Skip to 0 minutes and 9 secondsHello, again. I'm here with Dr. Tanya Wilson of the Adam Smith Business School of the University of Glasgow. And also, a graduate of the PhD programme at Royal Holloway, so she's OK.

Skip to 0 minutes and 29 secondsTanya, we're going to talk today about the gender pay gap. So how big is the gender pay gap? Well, that depends-- how do we define the gender pay gap? So according to our definition of the gender pay gap, it might be larger or smaller depending on that choice. And that becomes very, very important, because this is something that's in the news a lot, that people want to know. Is the gender pay gap, is it increasing, is it falling, do we need to do more about it? But that depends on how we measure the gender pay gap. So how do we measure the gender pay gap? Well, that's not that simple.

Skip to 1 minute and 10 secondsSo in the United Kingdom, for the past couple of years, there has been statutory requirements for all large companies-- so that companies with more than 250 employees-- they have to report their gender pay gap. So they report the mean hourly pay gap between male employees and female employees, the median hourly pay gap between male employees and female employees. So do those things amount to just different ways of saying the same thing or do they lead to rather different-- Well, when you think about it, when most of us think about the gender pay gap within a company, what we're interested in is the typical pay gap between a male worker, a typical male work, and the typical female worker.

Skip to 1 minute and 59 secondsNow let's define typical. What is the typical? Is it the average that we want to go and look at? So we want to look at an average worker. Well, who is the average worker? And that's when statistically, we have different measures of an average that we need to go and think about. So the average that most of us, I think, we think about is the one that we learn at school, the mean. Where you take the hourly pay received by each employee. For all of the men, you add that up and divide that by the number of men, for all the women, you add that up and divide that by the number of women.

Skip to 2 minutes and 37 secondsSo you're going get an average number across all men and across all women. And that's one way of thinking about the average pay gap. But one of the issues in that is if I was to take the average pay between you and I, I would come up with a figure that lies somewhere between your pay and my pay, right? But say someone else walked into the room, and it was a Jeff Bezos type, someone with a very, very high income. Then the average pay of three of us now is far away from your and my actual pay.

Skip to 3 minutes and 15 secondsIt goes and gets distorted far above what we were actually going get paid, because we have this one person who gets paid a really large amount. Of course, or the other way, if you have someone who comes in who earns a lot less than we do, then obviously, that goes and gets distorted downwards. OK, so Jeff Bezos aside, I guess an important point here is that you're going to have a CEO, for example, and if it's a reasonably successful or big company, the CEO is probably making a lot of money, that's just one person.

Skip to 3 minutes and 55 secondsBut the income of the CEO, the hourly wage if we put it that way, of the CEO is going to be way higher than-- Most of the employees in the company, exactly. So in most companies, as we know, the most employees will be working at more junior positions. And then there will be a smaller tier of middle management positions, potentially. Then executive level. And then you have potentially, the CEO of a very small group of people at the top of the company who generally have far higher wages than the people who are working at the lowest tiers within the company.

Skip to 4 minutes and 33 secondsSo if there's a huge disparity between those wages, a few people at the top being paid mega salaries can make the average pay within a company look far higher than most of the employees actually face. So another way of looking at average pay within the company is to go and look at what we call the median. And to think about the median, you can imagine that we take all of the employees in the company, we put them in the line. And we rank them according to their salary. Then you go down the line and you take out the person who's right in the middle, and you ask them what is their hourly pay.

Skip to 5 minutes and 17 secondsAnd that is then what we would call them the median hourly pay. So it's the person in the middle. And you can think in that situation, if Jeff Bezos walks in the room, well, actually, that median is not going to go and change. That person in the middle is not going to go and change. He's just one person. He's one person, and he's not going to go and change that gap. So in terms of the median hourly pay gap, again, think of all of the male employees or female employees put into two lines, ranked according to income. And you take the middle man, the middle women, and then you compare that pay gap.

Skip to 5 minutes and 55 secondsSo does the median pay gap tend to be smaller or larger than the mean pay gap? That depends. So in many companies, it is the case in many companies that the top executive structure is dominated by men. And in that case, you generally have a mean gender pay gap that is far larger than the median pay gap. However, there are some companies where the top CEO is female and more of the senior positions are filled by women in comparison to the lower positions. And in that case, you may actually go and have a gender pay gap that is in favour of women measured by the mean. Although by the median, that may not be the case.

Skip to 6 minutes and 47 secondsSo it depends on the structure of the company.

The Gender Pay Gap - Part 1

You can set up a bunch of on-camera interviews but you can’t predict exactly what will come out the other end.

One pleasant surprise for me was the discovery of just how much my colleagues care about good measurement. Measurement issues came up as a central issue, in all four interviews including this one with Dr Tanya Wilson.


Tanya Wilson is a Lecturer in Economics at the Adam Smith Business School, University of Glasgow. She is the recipient of the 2016 Royal Economic Society Prize, and was awarded the Sir Alec Cairncross Prize from the Scottish Economic Society in 2017. Her research interests are in the economics of labour markets and the economics of the family. Tanya has contributed reports to recent Scottish Parliamentary inquiries on the Gender Pay Gap in Scotland, and Scotland’s Economic Performance as well as the Royal Society of Edinburgh inquiry on women in STEM – Tapping all our Talents 2018.


Many researchers blithely assume that they accurately measure exactly what they are trying to measure. But in reality they often measure something else and don’t know it.

Tanya makes an important point. There are two commonly used ways to measure averages that can lead to very different answers. I’ll expand on this point and you can see still more detail by consulting this interesting website.

Here are 9 numbers:

2, 7, 10, 36, 4, 7, 8, 6, 1

The mean of these numbers is 9 which is calculated as:

(2+7+10+36+4+7+8+6+1) / 9

The median is 7 which we get by lining the numbers up in order and picking out the one in the middle:

1, 2, 4, 6, 7, 7, 8, 10, 36

The number 7 appears in position 5 out of 9.

(A very poor version of) Jeff Bezos would correspond to the 36 in this list of numbers. If we turn that 36 into a 360 the median will not change but the mean will go way up.

You might enjoy this take on the Distracted Boyfriend meme

Survival Statistics Extra

Here is part 2 of this interview just for those of you with a particular interest in gender pay equality. It’s very interesting but it doesn’t fit all that well with the flow of the course so I’ve turned it into a bonus item.

This is an additional video, hosted on YouTube.

Discussion

Why is mean income at the national level pretty much useless for understanding poverty dynamics? Is median income much better?

Share this video:

This video is from the free online course:

Survival Statistics: Secrets for Demystifying Numbers

Royal Holloway, University of London