Let the facts speak for you

My latest from Woroni - the audience I had in mind was me as an undergraduate. 

Too often as students we fervently debate questions as though they were strictly theoretical or normative when in fact they are empirical in nature. We can also be sceptical of the ability of empirical techniques to provide strong answers to some types of questions and dismissive of people who ask for data. This is unfortunate as it leads many of us to hold on to incorrect views for ideological reasons when a little research would reveal them to be invalidated by the evidence.

Some examples will help illustrate. Feminist theorist Kuhn is quoted saying that, in pornography, “the vagina stands in for the enigma of the feminine”. One could spend hours debating the anecdotal evidence for this claim—hardcore tumblrs for, interviews with models in Playboy against—without reaching a compelling conclusion. What you need is some cold, hard evidence.

But surely this claim is not accessible to empirical scrutiny? Comments about culture and sexuality can’t be reduced to numbers! Well, maybe they can, at least in this case. A feasible study would involve rounding up a suitably large and diverse sample of consumers of pornography and asking them some questions. For example: “when using pornography, are you interested in the model’s personality and biography, or just his/her body?”

A few hundred respondents later and you would have some pretty compelling data one way or the other. A longer, richer survey could yield more detailed information that might lead to more insightful observations than just yes/no answers; for instance, regarding the pornographic preferences of lesbians. 

Now when you’re discussing the latest controversy over a goon sack on a Thursday night you can’t go running off to conduct an expansive survey. But often you don’t have to – Google can point you in the direction of reputable studies on a wide range of topics in seconds, especially Google scholar.

For example, a friend recently suggested the US has low corporate tax level. A two minute search of OECD records via google reveals that the US actually has some of the world’s highest corporate tax, at around 40 per cent (though loopholes are plentiful).

There are times when anecdotes are more convincing than statistics, and more powerful in a casual discussion because of their ability to articulate larger ideas than the pinpoint questions typically answered by empirics. For example, one of the more convincing arguments in favour of liberalism over socialism is this one offered by Niall Ferguson. After the War, the homogeneous German people were split into two groups. The liberal capitalist group produced the class leading Mercedes Benz; the socialist group produced the Trabant, which can accurately be described as a shit box. Reams of data would be needed to make a similarly compelling argument using statistics.

Similarly, some valuable theoretical treatises would be ill served by the inclusion of statistics. Foucault’s discourse theory, for example, would just be diluted by the inclusion of data without gaining much in the way of veracity or articulation.

Disraeli’s quote also bears mentioning: “there are three kinds of lies: lies, damned lies and statistics”. Statistics can be deceptive. Some examples include misleading survey questions, not reporting the standard error, not including relevant variables in a regression and selecting a very biased sample.

So statistics are imperfect, but something worth keeping in mind. Importantly, empirical data, alongside the historical record and the rules of math and logic can be said to constitute ‘facts’. In our student years we spend our energy forming opinions. The most strongly held opinions are those that survive fierce discussion. But what make a discussion not only fierce but also rigorous are its factual foundations. Everyone is entitled to their own opinion, but not their own facts. And when the facts change, we seem silly if we don’t change our mind.