The CNN effect is the idea that television makes governments act. Put enough starving children or bombed apartment blocks on live rolling news, the theory goes, and the public becomes unbearable, and the politicians have to do something. It is one of the most flattering things ever said about journalism — and one of the least supported by the evidence. * The phrase comes out of the 1991 Gulf War, the first conflict beamed home in real time by a 24-hour news channel. For the first time, viewers, presidents and enemy governments were all watching the same footage at the same moment, and a whole generation of officials and journalists concluded that something structural had changed: that live pictures had become an independent force in foreign policy. The 1990s gave the theory its case files — northern Iraq in 1991, Somalia in 1992, Bosnia in 1995 — each widely described at the time as governments reacting to what was on screen. * Then researchers went and checked, and the story got a lot less flattering. * Take Somalia, the single most cited example. The received version is that pictures of famine forced George H. W. Bush to send in the Marines. Steven Livingston and Todd Eachus tested that in Political Communication in 1995, using content analysis of the coverage plus interviews with decision-makers in Washington and in Africa. What they found was the sequence running the other way: the decision to intervene came out of diplomatic and bureaucratic processes inside government, and the surge in news coverage came in response to those decisions, not before them. The cameras did not summon the policy. The policy summoned the cameras. * That is the pattern that keeps turning up. Piers Robinson, who wrote the standard book-length study of the subject, argued that the classic ground-troop interventions of the era — northern Iraq in 1991, Somalia in 1992-93 — were not products of a CNN effect at all, and that where media pressure does bite, it bites only under a specific condition: when policy is genuinely uncertain and elites are divided. When a government knows what it wants to do and speaks with one voice, coverage tends to follow the official line rather than push against it. Media power, in other words, is real but conditional — it fills a vacuum; it does not create one. * The most useful piece of the literature is a typology. In a 1997 Shorenstein Center paper at Harvard, "Clarifying the CNN Effect," Livingston pointed out that people arguing about this are usually arguing about three different things. Coverage can act as an agenda-setting agent, pushing an issue up the list of things officials have to have a position on. It can act as an accelerant, compressing the time available to decide. And it can act as an impediment, making a policy harder to carry out once it is under way — casualties on screen, hostages on screen, a plan that no longer survives contact with an audience. All three can happen to the same issue at different stages, which is exactly why the debate stays muddy. * And the impediment version is the one with the strongest evidence behind it, because television is much better at stopping things than starting them. On October 3-4, 1993, US forces fought a Somali militia in Mogadishu; 18 American soldiers were killed and 84 wounded, two Black Hawk helicopters were shot down, and Somali deaths were estimated between 700 and 1,500. On October 7 — three days later — President Clinton announced that America would withdraw from Somalia by March 31, 1994. Whatever role the pictures played going in, they were unmistakably present going out. * Now hold that against the case that should end the theory outright. Between April and July 1994, roughly 800,000 people were killed in Rwanda in about a hundred days. Western media largely failed to report that a nationwide extermination campaign was under way until almost three weeks in — by which point around 250,000 Tutsi had already been murdered — and much of the early coverage framed it as tribal warfare and spontaneous chaos rather than an organised political project. There was no wall of live footage, so there was no wall of public pressure. Meanwhile the US government was so anxious about the word itself that on June 10, 1994, State Department spokesperson Christine Shelly told reporters that "acts of genocide" had occurred, and, when pressed on how many acts of genocide add up to genocide, declined to answer — because naming it as genocide carried obligations under the 1948 Genocide Convention. * That is the whole argument in one comparison. The CNN effect is not a law of physics that converts suffering into intervention. It is a description of what happens to the small number of crises that get cameras — and a silence about everything that does not. Which means the real question was never "does coverage move governments?" It is: who decides what gets covered in the first place? * That decision is made by bureau budgets, visa access, airport logistics, insurance premiums, whether a country is legible to a Western audience, and whether an outlet still has anyone within a thousand miles. A conflict where reporters can reach the story cheaply and safely produces images; a conflict where they cannot produces silence — and to the viewer at home, silence is indistinguishable from nothing happening. Rwanda in 1994 was not less appalling than Somalia in 1992. It was less filmed. * So when someone tells you that public outrage forced a government's hand, treat it as a claim to be checked, not a compliment to journalism. Ask what officials were saying before the coverage spiked, and whether the coverage led the policy or trailed it. Ask whether elites were actually split, because that is the gap where media pressure has been shown to work. And ask the harder version of the question — which crisis, right now, has no camera near it, and what would you be demanding if it did?