While the ANC were the biggest losers from last year’s election, one of the biggest winners were South Africa’s pollsters.

I am the research director at the Social Research Foundation (SRF). In 2024 the SRF called the national election result as follows: “ANC 41.6%, DA 23.4%, MK 12.1%, EFF 11.8% and IFP 2.1%”.

Yet despite its precise accuracy our work endured sustained criticism, insult, and insinuation, before and after the vote.

Jonny Steinberg asked: "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?". That misinformation was published not in the National Enquirer, but in Business Day!

Michael Beaumont, who works for Herman Mashaba, accused the SRF of shilling for the Democratic Alliance, when it became apparent that their party was getting nowhere across much of the country.

The ANC essentially suggested that the polls were fake.

Calls were even made for the state to intervene and curb polling activities.

Months later, even after our numbers proved accurate, one of the loonier of all the barbs was published by the Daily Maverick that the SRF was part of the: "wingspan of that great bird of stubborn and flawless whiteness."

Political polling is costly and exacting. It requires careful questionnaire design and planning. Once a poll questionnaire enters the field we cannot control how respondents answer. We publish responses without post hoc adjustments. If the findings unsettle readers, that reflects public sentiment rather than a defect in the data.

In addition to have polled political opinion in detail at several junctions through 2022 and 2023, in May 2024 the SRF launched a tracking poll for the 2024 election. This began with a baseline survey of 1 800 interviews and then added 200 to 300 fresh interviews daily while dropping the oldest interviews to produce a rolling daily average. This was the first time a tracker had run in full public view in South Africa. At peak interest nearly 57 000 people visited the SRF website daily. Commentators speculated about the effect the tracker was having on the rand.

The SRF tracker initially had the ANC at 37% with the DA near 30% and the EFF and MK at around 10%. Strong ANC campaigning produced an early lift for that party and for a time the ANC gained roughly 0.5 percentage points per day. Two weeks before the vote it had risen to about 47% while the DA fell to roughly 22%.

Two events reversed the ANC lift.

President Cyril Ramaphosa signed the National Health Insurance bill into law less than two weeks before the election. Within days a constitutional court decision barred Jacob Zuma from standing. The combination shifted public sentiment. Between May 16 and May 23 the ANC fell from 46.4% on a 58% turnout model to 39.8%.

A day before the election our tracker read ANC 41.6%, DA 23.4%, MK 12.1%, EFF 11.8% and IFP 2.1%. The track’s margin of error was 2% and nearly every party fell within that range. Given the extent of the political change that was marked by that election, and also the very strong showing of MK, the SRF’s efforts stand as a particularly powerful case study of the accuracy of modern polling methods.