.gif)
Some of them treated the outcome itself as the evidence, according to a review by the Centre for Analytics and Behavioural Change (CABC).
Posts inferred fraud from assumptions about how racial groups would vote, not from a discrepancy in the count.
The CABC reviewed online conversation from 6 to 11 November and recorded 49,354 mentions from about 12,810 unique authors.
Volume peaked on 6 and 7 November, while counting and objections were still underway, and made up about 71% of the period’s mentions. It fell to 7,020 on declaration day.
The IEC declared the elections free and fair, reporting 52% national turnout, up from 45.9% in 2021.
This is according to CABC's third LGE factsheet, Online conversations after LGE's in SA.
“A claim that a result must be fraudulent because of who people assumed would vote for whom can never be proved wrong.
“Every result that disagrees with the assumption becomes more ‘evidence’,” says the CABC.
“Real disputes look different. They name a ward, a station, a result slip that does not match a published tally.
“When a claim cannot name any of those, it is a belief about voters, not a finding about the count.”
The CABC found that:
One post claimed that Black voters had not voted for the Democratic Alliance (DA) and called for an IEC investigation.
Another compared the DA’s support with the white share of the population and inferred rigging.
A third treated the DA leading in townships where its author believed it had not campaigned as evidence of manipulation.
Published totals show where a party performed strongly. They cannot show how individual voters in an area voted by race.
On 7 November, the IEC reported 197 objections, two of them upheld, and recounts ordered in five wards.
The IEC also ordered a revote in one Northern Cape voting district after cast ballots were destroyed before counting.
A recount checks ballots already cast. A revote asks voters to cast them again.
A petition on X calling for a rerun of the entire election was still circulating on 11 November. It made a far broader claim than any documented dispute supports.
Posts assigned guilt, danger or political entitlement to whole groups. One praised a politician for avoiding coalitions with white people.
Others portrayed migrants as a group intent on taking over the country and harming its citizens.
The CABC distinguishes this from ordinary political discussion: criticism of a party’s record on race, and debate about immigration enforcement, are legitimate.
One post coupled a rigging allegation with language welcoming the prospect of the country burning.
A reposted message predicted that Tshwane would burn after claiming the result was rigged.
A reply on 10 November threatened arson if two parties were brought into a coalition.
These examples show that threats circulated online. They do not show that the posts organised protest or led to physical harm.
No party won an outright majority in Johannesburg, Tshwane or Ekurhuleni.
In Johannesburg, the DA took about 35% of the vote, the ANC about 20% and ActionSA about 16%.
Under the Municipal Structures Act, councillors elect the mayor, so a party with fewer than half the seats still needs enough council support.
Confusion ran in more than one direction.
One post called coverage describing the DA as the winner “misinformation” because the party lacked a majority.
A TikTok caption said the ANC had won a majority, although it led nationally with about 31% of the vote.
The CABC examined dashboard totals, recent and randomised mentions, and keyword searches on rigging, nationality and threats.
It reviewed TikTok posts separately using SentiOne and checked factual claims against the IEC and published reporting.
The review identified unsupported claims and threatening language. It did not measure how widespread any of them were, establish intent to deceive or verify offline effects.
This is the third factsheet in the CABC’s election series.
The first covered pre-election fraud claims and anti-migrant narratives.
The second covered AI-generated and misleading media during voting.
This week's full factsheet LGE is available here.
Want to change how you receive these emails?
You can update your preferences or unsubscribe from this list.
