Did Google Trends predict the EU election results?

Posted on May 29, 2019

 

Last week, I looked at Google Trends traffic to see if it could give us an insight into what would happen at the EU Parliamentary Elections that took place last Thursday.

I felt I should follow up on this regardless of whether the predictions were good, bad or indifferent. As it turned out, they were a bit iffy, but let’s take a look at them and work out what did and didn’t go well.

What Google got right

Brexit Party success

Google Trends gave us a good idea that the Brexit Party would perform well. Search volumes over the last few months have shown that the protest vote party, led by Nigel Farage, has been opening up a gap between itself and its rivals.

It was no surprise that this online interest translated to polling booth votes, with the party winning 29 seats – 13 more than its closest rival.

Lib Dems outperform Labour and Tories

Since forming a coalition with the Conservative Party in 2009, the Liberal Democrats have become something of a forgotten party. That has perhaps changed in this election, where the Lib Dems earned the second most seats.

Google Trends data even indicated the correct order of how many seats the traditional ‘big three’ parties would win, as did the bookmakers, with Labour losing seats and the Conservatives faring even worse.

Poor showing from UKIP

Having had their thunder stolen by the Brexit Party, UKIP had a miserable time. On the Trends graph, the party barely registered at all compared to the other main ones, which reflected the lack of public interest in a party that won just a 3.3% share of the vote.

What Google got wrong

Change UK flops

The big curveball thrown by the Google data was the splinter pro-Remain party Change UK getting more traffic than anyone except the Brexit Party, despite polls indicating they had little support. The party performed largely as predicted though, winning no seats and getting only a fraction more votes than UKIP.

So why the high search volume? It may be simply down to the interest factor of a new party, and the fact that it’s such a normal name that could be confused with the petition website Change.org.

Good campaign for the Greens

When the Green Party was added to the graph, its data was lower than that of any of the main five we looked at, but the environmentally conscious party had an encouraging campaign, winning seven seats and boosting its share of the vote. This was largely predicted too, so it’s a surprise that it wasn’t reflected in search volumes.

Overall, not a disgraceful showing, but Trends didn’t really ‘predict’ anything that wasn’t widely expected, and it’s surprise statistics turned out to be off the mark. Still, even for search volumes to mirror election votes (such as correctly indicating the order of the Liberal Democrats, Labour and Conservatives even though the latter two get much more airtime) shows a correlation between Big Data and human behaviour.

John Murray

Content Team Leader at Engage Web
John works for Engage Web as a Content Team Leader and regularly contributes to the website and programmes of his beloved Chester F.C.

Like us on Facebook to see more posts like this

You might also be interested in:

No Comments »

There are no comments on this yet, be the first to write a comment.

RSS feed for comments on this post. TrackBack URL

Have your say!

Call Now Button

We have worked with:

minute-man-press-image
TEL: 0345 621 4321