How a Poll Got Me Off Twitter

Back in January when the current outbreak was still largely a China phenomenon, someone put up a poll on Twitter asking whether people thought it was likely there would be more or less than 1,000 deaths worldwide. Given the virus’s infectivity characteristics1Asymptomatic spread, high percentage of mild cases, and efficient access to key receptors. I had developed my own mortality model, but I was reluctant to share it lest I look like a gloomy crank. Instead, I simply took the “over” on that bet—which more than 50% of respondents didn’t.

That now looks ridiculous, but it did do something useful. I realized that by seeing all those responses centered on the 1,000 figure it had caused, however briefly, an anchoring bias in my thinking. That is when an early piece of information, often a price, but potentially any figure, causes you to root your subsequent thinking in the area of that number. My own early mortality estimates were 1-2 orders of magnitude larger, but there I was newly thinking in terms of more vs. less than 1,000. 

It was yet another example of how social media breaks brains. In addition to inducing a constant action bias—the feeling that you need to do something when confronted with a constant flow of information—it also causes this dangerous anchoring problem. Independent thought becomes almost impossible amidst an instant crowd consensus for something else.

This realization was, in part, why I dropped off Twitter in late January. Having seen the storm clouds coming, and not wanting to feel compelled to take pointless action in a constant state of cognitive inflammation, but also wanting to maintain independent thought as the storm rolled in, it was time to go. So, I left.