When a 4.7 magnitude quake woke Yanli Duode, the deputy headmaster of Yushu county's No1 Minorities Middle School at 5.39am on April 14, he decided to evacuate the boarding school's more than 800 students and staff to its playground. The quake also set off an alarm at the Qinghai province earthquake bureau, which ordered staff to return to their posts immediately. At 7.49am, as the researchers sat in a meeting room discussing the possibility of an aftershock and Yanli Duode asked his students to begin classes in the open, a devastating 7.1 magnitude quake struck the county. One building at the school collapsed and two others were badly damaged, but no students died or were injured. Away from the campus, more than 2,000 people were killed. The miracle at the school and the stark contrast with the authorities' passive response have prompted a fierce debate about earthquake prediction on the mainland. While most scientists agree that there is no reliable method for predicting earthquakes, psychics and pseudo-scientists claim to have warned the government about the disaster. On the internet, public opinion strongly favours the latter group. Dr Fang Shimin, an independent science critic in Beijing, said the mainland's debate about earthquake prediction was unique, having little to do with science and more to do with the country's history, culture and social problems. Zhang Deliang, a farmer in Dancheng county, Henan , is one of the most popular earthquake psychics on the mainland. He claims to have predicted the Sichuan quake two years ago with almost 100 per cent accuracy in terms of time, location and magnitude, and to have written warning letters to the China Earthquake Administration, the National People's Congress Standing Committee and the State Council four months before the disaster. This week, he also claimed to have predicted the Yushu quake, although less accurately, and to have warned the authorities. 'The location was more than 200 kilometres off, the date 23 days off and the magnitude 0.7 degree off,' he said. 'But 200 kilometres is a short distance in Qinghai, a vast area. With more funding and government support, I can improve the accuracy. Currently, all my estimates are made with a pen and a piece of paper.' While Zhang draws heavily on astrology when making his predictions, others have employed metallic balls hanging from the ceiling, a parrot imported from Australia and even headaches. All claim to have forecast the Yushu earthquake and to have warned the government. The chairman of the US Geological Survey's National Earthquake Prediction Evaluation Council, Professor Terry Tullis, said a small earthquake, like the one that occurred at 5.39am in Yushu, could not be used to determine whether a big earthquake was coming. 'Some earthquakes have recognised foreshocks, whereas others do not, and in any case it is not generally possible to know if an earthquake is a foreshock of a larger one to follow or just a modest-sized earthquake that signifies nothing to follow,' Tullis said. 'The status of earthquake prediction in the world is that we know the locations of the zones where earthquakes will tend to occur. However, we do not know very well when earthquakes will occur, we cannot predict exactly how large the next earthquake in a given region will be.' However, most mainlanders believe that earthquakes can be predicted and the government's failure to do so means that it is not doing its job properly. Fang said it was a situation largely of the government's own making. In the 1970s, the entire nation was mobilised to participate in earthquake prediction. 'Prediction comes from the people' became a slogan of the Cultural Revolution. After the 1974 Haicheng earthquake was 'predicted' on the basis of foreshocks, the propaganda machine heralded it as a success of Maoism. Even scientific communities in the US and Japan felt encouraged by the news and launched ambitious programmes to study earthquake prediction. But then the famously unpredicted Tangshan earthquake struck two years later, killing more than 240,000 people. Japan and the US have since given up trying to predict earthquakes, leaving the China Earthquake Administration as the world's only seismic authority that uses earthquake predictor as a formal job title. 'I can't imagine a worse job,' Fang said. 'On one hand, you can't make a prediction because there's no reliable method to make it. On the other hand, the nation will blame you whenever an earthquake strikes. 'The administrative structure of the Chinese quake authorities is built on the assumption that an earthquake is predictable.' In practice, the mass mobilisations to gather precursory observations that stemmed from the 'prediction comes from the people' edict have long been abandoned. Former premier Zhu Rongji issued a regulation in 1999 to require a 'high standard of scientific reasoning' for all earthquake predictions after more than 30 unofficial earthquake warnings in the previous three years led to the evacuation of thousands of residents and brought production lines and businesses to a standstill. However, the Sichuan earthquake, which was followed by many claims of prediction on the internet, rekindled the public's belief of quake prediction. That social pressure saw the country's lawmakers retain 'prediction comes from the people' in the new Earthquake Law, despite strong opposition from the scientific community. Zhang has been invited to give a lecture at the headquarters of the China Earthquake Administration. He said it would be closed to media. 'I won't release any predictions of future quakes to the public, however large,' he said. 'It's a state secret.'