Digital divination – seeing data-driven analytics as a matter of faith rather than a matter of fact?

At a recent workshop, Tim Neale raised the idea of approaching digitally-driven ‘prediction’ in terms of ‘divination’

During the course of his current research he is examining how fire-fighting agencies engage in the prediction of fires – be it a few weeks’ time or many decades into the future.

As might be expected, anticipating the outbreak of fire is an inexact science. When fire prediction practices are examined in situ, any idea of complex big data analytics and data-driven pattern analysis gives way to a couple of professionals engaged in low-level ‘spreadsheet work’ and a great deal of supposition, instinct, hunches and other forms of expert intuition.

The parallels between Tim’s work and our own studies of teachers’ in-school instantiations of ‘learning predictions’ are clear (‘just playing around with Excel and pivot tables’).

As such, seeing the technologies of fire prediction in this guise of ‘divination’ raises a number of non-scientific connotations not usually associated with analytics 

For example, describing digitally-driven prediction and analytics as process of ‘divination’ raises the idea of digital analytics as a standardised process that this nevertheless mystical, magical, a form of sorcery, ritualistic practice and beliefs rooted in superstition and faith. 

It raises the idea of digital analytics as little more than a pseudo-science or quack practice. In short, it unsettles our usual assumptions of computer-based analysis as infused with qualities of precision, accuracy, and objectivity.

A few people working in the educational data science are also thinking about how to present their outputs in less certain terms – for example, having any visualisations appearing to be hand-drawn in slightly unsteady manner, or presenting text in a cartoon font so not to be taken completely seriously.

We need to think further about how we might disrupt the appearance and description of data in schools so that it loses some of its aura of certainty, comprehensiveness, and all-seeing qualities. Three decades into the twenty-first century, data analytics is perhaps best seen as something that you might choose to dabble in … but is not something to be blindly followed and believed in as a divine source of truth. Schools are no place for what Leon Wieseltier has termed ‘the idolatry of data’.