LONDON • Many businesses will be affected by the Brexit vote. But one sector of Britain's economy has already been dealt a grievous blow: the opinion polling firms that were spectacularly wrong in predicting the European Union referendum's outcome.

The fact their error came after the same pollsters also failed to predict the results of the British election last year means that an industry once admired for its clairvoyant qualities is now derided as hopeless in measuring the pulse of the people.

Some problems facing British pollsters are well known from the experience of other countries. It's difficult to predict turnout, which is particularly important when votes are counted nationwide in a referendum, rather than by constituencies, as in parliamentary elections. People are also reluctant to tell pollsters how they are voting if they believe others would regard their choice as odd or unfashionable.

Newer, deeper trends also render the pollsters' jobs far more difficult. People now have many more overlapping loyalties, and ideologies no longer matter; someone who votes for left-wing parties because of economic considerations can also be a conservative on social issues, so it's more complicated for pollsters to put together the representative sample of voters.

A bigger share of voters now cast their ballots well ahead of voting day, so they respond to different events. Those who voted early in Britain were influenced by the assassination of a young MP during the campaign, while those who voted on referendum day largely forgot that episode.

But it is the Internet and social media that have upended pollsters' profession. Online interactions decide what an electoral campaign is ultimately about: Britain's politicians wanted the referendum to be about economics, but the electorate decided it was about immigration. Social websites also allow fringe politicians to reach wider audiences in ways pollsters don't understand, so someone dismissed as marginal can suddenly become important.