happy to find a recent article by Savolainen, Casey, McBrayer & Schwerdtle (2023) in which authors perfectly formulate my unease with these kind of statements. The authors summarize their criticism about positionality statements as follows:

Compared to a statement of conflict of interest, there are no clear guidelines with regard to the structure and dimensions which should be included in a positionality statement. This leaves the decision what to reflect on in such a statement fully to the authors of such statements leading again to a potential source of bias.
Dismantling authors positionality leads to an unresolvable conflict to recognize the limits of reflexivity. What was the position from which the positionality statement has been drafted? Can we be sure that there is not positionality factor for positionality statements? This leads to some Escher-esque confusions (whose work will by the way enter the public domain on March 23 this year - yeah!).
Bias can not be resolved by communication of individual characteristics of authors. The scientific process is based on conrol mechanisms and verification which make sure that knowledge can be reproduced and controlled by others without taking into account any personal characteristics. Furthermore, providing details about personal characteristics of authors can again lead more likely to more bias in for example review processes.

Leaving these very good remarks aside, I see the trend for such statements as a superficial approach to resolve biases and lack of fairness purely by statements and not by actions similar to “greenwashing” approaches.

References

Savolainen, J., Casey, P. J., McBrayer, J. P., & Schwerdtle, P. N. (2023). Positionality and Its Problems: Questioning the Value of Reflexivity Statements in Research. Perspectives on Psychological Science, 17456916221144988. https://doi.org/10.1177/1745691622114

Twining, P., Heller, R. S., Nussbaum, M., & Tsai, C. C. (2017). Some guidance on conducting and reporting qualitative studies. Computers & education, 106, A1-A9. https://doi.org/10.1016/j.compedu.2016.12.002

Why AI-based tools are never neutral

Sat, 28 Jan 2023 04:23:30 +0200

„ChatGPT itself is a neutral tool, and how it is used depends on the intentions of those who use it.“ (Cohen, 2023).

This messages is commmunicated by the US-based Association for Supervision and Curriculum Development (ASCD) to its 125 000 members (estimaton) regarding an AI-based tool which is currently pushing the discourse on AI in education. While the one camp is calling out the next revolution for education, it was predictable that the other camp will soon say that technology has no effect itself on education. Both positions are signs of a technological determinism (Chandler, 1995) and both are dangerous as take-home messages for educators. I wish every educator would deal with the phenomenon of tecnological determinism to avoid these easy conclusions.

While these standpoints between the extremes “revolution” and “complete neutrality” are problematic for all learning technologies, for AI-based tools those are especially problematic since their reliance on datasets and algorithms leads to many ethical issues arising before we can even think about implementing them in an educational context. Leslie et al. (2021) speak about “cascading effects” when we take existing inequalities in the world as a starting point to build potentially discriminatory data which lead to biased designs and an increase of existing injustice.

Cascading effects of health inequality and discrimination manifest in the design and use of artificial intelligence (AI) systems (Leslie et al., 2021)

The Independent High-Level Expert Group on Artificial Intelligence (European Commission, 2019) discusses 7 interrelated dimensions which contribute to the establishment of a trustworthy and ethical AI.

Interrelationship of the seven requirements: all are of equal importance, support each other, and should be implemented and evaluated throughout the AI system’s lifecycle (European Commission, 2019)

These ethical concerns have led the European Commission to a proposal to qualify the use of AI in education as a high-risk usage scenario whith implications for their implementation.

„AI systems used in education or vocational training, notably for determining access or assigning persons to educational and vocational training institutions or to evaluate persons on tests as part of or as a precondition for their education should be considered high-risk, since they may determine the educational and professional course of a person’s life and therefore affect their ability to secure their livelihood. When improperly designed and used, such systems may violate the right to education and training as well as the right not to be discriminated against and perpetuate historical patterns of discrimination.“ (European Commission, 2021).

EU Artificial Intelligence Act: Risk levels (Telefonica, 2022)

I hope this makes clear what is at stake if we introduce AI in education and that the ethical dimensions are not comparable to for example those of introducing a learning management system. All in all, technology is never neutral and educators need the knowledge and skills to assess potential risks and ethical implications. For this purpose, existing frameworks for the development and implementation of AI-based tools can play a role (Vakkuri et al., 2021).

References

Chandler, D. (1995). Technological determinis and media determinism. Available under (http://visual-memory.co.uk/daniel/Documents/tecdet/)

European Commission (2019). Ethics Guidelines for trustworthy AI. High-Level Expert Group on Artificial Intelligence. doi:10.2759/346720

European Commission (2021). Proposal for a REGULATION OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL LAYING DOWN HARMONISED RULES ON ARTIFICIAL INTELLIGENCE (ARTIFICIAL INTELLIGENCE ACT) AND AMENDING CERTAIN UNION LEGISLATIVE ACTS.

Leslie, D., Mazumder, A., Peppin, A., Wolters, M. K., & Hagerty, A. (2021). Does “AI” stand for augmenting inequality in the era of covid-19 healthcare?. bmj, 372. doi: https://doi.org/10.1136/bmj.n304

Vakkuri, V., Kemell, K. K., Jantunen, M., Halme, E., & Abrahamsson, P. (2021). ECCOLA—A method for implementing ethically aligned AI systems. Journal of Systems and Software, 182, 111067. https://doi.org/10.1016/j.jss.2021.111067

The downgrading of knowledge by education hipsters

Thu, 26 Jan 2023 04:23:30 +0200

„This has become a fashionable platitude, which…would result in a…content-free curriculum […]. This downgrading of knowledge is, ironies of ironies, to be implemented in the interest of creating a knowledge-based economy.“ (Coffield et al., 2004).

It is a trend in nowadays discussion on education to criticize factual knowledge and the teaching and testing of it in educational institutions. It seems also a natural reaction to follow the general assumption that you do not need to know what can be easily looked up (Let me Google this for you). But this makes only sense if you are not an expert in educational science and pedagogy. It is also nowadays not surprising that even high-level political forums such as the World Economic Forum invites non-experts in education to provide input on the future of educational systems. In this video, the founder of Alibaba recommends that the importance of knowledge will go away in the future and that we should better focus on the arts or dancing. How obscure such a recommendations look like becomes easily visible if you reflect that arts and dancing are creative activities that combine knowledge of diverse types so that these are clearly no separate categories.

In recent times this assumption that we should steer the educational institutions away from the transfer of factual knowledge and replace this by transversal skills and method training is repeated by a group of people who are very present when it comes to criticizing what´s wrong with schools and the educational system in nowadays societies. I call those people “education hipsters” because they are leading a trendy discussion against the mainstream that is from my perspective building on a consciously wrong interpretation of knowledge structures and their connection. But stardom is more important than facts - it fits to the overall approach of the downgrading of knowledge.

If you want to become an expert in anything you need a basic knowledge basis because otherwise you have nothing to apply or no basis to analyze anything. This cannot be replaced by experiential learning or Barcamps or anything else. Knowledge in different forms (factual, conceptual, procedural) is essential to deal with anything on a higher complexity level of any model of cognitive objectives (Krathwohl, 2002). Of course we can all teach design thinking and neglect the content expertise, but how do we decide what´s right or wrong then? By agreement of people who lack content knowledge?

In the recent discourse on ChatGPT you can also see a flavor of this assumption: If an AI will be able to answer factual knowledge questions then teachers should be asked to change their assessment? Seriously? Just because an AI can answer the questions it shifts factual knowledge questions into oblivion? We can of course discuss the issue of relevance or problems of inert knowledge (Whitehead, 1967) and of near and far-transfer of this knowlegde (Barnett & Ceci, 2002) but we should in all cases not contribute to the disdain of individual knowledge construction and since the necessity of factual knowledge is most likely even increasing in a hypercomplex world there will still be a need for assessing this knowledge. Letting learners understand the importance of it is a question of framing and supporting the knowledge acquisition process. If you read about that (factual) knowlegde is not important anymore, you better check the expertise or agenda of the writer.

References

Barnett, S. M., & Ceci, S. J. (2002). When and where do we apply what we learn?: A taxonomy for far transfer. Psychological bulletin, 128(4), 612.

Coffield, F., Moseley, D., Hall, E., Ecclestone, K., Coffield, F., Moseley, D., … & Ecclestone, K. (2004). Learning styles and pedagogy in post-16 learning: A systematic and critical review.

Krathwohl, D. R. (2002). A revision of Bloom’s taxonomy: An overview. Theory into practice, 41(4), 212-218.

Whitehead, A. N. (1967). The aims of education. 1929. Reprint.

This is not a post about ChatGPT but agenda setting

Sat, 21 Jan 2023 06:23:30 +0200

Since I am currently often asked about my opinion about ChatGPT I want to share why I do not write about ChatGPT.

The discussion about ChatGPT (Generative Pretrained Transformer) is at least as annoying as the Wordle wave that has hit my social media channels some years ago. Although we can put both examples under the category of “language games” (maybe in a different sense as Wittgenstein has described them) we can learn some more lessons about the public and partially scientific discourse of ChatGPT compared to the word game.

First of all, a lot of what I read are superficial reflections which seem to forget that AI in education is not a novelty. If you want to learn about the history of these kind of systems I can recommend recent work by Mike Sharples (Sharples, 2022) and bis book “Story Machines” (Sharples & y Pérez, 2022). If you want to announce the next revolution, read this first.
Second, it would be good if users are becoming aware that they are becoming potential co-designers by their input and that feedback on the quality of answers or serious shortcomings is helping to improve the model and in the end the product which will most probably be commercial. It is by the way a special flavour that people are advocating rigorously for open and free learning material and that OER is the only solution for education and do not see the paradox that they are contributing to the promotion of a prototype of what will most probably become a commercial product and competitor for the most used search engine (Microsoft has announced to invest 10 billion US $ into the company behind it, it costs around 3 Million US $ to run the prototype). It is of course absolutely valid if researchers and teachers are “going with the flow” and show their involvement in the so called “hot topics” in educational technology. The problem is, that this discourse is pushing away the real hot topics in education: Access to education, quality of education and inclusion of non-standard learners into educational systems. The problem with these topics is that they are not newsworthy and you cannot play with it. It is so pervasively know that these are very demanding issue for education that nobody will be regarded as being “innovative” or “future oriented” in discussing these topics.

This leads me to my lesson learned from the current discussion on ChatGPT: We can see the appearance and public but also scientific reaction to it as a special case of agenda setting or agenda melding. While agenda setting theory (McCombs, Shaw, & Weaver, 2014) focuses on how (mass) media select topics and transmit them to the public with political or economical objectives in mind, agenda melding refers to the process by which members of the public chose and blend media agendas to fit their individual preferences. By firing the discourse on ChatGPT the educational community is actively pushing the agenda of companies with a commercial agenda and we should reflect if this is fitting to our individual preferences.

References

McCombs, M. E., Shaw, D. L., & Weaver, D. H. (2014). New directions in agenda-setting theory and research. Mass communication and society, 17(6), 781-802.

Sharples, M. (2022). Automated essay writing: an AIED opinion. International Journal of Artificial Intelligence in Education, 32(4), 1119-1126.

Sharples, M., & y Pérez, R. P. (2022). Story Machines: How Computers Have Become Creative Writers. Routledge.

Backup strategy for Knowledge-workers under MacOS

Sun, 28 Aug 2022 09:36:59 +0000

If your productice work is 100% happening in digital environments the secure and reliable backup of data becomes a necessity and data-loss can have serious implications for all kind of processes. In this posting I will briefly describe my approach for the backup of data under the MacOS operating system. Actually, the easy backup and transfer of settings and data is one of the major benefits of working with Mac OS. My backup-strategy has three components.

1.) Hourly backup of core files

Already some years ago I started to store core files and folders in an incremental backup approach each hour. While I used to do this on a local Synology-NAS with BitTorrentSync/Resilio Sync I have nowadays switched to a complete cloud-based approach. For this purpose I use the software Arq (7) in combination with cloudhosting included in a O365 subscription. Arq supports end-to-end encryptions and you can restore files states back in time. After defining the files and folders which should be included in your backup, you do no thave to interfere with the application anymore. Sometimes a re-authentication with the cloud-provider of your choice is required. The Arq symbols signals its activity and by that I can be sure that no progress is lost no matter what will happen. Hourly is not meant literally since I am not online and working all the time. But as soon as I have internet access the adaptes files will be stored.

2.) Weekly full backup

To complement this backup strategy of core files and folders I am aining at having at least one full backup with the help of the built-in TimeMachine backups. For this purpose I connect an external SSD device (Samsung T5 1 TB) via USC-C directly to my machine. This ensures a fast and smooth production of the latest image. At work I have a SSD directly connected to my external monitor so I event do not have to think about connecting the backup device and I have another image I can rely on.

3.) Monthly image for restorability

Last but not least I would like to have another image avilable as a last resort if all of the solutions above will fail and I can restore a fairly recent working state in an acceptable amount of time. For this purpose I am using Carbon Copy Cloner with another external harddrive.

By this strategy, the worst that can happen is that my full backup of applications and settings is around 4 weeks old, but at the same time I have the most recent work backed-up as described in Nr. 1.) so the combination of restoring with TimeMachine or Carbon Coyp Cloner and the current files will lead to an accepable restored working state in a short amount of time.

Window management for professional knowledge workers

Fri, 19 Aug 2022 06:22:31 +0100

Since many years I am using at least one external monitor for work and it has supported my productivity a lot. During an intensive time of online-meetings and online-teaching in the last 3 years I have extended my setup with another external screen so that I have three screens available. On a normal working day I am switching many windows from and between these screens and sometimes I am also disconnecting my machine (a Macbook Air M1) completely from the setup. This means a lot of shuffling and moving of application-windows between the screens. For this purpose I have bought a version of StreamDeck to see if it is possible to define scenarios which allow to automate the switching between different productivity modes easily with an external button.

My setup for the use of three monitors

I could now finally realize this scenario with a relatively easy set of ingredients after a lot of trial-and-error and I would like to share the setup here:

Elgato Stream Deck MK.2 – Studio-Controller with 15 buttons
A small application called Moom which costs around 12 EUR
The Moom profile activator for Stream Deck

After installing all applications including the Stream Deck software for defining the functions of the buttons, I have ordered in Moom now the windows in the different screens according to productivity scenarios: A scenario for teaching and presenting online, a scenario for writing, a scenario for information aquisition and so on. In Moom it is very easy to save these scenarios. After all windows have been positioned, I can just chose “Save Windows Layout Snapshot” and the settings are stored. In the Stream Deck software I can now select the “Moom profile activator” and shift it to one of the buttons. After selecting now the required profile and naming the button accordingly, I can switch and position all application windows with a click of a single button. As a nice extra functionality, Moom allows me also to ignore the windows in the background and to select a standard scenario depending in how many external monitors are connected.

Call for Papers Special Issue for Educational Technology Research & Development (ETR&D)

Thu, 27 Jan 2022 06:22:31 +0100

Download the PDF of the CfP

Special issue editors

Dr. Michael Kerres, University Duisburg-Essen, michael.kerres@uni-due.de, http://kerres.name

Dr. Pavlo Antonenko, University of Florida, p.antonenko@coe.ufl.edu, https://education.ufl.edu/faculty/antonenko-pavlo-pasha/

Dr. Marco Kalz, University of Education, Heidelberg, kalz@ph-heidelberg.de, https://kalz.cc

Important Dates

• March 15, 2022 —outlines of proposed papers due to the editors. Submit in this form

• ~~May 1~~ June 20th, 2022 – Papers due in the Editorial Management system

• ~~July 1~~ September 1st, 2022 — Reviews completed & authors notified of decisions

• ~~September 1~~, November 1st, 2022 — Revised manuscripts due to the Editorial Management system

• ~~November 1~~ January 15th, 2023 — feedback on revised manuscripts

• ~~December 1~~ February 15th, 2023 — final manuscripts due by authors to the Editorial Management system

• ~~January 15~~ March 15th, 2023 – final manuscripts sent to the publishers

• ~~Early~~ Mid 2023 publication

Submission procedure

Authors should submit a 3-page outline proposal including a tentative title, information about contributing author(s), abstract, keywords and key references by March 15, 2022. Early submissions are encouraged. All proposals will undergo a rigorous review of the special issue review board who will recommend full submissions from among the proposals. All full manuscript submissions will undergo rigorous double-blind peer review by at least three reviewers of the special issue review board and regular ETR&D reviewers who will recommend revisions or acceptance. Instructions for submission of full papers are sent after proposal acceptance. Submissions and Questions regarding the special issue should be directed to: etrd@online-campus.net

Background

This Special Issue aims at fostering the discussion on research methodology in research on educational technology. Reeves & Oh (2017) have analyzed research methods of papers published in Educational Technology Research and Development (ETR&D). Most of the studies reported rely on traditional, mostly quantitative research methods, and the authors strongly recommend that the range of methods needs to be broadened. And within the quantitative realm of research methods, Bulfin et al. (2014) have criticized serious shortcomings in the methods applied in Ed Tech research papers: “Data from the survey highlight a preference for relatively basic forms of descriptive research, coupled with a lack of capacity in advanced quantitative data collection and analysis.” Often, studies on educational technology are based on the assumption that a certain technology could improve learning as such – without considering the individual and contextual conditions of learning and the instructional decisions for using an educational technology. Interventions often rely on a comparison of the use of educational technology with a control group of “traditional education” reproducing a paradigm of media comparison studies on a higher level. In a summary on statements from journal editors in the field, they clearly encourage authors to abandon these approaches (Hartsthor, Ferdig, & Bull, 2021, Johnson et al, 2021). More recently, Honebein & Reigeluth (2021) even report a striking increase of such comparative research studies published between 2010 and 2019 in ETR&D. They deliberately urge for research that overcomes these formats that have been criticized by seminal researchers for many years but, obviously, with not much success. The authors argue for research questions that rely on a general framework of instructional theory that take into account the complexity of instructional conditions and decisions and focus on the improvement of educational measures instead of remaining in a comparative paradigm which sets of contexts of learning or modality differences. Reeves & Lin (2021) conclude that “the research we have is not the research we need.” and demand to shift the focus from things (e.g. technologies) to problems to be solved in education (with technology).

The question remains, how these alternative research designs look like. When Kimmons & Johnson (2019) argue for pluralism in Ed Tech research the visibility of innovative approaches in scholarly journals seems sparse. For several years, design-based research (DBR) has been advocated as a new paradigm for educational research. Although often citied, a methodology of DBR (relying on iterations of improvements) still seems to be in need of thorough improvements. Anderson & Shattuck (2012, 16) “conclude that interest in DBR is increasing and that results provide limited evidence for guarded optimism that the methodology is meeting its promised benefits.” The systematic review from Zheng (2015) hints at many shortcomings of DBR studies which mostly rely on a single cycle of optimization (cf. McKenney & Reeves, 2013). Furthermore, current developments in Ed Tech raise the question if approaches for data acquisition and analyses are still suitable for studying emerging technologies and the new research problems related to the digitalization in the various fields of education. Traditionally, Ed Tech research most often relates technology to learning outcome measures neglecting the broad variety of indicators of human action and performance. Also, they focus on the individual’s learning neglecting the social and cultural dimensions of (institutionalized) learning. In the meanwhile, several perspectives in Ed Tech can tentatively be foreseen (Huan, Spector & Yang, 2019). The discussion about post-digital culture (Cramer & Jandric, 2021) informs us that the boundary between digital and analog worlds becomes blurred and seems misleading in our thinking about education. Comparisons of Ed Tech vs. No Tech will not be helpful in future research. Research shifts to the conditions, processes and the implications of technology relating to learning and education. Beyond quantitative research methods, a large variety of qualitative approaches developed in the social sciences still needs to be exploited for Ed Tech research (Willis, 2008). Emergent technologies for learning deliver a broad range of data, providing insights based on objective data, either direct measurement of biological indicators or of observational data (e.g. log files, eye tracking, motion tracking, cf. Gibson & Iffenthaler, 2017). Learning outcomes in technological environments, like maker spaces, relate to digital artifacts that need to be analyzed qualitatively. With the increasing proliferation of Ed Tech in all educational sectors, research is shifting from a (mainly) psychological perspective to a view of digital learning relating to social interactions, learning in and of organizations, institutions, and policy-making. With this, stakeholders and their role in the research process change: They participate as subjects – not objects – of research and reveal practices of sense making in education. Multi-level analyses in education emerge as the use of Ed Tech is understood as part of the society’s adaption to the more comprehensive trend of digitalization. To summarize, we might foresee a shift from subjective to behavioral (objective) data, from learning outcomes to processes and from micro to macro perspectives.

Focus & Scope

Ed Tech research can be understood as an interdisciplinary research endeavor relating to many disciplines. The focus of ETR&D can be located in the learning and educational sciences, which in itself are considered as interdisciplinary fields of studies. In the context of this discussion, the Special Issue will showcase new approaches for research on educational technologies and present methodological approaches that have the potential to provide a fresh perspective and ignite a new discourse for the advancement of Ed Tech research. Authors are invited to present their view on the future of research methodologies and methods in Ed Tech research. Submissions that successfully present the research in 5,000 words are particularly welcome but should not exceed 8.000 words. We particularly welcome submission like the following (list not extensive):

• a conceptual paper explaining an innovative research method in Ed Tech which provides examples from studies published before (at least one of the examples should be originated by the authors).

• a systematic review analyzing the state of research methods in the field of Ed Tech providing new insights into trends and perspectives

• a theoretical paper outlining emerging methodologies and methods for research on Ed Tech with a positioning into different research paradigms.

• a current study that focuses on emerging research methods or incorporates multiple research methods.

References

Bulfin, S., Henderson, M., Johnson, N. F., & Selwyn, N. (2014). Methodological capacity within the field of “educational technology” research: An initial investigation. British Journal of Educational Technology, 45(3), 403–414.

Cramer, F., & Jandrić, P. (2021). Postdigital: A Term That Sucks but Is Useful. Postdigital Science and Education. https://doi.org/10.1007/s42438-021-00225-9

Gibson, D. C., & Ifenthaler, D. (2017). Preparing the Next Generation of Education Research-ers for Big Data in Higher Education. In B. Kei Daniel (Hrsg.), Big Data and Learning Analytics in Higher Education: Current Theory and Practice (S. 29–42). Springer International Publishing. https://doi.org/10.1007/978-3-319-06520-5_4

Hartshorne, R., Ferdig, R. E., & Bull, G. (2021). What Journal Editors Wish Authors Knew About Academic Publishing (S. 1–117). Association for the Advancement of Computing in Education (AACE). https://www.learntechlib.org/primary/p/219093/

Honebein, P. C., & Reigeluth, C. M. (2021). To prove or improve, that is the question: The resurgence of comparative, confounded research between 2010 and 2019. Educational Technology Research and Development, 69(2), 465–496. https://link.springer.com/article/10.1007%2Fs11423-021-09988-1

Huang, R., Spector, J. M., & Yang, J. (2019). Emerging Issues in Educational Technology. In R. Huang, J. M. Spector, & J. Yang (Hrsg.), Educational Technology: A Primer for the 21st Century (S. 231–241). Springer. https://doi.org/10.1007/978-981-13-6643-7_13

Johnson, T. E., Lin, L., Young, P. A., Ilgaz, H., Morel, G., & Spector, J. M. (2021). Thinking from Different Perspectives: Academic Publishing Strategies and Management in the Field of Educational Technology. In R. Hartshorne, R. E. Ferdig, & G. Bull (Hrsg.), What Journal Editors Wish Authors Knew About Academic Publishing (S. 37–48). AACE-Association for the Advancement of Computing in Education. https://www.learntechlib.org/primary/p/219093/

Kimmons, R., & Johnstun, K. (2019). Navigating Paradigms in Educational Technology. TechTrends, 63(5), 631–641. McKenney, S., & Reeves, T. C. (2013). Systematic Review of Design-Based Research Progress: Is a Little Knowledge a Dangerous Thing? Educational Researcher, 42(2), 97–100. https://doi.org/10.3102/0013189X12463781

Reeves, T. C., & Oh, E. G. (2017). The goals and methods of educational technology research over a quarter century (1989–2014). Educational Technology Research and Development, 65(2), 325–339. https://doi.org/10.1007/s11423-016-9474-1

Reeves, T. C., & Lin, L. (2020). The research we have is not the research we need. Educational Technology Research and Development, 68(4), 1991–2001. https://doi.org/10.1007/s11423-020-09811-3

Willis, J. W. (2008). Qualitative Research Methods in Education and Educational Technology. IAP.

Zheng, L. (2015). A systematic literature review of design-based research from 2004 to 2013. Journal of Computers in Education, 2(4), 399–420. https://doi.org/10.1007/s40692-015-0036-z

Restarting my personal knowledge management

Sun, 02 Jan 2022 09:36:59 +0000

The break and transition into the new year is always a time for me to optimise my workflows or to explore some changes in the way I handle information, tasks or communication. Via some tweets by Doug Holton I have discovered Logseq as a new framework and tool to handle my information workflow. I have experimented with it in the last days and would like to summarise why I think this could become a game changer for my personal knowledge management.

Usage scenario of Logseq for my personal knowledge management

Outlining: First of all, for handling notes I need a very simple structure with some basic formatting and indentation and lists to structure information. Logseq is using an outlining technique which is very simple to learn and allows fast note taking.
Simple tagging: In Logseq you can either use an inline-tagging approach with double brackets [[]] or you can use a hashtag. The underlying theoretical basis for Logseq is the so called knowledge-graph, a semantic network of linked information entities. Via these tags, no hierarchical but rather a networked space of ideas and information is growing. Depending on the tags used, those can create a task list, a reading schedule, a communication plan - whatever you like.
Information integration and rendering: A third component which is vital to my knowledge management is that I need to be able to integrate different types of information into the same document. I heavily rely on my Twitter network and regularly spot information there which I mostly like and forget afterwards. In Logseq I can copy the tweet URL and integrate it into my note and it gets rendered correctly in the document. Besides Twitter I can also embed HTML, Youtube or Vimeo videos. Since some information is resting at a specific timestamp in a video, you can even embed a link to this timestamp. And of course I can take notes by indentation why this information is relevant or what I would like to do with it.

Embedded tweet in Logseq
PDF reader and annotation: For someone who is reading a lot of documents in PDF-format, each productivity solution needs to offer a way how PDFs can be read and snippets from these documents can be quoted, copied or commented. In Logseq it is easy to upload a PDF which is then opened on the left half of the window while on the right side.

Chose a PDF to upload

PDF reader left, Logseq right
The nice functionality is now that I can color-code text in the PDF and the put a reference to it in Logseq and can again indent and add notes and ideas to it.

Marking text

Putting a ref into Logseq
Zotero integration: Recently, Zotero has been integrated into Logseq and via this integration I can save publications via a browser-plugin and then access the files via Logseq and can follow the same approach like mentioned in 3.

If you want to start diving into Logseq I can recommend the video series below by Dario from OneStutteringMind.

Future development

I have already tested how to publish my knowledge graph to Github to easily access it from everyhwere and I am also testing the mobile app at the moment, but collaboration is not vital and my main work happens at the desktop. Logseq is open source and free but is apparently planning a payed pro-version for the future.

Data Analysis in the Cloud

Sun, 11 Apr 2021 18:23:30 +0200

In the last two years I have slowly deviated from using GUI-interfaces for statistical analysis towards a more open and transparent approach. Before and during the pandemic I have shiftet my data-analysis completely to R and I have offered in the last semester for the first time a course on “Data Science in Education” in which I have used R intensively with students. Since my usage of R has increased in the meantime also towards more complex scenarios like natural-language-processing or social-network-analysis I was loooking into an analysis environment that is easy to reach and that can run without disturbing my other tasks. This requirement has become even more important during the pandemic in which I hardly find time to let my computer run 1 hour for an analysis without interacting. My usage scenario is clearly only a fraction of any standard high-performance-computing so I was looking for a flexible solution which is cheap and scalable.

Cloud computing and elastic hosting combined

In the end I have found a solution which combines a pre-configured version of the R-Studio-Server with a very flexible and easy hosting provider. In this short summary, i will describe how you can set up your own analysis environment in the cloud.

First of all, you need an account with the hosting provider Digitalocean. If you follow the referral link you will receive a credit of $100 over 60 days. This means effectively that your experimentation will not even cost you a dime. If you hesitate to start an account with a company based in the US, I can point you the the very extensive information about adherence to GDPR rules. In addition, you can also decide to put your server into a European location (my server is in Frankfurt, Germany).
The second step is that you need to install a so called “Droplet” on your server. Virtualisation of servers has the big advantage that non-technical users get a readymade server package without the need to configure a lot of dependencies and update scripts on the commandline etc. You might have heard from Docker or VMWare which is also used to run for example two operating systems on one computer. The virtualization images with Digitalocean are called Droplets and there is a Droplet which exactly fulfilled my requirements.
Mauricio “Pacha” Vargas, a statistican from Chile, has created a Droplet that has all ingredients to run a lot of analysis options in the cloud. His droplet contains besides R and RStudio Server also the ShinyServer and a number of additional libraries and packages which should make the image relative future-proof. You can find a detailed step-by-step explanation how you can install the droplet and also add users to the Droplet on the Droplet page. You can even add multiple users and use the image in a research team.
The nice thing with the installation of the Droplet on the Digitalocean infrastructure is that you can at all times resize your Droplet and increase CPU capacity or memory. This comes of course with a price, but since you can also resize the Droplet back to the smallest image after a heavy operation (the basic image costs 5 US$ per month), the costs involved will be very limited. In my case, I was running into some memory problem during the collection of around 700 000 tweets and I resized the image just for this operation.

Your R-Server is after installation available under a unique IP and you can access it with a username and password. I am very satisfied with this solution and now I can start something on the server and come back after some time to continue analysis. I have not realised any delays due to the server-based access so far and it feels like I would be running R-Studio on my machine.

It would be nice if in the future a real multi-user access would be possible in which collaborative spaces could be defined on a project-basis without allowing access to all data in the environment. This would also make the cloud-environment an ideal environment for teaching statistics and data-science with R.

Data Analysis in the Cloud

Sun, 11 Apr 2021 18:23:30 +0200

Cloud computing and elastic hosting combined

First of all, you need an account with the hosting provider Digitalocean. If you follow the referral link you will receive a credit of $100 over 60 days. This means effectively that your experimentation will not even cost you a dime. If you hesitate to start an account with a company based in the US, I can point you the the very extensive information about adherence to GDPR rules. In addition, you can also decide to put your server into a European location (my server is in Frankfurt, Germany).
The second step is that you need to install a so called “Droplet” on your server. Virtualisation of servers has the big advantage that non-technical users get a readymade server package without the need to configure a lot of dependencies and update scripts on the commandline etc. You might have heard from Docker or VMWare which is also used to run for example two operating systems on one computer. The virtualization images with Digitalocean are called Droplets and there is a Droplet which exactly fulfilled my requirements.
Mauricio “Pacha” Vargas, a statistican from Chile, has created a Droplet that has all ingredients to run a lot of analysis options in the cloud. His droplet contains besides R and RStudio Server also the ShinyServer and a number of additional libraries and packages which should make the image relative future-proof. You can find a detailed step-by-step explanation how you can install the droplet and also add users to the Droplet on the Droplet page. You can even add multiple users and use the image in a research team.
The nice thing with the installation of the Droplet on the Digitalocean infrastructure is that you can at all times resize your Droplet and increase CPU capacity or memory. This comes of course with a price, but since you can also resize the Droplet back to the smallest image after a heavy operation (the basic image costs 5 US$ per month), the costs involved will be very limited. In my case, I was running into some memory problem during the collection of around 700 000 tweets and I resized the image just for this operation.

Implications of the new Twitter Academic Research product track

Fri, 02 Apr 2021 09:36:59 +0000

Twitter has recently announced that it will make the so called “full archive search” available for academic research. Until this change, it was either very costly (by paying for the enterprise version) or it took very long to build an archive of tweets (for example via the very good TAGS tool). With the new product for researchers and the advanced version of the Twitter-API (v2) it is possible to collect tweets from the whole archive even until the first day of existence of Twitter. This changes the potential to use Twitter also for retrospective research completely. Since the setup is not that easy I will lead you through the steps required for starting your own research with the help of Twitter.

Step-by-step instruction towards your corpus of tweets

To get started you need first of all an account with Twitter. Since you are planning to use Twitter in your resarch I expect that you will be already using Twitter and have an account. If you have a normal account you need to apply for a developer-account with the same credentials.
To get a developer account, you need to follow the application form over here. It is important to understand the policy that Twitter sets for its developer account and it is especially useful to understand the concept of a “use case” in their policy. A use case can be understood as a group of activities which could be repeated for a different purpose. Collecting data from the full archive search would be for example a use case. Embedding Twitter feeds into a website would be a different use-case. The purpose is to keep the number of use-cases as small as possible.
After having received a confirmation of the accepted developer account, you can apply for the academic research track on this page. In the application form you will be requested to provide links that prove that you are a researcher and you will need to specify some details about the purpose of your research with Twitter. An additional requirement is that your research should have a non-commercial purpose. After sending this form, you will receive in 24h a confirmation message that you will have access to the research track and via this an unbelievable amount of 10 Mio tweets which you can collect per month!
You will see the academic research product in your developer portal dashboard after acceptance. As a next step, you need to configure an app under the research track. Configuring an app provides you access to a number of keys and secret tokens. Your are not playing the Legend of Zelda here, but the these codes ensure that only authenticated users have access to the Twitter-API. The most important information here if the so called “bearer token” which you should copy to a textfile after you have configured your app. Ideally you make an additonal screenshot from it for your archive. Now your are ready to start collecting tweets. The only question is how you do this.
The answer to your question is: R and nothing else. I am really fascinated by the creativity and cross-disciplinary broadness that the R-project has developed over time. I know that for many the learning curve is high compared to statistical software suites, but R gives you so much power for handling all kinds of data that is has opened also new research options for me. There are many existing packages for R to collect tweets, but none of them were at the time of the opening of the academic research track prepare to use the new API. Christoper Barrie and Justin Chuntingho have invested some brains to give you the easies approach to collect large number of tweets in the research track.
The R-package academictwitteR allows you to collect tweets in three lines of code. First of all you should install the package via their Github repository:

devtools::install_github("cjbarrie/academictwitteR")

As a second step, you need to put your bearer-token in a variable.

bearer_token <- "" # Insert bearer token

As a third step your query will be defined.

get_hashtag_tweets("#twitterlehrerzimmer OR #twitterlehrerimmer OR #twlz", "2013-12-01T00:00:00Z", "2014-12-31T00:00:00Z", bearer_token, data_path = "TWLZ21/data14/", file = "TWLZ21/twlz14.rds")

I have collected in this example all tweets that have been sent for three hashtags and I have defined the timespan for the collection between December 2013 and December 2014. Last but not least you provide a folder name and a file name which should be used to store the output.

Some words on computational capacity

Since I have collected a larger number of tweets (around 700 000) and I have used a setup with a low amount of memory (1 GB), it was important to collect tweets in batches and then merge the data in a second step. I have during the process analyzed the performance of the computer and I have seen that I was running into memory limits. If you see this, you need to either choose a more powerful setup or make your timespans smaller to reduce the number of tweets per collection step. In the second featured post in this newsletter I will introduce a very flexible infrastructure for this kind of activities.

Starting a newsletter

Sat, 13 Mar 2021 06:28:35 +0100

I had already longer the plan, to develop a place which is somewhat slower compared to my Twitter account. In addition, I wanted to have a space in which I come back to pieces I mark as interesting but forget in the flow of the days. This is also reason for the title. Pieces which are lingering in your head for longer but then get overwritten by new pieces.

During my master studies in Nuremberg I have worked as a newsletter editor for an energy consultant. He sent me newspieces and I had to design a coherent piece of text from it. Since I had little knowledge of the domain, this was a challenge which I could overcome by feedback and some discussions with my employer.

My newsletter will be a collection of short reviews of academic publications, a mix of recommendations for academic workflows (productivity and methodology), short pointers to current calls and schlolarships, events and last but not least a personal collection of art and music recommendations.

I will try to publish a newsletter all 2 - 3 weeks. Depending on growth and feedback of my readers I will adapt the style and content. Each edition will be evaluated with a simple evaluation form.

Enough talking: If you think this could add something to your overcrowded inbox, feel free to subscribe here for the newsletter. Alternatively, you can just use the menu-item above to subscribe.

Bildungdigital21

Sun, 07 Mar 2021 08:05:11 +0100

A new direction for the German digital education strategy?

The German government has organized on the 22nd of February 2021 an online-event regarding a future strategy on digital education. The event was a mix of political intention statements and some advertisement of existing initiatives. What I found remarkable was that there is still the vision, that any overarching OER platform would solve any problems with education despite this has failed on the European level already many years ago. The video of the discourse is embedded below.

# We first install the Youtubecaption library
install.packages('youtubecaption')
# Now we import the captions from the video
library(youtubecaption)
library(quanteda)
library(tidyverse)
url <- "https://www.youtube.com/watch?v=WAlP39a5LUU"
caption <- get_caption(url, language ="de")
caption
# Now we need to make a document-term-matrix from the content
my_corpus <- corpus(caption$text)
summary(my_corpus)
# Extending the stopwordlist with a more extended one from Github
ger_stopwords <- read_lines("https://raw.githubusercontent.com/stopwords-iso/stopwords-de/master/stopwords-de.txt")
custom_stopwords <- setdiff(ger_stopwords, stopwords("german"))
# Constructing the
meine.dfm <- dfm(my_corpus, remove_numbers = TRUE, remove_punct = TRUE, remove = c(stopwords("german"), custom_stopwords))
meine.dfm.trim <- dfm_trim(meine.dfm, min_docfreq = 1, min_nchar = 3)
# Visualizing the matrix
textplot_wordcloud(meine.dfm.trim, min_size = 1, max_size = 2, max_words = 100)

This code takes the full transcript from youtube, builds a corpus, removes stopwords and visualizes the most frequent words in the resulting tagloud.