Annie Rorem (Batten ’13) conducts research on education and labor force participation with a particular interest in gender. Dan Addison

We are in the midst of a data movement. The prevailing sentiment among governments, companies and individuals appears to be that we could move the world if only we had enough data.

But the more data we acquire, the more careful we must be (and not only because we would rather our credit card numbers stay private). With more data comes a need for responsible analysis. We must ask critical questions, endure uncertainty, embrace nuance and sidestep the pitfalls of confirmation bias. It is not more data but rather good data, wielded with the lever of strong analysis, that will truly move the world. In my research, I apply this mindset to consideration of the gender wage gap.

Among full-time workers in the United States, men’s median wages are higher than women’s. Attention to this issue ranges from discussions about U.S. men’s and women’s national soccer teams’ earnings to initiatives from the White House. This attention reflects not only that we do not want gender alone to determine a worker’s wage prospects, and also that we do not want to suggest, through wages, that one gender is more valuable to the labor force than the other.

Although it is startling to see the extent of the difference in earnings—according to the Bureau of Labor Statistics, female full-time workers in 2013 had median weekly wages equal to 82 percent of those of their male counterparts’—this statement’s power is limited; it describes, rather than explains, the reality that men out-earn women at the median. This difference is certainly meaningful. However, without context, it invites speculation and incomplete causal claims, such as “Women are paid less than men because employers believe that women are less valuable employees.”

The reasons for the disparity are complex. For example, women are more likely to take time away from the labor force to raise children, which can stagnate earnings. Moreover, men and women often cluster in industries with different wage scales. Such realities merit critical analysis. We must hone our understanding of the difference between men’s and women’s wages by asking a critical question: “What role does gender play in determining the wages of workers who are otherwise similar?”

Many factors influence earnings. The amount a worker is paid reflects not only his or her value to an employer—based on his or her ability to fulfill the tasks of a position—but also the value placed on that industry relative to others. To isolate the effect of gender in determining wage, we must examine the effects on earnings of experience, education, rank and industry.

Of course, gender may correlate with the factors we seek to hold constant, including employment continuity, or industry itself. Individuals pursue professions for a variety of reasons, including interest, work environment and benefits. We cannot assume that men and women—or, for that matter, any two individuals—consider the same set of priorities when deciding on a job. We must also try to understand motivation and decision-making.

Furthermore, we can compare only things that can be consistently measured, so some aspects of compensation may go unevaluated. Productivity, for instance, is difficult to compare within or across industries. Our analysis of gender’s effect on wages across the entire labor force may always be subject to some uncertainty.

It is human nature not only to seek evidence to support what we already think to be true but also to find it. I am predisposed to shore up what I believe: that women have noticeably different workplace experiences from men’s and that salary is a component of this. Avoiding the pitfalls of confirmation bias poses a challenge equal to the mechanics of the analysis itself. Scrupulous analysis of a specific question is almost nearly guaranteed to reduce the shock of what is discovered. That analysis will help build a lever strong enough for the weight it must support: the weight of the world it is moving.

Annie Rorem is a policy associate for the Demographics Research Group at UVA’s Weldon Cooper Center for Public Service.