Common sense would dictate that the more information you have, the more prepared you are to make a better decision or draw a better conclusion. That’s why it was bizarre to me when the other day I was reading through articles on 538.com – one of my favorite sites – and one of them was called “Why we got the election the least wrong.” The article was as thought-provoking as it was depressing (given the results of said election). 538 predicted Clinton would win 70% of the time. This was compared to the Huffington Post and other outlets who had Clinton winning upwards of 97% of the time. All of these models focused on polling data from various polling institutions, which got me wondering, at what point do we have too much data? That is, are we so inundated with data that we can’t see a threat as foreboding as a storm cloud heading our direction because we’re looking at numbers and figures that normalize the event of a strongman narcissist running for president? Would we have voted differently had the polls not convinced so many that Clinton had it in the bag? And how much do we trust sites like 538 moving forward after they underestimated the human element of election day results?
At Thanksgiving dinner, my soon to be brother in law, recently engaged to my older sister, said to my family “I’m just never going to look at polling data again. They don’t know what they’re talking about.” I didn’t question him at the time. As an avid Clinton supporter, he still had a bitter taste in his mouth post-election, and I did not want to be argumentative. Underneath, however, I was curious about his comments. Is one failed measure of polling enough to make all polling, and statistics in a larger sense, useless? I don’t want to call Nate Silver’s life work valueless. I think there is so much to be gained from looking beyond media and cultural narratives and looking at the statistics and raw information firsthand in order to make a cogent argument. As was said during the data lecture when that first scientific book was published, letting the population at large be their own observers, read into the information as they want to – that process makes us a more intelligent society. The problems arise when statistics are taken like religious scripture – when analytics replace arguments and when someone becomes the mouthpiece of a data collector rather than forming any nuance in their views. I almost think the election of the demagogue that is Trump was in part a rejection to this kind of robotic thinking, for better or for worse (worse).
Moving forward, I’m still going to use 538 for my politics, rather than MSM where they just have ridiculous talking heads shouting over each other all day long. But the next time I see a poll I will take it with that added grain of salt. Data is critical, as if making arguments out of individual research, but not at the cost of a human element to all things that is so hard to predict.