A data scientist’s guide to ending the wage gap for good

More than 20 years that have passed since the National Committee on Pay Equity first called for action on the gender wage gap. But not much has changed. Women continue to earn less than men, and research shows that women often have less successful salary negotiations, sacrificing tens of thousands of dollars in future earnings. As a woman who works in the tech industry, I often find myself asking: What will it take to truly drive change and close the gender wage gap?

For me, the answer is data.

After I graduated with a PhD in Astrophysics from UC Berkeley, I was interviewing for a job as a data scientist in San Francisco. My prospective new boss said, “I know you make about $14,000 a year as a graduate student at Berkeley, I’m going to offer you more than that.” And he did! Imagine my excitement when my starting salary was much more than my graduate stipend.

At the time, I had no idea what I should be making, nor did I know how to negotiate, as my last “job” had been in a completely different industry. What’s more, I had no other comparable offers to use as a baseline. I attempted to negotiate a higher salary because I was told that you always should, but I was ultimately unsuccessful.

A year later, I was on a business trip with two men from my team when they started discussing salaries. We had all been hired around the same time, by the same hiring manager. We had the same job title and similar experience. Yet both of these men had a base salary that was $10,000 higher than mine. It turned out we had all been given the same initial offer. But they were able to negotiate a higher salary with our boss, while I was not. What will it take to truly drive change and close the gender wage gap? For me, the answer is data. 

This experience was incredibly frustrating. And sadly, it’s not unique. Numerous studies have shown that managers respond less favorably to women trying to negotiate than men, and it is a well-studied fact that overall, women are paid less than men for the same jobs.

Fast-forward to applying for my current job at Hired, a San Francisco-based start-up that connects job seekers with companies looking to hire. Unsurprisingly, I set my salary expectations based on what I was previously making. But Hired, which has a team dedicated to what one might call “career coaching,” alerted me to the fact that I was setting my salary expectations far too low.

It was at this point that I truly understood the patterns associated with the gender wage gap and identified my worth. My Hired advisor encouraged me to set my salary based on the market rate for my role, as opposed to my previous salaries. This ended up being a 25% increase over what I was initially asking for. Even better, my new boss agreed.

Discovering that I was substantially underpaid made me curious—exactly how widespread was this phenomenon? A unique feature of Hired’s platform is that every candidate must set a preferred salary. When taken in the aggregate, that allowed me to identify some interesting findings that haven’t been widely measured before. For example, I looked into how men and women value themselves across different job titles and found that as the ratio of men to women in a role increases, so does the salary expectation gap. This means that male-dominated software engineering roles have double the expectation gap of design roles, where more women work. While findings like these can be discouraging, we also found signs of change, such as women with less than two years of experience are asking for 2 percent more than their male counterparts – and getting it. Discovering that I was substantially underpaid made me curious—exactly how widespread was this phenomenon?  

Based on this analysis as well as my personal experience, I believe now more than ever that data must play an integral part in any solution to the wage gap.

Data has become prevalent in almost every industry and business function, from security to finance to customer engagement, commerce and logistics to mobile to communications. What’s interesting is that despite this wide embrace of all things data, decisions around compensation are still largely based on subjective processes that can be negatively influenced by bias, unconscious or otherwise. That is exactly why data needs to be integrated into every company’s compensation model.

A data-based approach to compensation, meaning a pay model that is based on key findings around market worth and skill set, clarifies the market and internal value for each job. It quantifies compensation costs, validates compensation strategy, and ensures pay equity. Companies can have a data-based approach to compensation by implementing a pay structure that includes a compensation philosophy and strategy, pay ranges, and adjusted pay grades that are in internal alignment. For individual women, a data-based approach to compensation is the key to education and information that women need to negotiate effectively. For companies, it provides a benchmark and level of transparency that otherwise isn’t available. A data-based approach to compensation clarifies the market and internal value for each job.  

Interestingly, research suggests that when you add transparency around salary data, it boosts productivity and the number of female job applicants. This in turn can increase your workforce retention and profitability. Take, for example, the restaurant business, where the traditional tipping model largely favors front-of-the-house staff and creates huge wage disparities across different parts of the restaurant.

A recent Freakonomics episode, featuring restaurant owner Danny Meyer of The Modern, brought this issue to light. Meyer decided to flip the traditional restaurant model on its head, removing tipping and instead implementing “hospitality included” pricing. This allowed Meyer to redistribute wealth across the restaurant’s entire staff and create a more collaborative and positive work environment. It also increased the restaurant’s productivity and profitability overall.

When I look back at my career thus far, I often wonder what my experience would have been like had I had valuable tools like PayScale and knowledge around my market worth from the get-go. Sometimes it baffles me to think that it took working at five different companies before I truly understood–and was paid–my market worth. I know that many women out there aren’t as lucky and continue struggle with salary negotiations. My hope is that in sharing my story, and continuing to release new studies that emphasize the power of data, I can help both women and companies rethink how they approach salary negotiations and the gender wage gap at large.

Follow Jessica on Twitter at @berkeleyjess. We welcome your comments at

home our picks popular latest obsessions search