Exploring the pay gap with data

While preparing a workshop on Open Data I'll be running in a few weeks, I discovered an interesting dataset on data.gov.uk showing gender pay gaps at large organisations.

At the end of the financial year, public and private organisations with over 250 employees will have to submit reports on gender pay gaps. The data submitted so far has been made available.. Taking a look through it I discovered a rich dataset, showing mean hourly pay, medians hourly pay, bonuses, etc. It also includes SIC codes. These are classifications used by Companies House to describe what companies do.

Using these classification codes I've averaged values for each sector represented in the dataset. You can use the arrow to cycle through the different figures:

Source: Government Equalities Office

Turns out the Guardian have been on the ball and have produced a few articles on the dataset already, including a handy explainer on the different types of figures.

Exploring the dataset, and following links to individual company reports, has given me a different impression of what pay gaps actually are. Beforehand the term 'pay gap' made me think of situations where men and women earn differently doing the same or similar jobs. In reality, it's a lot more to do with who's at the top. You can have a large equally paid workforce, but the gap comes along when the executives are mostly men, and earn mega-bucks plus bonuses. This is especially true when we use mean rather than median to calculate paygap figures.

You're probably wondering who were talking about here, so I've added in a chart for exploring the top/bottom 20 individual companies for each metric too:

Left to right, women paid more to men paid more. Source: Government Equalities Office