From Spotify to Netflix, artificial intelligence is curating our personalised music and movie playlists. It is blending our likes, shares, and comments to decide what we see, and what we don’t see, on social media.
But it can also be used to plug the gaps in our understanding of some of the biggest questions put out to public opinion.
In fact, just last year it was used to accurately predict Joe Biden’s US presidential election victory over incumbent Donald Trump.
In a first for opinion polling in Malta, machine learning has now been used to predict how the Maltese public would vote if a general election were to be held today.
Times of Malta commissioned independent market research firm Esprimi and their sister company Lobeslab and asked them to get the most accurate prediction of an election to date.
This is how they did it.
Tackling the non-responses
First, the pollsters carried out a survey of 600 people. But that first round of the survey gave Labour a much larger lead.
Labour enjoyed 38% support with just 15% saying they would vote Nationalist – a gap of 23 percentage points.
However, nearly a quarter of respondents, 24%, said they would either not vote or did not know who they would be voting for.
This figure, known as a non-response, is much larger than the real number of those who do not vote in an actual general election, and so can skew the accuracy of the survey results.
The survey scientists managed to bring the number of non-responses down from 151, to 29, or just 2% from the original 24%.
They did this by building models of likely voting intentions based on respondents’ answers to several other questions.
These answers were compared to the respondents declared vote in the 2017 election.
The method was tested, and the approach accurately predicted the 2017 election result with a negligible deviation of less than 1%.
Running the non-responses through this system gave the pollsters a much clearer indication of how they believe the country would vote if an election were to be held right now.
Independent journalism costs money. Support Times of Malta for the price of a coffee.
Credit: Google News