An interesting tension point emerging from our research relates to time and temporality – specifically, the temporal disparity between how schools could be making use of data, and how they are actually making use of data.
For many teachers, making use of data is an episodic and events-driven process. For example, data use might perhaps be triggered by a specific milestone or scheduled point in the academic year. A teacher might engage with her class data before a parents’ meeting or after an internal examination. School leaders might run a set of analyses before an external audit. As this teacher explains:
I would say that you don’t do it every day. I would say that, you know, you make time for it some time in the week or in the lead up to things like, you know, it might be parent/teacher interviews or it might be in the lead up to an assessment. … I think there’s more formal times when it happens. You know, it might be reports, it might be at the end of a learning task or whatever it might be. It’s pretty hard to do it on the run and you’d have to make a special effort really to do that.
In contrast, then, we are finding little appetite amongst school staff to be given immediate or continuous feedback on their schools’ data in the form of ‘real-time’ feedback or other such promises of instantaneous and comprehensive data monitoring and reporting.
In this sense, schools’ current approaches to data use appear to be tightly temporally-bounded. Teachers and schools work on the basis of wanting and/or needing to know specific things at certain times for discrete purposes. School staff will take comfort in the reassurance of ‘knowing that the data is there’, but also display a notable reluctance to engage frequently with what the data says other than on a ‘needs must’ basis.
From a data science perspective, this might appear to be an inefficient or even illogical approach. Yet from a ‘school’ point of view, maintaining an ‘arm’s length’ relationship with data makes good working sense. Teachers and schools have finite amounts of time to conduct large numbers of routine tasks. Any additional insights or surprising diagnoses might well turn out to be unwelcome intrusions into this tightly scheduled cycle of events. As school staff have told us in our studies so far, there is often insufficient time to follow-up any additional insights that might arise from a bout of data analysis. As one school data analyst put it, these tend to be put aside as ‘A different story for another day’ – i.e. a distraction from the task at hand and, therefore, a possible drain on one’s already limited time.
These tensions of temporality are touched upon in a new paper by Taylor Webb and colleagues (2019). They point to the contrast between how AI and data science “inflects anticipatory priorities in education (p.2) – for example, ideas of future-focused projections and predictive actions, continuous measurement, real-time feedback and so on. In contrast, they note that: “The dominant form of time in education is chronological … assum[ing] that time is … something that can be counted … calendars, schedules, deadlines and performance goals are the objects and practices of chronological education” (p.4).
This points to a fundamental temporal mis-match between data-driven technologies and the schools that they are being inserted into. Webb and colleagues make the point that analytics and automated systems clearly do not run on ‘clock time’ let alone ‘calendar time’ – yet these are the chronological logics that “educational governance is trapped within” (p.7). In this sense, the application of data science logics into school systems carries a threat of ‘rupturing’ long-standing and dominant understandings and habits of ‘school time’.
From what we are finding from our fieldwork so far, there is little expectation amongst schools that the dominant ‘habits’, ‘logics’ and ‘grammars’ of school-time are likely to be reshaped by the distinct temporal modalities of the data-driven systems that are being used. All three of our research schools are locked into deeply-rooted routines that continue to shape (if not often dictate) how things are done.
In this sense, it might be tempting to conclude that the all-encompassing potential of new data-driven technologies is destined to remain subsumed into the dominant ways of ‘doing time’ in school – i.e. only being referred to on occasion and/or on a ‘needs-must’ basis. It might well be, then, that schools will give the appearance of being data-driven while not radically altering their routinized practices, processes and perspectives.
Alternately, it might be that the temporal logics of data science will actually permeate into the external governance processes that schools are beholden to – such as assessment regimes. For example, school authorities might begin to require continuous forms of reporting or ‘any-time, on-demand’ testing and examinations. In this way, schools could be slowly reoriented around the logics of data-time by external coercion and official imperatives.
However, Taylor Webb and colleagues suggest two more radical possibilities. Firstly, taking the nihilist post-capitalist perspective of accelerationism as a starting-point, they raise the prospect of simply letting these alternate temporal logics loose within education systems – accelerating their implementation into school systems and seeing what alternate possibilities arise from abandoning traditions of human agency or deliberate governance.
Secondly, is the less chaotic (but no less antagonistic) option of planned resistance. This might involve working out ways to disrupt and ‘jam’ the logics of data-driven systems – attempting to make school ‘incomputable’ or ‘imperceptible’. Schools could engage in efforts to hack, jam and out-manoeuvre these systems – in other words, working to foster an ‘un-timely politics’ of schooling.
Webb, T., Sellar, S. and Gulson, K. (2019). Anticipating education: governing habits, memories and policy-futures, Learning, Media and Technology, DOI: 10.1080/17439884.2020.1686015