US election - how did polls get it so wrong?

Sometimes, in order to fit with the prevailing opinion, we may practice a little misdirection when surveyed. The justifiction might be that as we're not actually voting, our answers don't really make a difference in any important way. The impact of social acculturation in surveys is reflected by the fact that in this election as in others, online surveys have reportedly provided more accurate predictions than phone polls.

The latter are more personal, involving a clear interaction between two human beings. In them, people are perhaps more likely to be influenced by the tone of voice of the surveyor, or the way a question is phrased.

The former type are more "cool", to use the terminology of media lecturers; they involve less of a sense of flesh-and-blood interaction. After all, a modern political survey could easily be conducted - and perhaps sometimes is - by an algorithm rather than a person.

Meanwhile, the breakdown of institutional loyalty in western cultures means that people are less likely to have made up their minds well in advance of an election. Swing voters represent a growing constituency within many democratic systems.

Younger people in particular do not approach elections with the same institutional or dynastic party loyalties as did many of their parents or grandparents. They are more likely to base their choices on issues. In much of Europe, there is more outlet for this than in the US, because there are a greater number of parties and a more coalition-friendly system of government.

In this US election, there was also the phenomenon represented by Bernie Sanders. Many of his followers - a great number of them young people - felt that the deal done to enthrone Mrs Clinton as the Democratic nominee was unfair to Sanders and the positions he represented. This resentment, like that felt by many within the Reblican party who favoured a non-Trump candidacy, was another factor contributing to a more fluid and more unpredictable electorate.

Whatever the ultimate reasons for the poor predictions of pollsters, they will need to work long and hard to improve their performance if they are to be of any real value in helping people understand elections and their results.