April 8 was Equal Pay Day, the day by which women will have theoretically worked enough to catch up to what men made the year before. In honor of that, the Senate voted on the Paycheck Fairness Act, a bill aimed at giving women a little more power to fight wage discrimination, which Republicans unanimously blocked. While some Republicans claim they care about the wage gap and just object to what they see as burdensome regulation, other conservatives have been calling the idea of the gender wage gap itself into question.

It is a fair question to ask what causes the gap. While it’s true that women make 77 cents for every dollar a man makes when they work full-time, year-round, it’s also true that this figure can obscure various factors that aren’t purely discriminatory. Work experience plays a role. Industry and occupation play a role. Education can play a role.

In trying to figure out how much of the wage gap is discrimination and how much can be explained by other factors, nearly every statistician conducts regression studies that take measurable factors into consideration by holding them constant and seeing what’s left over. From government agencies like the Office of Personnel Management and the Government Accountability Office to women’s advocacy groups like AAUW to economists like Francine D. Blau and Lawrence M. Kahn, a similar group of factors are held constant to find the “unexplained” gap, or the murk where bias would rear its head if it does exist. One of those constant factors is race.

At first blush, this makes sense. All of these researchers are striving to compare the most apple-like of apples to apples—a woman and a man who look as identical as possible and therefore should be paid the same. Therefore, they compare a woman to a man with the same job tenure, seniority, occupation, marital status and race, or in other words, measurable differences. Discrimination will crop up when everything they can measure is stripped out but a gap remains.

But that means that race gets removed from the conversation about discrimination. It’s ends up in the “explained” category. In the study by Blau and Kahn, for example, they list their six controlled factors and note that 2.4 percent of the gap is explained by race. On the other hand, 41.1 percent of the gap remains unexplained, the part that is “potentially due to discrimination,” according to their paper, but not a part that includes racial disparities.

We know that race dramatically shapes wages—that’s part of why it gets lumped into the explained category. Using Census information, the National Women’s Law Center found that African-American men make 73 percent of what white men make, on average, and African-American women make 64 percent. The numbers are even lower for Hispanics: men make 61 percent of white men’s earnings and women make just 54 percent. Men of color even make less than white women.

So, yes, taking this measurable difference into account will surely help explain some of the wage gap. But does that mean we should remove it from the conversation about discrimination? Do we have a good explanation for why people of color of both genders make less than white people? There may be some mitigating factors shaping the racial wage gap as well, but there’s plenty of research indicating that our labor market still discriminates against people of color. Race may be factored into calculating the wage gap, but it’s pushed aside in the discussion about whether women are up against real life wage discrimination. It’s treated as a given.

Even some of the factors that sound objective and explainable could conceal discrimination. In economists Blau and Kahn’s study, the most recent to focus on measuring discrimination, occupation or the jobs women end up in, accounts for more than a quarter of the wage gap. One could see this as a choice, and some women may gravitate more toward teaching elementary school instead of college students. But there are plenty of barriers that keep women from top-earning occupations. And even when women do similar work compared to men, they often make less. Maids are paid less than $10 an hour at the median, but janitors are paid more than $12 on average. Low-skill women’s jobs pay nearly $150 less a week than men’s, on average, while high-skill women’s jobs pay $471 less. It may be hard to determine how much this determines the wage gap, but clearly it’s biased that society values work less when women perform it.

For their part, Blau and Kahn realize that some factors they hold constant may mask discrimination. They write that “if some of the factors controlled for in such regressions—like occupation and tenure with the employer—themselves reflect the impact of discrimination, then discrimination will be underestimated” in their study. Race isn’t terribly murky, though. There’s no objective explanation for why black women make less than white women. And when we divorce that fact from the larger conversation about the wage gap, we fail to challenge the fact that women of color are experiencing multiple forms of bias.