Quality, not speed, is what we need - A case for a sustainable transformation of digital education

German schools have been slower to embrace digital education than US schools over the past decade, due to concerns about the influence of commercial players and data protection. The Covid-19 pandemic has now significantly increased reform pressure in Germany. Sigrid Hartong (Helmut Schmidt University Hamburg) argues for a sustainable, pedagogically and socially oriented digitalization that extends well beyond mere functional skills.

This article is part of our dossier "Digital classrooms - Transatlantic perspectives on lessons from the pandemic".

Education policymakers worldwide seem to agree that the education crisis accompanying the Covid-19 pandemic can only be mastered by quickly and comprehensively installing digital "replacement" systems and making them accessible to all students and teachers. While previously, school platforms or learning analytics were primarily hailed as providing more individualized or efficient teaching and learning processes, the pandemic and its resulting prohibition of on-site teaching have dramatically shifted the focus towards digital technologies as the “only way” to provide ongoing instruction. With such massive pressure for action in education policy, the major emphasis is currently placed on the quick implementation of best-practice models , teacher training, etc.

With all due respect for emergency reactions in education policy, the author of this article nevertheless warns strongly against disregarding more differentiated, critical perspectives on the digitalization of education, especially in the current time of crisis. In fact, the digitalization of education – and thus a future-oriented approach to overcoming the crisis – can only succeed if it meets the following criteria:

  • digitalization in education must be sustainable, i.e. it must not be reduced to short-term equipment or operational issues;
  • it must be pedagogically oriented, i.e. it must not leap at every technological “evidence of a potential”; and
  • it must be socially oriented, i.e. it cannot underestimate, but must instead actively counteract the various forms of digital (e.g. algorithmic) inequality and discrimination.

International exchange and global mutual learning – beyond the short-term exchange of best-practice models – will play a decisive role in this.

To begin with, it is important to consider that scientific evidence on the hoped-for effects of digital educational technologies – the reduction of educational inequality, increased learning performance, personalization of instruction, promotion of cooperative learning – has, at least so far, proven to be far more contradictory than is often claimed.1

In fact, hardly any positive learning effects have been demonstrated in a reliable form to date that cannot also be achieved using analog methods,2 or that are scalable to whole classrooms or schools. Put differently, results so far show that students need very different scopes and types of (partly cost-intensive) technology in order to support, and not to harm their learning. Simultaneously, negative effects may include various other side dimensions such as screen time or psychological effects caused by gamified learning designs.3 Consequently, there is a pressing need to equally and systematically evaluate opportunities and risks of digitalization in the education sector, particularly in view of the enormous financial and ecological resources required to create and maintain digital infrastructure (which will then not be used for other forms of educational investments).

At the same time, such a risk-aware evaluation goes well beyond the (potential) effects of digital technology on learning outcomes, but also includes a better understanding of the education policy background and the broader societal context. Such broader relationships between digital, data-based educational governance (e.g., the global increase in educational monitoring, testing, and data-based school evaluations) may initially appear less relevant to the impact of digital education technologies; but in fact, the opposite is true. A better understanding of the growing power that comes with using digital data and the increasingly centralized data infrastructure reveals that privacy issues are only a minor part of the problems we need to address. Here, too, we are dealing with knowledge that is highly significant for designing risk- and value-conscious digitalization policy.

I would like to offer some examples of such relationships in digital education governance, showing how a comparative view of Germany and the United States can help develop such awareness.

A key similarity in education governance between Germany and the United States are their highly decentralized systems. In both countries, the jurisdiction for education policy and much of the funding is at the subnational level, meaning that education policy decisions are made by Germany’s 16 federal states, or in the United States at the state or municipal level. The consequences are not only strong intranational differences and strong regional educational inequalities, but also a relatively high degree of inertia with regard to system-wide reforms. For example, both countries struggle to design digital data systems for education that are compatible across German or US states (interoperability). There is also a lack of nationwide standards and financing channels beyond education policy jurisdictions. At the same time, both Germany and the United States have undertaken numerous initiatives in this area in recent years, or bolstered them significantly during the pandemic.4

Although both nations are in many respects on similar paths and face similar challenges, their design of these initiatives can differ greatly. In many cases, these differences have historical and thus systemic roots in values and structures in both countries.

The United States, for example, is not only more strongly oriented towards data and technology  in educational governance design – as can be seen, for example, in the much more intensive integration of standardized tests for assessing the performance of students and teachers – but business and non-governmental interests also play a much bigger role. With regard to the digitalization of education, this applies in particular to the so-called education technology, or EdTech, sector, which in the United States is one of the largest by global standards, but in Germany is still very rudimentary. In the United States, for example, many EdTech providers, particularly the big ones like Google, Microsoft and Apple, are taking advantage of the chance to establish their technologies in schools, with a relatively low threshold to entry. This is happening, for example, in state-funded, privately run schools. While some of these so-called charter schools proclaim their digital innovation or the strong performance of students, charter schools do not all fit this model. Many charter schools also adhere to educational reform programs and are far more critical of digitalization. At the same time, this sector, which is often supported by corporate foundations, is extremely well networked and correspondingly politically sensitive.

Market restrictions are (also) value decisions in Germany

Although in Germany the influence of the EdTech proponents has also increased significantly in recent years, there have been and still remain strong reservations, especially about the large global EdTech providers. There are, in contrast, greater efforts to create state-developed solutions or small-scale, locally tailored products.

This is also linked to the much stronger regulation of data flows and use of data, affirmed with the adoption of the European General Data Protection Regulation (GDPR). For many EdTech providers in Germany, this means in many areas a strong restriction of influence and market potential. Such market restrictions should also be understood as value-based decisions that are closely linked in Germany to the idea that digital education governance must put education first and not endanger pedagogical freedom and protections, especially the autonomy of teachers. Even if this idea has come under increasing pressure in Germany, especially during the current coronavirus pandemic, it is still an important facet of the debate.

The situation is similar with regard to establishing standardized data infrastructures, flows and exchange. In the United States, numerous non-governmental reform networks have launched comprehensive initiatives to standardize data systems over the last 15 years. These include the ed-fi alliance, the national Data Quality Campaign, the State Longitudinal Data Systems (SLDS) program established by the national education authority, and EdFacts. At the same time, standardization is not only about integrating digital education data across education policy levels (e.g., merging school data from the state to the national level) and thereby making it accessible to as many people as possible in the sense of a "public right to transparency." Rather, the goal is also to link educational and social data, data on individuals over the course of their lives, and to link cognitive data (e.g., results from achievement tests) with social behavior, the school climate, or emotional, motivational, and physical data (e.g., facial recognition, eye-tracking, brain wave measurements, etc.). Even though the US education system still has much less data integration of this kind than many Asian countries, this cross-sectoral, increasingly dense data collection is becoming more and more important – for example in the context of the abovementioned SLDS. This idea is to have the most comprehensive data collection possible of learning and teaching processes, but it also includes the hope of finding technical ways to address learning problems at the earliest stage possible.

Data-driven standardization reduces diversity in education

But such a perspective can also be consciously questioned if generating and combining data is understood not only as a functional task – in the sense of providing evidence about learning – but also as the transfer of highly complex educational realities into numerically modeled representations, for example so that algorithms can process them. The decision on how to model also means a consequential determination and appraisal of what and to what extent things are visible and thus ultimately doable in education policy, pedagogy or educational administration. This even more so, the more standards and definitions are set in terms of which data is (can be) produced in which way, and which relationships and formats it (may) have.5  In other words: with increasing data linkage – whether driven by state or private providers – the heterogeneity of models and thus the (possible) diversity of “valuing” education digitally is gradually reduced. What remains, on the other hand, determines how we "see" education and educational institutions, teachers and students, and which political or educational decisions are deemed "necessary" by these images. One example is the central production of school data sheets, in which data from different sources is brought together in a standardized way, thus enabling a supposedly "holistic" view of these schools. Accordingly – and here again the understanding of pedagogy and education plays a decisive role – the functional merging and standardization of digital education data can also be highly problematic beyond data protection questions.

Nevertheless, the number of initiatives for greater data integration, compression and publication has also risen sharply in Germany in recent years. Examples include the implementation since 2002 of a so-called Kerndatensatz (KDS), or core data record, by the German federal states’ education policy body, the Education and Culture Ministers Conference, and the increased establishment of centralized collection points for test data.6 Current debates about a national education register are also part of this (although the introduction of a centralized student ID failed due to massive protests), as well as various data integration initiatives by digitalization advocates such as the Alliance for Education. At the same time, the centralization of data is still far more time-consuming and is often deliberately incomplete. This is particularly true with regard to the cross-sectoral integration of personal data, but also, and especially, with regard to the centralization and publication of "school-related" data (e.g., school profiles with detailed data aggregated at the school level, such as on test scores).

Overall, it is thus evident that value systems and the legitimacy of certain arguments and perspectives have a decisive influence on how the digitalization of education is shaped. However, it is also important that these perspectives are always associated with certain risks or disadvantages.

In the United States, these include a much greater structural dependence on private actors.  However, recurring data scandals and problems with the educational quality of digitally oriented charter schools also indicate that large parts of the EdTech industry are driven by profits and not necessarily by educational motives. This dependency has increased even more during the pandemic, and the long-term educational consequences of this EdTech boom are not yet foreseeable. However, these comprehensive standardization and datafication initiatives are also meeting growing resistance in the United States, especially the increasing dominance of data usage in making education-related decisions. This not only pertains to the exclusion of alternative forms of assessment, but also numerous unintended side effects for schools, teachers or students in an environment of continuous assessment and growing monitoring.7

While Germany's tendency to be more cautious and reserved in its approach to datafication and digitalization as well as to the involvement of private actors has so far contained such problems, other challenges are becoming apparent. For example, there has been repeated criticism that pedagogically high-quality digital innovations are being held back because they require certain market structures to develop competitive alternatives beyond Germany.  At the same time, the disadvantages of a lack of standardization and data integration are becoming apparent, which makes working with digital tools cumbersome, especially for schools. It was clear in the pandemic that the German education system suffered more than many others due to the lack of infrastructure, systems and standards in most schools. These are now being massively expanded in Germany as well, but often in an uncoordinated manner, which makes them extremely cost-intensive and, in many places, not designed to be sustainable.

So while education systems around the world are suffering the effects of the pandemic in a similar way, it must not be forgotten in the face of the current exceptional situation: An understanding of heterogeneous value priorities, of various possible perspectives on education and thus also a better understanding of the very different effects of datafication and digitalization will be central to a well-considered, pedagogically sound digitalization strategy in education.

Accordingly, democratic debates about reforms must include as many of these possible perspectives and potential risks as possible, not only from a functional but also from a values perspective, and raised over and over again if necessary. Debates among policymakers and in public must include a wide variety of very different interest groups. Experience in the United States in particular suggests that it is necessary to avoid excessive dependence on the private sector in public education governance. Nevertheless, it should also be possible to incorporate pedagogically meaningful solutions from private providers. Both together are possible with sufficient regulation and points of control, as in other social areas, for example through algorithm audits. Digital tools should be pedagogically designed, and there should be a choice between digital and analog as well as different models, while taking into account the special need of protecting the educational sphere.

Finally, the demand to create or consciously maintain heterogeneity also applies to the design of the data infrastructure itself. Comprehensive centralization and standardization of digital data (systems) may seem technically possible and functional, but it always comes with a systematic reduction of the interpretation of the educational world. Yet this variety (and thus also contradictory data) can be central to reflecting the complexity of educational processes and tackling problems by means of trust-based, pedagogical debate.

Overall, we should therefore abandon the idea that sustainable digitalization will soon be achieved. Sustainable in this context does not only mean the ecological dimension, such as that associated with the purchase of appropriate hardware, but also the question of what kind of educational digitalization will enable active engagement and thus a values-oriented shaping of future society that is advantageous over the long term. In other words: while the question of "how" is at the forefront of swift digitalization, sustainable digitalization focuses on the question of "why." Anyone who consciously seeks sustainability should therefore make it a top priority to create the appropriate time and training resources for a (self-)conscious engagement with digital education technologies – Critical Digital Literacy – beyond merely functional applications. Not only in education policy, but also in every public office and every classroom.

This article is part of our dossier "Digital classrooms - Transatlantic perspectives on lessons from the pandemic".
 


1) For a comprehensive overview, see https://jesperbalslev.dk/evidence-of-a-potential-ph-d-thesis/ as well as  Reich, J. (2020): Failure to Disrupt. Why technology alone can't transform education. Harvard University Press.

2) See also Zierer, K. (2020): Lernen 4.0. Pädagogik vor Technik. Möglichkeiten und Grenzen einer Digitalisierung im Bildungsbereich. Schneider Verlag: Hohengehren.

3) At the same time, major evidence is still missing, such as on the pedagogical effects of artificial intelligence.

4) See Hartong, S. (2019): “The transformation of state monitoring systems in Germany and the US: relating the datafication and digitalization of education to the Global Education Industry.” In: Marcelo Parreira do Amaral, Gita Steiner-Khamsi and Christiane Thompson (Eds.), Researching the Global Education Industry – Commodification, the Market and Business Involvement. Palgrave Macmillan: 157–180

5) For example https://techdocs.ed-fi.org/display/ETKB/Ed-Fi+Data+Standard (USA), or https://www.kmk.org/fileadmin/veroeffentlichungen_beschluesse/2003/2003_05_08-KDS-Individualakten-Laender.pdf (Germany). 

6) Such as at the Institut zur Qualitätsentwicklung im Bildungswesen, IQB, www.iqb.hu-berlin.de.

7) See for example https://www.govtech.com/products/Facial-Recognition-Software-on-the-Ris…; https://www.theguardian.com/world/2019/oct/22/school-student-surveillance-bark-gaggle.


Translated from German by Ellen Thalmann.