Measuring the Gender Pay Gap
Yes, we’ve all heard the line, there’s lies, damned lies and then statistics. The power of it as an observation about the world is that by picking and choosing the numbers you quote you can prove just about anything that you want: which is why when people present us with carefully chosen numbers we need to be so wary.
For example, in the US the gender pay gap is normally stated as women earning 79 cents (or whatever the number is) for every dollar earned by men. Now while it’s true, it’s not actually very informative. For, for example, do men and women work the same number of hours? No, they don’t, not on average, they tend to have shorter work weeks and in common with their sisters around the world they also tend to take more time off for illness.
No, of course that is not all of that 21 cent gap, but it is some of it, which is why that particular statistic isn’t really all that useful to us in deciding firstly, whether this is a problem we want to do something about and secondly, what we might do.
One such number that caught my eye this week was from an MEP (a Member of the Euuropen Parliament and thus one of those with power in Europe):
And how has it come to be that the UK has the largest gender pay gap in the European Union?
Now I expect politicians to be ill informed but that’s ludicrous. As the figures from her own organisation show, the UK’s gender pay gap is actually below average for the European countries, a very far cry from being the worst.
So what is happening? She’s quoting from these other figures, which do indeed seem to show the UK has the largest gender pay gap in Europe. How can we have both a below average and the largest gap in the same country at the same time?
The answer is in which actual figures are being looked at. The first figures are made up of only people who are working full time (the numbers accord very well with those you can work out by looking at average hourly pay).
The second set of figures come from looking at all of those in work. Which, as you might imagine, includes those who work part time.
And this is where the problem comes in. Those who work part time get paid less per hour than those who work full time. This is true in every country, it’s true of both men and women. Work part time and you’ll get less per hour than your full time contemporaries.
Add this together with the fact that we have a very different structure of employment in the UK than they do in other European countries and that’s where the difference comes from. For the UK has many more women working part time than the others. This is normally though of as something beneficial: those women who want to can find part time work which allows them to have and raise their children, rather than being forced into either full time work or none. But when you look at the average pay rates of men and women you’ll then be including both the effect of the gender pay gap and the part time pay gap.
Which is what allows a politician to say that a country which has a below average pay gap actually has the largest one.
See, you can prove anything with statistics.

May 9th, 2008 at 8:17 pm
As a part-time teacher, I first objected to your statement about part-time workers earning less per hour. In the school district where I work, salaries are the same per-hour, depending on experience and education. So as a half-time teacher, I make half of what I would make as a full-time teacher. But then, I found this information on the U.S. Bureau of Labor Statistics: http://www.bls.gov/opub/cwc/cm20030814ar01p1.htm. In 2001, when the study was conducted, all full-time employees averaged $17.13 per hour, while all part-time employees averaged just $9.17 per hour.
But then again, I would love to see the divisions within this study. Remember – many of these part-time jobs are held by high school students and college students, and even if they are held by non-students, many are within minimum-wage-pay fields: convenient stores, restaurants, etc. I wonder if we looked at divisions between full-time and part-time, within individual fields, if the difference would not be so broad.
So I agree, Tim: it’s easy to do whatever you want with statistics, to fulfill any purpose.