The Spirit Level - Why More Equal Societies Almost Always Do Better11 Jan 2016
There’s been much discussion, mostly centered around politics but traces of this type of thinking could be found elsewhere, about the role of facts in political discussion these days. Examples like “True Enough” by Farhad Manjoo, Colbert’s seminal debut essay on “truthiness” and the constant undermining of facts in public discourse make it very difficult to have a meaningful discussion about any controversial topic. So many of my conversations with friends or family devolve into some form of the following:
- Me: here’s some evidence about why I believe the point I just made. Here’s the source. It’s not the whole picture, it only reflects a sliver of knowledge I have.
- Interlocutor: Well I don’t know the motivations of that source, and I have a counterpoint from another source whose worldview matches mine better. My counterpoint probably doesn’t even rightly address the point you just made, is tangential at best, and because we’re both motivated parties in this discussion, there’s no good way for us to evaluate our arguments. So both our arguments are equally as good, or alternatively, equally as bad.
- Me: Ok, seems we’ve reached an impasse.
This is a very frustrating way to get no where. But I’m also at a loss for how to remedy this sort of thing. How can I evaluate my facts, someone else’s facts, figure out which argument is stronger, which side is right, under a deluge of facts? So far my answer has been to resolve to some relativist position, “we’re both right, if you look at it in different ways,” which feels hollow and inapt to match the high stakes of the kinds of questions we’re talking about - inequality, climate change, housing policy, foreign policy, and so on.
As a data scientist, one of my go to aphorisms is: “It takes an order of magnitude more work to debunk bullshit than to create it.” The authors did more than an order of magnitude of work to debunk prevailing attitudes about inequality. The authors, and others, are working very hard to debunk all sorts of bullshit. But the hardest realization for all of us working on difficult questions like these is that just doing the work will not debunk the bullshit. There’s still a chance after you’ve done the work, that bullshit will persist.
This book will likely reify your biases. If you came in believing income inequality matters (like I did) then you will believe that doubly so. If you came in believing that the world works just how its meant to, and that levels of inequality are where market forces have led them, you’re probably not going to be convinced that we need to do something about inequality here and now. There is plenty of damning evidence, but this is not quite a mic-drop argument. The authors even take on their critics right in the text, either pre-empting their counter arguments or engaging with arguments that were presented after publication. But, somehow, I fear very few people’s minds will be changed.
Causal links are hard to establish between fuzzy measures like inequality and health, crime, violence, etc, especially without experiments. But a bevy of correlations are hard to ignore too. This was reminiscent, to me, of the smoking causes lung cancer debate. This debate is often taught in statistics as an example of causality established outside the bounds of controlled experiments. No experiments proved that smoking causes lung cancer (it would be unethical to do experiments), but lots and lots of research about the link between the two was enough to convince most public health officials of the causal connection. Similarly, a bunch of scatter plots with lines drawn down the middle, as presented in the first half of this book, do not make for the best evidence. In fact, it is hard to evaluate the statistical veracity of each of these plots without any measures of fit, variance or robustness to consider. But after enough evidence mounted against inequality, it’s hard to ignore the probable implications.