Arguably the biggest losers from this last UK election have been the polling companies. Throughout the election process they predicted the main political parties were deadlocked, while the election delivered a decisive win for the Conservative Party.
Historically there are plenty of polls that have missed the mark but in this age of data, our trust in polls seems to have grown out of proportion with their accuracy. The strategies the political parties adopted during the election were based on the poll data. The media coverage and focus was on dealing with a hung parliament rather than on the issues that the parties stood for. The UK election will be a case study in the pitfalls of trusting poll data.
But could we have done better, in this age of data? Perhaps. According to the Head of YouGov, Peter Kellner, the polls before the election were far off the mark because voters said one thing, but then did another once they got to the ballot box.
This phenomenon has been put down to the Shy Tory Factor, when conservative voters provide misleading answers to pollsters or refuse to participate in exit polls (where it is called “non-response bias” by pollsters). The prevailing theory is that these Shy Tory, or Shy Conservative, voters opt out of polling, or offer misleading answers, because they don’t view the elite media, who sponsor the opinion and exit polls, as truly neutral.
Google certainly see some opportunity to address this bias and discover what people really think. They employed a PhD student at Harvard University, Stephens-Davidowitz, after he investigated the potential of Google search data to gauge users’ views. People are more likely to be honest in searches than in polls, he reasoned, for example people Google the word “porn” far more than most ever let on.
So it seems like search could provide the answer to the poll accuracy in the future. Although whether Google wants to step into this influential role, in light of current EU Anti-trust legislation, is another question. In any case there are lessons here for aspiring data analysts – our focus should increasingly be on search intent.
We should be examining search data carefully to determine what our users think and prioritising this information over user surveys. The role of SEOs should be expanded into identifying and engineering these user journeys into positive experiences for users.