So we could calculate the height of the average human being. Say it is 160 cm. That is a simple fact – we have measured everybody.
As a scientist or statistician, would you want to put confidence intervals around that average: ‘we are 95% certain that the true average human is between A cm & B cm in height’? What would be the point, & how would we do it?
Suppose we repeat the exercise in ten years time. The height of the average human being is now 161 cm. Can we say that humans are getting taller?
Well again, we have a simple fact.
Or do we? Just how accurate are the estimates – how great is the measurement error on the estimate of height? Has something happened to the technology of tape measures which makes the number bigger so that people only seem to be taller?
We can perhaps set up experiments to establish the degree of measurement error in the estimate of height of an individual human today, & conclude that a difference of 1 cm in the measured average is not ‘statistically significant’. But how sure can we be that the errors were the same with the technology we used ten years ago?
And does any of this matter anyway? Would any decisions, any behaviours, any global politics, change if we are sure that people are now, on average, 1 cm taller? Or any scientific laws about human growth & development? Would we believe that in one hundred years the average human will be 10 cm taller & start to plan accordingly? And if a small number of obsessive people keep working away at this, how likely do we think it is that one day they will uncover some very important & widely applicable Law of Nature?
Let us go back to our first census. If we are going to go to all the trouble of measuring everybody it surely makes sense to record other information at the same time, so that we can make accurate estimates about the size of differences between various groups.
Well human beings come in two basic models, so it makes sense to record, alongside their height, whether they are male or female in form.
Of course we all know that men are taller than women, we do not need a census to tell us that, but it would be useful, for all sorts of reasons, to be able to put an exact number on the difference, which we expect to be round about 5 cm.
But – when the figures are compiled we find that the world’s average woman is 1 cm taller than the average man.
Somebody, somewhere in the process must have made a terrible mistake. Politicians, the press, ordinary people everywhere are delighting in the discomfort of the so-called experts.
Except it might actually be true, & if we could decompose, or analyse the figures in more detail, we could explain, if only people would listen, how it is that, even though it is true that in every age group, in every country, men are taller than women; the observer in outer space would be able to see that, taken altogether, earthling women are taller than the men.
We know that average heights vary according to which country & era you were born & raised in, for reasons which may be partly genetic & partly environmental. But the age structure, the proportions of men to women or even sheer numbers of people are not evenly distributed across the globe. We cannot decide, by reason alone, whether the relative contributions of tall women in countries conducive to growth, or tall young people in countries where recent birth rates have been higher, or superior survival of women in rich environments, have in fact led to this seemingly perverse result.
But is it ‘statistically significant’?
Why don’t people generally talk about height as being a political identity? … at any given time in any given place, we would not know where to find people of our own height & would not know how to find us … it is not entirely clear what we would talk about … our [shared] experience simply is not substantial enough to warrant a special meeting … for now they do not represent any kind of group from which one could claim a viable social identity
