When the terrible fire broke out at King's Cross Station, most people's instinct was to rush up the escalators, to get out. In fact, since heat rises, the best course may have been to go down into the station's lowest depths.
Science, very often, is counter-intuitive. But why is it that people find it so hard to apply even the facts that they know? Somehow, the theory taught in school (heat rises) is not connecting with people's everyday lives.
"It is very common for children to apparently hold two views at once," says Tina Jarvis, director of SCIcentre for initial teacher training at Leicester University and Homerton College, Cambridge. "They learn the science answer and give it in science situations while still interpreting the world in other ways." Pupils learn the facts but do not really understand them.
Mrs Jarvis and her colleagues argue that good, accurate science teaching in primary schools is an essential start if the general public is to become well enough informed to make some judgments about issues such as BSE (mad cow disease) or global warming, and for the system to produce enough scientists to find ways of dealing with them.
There seems to be a consensus about why science should be taught to all primary pupils, what type of science, and even how it should be taught.
Science is a core subject in the national curriculum. Seven-year-olds are assessed and given a level by their teachers; 11-year-olds are tested in it, and their results appear in the performance tables. Nevertheless, the subject could be under threat when the national curriculum is reviewed, certainly as a core subject, possibly as a requirement for younger primary children. So the experts' rationale is worthconsidering.
They say science contributes distinctive skills, concepts and perspectives, not offered by any other subject. Its ideas are often counter-intuitive, so they can't be acquired informally. Learning facts will not help people respond to new situations. They need to learn to think scientifically. Misconceptions need to be addressed early because, once entrenched, they are hard to change.
There is also broad agreement on what type of science should be taught in primary school. Children should be involved in practical investigations, and these should lead to explanation, not just observation; children should begin to develop the ability to create models for the way things work by explaining their findings.
One of the problems in primary science has been teachers' lack of knowledge; a guessed answer to a child's question is quite likely to be wrong. Mrs Jarvis suggests that children should learn, and spend time on, a limited set of concepts, which teachers would be trained in.
It is also important to demonstrate how science is relevant to everyday things. For this reason, Mrs Jarvis strongly opposes the introduction of specialist science teaching in primary schools. "If I am going to produce the child who can see the value of science, it can't be separate."
The integrated approach also enables science to support other areas of the curriculum, particularly maths and English. "To make these subjects relevant they need to be applied. Talking and writing about science requires high quality descriptive language. Measurement, calculation and communicating findings through graphs and tables can make mathematics understandable. "