The hard truth about Hillary and the polls
Pollsters will spend weeks trying to assess where they went wrong on election day, but there are other questions we should be asking
Donald Trump became America’s 45th president last week. Most of us didn’t see it coming, not in the least his own campaign. All of the data at our disposal and all of the national publicly available polls showed it would be close, but everything in our guts assured us that his rival would eke over the line. The polls didn’t lose this election – Hillary Clinton did. However, she didn’t do this alone.
There is a hard truth about data that people don’t generally understand: opinion data isn’t biased. Sure, there are sources of error that complicate the process, but the data itself has no lean – it simply is, and we interpret it. Humans, however, do have bias. So in a week where I, like many of my friends in the Democratic political class, have had to come to terms with some very hard truths, allow me to share a few of them here.
Hard truth #1: Donald Trump will be president because he won enough states to win the electoral college. While this might seem obvious today, it is important to restate as a hard truth, because most people weren’t reading the national public opinion polls on the US election thinking about the electoral college. Instead, they twisted and turned with every one-point increase or three-point decline in the overall ‘head to head’, every new data point revealing who was ‘leading’ the race. And when they looked at predictions of who would win as stated in a likelihood percentage, regular people weren’t asking why any of them were still showing a chance – however slight – that Trump would become president. Had more people paid closer attention and dedicated greater resources to other sources of information, from better state-level polling to more qualitative and secondary research, they might have seen the signs.
Hard truth #2: The United States is a tapestry of many peoples, of many backgrounds and many ethnicities. But it is still composed primarily of white middle-class voters, and Donald Trump won decidedly with that group. He won huge with middle-class whites because while we in the political class (Democrats and Republicans alike) took him literally but not seriously, they took him seriously, but not literally. And many continue to make the mistake of discounting the feelings he tapped into, refusing to consider legitimate concerns because of those few who engage in racist rhetoric. Considerable percentages of Trump voters found his language offensive and over the top, but voted for him anyway, because he spoke plainly and directly to the very real economic angst they were feeling.
Which leads to hard truth #3: Message matters. The Trump campaign was a metaphoric masterpiece; when he said “build a wall,” they heard “stop job loss”. When he said “dismantle Nato”, they heard “make everyone else pay a fair share”. When we saw him wear a ridiculous hat with “Make America Great Again” on it, they saw him wearing their national dream like a team uniform, a dream that, to them, was slipping away with no apparent concern from the powers that be. When he said “the system is rigged” we laughed. When they heard “they system is rigged”, they screamed because someone was finally being serious about their problems, if not literal.
In communications, if you do not understand what someone is feeling, or choose to discount it, you will always – always – fail to deliver a meaningful message. Donald Trump tapped into an emotional juggernaut and mined it beautifully if ineloquently. We wrote stories and sneered about how he spoke to supporters at a seventh-grade level, and smugly congratulated Clinton for speaking at a twelfth-grade level. And Trump made no bones about who his audience was. We derided him for it, daily. Hillary’s message was heard by her base but failed to connect to an audience that President Obama had (needed) some success with – white lower- to middle-class voters. And that failure to connect cost her dearly in the swing states where she needed to win: Michigan, Wisconsin, Ohio, Pennsylvania, North Carolina, and Florida – often by less than a percentage point.
Pollsters who conducted public presidential polling will spend weeks trying to assess where they went wrong on election day. But what we already know is that for any campaign to be successful, it must use any and all data at its disposal to identify its key audiences, make emotional connections on the issues that matter most to those audiences, and craft the right messages for those audiences.
Photo credit: Ted Eytan CC BY 2.0