Why the inequalities in our information diets matter (Part 1 of 2)

Part I: Visiting the United States in November 2016

As I write this, I am on a train travelling from Boston to New York City, at a time when people in the United States are still coming to terms with what they thought they knew about their country. The trip has been rushed because I need to be back in Sydney to catch up on work, which meant I didn’t have time to catch up with all the excellent people on the east coast. Here is a picture of approximately where I am right now.

img_2104
Between Boston and New York City, about an hour before New Haven.

Without the benefit of hindsight, it feels like the disintegration of knowledge continues to accelerate into the post-truth era; that societies around the world are struggling to hold onto their values across a populous that outsources its opinions and beliefs to some nebulous idea of whoever is both influential and accessible. While information overload is not at all a new idea, we do not yet entirely understand how information overload may actually restrict rather than open up what we hear, what we click on, what we soak up into our worldview; and how that can reinforce our echo chambers to create polarisation and conflict.

While it is not something we usually like to admit, the information to which we are exposed – our information diets – are an excellent predictor of our attitudes and opinions. This is not necessarily because our information diets causally affect what we believe and how we act, but because we surround ourselves with people who agree with us and rarely seek out opinions that are contrary to our own. And even when different opinions are put right there in front of us, we disregard them.

For researchers who like to observe the world while pretending not to be part of it, right now looks less like a post-truth era and more like a golden era for understanding the impact of news media and social structure on human decision-making and behaviour.

There has never been a better time to take advantage of massive streams of data about information search and exposure to measure their impact on opinion, attitudes, decisions, and behaviours. In my team, we already use social connections on Twitter to train machine learning classifiers that can predict whether you are likely to have negative opinions about vaccines. We are now close to finalising work showing that a population’s (relevant) information diet can explain the variance in vaccine coverage of individual states in the United States.

img_2087More of Boston walking from Back Bay to Boston Children’s Hospital.

Part II: Some news
While I am still looking at the bright orange colours of the United States, and before I head back to the purple jacarandas in Australia, I will stop here. But I will share some related news soon.