Whatever you think about the result of the US election, it’s widely acknowledged that there was a failure of polling.
In just a few hours, what seemed an assured but narrow victory for Hillary Clinton morphed into the landslide win of Donald Trump.
The reason – as with Brexit, where very similar happened – seems quite straightforward; that people minded to vote for a candidate or policy featuring non-liberal views or characteristics tend not to shout about it, and may well even lie.
Now, this is completely their right – but the issue comes when pollsters assume their population sample is a robust indicator and apply it to the population at large.
There’s a whole debate to be had about why so many people are disenfranchised, and the role played by journalism and social media in feeding this – but that’s not one for this blog.
Our interest is purely in the data – the data that was ultimately inaccurate. The data that is likely to put the whole polling industry’s credibility at stake.
While time may show us exactly what went wrong, it will have to be a factor that when data from one cohort is applied to all others, essentially when people are depersonalised from their own data, that errors can creep in. And that’s without reckoning with bias, margin of error and all the other fun that polling inherently brings.
Our data, our views – the thoughts, feelings and actions that belong to and make up each of us, need to stay with us for maximum accuracy. In the same way that we are the best guardians of our own health and financial data, when our polling (or any other important data) is associated with us, rather than pulled away to become part of a population average, it’s likely to be more accurate – and therefore more useful.
If both political polling and political journalism need to undergo a regeneration, then maybe polling and voter intention methodology does too.
One thing is sure – keeping our own data where we can share it, not anyone else, in a true Internet of Me, can only be a good thing in so many areas going forward.