April 9, 2018 – As all the major polls called a Clinton victory back in 2016, only one small Croatian polling company called the election and a Trump victory almost perfectly. And now the polling oracles of Oraclum have spoken once more.
A Facebook post from Oraclum chief Vuk Vukovic this morning:
And now…this: [Bias in polling and approval ratings]
I say that Trump’s approval ratings have the same problem as his pre-election polls – they are systematically biased.
=> The response rates of polls are down to only 9%, which means that if you want a sample of 900 people you need to call about 10,000 of them. This leads to a large non-response bias which is making it difficult for pollsters to capture a truly randomized sample, which obviously reduces their accuracy.
For one thing, it fails to capture the so-called Shy Trump voters who avoid giving out truthful opinions in polls.
=> Oraclum’s BASON Survey is one way to move past this. It is proven to yield much more accurate estimates of election outcomes when taking into account that people distrust polls by asking them what others around them think. It leaves people well within their own comfort zones and gives them a chance to think and without pressure express their opinion on a subject. The way we ask the questions also gives people further incentive to think about the questions and self-correct their own answers. This delayed judgment has been proven by behavioral scientists to improve accuracy of people’s own forecasts.
If you wanna read more, we’re launching the Oraclum White Paper section, and we’re gonna start with number 9 (so somth like Star Wars, but not really 🙂 ). Read the full paper here.
To sum it up:
Trump’s 2017 approval ratings are very close to his 2016 pre-election polls in the key swing states, which means they are uninformative about his party’s chances at the 2018 midterm elections, and should not be used to predict midterm election results.