One of the most distasteful things about big data is top-down-ness. We all submit information through our social media habit, and the powerful aggregate it all with sophisticated algorithms and work out new ways of controlling us. There’s a kind of TINA spirit to the whole thing (which no doubt is manufactured by elites) – big data is the ‘science’ of the future; data presents the future path for government; surveillance is inevitable.
Technology is a threat to civil liberties. But it may also be our only hope for emancipation. In order to conceive of alternatives and to realise hope, we need to critique the extent to which big data analysis is indeed scientific. My contention is that the ‘big data’ graphs and pictures are not at all scientific because there are no regularities which are explained by it; big data operates rather more in accord with the principle of sympathetic magic than science. The images hypnotise and everybody becomes imprisoned by it.
We should learn from the ecologists. Statistical ecology is an important and growing field which studies the dynamics between different organisms, food chains and habitats. Ecological analysis is used to explore the factors that contribute to the health of an ecosystem, to warn of threats to ecological diversity and to help identify interventions which might be beneficial to the management of overall ecology.
Ecological thinking in the social sciences is not new. The American sociologist Everett Hughes wrote about the “Ecology of institutions” in 1936 - just before Western Europe would unleash forces that would destroy its ecology for a generation (Hughes was particularly interested in studying Nazi Germany). Hughes points out that it is absurd to concentrate measurements on particular activities or even particular institutions. What needs to be done is to explore relations between activities and institutions and understand the constitution of their diversity. Of course, in the 1930s the tools for doing this didn’t exist. Attempts to grapple with entireties and relations were however attempted. Perhaps one of the most celebrated examples is the Mass Observation project of the 1930s which focused on life in everyday Bolton (or Worktown as it was called).
If we see big data as mass observation of our social ecology we can start to ask powerful questions about the role of government. If governments destroy ecologies they are probably not doing a very good job and we should find other people who could do a better job.
But I want to start closer to home. Universities are ecologies. The different roles, responsibilities and personalities appear to constitute different ‘species’! How do they work together? In what ways do they not work together? What are the effects of managerial intervention?
Here we see the key problem with big data. Because big data results in analyses which are available to the elite, it results in decisions based on a particular elite interpretation. Consequently, decisions based on big data are attenuative according to the particular interpretation in operation. As a result, despite the potential richness of the data, social ecologies are effectively squeazed to conform to a particular ideal.
Robert Ulanowicz calls this squeazing “mutual information”, borrowing the term from Shannon. Mutual information is the coordination of the stuff which we all know about: the mission statement is a classic example of 'mutual information'. Some degree of mutual information is necessary in any ecology, because otherwise it lacks coherence. However, mutual information does not have to be imposed from the top to the bottom. It exists in teams, departments, and so on. Ulanowicz argues that we should consider the overall health and richness of an ecology by considering mutual information along with what he calls ‘flexibility’. Flexibility is ‘not information’ – or certainly it is ‘not mutual information’. It is the ‘redundant’ stuff that people do and think about which isn’t accounted for, which isn’t directly useful, which serves no apparent purpose - which drives the accountants mad or leads senior managers to accuse staff of being 'lazy' or 'unproductive'. All ecological systems exhibit both mutual information and flexibility. Without flexibility they become brittle and die as they become unable to adapt to shocks and changes in the environment.
In most institutions since the economic crash, austerity has resulted in the ramping up of ‘mutual information’ and the elimination of flexibility. My own institution conducted what it called (horribly) a ‘delayering’ exercise, removing autonomy from departments and concentrating power at the top. The tests of health are simple. How many times do senior managers say “no” to the ideas of junior staff? How many times do they refuse resourcing or funding requests? How many times do they say “yes” to their own ideas? How many times do they say “no” to their own ideas? And my favourite: How many times do people throughout the institution utter (for whatever reason) “What the Fuck?!” to things that happen per week? The WTF count is very reliable: it seems to be quite high where I am!
If we use our data right, we can ask these questions. We can demand from our managers that they act as proper custodians of educational ecologies, and not as the self-important “CEO’s” that only hubris and covetousness delude them into believing themselves to be.