|
Ein Gastkommentar von Chris Zomer und Luci Pangrazio
The datafication of education is an ongoing concern. With datafication we mean the increasing translation of educational practices into digital data, as well as the implementation of processes and tools necessary to support these translations. Some concerns addressed by us and other scholars are the decontexualisation of learning (learning cannot always be translated into measurable data) but also the issue of data privacy, especially since student data is often collected by commercial platforms and can potentially be shared with third parties.
Little has been said, however, about the role that dashboards play in the datafication of education. There are mainly three forms of dashboarding:
(1) the use of dashboards in learning platforms that provide information about student activities and outcomes, such as their learning progress or their engagement.
(2) the use of dashboard for regulatory purposes, implemented by government authorities to track local or national school metrics
(3) the use of dashboards for individual school governance, to be used by all staff (or students).
In Australia, where we do our research, we have come across some far-reaching instances of dashboarding of the latter kind.
One of the schools we worked with implemented a bespoke dashboard using Power-BI. Leadership told us that the dashboards were used to judge teacher performance. This was partly based on a battery of standardised tests that the students were subjected to, and which were visualised on the dashboards, as well as fine-grade comparisons of GPA averages per cohort. Rather than disciplining teachers directly, the school envisioned teachers would internalise the metrics of the dashboard and would ask for help if ‘their’ results were not up to scratch.
Another concerning example of dashboarding we found in the literature. In a 2021 article is described how a school in Sydney implemented a student dashboard to encourage ‘goal setting’ and ‘self-regulation’. A target GPA for each subject was calculated based on a variety of factors, such as standardised test results and previous years’ GPA. Students were then provided with a dashboard filled with graphs and charts, so they could ‘track’ any progress over time but also where they sat in relation to other students. In this case, we feel that students were taught to see themselves in terms of numbers, with the constant reminder that their ‘targets’ needed to be met.
Both examples point to the increasing surveillance of teachers and students using data. Dashboards facilitate a form of self-internalised control in which students and teachers are ‘trained’ to compare themselves to others and adjust their behaviour accordingly. Of course, it is a noble aim to increase student learning, but the data presented on the dashboard only shows part of the picture of a student’s learning.
Both examples are also illustrative of the ‘professionalisation’ of schooling, or, to put it more bluntly, to make schools operate more like businesses, using ‘data-driven’ decision-making. It is worth noting that both examples are drawn from private schools, which play an important role in Australia’s educational landscape: around 30 percent of Australian secondary school-aged children attend a private school. Some of these schools charge considerable fees. It is fair to say that these schools operate very much as businesses, looking for ‘new markets’, for instance, by recruiting international students from South-East Asia. To be successful in this highly competitive market, schools must have good results that they can market to parents of potential students. As such, it is only logical that the dashboards they use focus on cohort views, suggesting that the main objective is to manage collective performance, rather than individual students’ growth.
The dashboarding of schools also raises the risk of algorithmic bias. On the Australian markets there are a number of providers that offer plug-ins to schools such as IntelliSchool and TrackOneStudio that offer data analytics visualised on seemingly easy-to-understand dashboards. These companies rely on large datasets from participating schools, which some use to train their AI-model. One of the features promoted informally by these companies is the prediction of student’s final year’s outcomes. In a highly competitive educational system this can be used to disqualify disadvantaged students on the basis of biased data.
The dashboarding of education is not unique to Australia. In the Netherlands, for instance, around 20 percent of primary schools make use of Leeruniek, an edtech provider that promises useful insights into student progress and ‘trends’. Companies like Leeruniek, generate the data but are also providing the means to make sense of it, both highlighting the problem (there is too much data) and offering a solution (the dashboards). Leeruniek shares aggregated data with government and researchers and as such the data may be used to inform policy. We are concerned that third parties are becoming influential intermediaries influencing policymakers with enormous sets of rich but decontextualised data.
We argue for restraint using dashboards and data analytics, as they often only sketch part of the picture of students’ learning. Furthermore, dashboards enable a perversive form of control that is more about number management then about creating genuine educational opportunities.
|
