Tomorrow’s South African election represents 20 years of ANC rule. This election represents the first opportunity to vote for the so called “born frees” – those born after the first democratic election in 1994 – to vote.
In 2012, Nate Silver successfully predicted the results of the US elections by consolidating and extrapolating the results of multiple polls.
Can similarly accurate predictions be made for this South African poll?
The substantial difference – a much smaller number of polls in South Africa versus a large number of well established polls in the US.
All South African polls predict an ANC victory, with a reduced majority, increased support for the Democratic Alliance, and declining support for smaller parties – with the possible exception of Julius Malema’s radical EFF. Unlike US polls, however the predicted victory margins vary wildly:
Some of the wild cards: How much will the radical left eat into the ANC vote? Will Kasril’s No Vote campaign have an impact? Will the born frees vote, or will they simply not bother? Will the DA be able to grow its limited share of the popular vote?
Go out and exercise your right to vote tomorrow!
No matter what the result, I will be interested to see whether a South African Nate Silver will emerge victorious from the predictions race. Or will poor sampling and other data quality issues leave all the pundits way off the mark?
Image sourced from http://www.gov.za/issues/elections_2014/index.html