The Data behind Donald Trump’s Victory
By Aditi Bhat
Nobody saw it coming. Not the media. Not the data scientists. Certainly not Hillary Clinton. Many came out to say, “Data is dead” after the victory of Donald Trump, this US Elections.
If you were following this election, you would have heard how most of the early trends were in favour of Hilary Clinton winning the presidency hands down. So, where did things go wrong? Where did it take an unexpected turn that shocked not just the people of the United States of America, but the entire world?
In fact, data scientists, pundits and pollsters across the board predicted a Democrat for president (FiveThirthyEight gave Clinton a 71% chance at victory; The New York Times‘ Upshot had her at 91%). Even Donald Trump’s analysts were expecting her to win by such large numbers.
To understand where things went against popular predictions, let’s break it down:
With Donald Trump’s offensive and ‘grabbing’ comments about women, analysts expected the female majority to vote in favour of Hilary Clinton’s pocket.
However, when the final numbers came in, there was only a 12% difference. Back in 2008 and 2012, the same percentage of female voters were in President Obama’s favour. This came as a rude shock not only to Hilary Clinton whose campaign was all about women empowerment, but also for all the voters who were expecting their first female president. And as expected, the majority among the male voters remained in favour of Donald Trump.
With Donald Trump’s tough talks about building walls and strict immigration policies, Hilary Clinton unsurprisingly drew far larger numbers of Hispanic, Latino and black voters to the polls as rightly predicted by analysts.
However, for a Republican nominee who received backlash within his own party, Mr. Trump held on to roughly the same share of Hispanic and black voters as Mitt Romney had claimed four years ago. Also, as predicted the white majority stood for Donald Trump.
With Brexit being a recent occurrence, fuelled by Islamophobia, analysts predicted older voters to be in favour of Donald Trump. Just not in such large numbers.
2016 saw the highest Republican voter turnout since the 1980’s; the 45+ year-olds stole the show, voting in favour of Donald Trump. These older, whiter Republicans were a surprise to analysts who hadn’t accounted for them while predicting the polls.
This trend didn't vary as much from what was predicted by analysts. Voters who had college and postgraduate education were far more likely to support Hilary Clinton.
Those with high school and an associate degree showed their support for Donald Trump, but only by the slightest margins.
One of the popular notions explaining Trump's victory was the support from the working class. However, exit poll data shows that over half of voters with incomes below $30,000 per year supported Hilary Clinton.
However, the higher income brackets favoured Donald Trump, even though it was only by narrow margins. This data was particularly surprising as Hilary Clinton was earlier criticised for having ties with the Wall Street.
So, what happened with the election data analytics and algorithms? The answer, it seems, is a combination of the shortcomings of polling, analysis and interpretation. Perhaps both in how the numbers were presented and how they were understood by the public. This leaves a lot of scope to learn and grow as the field develops.
All images are sourced from: https://goo.gl/HFEGY1