Political polls often get voters and commentators excited. But they are often a sign of evanescent emotion rather than considered decisions, and rarely match ballot results. But that doesn’t stop them being used to sway opinion.

Overseas in the West, where voters are highly responsive to this form of persuasion (and at home, where whites are most easily affected), polls are often deliberately designed to elicit certain responses, through psychological strategies like priming effects, which in the recent past have even included the use of pathway-dependent question ordering with automated online questionnaires - if you respond a certain way to one question, you get the questions in a different order, or get a different set of questions designed to nudge you toward the “right” answer.

Here in South Africa, we can’t afford to be that sophisticated. We also can’t afford to be quite that tricky - pushing a silly outlier result can often get you ridiculed, especially if it supports an existing bias known to affect your organisation.

‍

Rainbows and sunshine

With that in mind, the DA-supporting IRR will be releasing their new poll in full on Tuesday, which has been released as an early exclusive to Netwerk24, likely in the hopes of boosting white conservative confidence in the DA. The results are clearly dramatic:

These boasts have been made before on the basis of similarly outlying polls. Greg Krumbock (the late DA campaign strategist) claimed a 32% DA, 39% ANC result, while Leon Schreiber touted similar results before the election, including that the DA is the most favourably-viewed party in the country.

Much of their data is prima facie credible, but the last Brenthurst Foundation poll was much less dramatic, and was made at the same time as the IRR’s. From our article on the poll:"

“Voting intentions have remained stable for the most part, with the ANC steady at 41% of the vote, rising to 43% when adjusted for turnout […] The DA has seen an unexpected national bump up to 26%, or 27% with turnout factored in, but has been weakened in the Western Cape, falling 9% to 46% since 2021…”

If that is true, the DA will likely invite the ANC to govern the Western Cape with them next election, rather than partner with market-liberal nonracialist parties like the VF+, who they have threatened to kick from every municipal coalition in the province for daring to take them to task for corruption and ANC collusion.

But it all begs the question, why do the two polls deviate from each other so much?

Most of the time, polls have a margin of error of 3-4% (much like the IRR’s), meaning that results could be over or under the real sentiment by as much as that figure. And even taking this into account, in every election since exit polls were conducted, the ANC has consistently overperformed opinion poll estimates by 3-4%.

My suspicion is that this is accounted for by failure to find an accurate methodology for adjusting for turnout, since no other particular parties seems largely unaffected by this blip. With those assumptions in mind, the IRR’s results give us the ANC at 30-36%, and the DA at 27-33%, while Brethurst gives us 44-50% for the ANC and 26-32% for the DA.

Still impressive, but why does the IRR deviate rom Brenthurst so much? Is it sampling bias? Cities are easier to poll in person, for sure. Phone and internet polls are easier to conduct, but easier to ignore from the target’s perspective. Brenthurst are certainly known to favour a lot of travel and legwork, moreso than the IRR.

But ultimately, we don’t really know. Certainly the DA has gotten a boost in black support, likely driven by a combination of the Department of Home Affairs turnaround, and their increasing defences of BEE and land reform, including defending the Expropriation Act from attacks by conservative organisations, and blaming AfriForum for Trump’s sanctions.

The trouble with these sorts of polls is that they are often affected by what we shall call “protest sentiment”. If people are dissatisfied with a certain party they’re used to voting for, they might be more likely to say they won’t vote, or will choose someone else, whether or not they actually end up voting for them later. Another effect is controversy - if they believe they might be judged for their opinion, they will hide it, and say something less controversial.

But that doesn’t even account for changes in sentiment. Tracking the polling leading up to the 2024 election, the SRF’s results in April showed the DA at 32%, but they fell to 23% by the time of the election, and got 21% in the end. A number of such polls have overestimated the DA’s support by anything from 5-10%.

‍

‍

Hope?

And what is most interesting, is what effect hope has on the electorate. When Ramaphoria was in the air, the DA lost a significant chunk of their most retarded supporters, who praised Ramaphosa for showing up and making promises, and because they have their eyeballs glued to News24’s nonstop arse-kissing extravaganza.

But that soon died away. Now we have hope in the GNU, mainstream parties have gotten a bump. The SRF’s poll from September last year showed an increase for both the status quo parties. And of course, the ANC has reverted to type, and people have continued to desert it, leaving this GNU bump only favouring the DA - for now.

And lastly, this doesn’t take into account the possibility of unforeseen changes. Before MK rocked up on the scene, the picture was different. They absorbed votes from both the ANC and the EFF, as well as a little from the IFP, who at the time were slated to win KZN in coalition with the DA. The SRF’s poll from February 2024 showed:

Their March 2025 poll looking at the perception of the “progressive caucus” (MK/EFF) showed that black perceptions are split more or less down the middle, with a third undecided, a third opposed, and a third in favour.

But this is no different for the Cape independence movement. The movement’s popularity rose with the fall of Ramaphoria and the realisation of the permanence of the ANC’s programme and the general state of decline. The GNU’s offer of hope for inclusion of minority interests into the national programme has also put a damper on people’s desperation.

The latest poll from the CIAG has seen a dip in support for secession, though primarily among whites, many of whom have always felt a guilt-based compulsion to show loyalty to the new national project:

But these feelings depend on the ability of the GNU to deliver. Phil Craig and his team have seemingly shown no reluctance to publish this poll despite the dip in support, likely driven by their confidence that support will bounce back as the GNU fruits rot on the branch.

While some areas have seen recovery, the overall picture is that we are travelling in the same direction, and sinking at the same speed - no coalition will be able to justify stripping out 30% of state expenditure, and taxes are already being extracted well beyond the Laffer Curve.

The ANC is also seeing the rise of the radical faction, ahead of the expected unseating of Ramaphosa as head of the ANC.

While the ANC may be capable of swallowing a tactical retreat for now, the likelihood that they will be scrapping the National Democratic Revolution just as the last elements are being passed in Parliament seems highly unlikely.

In that case, the DA will neither be able to stop them, and their opposition to secession will guarantee that everyone is dragged down together.

‍