The most cited channel was word of mouth, named by 47% of respondents. Facebook followed at 32%, then a cluster of community-level settings: social events (24%), social mobilisation activities (23%), WhatsApp (20%), religious events (20%), television (20%), and conversations with health practitioners (19%). The top ten channels mix online and offline roughly evenly.

In almost every country in the top ten, respondents reported hearing rumors through at least one online platform and at least one offline setting. The cross-channel pattern held across very different information environments.

That pattern is what the headline number reflects. Rumors do not sit in a single channel: a claim first encountered on TikTok may be mentioned at a religious gathering days later; a conversation at a social event may circulate back into a WhatsApp thread by the weekend. For 41% of respondents, the rumor had already moved.

This is not an exception. It reflects how information ecosystems function in practice.

For practitioners, this has a direct implication: listening that focuses on a single channel, even at scale,  will systematically miss how narratives emerge, evolve and gain traction.

Respondents described a wide range of narratives. After thematic coding of the open-text answers, three themes stood out.

The most common, named in 17% of responses, was conflation with other vaccines entirely. When asked about polio, respondents named COVID-19, measles, BCG, HPV, HIV or mpox. In practice, respondents describe this as a generalised “injection risk” framing — where different vaccines are discussed interchangeably, for example: “this is like the COVID vaccine,” “all these injections have side effects,” or “this is another experiment.”

This suggests that in many communities, vaccine concern is not framed in polio-specific terms but as a broader idea that cuts across products.

The second most cited, raised by 14% of respondents, was the persistent sterility and population-reduction narrative. It appeared across Nigeria, Malawi, Ethiopia and Pakistan (not confined to any single country) and is a decades-old claim that this poll confirms is still moving. Respondents described this in concrete terms such as: “the vaccine reduces fertility,” “it is used to control population,” or “it targets specific communities.” These framings often appear in both religious and political narratives, and tend to resurface during campaigns or in moments of heightened attention.

The third, at 11%, was generalised "vaccines are harmful" framings, often anchored to specific symptom concerns (fever, paralysis, tumors, hormone changes). Conspiracy framings such as government plots, Western interference, "testing on our children”, appeared in around 6% of responses, often clustered in specific local contexts. These concerns are frequently conveyed through personal or second-hand experiences — for example, linking vaccination to visible symptoms or long-term health effects — which can make them more credible within peer networks.

Conspiracy framings — government plots, Western interference, "testing on our children" — appeared in around 6% of responses, often clustered in specific local contexts.

Taken together, these patterns show that narratives are not only moving across channels, but also across topics — shifting between specific vaccines, broader health concerns, and political or social interpretations.

Of the top-ten countries, Pakistan stood apart. Pakistani respondents described a Facebook-led information environment: Facebook at 53%, word of mouth at 18%, Instagram and YouTube each at 24%. Half of Pakistani respondents (50%) linked the rumor to a specific recent event or public figure, compared with 27% in the rest of the sample.

Even in Pakistan, though, 38% of respondents still report hearing the rumor through both online and offline channels. The channel weights shift. The cross-channel pattern does not.

This reinforces a consistent finding across contexts: channel dominance may vary, but cross-channel transmission remains constant.

Asked what would help their communities stop believing these rumors, respondents stacked categories rather than selecting between them.

54% named education and awareness campaigns. 9% pointed to trusted messengers — religious leaders, community leaders, elders. 6% called for community mobilisation and door-to-door engagement. 5% named frontline health workers. 6% mentioned social media and digital platforms. A further 4% emphasised testimonials and real stories; another 4% explicitly named dialogue and two-way listening.

Social media did not appear in responses as the problem to be solved. It appeared alongside human outreach, not in place of it.

The response respondents describe is integrated — education and trust, humans and platforms, online and offline.

In other words, the way communities expect responses mirrors the way rumors spread: across channels, through both people and platforms.