True Meaning of Figures

This week has seen widespread outrage over the new Jersey pay statistics, particularly the £620 average weekly wage. This comes of using the wrong type of average for the sake of spin.

There are different ways to calculate averages to give the most meaningful value in different contexts, but they can also be misused to give a misleading value in a context where the truth could be inconvenient.

The commonest type of average, the first one most people learn about at primary school, is a ‘mean’.

Where figures cluster around a single central value, a good approximation of that value can be made by adding the values of every datum up, and then dividing the total by their count. The trouble with processing wage statistics in this way is that earnings do not follow the ‘bell-curve’ distribution around a medium value that means are designed for.

Instead, they have a ‘power law’ distribution in which relatively low figures are commonplace and ever higher figures become ever rarer. The meaningful average for a power law distribution is the ‘median’ in which just as many data have a higher value as a lower one.

Now if you were a cynically dishonest government wishing to tell the world how prosperous your policies were making your people, you could instead calculate the mean wage and pass that off as the average.

But it would not be; the tiny number of very large figures would distort and inflate the mean to well above any sensible concept of the average.

Wouldn’t it make the government look good?

However, there is a serious downside to the puffing up of the statistics: By making all the people who are actually doing all right think that they are a lot further behind than they really are, the misapplied average spreads discontentment and unhappiness.

Worse still, from an economic management viewpoint, it creates an aspiration among genuine medium earners to seek hefty pay increases to restore their apparent position, an inflationary pressure that we could well have done without, in these troubled times.

The Statistics Unit appear to see themselves as spin-doctors to the Council of Ministers, rather than information providers to the Island as a whole. They are letting us all down by this approach. I for one would like to see a change of heart, and the provision of useful and helpful information to become their new objective.