Polls – we knew what the 18-45 year olds were going to do!

04 Jul 2017 | Research & Business Knowledge

Well as you asked here is a short story…

A couple of Fridays ago I was having a virtual beer with a couple of ICG stalwarts, and we thought it might be fun to do a bit of an election poll. 

We wanted to test out if using the Field Agent crowd sourced model (with no complex  sample selection and weighing)  would produce a relatively accurate indication of was likely to happen….

So we put out a survey with a prize draw of £100 open to everyone that responded. On Friday, 27 May… within 48 hours we had over 1,000  responses from 18 to 45-year-olds  across mainstream Britain with about an even split between men and women.

This suggested  we could be in for a 3rd surprise of Trump-like and Brexit proportions.  If the election was held the next day 49% of this group would vote Labour, 24% Conservative showing a swing to Labour of roughly 20% since the last election And the biggest gain for Labour was expected to come from those that did not bother to vote last time ( amongst the 18 to 45 yr).

The resulting article was published on LinkedIn.

We predicted that if Corbyn captured the hearts and minds of the 18 to 45-year-olds we could be in for a surprise.

Last Friday we thought would run the survey again to see if what had changed, this time a slightly smaller sample of around 500 was captured before the news of the tragedy in Borough market was broadcast and we closed the fieldwork. The shift to labour had continued to gather momentum and an increasing number of 18 to 45-year-old’s reaffirmed their intention to vote

We also captured some qualitative comment asking people to explain the reasons for  for their choice. In answer to the question "what should to Teresa May do to win the election our “crowd” suggested she  should "stop lying, stop talking, stop attacking the working class, stop hiding !“

The other interesting statistic  was the wisdom of the crowd as despite the strong intention amongst the 18 to 45-year-olds to vote Labour, they recognised that the Conservative party would be delight the winners of the election,  but at the time of the initial survey , they did not predict the narrow margin.

One of the reasons we did this exercise was to explore how accurate the prediction of  relatively representative crowd but with very unsophisticated sampling, would prove to be in practice.

Yes it took a few hours to put the questions together and analyse the output, but with an external cost of only £200… perhaps this goes to show with today’s technology and methodologies there are lots of opportunities for low-budget research to deliver robust research??