So, we find ourselves here again. The polls said one thing, the election the other. First it was the UK election; then Brexit; and now the US election of Donald Trump to being the 45th American President.
Despite the polls suggesting the election would be close, they never the less predicted a Clinton win, which wasn’t to be. More so than in almost any election, the US media – and so the world and the electorate – live and die by almost daily polling of opinions.
At one point, I did think about copying and pasting a previous post of mine about Brexit, just find+replace Brexit for Trump and it’s done. But then I thought it about some more and realized, I really don’t know what is asked in a poll. So much weight is given to them, in the media at least, that before I begin to criticize them again I should at least know what is asked.
After a quick look online I found a good range of polling questionnaires from mainstream media; research houses or universities, all of which I took as a sign of robustness in their sampling and analysis (one of the charges against the faulty prediction of the Conservative election in the UK win was inaccurate polling of British voters – though it seems the US may not be immune to this). The thing that surprised me first was how similar the questions were.
First, a question on is the country headed in the right direction or not; then a favourability question of each major candidate; then a question of who you would vote for if the election was today; then follows a series of demographics questions like age, race and gender.
But that’s it – is the country going well; who do you like the look of; and who would you vote for.
Now, let’s take a step back. Imagine this was any brand – Cadburys; Samsung; Google; Pepsi – and the study was an equity study; the perceived value of the brand in the mind of consumers. Well, really, isn’t that polling – the perceived value of a candidate; the likelihood that you will endorse them and vote for them in the coming election?
I think back to any brand equity presentation I’ve presented, if I was to simply stand up and give the top number. Your brand equity is x%, your competitor’s is y%, I’d be shouted out of the room. Brands don’t just want to know where they are. They want to know what got them there, what’s holding them back, what are they seen to stand for that their competitors aren’t and vice versa, what are the open territories that no brand owns, where’s the opportunity moving forward?
And yet none of this is in the polling questionnaires. Doesn’t the electorate deserve to know what people think about the candidates, isn’t there a better news story to be made from ‘these are the latest polling figures and now let me explain why’? Quantitative numbers are a blunt instrument that can easily conceal the truth; beyond simply asking who would you vote for, we need more detail of the why – of the reasoning behind the numbers.
Perhaps the biggest opportunity in polling is to open the survey up to open responses – allowing respondents to answer a question with their own words and not select an answer. Of Kipling’s six honest men, we want to know why. Why are you voting for Trump? Why are you voting for Brexit? The key to a great story is not who, or what, or when, but why.
Now I know this would require asking an open question, which has regularly been lambasted as ‘hard to ask, difficult to code and impossible to interpret’, but the other view is that it is the best way to get an insight into the mind of respondents. And yes it may take more time to analyse the results, but ultimately it may give more insight into the votes and so more credibility to the numbers and respect to the research industry.
Perhaps asking open questions within the context of what is now the fast-paced, 24-hours news cycle of an election is a step too far. But surely we should respect the research industry, the respondents, and the idea of polling enough to give it more airtime than the 2-minute survey that it currently gets. After all, the future doesn’t look good for the polling industry and the next big vote – but at least next time we might know a little bit more about people’s votes rather than just the topline.