Opinion polls are set to play an outsized role in the run-up to the 2026 local government elections, but their rising prominence brings a familiar tension: are polls neutral snapshots of public opinion or instruments of political persuasion?
Measurement, not prophecy
Polls capture a moment in time. They report what a sampled group of voters say they will do or feel at that instant — with assumptions about turnout, sampling and weighting built into the result. That distinction is especially important for municipal contests, where local personalities, ward-level dynamics and turnout fluctuations can quickly render a national or broad survey misleading.
After the 2024 national poll, several reputable surveys were broadly close to the final outcome. The official figures showed the African National Congress (ANC) on 40.18%, the Democratic Alliance (DA) on 21.81%, the MK Party on 14.58% and the Economic Freedom Fighters (EFF) on 9.52%. These numbers underline that good polling produces useful information, but they also expose the limitations when applied without local context.
| Party | Share (2024 national result) |
|---|---|
| ANC | 40.18% |
| DA | 21.81% |
| MK | 14.58% |
| EFF | 9.52% |
Why local polls are harder
Municipal elections are not a single national contest; they are thousands of ward battles and proportional-representation calculations. Three features make them difficult to poll accurately:
- Local variation: Voters in different wards react to local service delivery, councillor performance and community issues in ways national polls cannot capture reliably.
- Turnout uncertainty: A national supporter base can translate into very different outcomes depending on who actually votes on election day.
- Timing and late swings: Attitudes can shift rapidly in the final weeks, so early surveys may miss decisive movements.
These factors create openings for parties and interest groups to use favourable numbers selectively and to cast doubt on unfavourable surveys. The result is predictable: campaigns will amplify findings that suit them and attack those that do not — often before the methodology has been closely examined.
How polls become political weapons
Polling can be weaponised in several ways. A party may:
- Publicise a local or national poll showing momentum to attract donors, volunteers and media attention.
- Attack the credibility of opponents by discrediting pollsters whose results are inconvenient.
- Use selective figures to shape voter expectations, encouraging supporters to stay home or to mobilise, depending on the narrative they want to create.
This horse-trading of narratives benefits those who can most quickly and convincingly manipulate the story around the numbers: party strategists, sympathetic media outlets and external actors with resources. Voters and journalists who do not interrogate methodology risk accepting headlines that overstate what a poll can legitimately claim.
What to watch for
In the months ahead, readers should pay attention to at least three elements when assessing polls:
- Methodology: sample size, mode (telephone, online, face-to-face), weighting and the dates of fieldwork.
- Geography: whether the poll is national, provincial, metropolitan or ward-specific — and whether it attempts to aggregate results from diverse local contests.
- Turnout assumptions: how likely voters are identified and modelled to vote, and how the poll treats undecided respondents.
Good polling firms will publish those details. Parties and media that do not examine them are inviting misinterpretation.
Polling will remain a valuable tool. Properly contextualised, it can alert parties to emerging trends, help journalists prioritise coverage and give voters information about the broader political landscape. But in the 2026 municipal campaign, where ward-level nuances often decide outcomes, the raw headline number will be less and less decisive on its own.
Expect the next year to be as much about narrative management as it is about numbers. Those who benefit are the campaign teams that can turn favourable figures into momentum; the price is paid by a public that must sort signals from noise.
