A poll putting the ANC at just 18% in Johannesburg has put the spotlight on the research behind the figure, including who commissioned it, who conducted it and whether the findings accurately reflect the party’s support ahead of the November local government elections.
The poll, commissioned by the Social Research Foundation (SRF) and conducted by Victory Research, puts the DA at 42% in Johannesburg, with the MK Party and ActionSA also recording double-digit support.
The ANC’s 18% represents a 12-percentage-point drop from the previous Johannesburg poll, which placed the party at 30%.
Nationally, the SRF poll puts ANC support at 34%, down from about 40% in the first quarter, while the DA is on 27%, the MK Party on 15% and the EFF on 7%.
In eThekwini, the ANC has fallen to third place behind the MK Party and DA, while Tshwane remains its strongest metro, with support just above 30% but still below the DA, which is above 40%.
The poll was conducted between July 8 and 31 among 504 Johannesburg respondents, with a 4.4% margin of error at a 95% confidence level.
It also uses a 53% turnout level for the Johannesburg scenario, a factor that has become central to questions about how the figures should be interpreted.
Who commissioned the poll?
Speaking to IOL,Victory Research chief executive Gareth van Onselen confirmed that the SRF commissioned the survey and that the foundation is a client of Victory Research.
“The Social Research Foundation commissioned the poll. The Social Research Foundation is a client of Victory Research,” Van Onselen said.
The relationship between the two organisations is significant because questions about the independence of Victory Research and its political associations are not new.
In 2024, ActionSA national chairperson Michael Beaumont questioned SRF polling after it emerged that Johan van der Berg, Victory Research’s founder and managing partner, was also employed by the DA as its head of research.
DA leader, at that time, John Steenhuisen confirmed at the time that van der Berg worked for the party and handled its polling.
“He works for the DA and does our polling, he obviously contracts out because of the specialist skills that are required,” Steenhuisen said.
Beaumont described the arrangement as a “severe conflict of interest” and argued that voters should know who was behind polling presented as an assessment of public opinion.
The criticism did not, however, establish that the polling was inaccurate. The final SRF tracking poll before the 2024 election put the ANC at 42.2%, the DA at 21.6%, the MK Party at 12.4% and the EFF at 10.8%. The actual results were 40.2%, 21.8%, 14.6% and 9.5% respectively
Van Onselen has previously worked for the Democratic Alliance, while Victory Research founder and managing partner Johan van der Berg has also been associated with the party.
Ahead of the 2024 elections, concerns were raised about the political links between the polling company and the DA, including van der Berg’s role in the party’s research and polling.
Van Onselen dismissed concerns over his previous senior roles within the DA and their potential impact on the independence of the latest research.
“I don’t care about them,” he said.
Asked whether he was currently affiliated with, or providing polling or research services to, any political party, Van Onselen said private clients were not disclosed without their approval.
“No private company will reveal its private clients unless its private clients approve. Victory Research is no different,” he said.
The SRF was founded by political analyst Frans Cronje, who previously served as the chief executive of the South African Institute of Race Relations.
The political backgrounds of those involved have become part of the debate around the latest poll, but analysts differ on how much weight should be placed on them.
Professor Theo Neethling said the reliability of the poll should be assessed primarily through its methodology, sampling, weighting and turnout assumptions rather than the political backgrounds of those involved.
He also pointed to the performance of the SRF’s polling during the 2024 elections.
The final SRF tracking poll put the ANC at 42.2%, the DA at 21.6%, the MK Party at 12.4% and the EFF at 10.8%.
The actual results were 40.2% for the ANC, 21.8% for the DA, 14.6% for the MK Party and 9.5% for the EFF.
“I consider Gareth van Onselen to be a reliable and seasoned polling analyst, especially in light of the fairly robust results of the Victory Research/SRF polling in 2024,” Neethling said.
How much should voters read into the 18%?
Neethling cautioned against treating the Johannesburg figure as a prediction of the November election result.
“There is an important methodological issue here: turnout assumptions matter enormously,” he said.
He said a poll of registered voters was not necessarily a prediction of the eventual electorate, with turnout and tactical voting capable of producing sizeable differences between support recorded in a poll and the final result.
Neethling said South African polling had generally been effective at identifying broad electoral patterns, although it had not consistently forecast precise outcomes.
He said the 2024 election was an example of both the strengths and limitations of polling, with many reputable polls correctly predicting that the ANC would lose its parliamentary majority but generally overestimating its support and struggling to assess the eventual strength of the MK Party.
“For the present polling, I would therefore have more trust in the alignment of multiple independently conducted polls rather than in any individual pollster or analyst,” Neethling said.
Van Onselen defended the 53% turnout scenario used in the Johannesburg poll.
“There are no assumptions that underpin the model. 53% is the turnout level that the data allows for, without compromising the sample. It is simply an illustration of what happens to party support when turnout drops,” he said.
He also said Victory Research could not explain why ANC support had fallen by 12 percentage points between the two Johannesburg polls.
“I cannot say what is responsible for the drop. Victory Research simply measures support. The SRF will have some analysis on what is driving that decline,” he said.
Analyst questions 18% figure
Independent political analyst Goodenough Mashego was considerably more critical of the findings.
Mashego said the SRF research was “more ideological” than conventional social research and described it as a “psychological operation” intended to influence perceptions of the ANC’s prospects.
He questioned the use of telephone polling and whether it adequately reflected Johannesburg’s population, arguing that people who did not answer calls could affect the representativeness of the sample.
Mashego also questioned the relationship between SRF and Victory Research, saying Victory Research was more of a commissioned research company than a think tank conducting independent research.
He disputed the scale of the ANC’s decline suggested by the latest poll.
“The ANC will haemorrhage, but it won’t haemorrhage up to 18%,” Mashego said.
He said the ANC was likely to lose support but argued that the 18% figure should not be treated as reliable.
ANC faces wider metro threat
University of KwaZulu-Natal political analyst Zakhele Ndlovu said the ANC nevertheless faced a serious electoral threat in Johannesburg, eThekwini, Ekurhuleni and Tshwane.
He said a significant loss of support in November could indicate that the ANC could fall below 35% and even 30% in 2029, although much would depend on the outcome of the party’s 2027 elective conference.
Ndlovu said the possibility of a losing faction breaking away after the leadership contest could create further problems for the ANC.
He also said the party’s 2027 leadership race would take place against the backdrop of a party no longer guaranteed to produce South Africa’s president.
“The ANC’s credibility is on the line,” Ndlovu said.
He warned that the party’s long-term survival would increasingly depend on its ability to attract younger voters as its older support base declines.
“The cancer is at an advanced stage and it is staring death in the face,” he said.
The SRF poll also found that 65% of voters nationally support continued ANC-DA cooperation in municipalities where neither party wins an outright majority.
Support for such cooperation rises to 80% among ANC voters and 69% among DA voters.
While the 18% Johannesburg figure has intensified scrutiny of the ANC’s position in the country’s economic hub, analysts caution against treating the number as a forecast of the November election result.
Neethling said greater confidence should be placed in consistent trends across independently conducted polls than in any individual survey.
