Is SA getting a dose of fake news in the run-up to the election? And is this fake news coming not from Julius Malema or Jacob Zuma, but from pollsters and academics?
I frame this as a question because I’m not sure of the answer. But I’m wondering hard enough to raise the matter here. You can see the temptation. Those who conduct polls are aware that their findings could affect the outcome of the election.
Announce that the ANC will probably get 45% of the vote and govern in coalition with small parties, and you may discourage opposition voters from going to the polls. But tell the world that the multiparty charter (MPC) is just a few percentage points behind the ANC; then say the ANC can only stay in power by going into coalition with neo-fascists ... well, now the stakes are existential. Now your work may actually push the opposition over the line.
The first poll that looked a little odd was commissioned by the now dormant Change Starts Now in February. It told us that the ANC was coming in at 39%. But the poll, conducted by veteran pollster David Everatt, neither excluded nor reallocated non-responses and so in effect fabricated a large, ghost party.
Recalibrating Everatt’s data in an article in the Daily Maverick, former statistician-general Mark Orkin concluded that his poll should have had the ANC at 47%. Change Starts Now was hoping to rob the ANC of several hundred thousand votes. I guess it could be coincidence that the work its funders paid for sang a hopeful tune.
The Change Starts Now poll was conducted before the spectacular rise of Jacob Zuma’s MK party, which is actually threatening to steal a lot of votes from the ANC and others. The first poll published after the MK party’s rise was by the Brenthurst Foundation and showed MK at 25% in KwaZulu-Natal and 13% overall, with the ANC down to 39%.
But do some arithmetic on the Brenthurst Foundation’s data and you soon find something strange. If MK is at 25% in KwaZulu-Natal and 13% nationally, it is polling at an average of 9% in the eight provinces other than KwaZulu-Natal. Given that it has little support outside of Gauteng and Mpumalanga, it would have to be polling spectacularly in these two provinces.
But in Gauteng the Brenthurst Foundation had MK at a mere 6% on the provincial ballot. This means a staggering number of Gauteng voters are planning to give their national vote to MK and their provincial vote to others. This is odd to say the least, breaking violently with voting patterns over the past 30 years and suggesting an unprecedented new mentality in the SA electorate. Who knows, perhaps that is what’s happening. It would be strange though.
The latest poll suggesting that the stakes are existential was released last week by the Social Research Foundation. It puts the ANC at 37% and MK at 13%. Add the two together and you get 50%, exactly enough to govern. Another lucky coincidence, perhaps, that the doomsday scenario is written perfectly into the numbers.
In the UK, Canada, Australia and elsewhere polling is highly regulated; pollsters must open their methodology up for scrutiny before publishing their findings. Not in SA. All the pollsters cited above have shown us no more than cryptic fragments of their methods. When independent social scientists have written to ask for more information, they have been answered with silence.
Maybe these polls are all correct and by early June I will be eating humble pie. But if the ANC comes in at, say, 45%, or MK at 6%, the credibility of polling in SA will have been horribly damaged.
White pollsters using their hard skills to skew the data to back the white party? That’s supposed to be bullshit sprouted by Jacob Zuma and Julius Malema. It isn’t meant to be true.
• Steinberg teaches at Yale University’s Council on African Studies.
Would you like to comment on this article?
Sign up (it's quick and free) or sign in now.
Please read our Comment Policy before commenting.