The View From Here

We are, it seems, gripped by the science of small numbers. I think of it as the last legacy of Thatcher and the "me generation". Faced with a report based on a sample marginally greater than one, but which tells us something we would like to believe, we immediately seize it as a scientific truth.

The packet of cereal on my breakfast table tells me that a serving of Cheerios contains 169 calories and 281 grams of fat. As I dig into my bowlful I'm happy to know that I'm eating so sensibly - in spite of the fact that the estimate was for a 30-gram bowlful (about three tablespoonfuls with semi-skimmed milk, and I'm eating ... well, a lot more, with full- fat yoghurt and honey. And I'm outraged along with everyone else to find that the 0.4 grams of fat in a serving-size can of baked beans is probably wildly underestimated. I want my tin to contain precisely what it says - not some fluctuating figure averaged over the vat from which it came.

Most of us are poorly trained at interpreting statistics, and yet we regularly accept numbers as more convincing evidence than words. Having walked around a school on open day, talked to the teachers, quizzed the head, scrutinised the facilities, we go home, turn our backs on the evidence of our own eyes, and tot up the numbers who got grades A-C in GCSE maths instead. We tour university campuses with our daughters and sons, checking how many of the students are on e-mail, and whether the science labs have the latest in hi-tech equipment, then decide which is "better" on the basis of which university asks for 24 rather than 20 A-level points to gain entrance.

This week we were happy to be told on the basis of a scientific study that girls were more accommodating and amenable than boys. Statistically the difference was so significant, we learnt, that the scientists who had conducted the survey were postulating a gene for nice behaviour. Girls had it, boys didn't.

Anyone who had read Michelle Stanworth's Eighties study of sexual divisions in the classroom, Gender and Schooling, should have recognised the scenario: girls behave nicely, are less demanding, perform adequately without pressure and therefore get less of the teacher's time; boys are disruptive, uncooperative, and monopolise the teacher's attention. Except that in the Eighties we were told that those traits dramatically disadvantaged girls, contributing to their under-performance in class and generally in education. Female pupils were less likely to be challenged and stretched academically, with implications for their future performance.

Stanworth told us sternly that we were neglecting our girls because they were too polite. Now we are being told that the girls are fine - it is the boys who are neglected, because they lack social flexibility, so now we need to pay them special attention. Aren't we actually being told the very same thing as before, only in a different tone of voice? Because the girls are being congratulated we accept the spurious statistics of what turns out to be a thoroughly flawed, quantitative study.

If the special characteristics girls are supposed to demonstrate are ones we want to hear about - ones that make them better equipped than boys to function in today's world, more likely to achieve - then, since they are statistically presented, we accept the study's findings as the truth. And since the most immediate consequence of those findings is to suggest that boys need extra attention, extra time from the teacher, extra opportunities ... well, we haven't heard many traditionalists complaining about the figures on which the study was based either.

It is very tempting to take hold of bits of "scientific" study whenever they support our own point of view. I'm utterly committed to state education. So I'm also committed to labelling as misleading the league tables that show schools to which I didn't choose to send my children are producing better exam results than the one I did. Those schools don't enter children for an A-level unless they are sure to get an A or a B, or they insist on parents paying to enter their child privately, so that their grade will not damage the school's statistical record.

Meanwhile, colleagues who are paying to educate their children are equally determined to allow the figures to prove the opposite. And we both determinedly generalise on the basis of our sample of one or two. I insist that my son is stimulated and challenged at school so, by implication, all bright children are likewise. They maintain that their child was bored and underestimated in year seven, and so, by extension, are all able 11-year-olds.

A physicist friend pointed out to me recently that the phrase "a quantum leap" is now consistently wrongly applied in English usage. A quantum is a measure of extreme smallness, a quantum leap an extremely small increment - whereas we all use it to mean a magnificently large step forward, a major advance in whatever is being described. Which goes to show, I suspect, just how curiously we all behave when it comes to numbers. Big? Little? Who cares, as long as it's a quantum leap towards something we really want to believe to be true

Lisa Jardine is Professor of English at Queen Mary and Westfield College, University of London.