The future of polling: listen more and better
Por Maurício Moura
The year of 2016 brought many electoral surprises. The unlikely victory of the Brexit in the United Kingdom (the decision that lead to the process of leaving the euro zone) and the unthinkable victory of Donald Trump in the US were the most shocking ones. The overall sentiment was that most of polling predictions mislead the public opinion. Furthermore, the press promoted a significant debate about the effectiveness of current polling methods. We don’t agree with the general assessment about lack of polling precision. However, there is no doubt that 2016 was a year full of lessons about the future of collecting data and measuring public opinion.
First of all and for the record, by all pure technical standards (what pollsters call margin of error), the elections prediction outcomes in the US and in the UK were correct. The heavily number of polls had predict that Hillary Clinton would won the popular vote (as she did by almost 3 million votes) and Donald Trump would win the electoral vote by an average probability of 30% (imagine, if someone says that you have a 30% probability of dying in a airplane flight, would you consider that low or high? Of course it is not a low chance).
What happened in both cases was that analysts (including very experienced pollsters) were not able to fully factor the “anger” vote into their analysis and voting models. Leaving Europe for the Britons and voting for “making America Great Again” were fundamentally based on “anger” against the current political system that are not fully responding (or even neglecting) structural issues such as the effects of globalization and immigration on the employment and local development. What really happened in 2016, and that’s the lesson learned, is that we should not confuse bad polling with bad analysis.
On the other hand, those close races have thought us that more elements are needed to drive better analysis and conclusions. The current approach of polling — face-to-face, telephone, many questions, lots of people to execute surveys and on — is certainly dying (and dying fast). The future of polling relies on a cultural and technological shift. The shift to mobile and internet means that we are facing a more fluid situation. People change their minds much faster today. In a daily basis, most of voters are exposed to a huge amount of data and information. It is much easier to move opinions and much harder to follow. Face to face surveys become obsolete very fast. Such technological change is turning into a cultural shift. The opinions are as fluid as Facebook posts.
On top of that, people lie on answering surveys. Many Trump voters were not exactly “proud” to say which candidate they were intending to vote — what we call the “shadow vote” (that happened as well in 2015 during the UK general elections — the conservative vote was harder to capture). Some pollsters are incorporating different questions to try to corroborate some key answers. One example is the following question: “Would happen to know who is your closest neighbor will vote for?” The neighbor has always been a good proxy to public opinion. People that live in same neighborhoods tend to have similar opinions. Location is still the most variable to predict vote.
Hence, it is imperative to incorporate into polling analysis new technologies and approaches such social media listening and big data elements — especially in close races. The future of polling relies on “listening” the conversations on the social media environment and “knowing in the advance” the respondents, i.e, developing resourceful big data tools to decrease the need of asking questions. The ideal survey in the future will be constant and will not rely on asking questions.
In summary, the future of polling relies on ask less, listen more and listen better. Anything other that would be a new avenue to criticism and discredit.
O que achou deste post? Se gostou, clique no coração ❤ abaixo e ajude este texto a chegar a mais pessoas. Deixe também seus comentários e compartilhe com seus amigos. Se quiser receber mais conteúdo sobre marketing direto, big data, pesquisa e inteligência de mercado, siga nosso canal!