Tom Simpkins

Statistics: the answer to gender inequality in businesses

Tom Simpkins, Consultant at data analytics firm Concentra, outlines how to avoid the pitfalls of misinterpreting data and why careful analysis can transform a business into an industry leader

This month, the government announced a new initiative for organisations with more than 250 employees to disclose the average pay of both male and female employees. In accordance with this new regulation, employers will be expected to publish differences between male and female starting salaries, average basic pay and total average earnings of men and women with a breakdown of grades and job types.

The aim is to close the gender pay gap by bringing transparency to those companies paying female employees significantly less than their male counterparts doing the same job. However, while this is an extremely welcome measure, companies and commentators must be very careful about the conclusions they draw from broad statistical measures.

The misinterpretation and presentation of statistics is a real issue for businesses in today’s data-driven culture, and organisations must be sure that they are interpreting the correct information to get the best business results.

Drawing false positives

One particularly common trap that businesses can fall into when analysing such data is known as the Simpson’s paradox – in essence this means that data may not give the full picture or even display an accurate truth when other factors are considered.

This can be readily explained using the issue of gender equality in the workplace.

For example, a business with 150 male employees and 150 female employees may show that the average pay for each gender is £50,000 across the business. This gives the false impression that a business is a perfect example of pay equality, but when different skill levels and pay grades are investigated the picture can become much less rosy.

Understanding the principle

Perhaps the most famous example of Simpson’s paradox is the Berkeley gender bias case, when the University of California, Berkley was sued in 1973 for gender bias against women who had applied for admission to the school’s graduate course.

Initially, the admission figures for the 1973 academic year demonstrated that men applying for courses were more likely to be accepted than their female counterparts and that the difference was so large that it was unlikely due to chance.

Yet when the data on the departmental breakdown was assessed in greater detail, it showed that the majority of departments had a “small but statistically significant bias in favour of women.” The reason that the high-level figure suggested gender bias was due to the fact that, in this instance, women were applying to more competitive departments than men.

Beware the hidden variables

The proliferation of data means the demand for reporting and transparency will only increase of the coming years for almost all businesses.

Initiatives such as reducing the gender pay gap need to be underpinned by thorough and accurate exploration of the available data, which then needs to be clearly presented so that it can be fully understood.

Failure to do so could have serious implications, missing governmental or industry targets and reducing the overall effectiveness of your workforce. Data holds the key to truly unlocking the potential of a business, identifying those areas where improvements can be made and those were the company is excelling.

The key to understanding gender inequality in the workplace and laying the foundations for making meaningful change lies in unlocking the secret of business data. Getting this right could mean the difference between being an industry leader or damaging your brand’s value and reputation.

Join the discussion as a guest or using , or Google

Top Ten Most Read