One thread that runs through all the major events this year, from #FeesMustFall to Brexit to the US election, is that emotion trumps empirical truth, that feelings have conquered facts. We are living, it seems, in a post-empirical apocalypse.
Despite sound evidence that zero fees are an impossible request or that Brexit will hurt the poor, or that Trump can be dangerous, those making the decisions ignored the facts. The protests continued, Brexit happened and Trump won. Tired of the promises of the ‘establishment’ or the ‘elite’, the protagonists acted on their feelings, their emotions. Voters who favoured Brexit or Trump talk of ‘feeling’ excluded, frustrated for not being heard. Mention facts that oppose their views (eg. that Americans have never been this well off or that immigrants contribute significantly to the economy) and it is often dismissed as creations of the ‘mainstream media’ or ‘experts’. In South Africa, discussions of the university pass-through rate and budgets are met with sniggers of ‘neoliberalism’ or ‘white monopoly capital’. It’s as if the general mood is shouting: Don’t try to convince us with your fancy numbers, we’ll say and do what we like regardless.
Granted, this is an oversimplified interpretation of what were (and are) far more complex events. But my sense is that on both the left and the right, empiricism – the quest for evidence to support a hypothesis – is dying. And this scares me.
Let’s take another example: the demand for decolonised education. Let me be frank: there is a need to decolonise the curriculum. How is it possible that a student can graduate at a South African university with a Masters in Economics and not know about M-Pesa, or the economic consequences of slavery in Africa, or South African competition policy regulations? Curricula must be contextualised to the African situation.
But the thing is, students – even those vehemently demanding decolonised curricula – are often ignorant of their own continent, and its history. In a second-year class test earlier this year, a third of my students could not identify Botswana on a map. Seriously? Botswana? Decolonisation demands more empirics, not less.
Decolonisation done right can be a wonderful thing: using the best methodologies and techniques that science has to offer applied to African questions. I think of the wonderful work my colleague Nox Makunga is doing by using Khoisan knowledge to identify fynbos species that help with the treatment of cancer. Or the work of Servaas van der Berg and his team in identifying the binding constraints in South Africa’s schooling system. Or let me take my own area, African economic history: without statistics uncovered in colonial sources we would not know that wages in many African cities in the mid-twentieth century were higher than almost everywhere in Asia. Yes, Africans have not always been the poorest: it is a myth we have now overturned with empirical evidence.
Once we start unearthing verifiable facts, the world becomes slightly more understandable, and even predictable. Here’s another example from economic history. Nico Voigtlaender and Hans-Joachim Voth argue in a recent paper that Hitler’s popularity was a result of the construction of the Autobahn, Germany’s impressive highway system, after his election in 1933. Using newly collected data, they show that building highways was a major contributing factor in boosting support for the Nazi dictatorship. And this was not because the new construction projects employed a lot of people: instead, the highway project signalled ‘economic competence’ and an end to austerity measures, so that many Germans accredited the recovery to the Nazi dictatorship. They show, for example, that support for the Nazis increased particularly strongly in areas with greater radio availability, which was a major source of propaganda.
Empirics make the past understandable, but it is also a warning for today. One of the few programmes Trump has insisted on since his election is a focus on America’s crumbling infrastructure, particularly the highways. And we know his popularity is partly a consequence of his social media campaigns.
The only way our understanding of the world around improves is if we collect evidence – hard, empirical truths that can be verified and replicated by other scholars. That is the scientific approach. Without evidence, we are left with only conjectures and our feelings about those conjectures. And this is dangerous, because it means that only the feelings of those who are able to shout the loudest – the bullies – count.
Data science is not perfect. When data analysts make mistakes – like wrongly predicting Brexit and or Trump’s win – data is often dismissed as pure hocus pocus. The old joke goes: when a climatologist wrongly predicts the weather, she gets a stronger computer; when an economist wrongly predicts the economy, she gets fired. Data scientists should instead learn from their mistakes, improve their models and algorithms and continue to make progress in understanding the incredibly complex thing that is human behaviour.
To avoid a dictatorship of emotion, we need to keep a critical mindset. We need to question, investigate, think critically, and make sure we don’t just accept or retweet or repost statements without verifying they are, in fact, truths.
But more importantly, we need to equip our next generation with the skills to participate in the world of empirics, in the world of Big (and small) Data. We need to embrace statistics – and not only in the hard sciences. We especially need more and better statistics courses in the social sciences and humanities, where the actions of humans require us to be more careful in our predictions, but not less rigorous in our empirics. In all fields from psychology to history to political science, the benefits of a more statistical approach – sometimes called the digital humanities – is gaining ground.
This means that we need to encourage high school students to choose the more difficult mathematics. We need them to learn code, because they will need it for basically every degree that will be in demand in the next decade. More fundamentally, we need to give them access to an ever-expanding world of online education where they can learn these things easily and affordably.
To understand the complex world we live in, and make it better, we need empiricism to rise.