This is a photograph of Mt. Solitary, in the Blue Mountains of Australia, after which this website was named. In this strange and uncertain time in the history of the world it’s hard to live far from home, I am often grateful that I’m the kind of person who takes lots of photographs.
I recently finished reading Lost Connections by Johann Hari, a book which tries to unravel the reasons behind the alarming prevalence of anxiety and depression in the modern West. The book really resonated with me and I wanted to share the main things I took away from it, as someone who thinks about these things a lot, and certainly has skin in this particular game.
In the book, Hari, who suffered from depression for many years starting from age eighteen, contends that the idea that depression and anxiety are largely due to chemical imbalances in the brain and should be treated with drugs is wrong. The “lost connections” of the title are what Hari believes to be the true causes: depression and anxiety are sicknesses caused by the relentless and unnatural pace and style of modern life in the West, where human beings are living in conditions further away from their natural evolutionary habitat than ever before in our history. This loss of connection, with meaningful work, meaningful relationships with others, nature, and more, creates a powerful sense that something is not right, a longing that cannot be met by continuing to live the way we live. These connections are so intrinsic to the human condition, so essential for our contentment as animals, that without them we flounder and become lost.
The first half of the book, in which Hari describes in detail what he believes to be the seven different types of connection we have lost, was compelling. There’s plenty of studies to back up what he’s saying and he presents the material in an easy-to-understand and logical way, peppered with personal anecdotes and descriptions of experimental milestones in the study of depression in the life sciences. His description of the DSM definition of depression, which contained all sorts of illogical exceptions in certain cases (for example that depression is not considered a mental disorder if the person is grieving a loved one, so long as that loved one was sufficiently “close” and so long as “not too much” time has passed), demonstrates the absurdity of treating it as a purely neurological disease. There is a solid and undeniable accumulations of experimental evidence and experiential anecdotes to convince the reader that there’s probably another reason why people (including in all likelihood the person reading the book) are depressed.
Many of the things Hari claims we have lost connection with I can certainly identify with, or have felt the lack of many of them in my life, or can see the effects writ large in the structure and essence of the society I live in. But I also felt that his attempt to offer a solution comes up short, since it is probably only possible to implement if you are lucky enough to have wealth, health, education and privilege on your side.
For example, in order to reconnect with meaningful work one needs to have the ability to change jobs to something one finds more meaningful. Hari gives the example of an acquaintance who works a painfully unfulfilling job in a paint shop and dreams of teaching people how to fish. Anyone who has suffered from depression can recognise the subtle redesign of the age-old “cheer up” quick fix here. The fact is that most of the reasons we have lost connection with the things Hari mentions in his book is because society has been deliberately structured that way by vested interests (Hari does mention this often in his book to his credit). This is a societal problem that requires a societal solution: a rejection of neoliberal capitalism and a full embrace of an empathetic and comprehensive welfare state. A few tech bros reading a self-help book and moving to the country to “reconnect with nature” is not going to reverse the depression pandemic.
I guess I’d like to see more people with influence and audience like Hari go a little further in pressing for the deep and structural societal change that the world so desparately needs to cure itself of the sickness of inequality and division. People have become too afraid of spruiking socialism and the only way to get around that is to spruik it hard enough and often enough that it doesn’t scare people anymore. Possibly things are too far gone in the United States but I think there’s hope in other parts of the West. This is the connection we all truly need to rekindle: the connection with each other: throwing away the hideous Thatcherite doctrine that “there’s no such thing as society” and deciding to actively make things better by lifting everyone up.
I’m very interested in using predictive modelling to forecast elections and elections don’t get any bigger than the US presidential election. And of course the 2020 election is a once-in-a-lifetime election, the choice between four more years of Donald Trump, about whom enough has been written that I don’t need to dignify him with another word, or whichever sentient human being has been chosen as his alternative, which turned out to be Joe Biden.
Most people without a passing understanding of Bayesian statistics and the nature of uncertainty, and the fact that it can be quantified, brush off the notion of forecasting the upcoming election, usually saying things like “you can’t trust polls anymore after Trump and Brexit” (polling averages in both of those events were well within the margin of error), or making blanket statements like “I think he’ll be re-elected” without offering any sort of evidence to back it up (the more self-centered will remind you they predicted a Trump victory in 2016).
Broadly speaking there are two camps of people forecasting US elections: those who rely on “fundamentals” models, like Allan Lichtman, which use broad indicators like GDP, the existence of civil unrest, and incumbency to predict whether or not there will be a change of party holding the presidency; and those who rely on poll-based regression models, like Nate Silver from FiveThirtyEight.
There’s value in both approaches but the powerful thing about sophisticated models like Silver’s is that they are able to express the amount of uncertainty in the race. This is especially important this year, where the election will take place during an unprecedented pandemic. I personally think this makes “fundamentals” models much less useful this year, as instead of using “is GDP healthy?” as the predictor they should be really using “is GDP healthy for a pandemic”, which of course has very few historical data points.
Of course, uncertainty is an unnatural, scary and unintuitive thing and it’s a lot easier for people to say “but FiveThirtyEight said Hillary was going to win in 2016, they were wrong” rather than “FiveThirtyEight forecasted a 65% chance of Clinton winning and a 35% chance of Trump winning”. Part of this is human nature: it’s a binary contest with a single outcome and post facto that 65% looks a lot bigger than that 35% so it’s tempting to go for the first interpretation. But I hope that this year people appreciate what is at stake: the integrity and function of the most important democracy in the world, and take a little more care in interpreting uncertainty when they see it in forecasts.