Social Media, Misinformation, and Public Communication

Why this matters in public value

For most of the twentieth century, a government agency, a national broadcaster, or a handful of established newspapers formed a shared authoritative channel: one version of events reached most citizens through a small number of gatekeepers (see the Wikipedia article on gatekeeping in communication, referenced below). That arrangement is gone. Citizens now encounter public information alongside rumour, satire, advertising, and deliberate falsehood in the same social-media feed, often from sources with no editorial accountability and no obligation to correct an error. A public body that once needed only to issue a press release now competes for attention inside an environment it does not own and cannot moderate.

This matters for public value because Mark Moore's strategic triangle treats legitimacy and support as one of the three legs an organization needs, alongside a value proposition and operational capacity (see Topic 1.2 — The Strategic Triangle: Legitimacy, Value, and Capacity). Communication is how an authorizing environment learns what an institution is doing and why, and it is also how that environment learns to doubt it. When false claims about a vaccine's safety, an election's integrity, or a flood warning's accuracy circulate faster than an official correction, the operational damage is measurable: lower vaccination coverage, lower compliance with evacuation orders, and higher costs to re-establish facts that were once taken for granted. Communication failure is not a reputational inconvenience sitting alongside the real work of public value creation; in a fragmented media environment it has become part of the real work.

The stakes are asymmetric. A false claim can be written and published in seconds; an institutional correction typically requires clearance, legal review, and a communications sign-off before it reaches the same audience — by which time the false claim has often become the more familiar version of events. Public and social-sector leaders who treat communication as an afterthought to policy design are, in practice, ceding the first and often the most durable account of their own decisions to whoever moves fastest.

Core concepts

Media fragmentation describes the shift from a small number of shared, high-trust channels to a large number of competing, algorithmically curated ones, each showing a different citizen a different version of public life. Fragmentation weakens the gatekeeping function that once filtered claims for accuracy before publication and replaces it with platform ranking systems optimized for engagement, not veracity.

Misinformation is false or misleading information shared without an intent to deceive — a citizen who shares an outdated statistic in good faith is spreading misinformation. Disinformation is false or misleading information deliberately created or spread to deceive, for political, financial, or strategic gain; the distinction lies entirely in intent, not in the content itself, which is why the same false claim can start as disinformation from one source and spread onward as misinformation from many others. A closely related term, malinformation, describes information that is factually accurate but deployed out of context to cause harm (for example, a genuine document leaked to mislead rather than to inform); it is worth naming because official fact-checking built only to catch falsehoods will miss it.

An infodemic — a term the World Health Organization (WHO), the United Nations agency for international public health, adopted and formalized during the COVID-19 pandemic — is an excessive volume of information, both accurate and inaccurate, that makes it hard for people to find trustworthy guidance when they need it. Infodemic management treats the volume and velocity of information itself, not only its truth value, as a public health and public safety variable that institutions can monitor and respond to.

An echo chamber is an environment in which a person mainly encounters information that confirms views they already hold, typically through the people they choose to follow; a filter bubble is the related but distinct effect of a platform's own recommendation algorithm narrowing what a person sees without their choosing it. Both reduce a citizen's exposure to correction, but for different reasons — one social, one technical — and so require different responses.

Prebunking, grounded in inoculation theory, is the practice of exposing people to a weakened, pre-emptive version of a misleading argument or technique before they encounter the real thing, so that they build resistance to it in advance — the communication equivalent of a vaccine, in contrast to debunking, which corrects a falsehood after it has already spread and already done some of its damage. Media literacy is the broader, longer-horizon capability of being able to access, evaluate, and critically assess information regardless of its source, which prebunking exercises and platform-specific corrections both depend on citizens already having to some degree.

Crisis communication is the practice of communicating quickly, honestly, and consistently while an emergency is still unfolding and facts are still incomplete; it is one thread of the wider discipline of emergency and resilience management (see Topic 3.12 — Crisis, Emergency, and Resilience Management), but it is treated here specifically as a communication problem: what to say, through which channel, and how fast, when the institution does not yet have all the answers.

Legitimacy, in the political-science sense, is the public's acceptance of an authority's right to govern or to be believed, independent of whether every one of its actions is popular. Misinformation and disinformation damage legitimacy not primarily by winning any single argument but by cumulatively teaching citizens that official sources cannot be relied on to be first, fast, or complete — the same trust deficit examined at greater length elsewhere in this part (see Topic 4.2 — Trust in Government and Civic Engagement).

Best practices

1. Publish through owned channels first, and treat platforms as distribution, not the record. Maintain a website, mailing list, or other channel the institution fully controls as the authoritative version of any statement, and link every platform post back to it, so that a citizen who wants to verify a claim has somewhere stable to go.
2. Prebunk the falsehoods you can already predict. Before a election, a vaccination campaign, or a severe-weather season, identify the two or three false narratives most likely to recur and publish a short, plain-language explanation of the manipulation technique in advance, rather than waiting to react to each individual claim as it appears.
3. Staff a real-time listening function during live events. Assign a named person or small team to monitor what is actually circulating during a crisis or a major announcement, so that communications leadership is responding to the rumours citizens are actually seeing rather than the ones the institution assumes are spreading.
4. Set a public, standing correction protocol and use it consistently. Decide in advance who can approve a correction and how fast, publish corrections in the same format and on the same channels as the original claim, and avoid amplifying a fringe falsehood by correcting it so prominently that more people see the correction than ever saw the original claim.
5. Build one shared operating picture across every agency involved in an event. When multiple public bodies are communicating about the same crisis, agree a single set of facts and a single spokesperson structure before the first statement goes out, so citizens do not receive contradictory information from agencies that are, in practice, on the same team (see Topic 3.12 — Crisis, Emergency, and Resilience Management).
6. Invest in media literacy as infrastructure, not as a campaign. Fund media literacy work — in schools, libraries, and community organizations — on a multi-year basis, because the ability to evaluate a source is a durable civic capability that a single seasonal campaign cannot install.
7. Meet citizens on the platforms they actually use, not the one the institution prefers. Track where the audience the institution most needs to reach — by age, language, or region — actually spends time, and resource communication there, rather than defaulting to whichever channel is easiest for the communications team to run.
8. Separate the falsehood from the correction clearly enough that the correction is what people remember. State the accurate fact first and most prominently, name the falsehood only briefly and clearly as false, and repeat the accurate fact again at the end, so that a skimming reader is not left with the myth as the most memorable part of the message.
9. Work through locally and communally trusted messengers, not only official spokespeople. Community leaders, local clinicians, faith leaders, and community organizations often carry more day-to-day trust with a specific audience than a national agency does, and a message co-delivered through them typically travels further and is believed more readily.
10. Measure belief change and behaviour, not just reach. Track whether a corrective campaign actually shifted what a sampled audience believes or how they act — vaccination uptake, evacuation compliance, voter-registration accuracy — rather than only how many people saw or shared it, since high reach with no belief change is a wasted budget.
11. Publish the underlying data and methodology behind a contested claim, not only the conclusion. When a statistic is disputed, releasing the data and the method behind it, in a form a technically literate journalist or citizen can check, does more long-run reputational work than repeating the conclusion more loudly.
12. Rehearse the crisis communication plan with the people who will actually use it, before a real crisis arrives. A communication plan tested only on paper reliably breaks down under the time pressure of a live event; a plan rehearsed with a realistic simulation, including a misinformation injection, exposes the gaps while they are still cheap to fix.

Questions to discuss with your team

Do we have a channel citizens would trust ahead of a fast-moving false claim, and do we know how many of our own citizens actually use it? Many public bodies assume their website or official social-media account is the default source citizens will check, but usage data often shows a much smaller reach than assumed, concentrated in an audience that was already engaged. If a false claim about your organization's decisions went viral tomorrow, discuss honestly how many of the people affected would see your correction within the first hour, through which channel, and in what language. Where the answer is 'not many,' the team should treat channel-building as unfinished infrastructure rather than assuming it already exists. This question is best answered with real analytics, not assumption.

What is our prebunking programme for the misinformation we can already predict, and who owns it? Most institutions can name the two or three false narratives most likely to recur around their next election, health campaign, or infrastructure decision, yet few have a standing, resourced prebunking effort rather than a one-off campaign. Discuss which predictable falsehoods are coming in the next 12 months and whether anyone has explicit responsibility, budget, and sign-off authority to prebunk them now, before they spread. Consider also whether your organization has the internal expertise to design a prebunking message well, since a poorly designed one can backfire by repeating the falsehood more than the correction. If no one owns this today, decide in the room who will.

How fast can we issue a correction, and who has the authority to approve it outside office hours? Many correction protocols exist on paper but collapse the first time a false claim spreads on a weekend or overnight, because approval sits with someone unreachable. Walk through, as a team, exactly what happens from the moment someone on your staff first sees a viral falsehood to the moment a correction is published, naming every approval step and every person involved. Identify the single slowest step, and discuss whether it can be pre-delegated for time-critical cases. A protocol nobody has tested under time pressure is not yet a real capability.

How do we reach citizens who do not follow us on any platform and do not read local news? Every institution has an audience it structurally struggles to reach directly — often the people with the least time, the least trust in institutions, or a different first language — and misinformation reaches them just as easily as it reaches anyone else. Discuss who, concretely, this audience is for your organization, and what non-digital or community-mediated channels (a community organization, a local pharmacy, a faith leader, a workplace) might reach them where your own channels cannot. Consider whether current budget reflects the difficulty of reaching this audience or simply the ease of reaching the audience you already have.

What do we actually know about why citizens distrust us, and does our communication strategy address that cause or only its symptoms? Trust deficits have specific, traceable causes — a past broken promise, a slow crisis response, a policy that visibly failed a community — and communication built without understanding the cause often addresses the wrong problem (see Topic 4.2 — Trust in Government and Civic Engagement). Discuss what evidence you have about the specific sources of distrust in your service area, as opposed to a general assumption that 'people don't trust government.' Consider whether your communication plan is designed to rebuild trust in the specific area where it was lost, or is a generic messaging exercise applied everywhere equally. Communication cannot fix a trust problem it has not correctly diagnosed.

How do we know whether our media literacy investment is actually working? Media literacy funding is easy to announce and hard to evaluate, and many programmes run for years without any evidence of whether participants' ability to evaluate information has actually changed. Discuss what evaluation method, if any, currently exists for your media literacy work, and whether it measures belief and behaviour change or only attendance and satisfaction. Consider whether the programme should be redesigned around a testable claim, such as improved accuracy in identifying manipulative content, rather than around a general aspiration. An unevaluated media literacy programme is a belief, not yet a fact.

In practice: a public value example

The public health directorate of a mid-sized coastal county government received laboratory results late on a Friday afternoon indicating elevated bacterial contamination in one section of the municipal water supply, serving roughly forty thousand residents. Before the directorate had finished drafting its precautionary boil-water notice, a resident's post claiming the water was 'poisoned' and that officials were 'covering it up' had already been shared several thousand times on a regional social-media group, mixing the accurate underlying concern with an inaccurate and more alarming account of the cause.

The directorate's communications lead, who had run a tabletop exercise on exactly this scenario eighteen months earlier, activated a pre-agreed protocol rather than starting from scratch. A single spokesperson was named within thirty minutes, avoiding the contradictory statements that had damaged public confidence during an unrelated incident the previous year. The directorate published the boil-water notice simultaneously on its own website, by text alert to registered residents, and as a plain-language social-media post that named the viral claim directly, corrected the specific inaccuracy (the contamination was a treatment-process lapse, not deliberate), and stated clearly what remained unknown, including the expected duration of the notice.

Recognizing that its own following on social media reached only a fraction of affected residents, the directorate also contacted three community organizations and two local pharmacies it had built relationships with during a prebunking exercise the previous year, asking them to share the same core facts in their own words. Within four hours, county data suggested more residents had seen the corrected notice through a community channel than through the county's own accounts. A short follow-up notice a day later, once the treatment fault was fixed, again used the same channels and again named what remained precautionary rather than confirmed.

An internal review afterward credited three specific choices with limiting reputational damage: the pre-agreed single-spokesperson protocol, the decision to name and correct the false 'deliberate poisoning' claim directly rather than ignoring it, and the reliance on community messengers the directorate did not have to build from scratch during the crisis itself. The review also noted a genuine gap: the directorate had no equivalent relationships in the county's second-most-spoken language, and the corrected notice reached that community roughly a day later than everyone else. This scenario is illustrative and does not describe any real jurisdiction's water system, agency, or incident.

Four sector lenses

Local government

Local government communication operates at the scale where trust is most personal and most fragile: a resident's opinion of the whole institution is often set by one interaction with one local service. Local bodies typically have far smaller communications teams and budgets than national ones, which makes relationships with community organizations, local media, and informal community leaders disproportionately valuable, since they extend reach the institution cannot buy directly. The corresponding risk is that a small team with no dedicated monitoring capacity is often the last to know a false claim is spreading locally, precisely where it can do the most immediate harm to compliance with a safety notice or a service change.

National government

National government has historically relied on a small number of national broadcasters and newspapers as a de facto shared channel, and the loss of that arrangement has been the most visible of any sector, particularly during contested elections and public health emergencies. National bodies are also the most frequent target of coordinated, cross-border disinformation campaigns, which the Organisation for Economic Co-operation and Development (OECD) — the intergovernmental body whose thirty-eight member countries coordinate economic and public governance policy — has documented as a growing threat to democratic information ecosystems, distinct from the everyday misinformation any institution also has to manage. National governments generally have the resources to build sustained prebunking and media literacy programmes but must guard against a different risk: official correction efforts that look like political messaging lose the very neutrality that made them credible in the first place.

Social sector and nonprofit

Nonprofits and non-governmental organizations (NGOs) often carry a level of public trust that outpaces government's, particularly on issues like public health or humanitarian response, which makes them valuable communication partners but also a target for disinformation designed to discredit them specifically. Social-sector organizations typically communicate on far smaller budgets than government agencies, which pushes them toward reliance on a single platform for reach — a structural vulnerability if that platform changes its policies or algorithm. Because much social-sector funding is mission- or campaign-specific, nonprofits can also struggle to justify the multi-year, unglamorous investment that media literacy and prebunking work requires, since funders often prefer to support a visible programme over invisible resilience-building.

Multilateral and international

Multilateral bodies face an infodemic that does not respect borders or languages, coordinating corrective information across dozens of jurisdictions with different legal regimes for speech and different levels of institutional trust. The WHO's infodemic management work, developed through the COVID-19 pandemic, is the clearest example of a multilateral institution treating information itself, not only disease transmission, as a variable it must actively monitor and manage. International bodies must also navigate a genuine tension that national and local bodies rarely face at the same scale: sharing prebunking research and correction methodology globally while respecting that decisions about what speech to regulate remain, appropriately, a matter of national sovereignty rather than a multilateral mandate.

Common failure modes

The silence trap. An institution that waits until it has complete, verified facts before saying anything leaves an information vacuum that rumour and false claims fill first; by the time the official version arrives, it is competing against an established narrative rather than setting one. The cost is not only slower correction but a lasting perception that the institution is either slow or evasive, even once it does speak.

The debunking treadmill. An organization with no prebunking capacity spends all of its communication effort reacting to each new false claim individually, never getting ahead of the pattern, and burns out its communications staff while never reducing the volume of falsehoods it has to answer. Because debunking a claim can inadvertently increase its familiarity and reach, a purely reactive posture can leave the institution worse off than doing nothing at all on lower-stakes claims.

Single-channel dependency. An institution that builds its entire direct-to-citizen communication capability around one platform is exposed whenever that platform changes its algorithm, its moderation policy, or its terms of access, none of which the institution controls. This has repeatedly forced public bodies to rebuild an audience from near zero with no warning, precisely when they most needed a reliable channel.

The data dump. Publishing raw data or a lengthy technical rebuttal in response to a viral falsehood satisfies an institution's instinct to be thorough but rarely reaches or persuades the audience that saw the original claim, because almost no one reads past the headline. The cost is a correction that is technically accurate and practically invisible.

Over-correcting a fringe claim. Responding loudly and repeatedly to a falsehood that had only a small, fringe audience can amplify it to an audience many times larger than the one that saw the original claim — a well-documented risk in misinformation research. The cost is a self-inflicted spread of exactly the claim the institution was trying to contain.

Maturity model

Checklist

• A named owned channel exists and is treated as the authoritative record, with every platform post linking back to it.
• The two or three most predictable false narratives for the next 12 months have been identified and assigned an owner.
• A prebunking message has been drafted and tested for at least one predictable narrative.
• A correction protocol exists in writing, naming who can approve a correction, including outside office hours.
• A real-time listening function is assigned to a named person or team for major events and announcements.
• Relationships with at least two community-trusted messengers, covering the languages spoken locally, are documented and maintained between crises.
• The crisis communication plan has been rehearsed with a realistic simulation within the past 12 months (see Topic 3.12 — Crisis, Emergency, and Resilience Management).
• A shared single-spokesperson structure is agreed with every other agency likely to be involved in the same event.
• Media literacy work is funded as a multi-year programme, not a single campaign.
• An evaluation method exists that measures belief or behaviour change, not only reach or impressions.
• The underlying data behind any contested public claim can be published within a defined time limit.
• The specific, evidenced causes of local distrust have been identified rather than assumed (see Topic 4.2 — Trust in Government and Civic Engagement).

Key sources

The World Health Organization's infodemic management programme, developed through the COVID-19 pandemic, is the most fully worked example of a public institution treating information volume and velocity as a variable to be actively managed rather than a side effect of a crisis. The OECD's information integrity work, including its 'Facts not Fakes' assessment of how member governments are responding to disinformation, grounds this topic's claims about the institutional and cross-border dimension of the problem. The academic literature on inoculation theory and prebunking, particularly the research associated with Jon Roozenbeek, Sander van der Linden, and their collaborators, grounds this topic's claims about what has been shown, in controlled research, to build resistance to misinformation before it spreads, as distinct from correcting it afterward.

References

1. Misinformation — Wikipedia — https://en.wikipedia.org/wiki/Misinformation
2. Disinformation — Wikipedia — https://en.wikipedia.org/wiki/Disinformation
3. Inoculation theory — Wikipedia — https://en.wikipedia.org/wiki/Inoculation_theory
4. Media literacy — Wikipedia — https://en.wikipedia.org/wiki/Media_literacy
5. Infodemic — Wikipedia — https://en.wikipedia.org/wiki/Infodemic
6. Echo chamber (media) — Wikipedia — https://en.wikipedia.org/wiki/Echo_chamber_(media)
7. Filter bubble — Wikipedia — https://en.wikipedia.org/wiki/Filter_bubble
8. Fake news — Wikipedia — https://en.wikipedia.org/wiki/Fake_news
9. Crisis communication — Wikipedia — https://en.wikipedia.org/wiki/Crisis_communication
10. Political legitimacy — Wikipedia — https://en.wikipedia.org/wiki/Legitimacy_(political)
11. Social media — Wikipedia — https://en.wikipedia.org/wiki/Social_media
12. Gatekeeping (communication) — Wikipedia — https://en.wikipedia.org/wiki/Gatekeeping_(communication)
13. Infodemic — World Health Organization — https://www.who.int/health-topics/infodemic
14. Facts not Fakes: Tackling Disinformation, Strengthening Information Integrity — OECD — https://www.oecd.org/en/publications/facts-not-fakes-tackling-disinformation-strengthening-information-integrity_d909ff7a-en.html
15. Prebunking interventions based on 'inoculation' theory can reduce susceptibility to misinformation across cultures — Roozenbeek, van der Linden, and Nygren, Harvard Kennedy School Misinformation Review — https://misinforeview.hks.harvard.edu/article/global-vaccination-badnews/