The Average Story

The JEP has again printed wonderful headlines about the "average wage" in Jersey, making out that workers are doing wonderfully. I wrote on this subject last year, and am pretty fed up with the ignorance of the copy-writers, and I assume the way in which the States Statistics Unit does not stress most strongly the limitations of the average (or the JEP chooses to ignore it).

Two obvious points:

a) Averages are useless without a measure of the spread - are most wages clustered around the average, or are they well spread apart from it. One measure is the standard deviation. You may hunt in vain for that in the JEP article. If the brief of a paper is to inform and educate the public, this should be given, and explained in simple terms.

b) Most wages are a skewed distribution, rather than what is called a normal distribution, i.e. they are not a "bell shape", but are like a bell squashed out of shape on one side. The result: most wages are less than the "average" (arithmetic mean), and the fewer high wages push the average higher than one might expect. This is why most countries have moved from using the average to using the median (or middle point) as this is much better for showing where people are (half above, half below) as best practice. The States Statistic unit have said they can't get that kind of detail from employers (as for instance is managed in the UK), and that means that any comparisons with the UK "average" (i.e. the median) are grossly misleading.