Rather than moaning about pay inequity, research organisations should use their data skills to solve the gender salary gap, according to two New Zealand scientists who have created a statistical model to do just that.

In a simulation exercise based on their knowledge of employment arrangements at New Zealand¡¯s crown research institutes (CRIs), the pair showed that the model ¨C based on Bayesian statistics and causal inference ¨C was able to reproduce real-world pay disparities using publicly available information about staff salary bands.

The same approach can be used to test whether pay equity policies have the desired effect, the two researchers argue. ¡°There is little point in identifying a pay gap if no action is taken to address it,¡± they point out in the .

In the exercise, the?researchers generated a league table of salaries in a ¡°hypothetical¡± organisation with 500 staff, producing pay inequities that resembled disparities in existing CRIs.

51¹ú²úÊÓƵ

The pair also modelled the impacts of simplistic interventions ¨C such as giving every female staff member a flat pay rise ¨C and found that the disparities persisted.

Corresponding author Camilla Penney said public sector organisations were obliged to report on gender pay gaps, but their proposed remedies often proved expensive and ineffective. ¡°We¡¯re not good at testing¡­the effects of an intervention,¡± she said. ¡°We¡¯ve got these awesome statisticians. Why aren¡¯t we using them to solve this problem?¡±

51¹ú²úÊÓƵ

Penney, who models earthquakes as seismology lecturer at the 51¹ú²úÊÓƵ of Canterbury, said policymakers tried to tackle pay inequities by tweaking existing salary structures. But an alternative approach would be to design better pay structures from scratch.

¡°One of the powerful things about the method is you could actually use it to say, ¡®what do we want our salary structure to look like?¡¯ You could use a model like this to set up your pay scales as equitable to start with.¡±

She said a looming restructure of the seven CRIs, which will merge into three ¡°public research organisations¡± from July, provided an ideal opportunity to put the theory into practice.

First author Tom Moore, a quantitative ecologist with the Te P¨±naha Matatini Centre of Research Excellence for Complex Systems, said CRIs had both the data and the expertise to ¡°lead the way on pay equity¡±.

51¹ú²úÊÓƵ

¡°We¡¯re hoping this method can form a focal point for collaboration between human resources teams and data scientists, as the new public research organisations form,¡± Moore said.

john.ross@timeshighereducation.com