Surf to the site of the Walk Free Foundation, created by Australian billionaire Andrew Forrest, and you’ll learn that slavery is one of the great evils of our time. In fact, the organization produces the Global Slavery Index, which says that, as of 2014, there were 35.8 million slaves in the world.

But the State Department has its own report on slavery, the Trafficking in Persons Report. In 2013 this annual volume said that “social scientists estimate that as many as 27 million men, women, and children are trafficking victims at any given time.” But in 2014 the State Department said the report “will open the door to reach change—not just on behalf of the more than 44,000 survivors who have been identified in the past year, but also for the more than 20 million victims of trafficking who have not.”

Seven million “trafficking victims” disappeared over the past year? What happened? Washington Post Fact Checker columnist Glenn Kessler investigates in this column.

He breaks down the problem into two components. First, what exactly do we mean by “trafficking?” Second, how does the Walk Free Foundation determine that there are 33.8 million slaves in the world?

American University law professor Janie A. Chuang addresses the problem in a 2014 article in the American Journal of International Law. In 2000, the United Nations decided to expand the definition of “trafficking” from that of women and girls being kidnapped for sexual exploitation into something that was vaguer and included men. As a result, Chuang writes, “ever since, diverse advocates have appropriated the ‘trafficking’ label so that the activities covered by the term trafficking remain very much in the eye of the beholder.”

One of the areas of debate is whether or not “forced labor” is the same as slavery, even if the forced labor does not involve a victim moving from one place to another. The International Labour Organization calculates that in 2012 there were 20.9 million victims of trafficking in the world—and 56 percent, or 11.9 million, of them were subject to forced labor.

All of this leads to a question—do we actually know how many “slaves” there are in the world, and if we don’t know, how can we figure out if philanthropy is effective?

A 2014 piece in The Economist says that when Andrew Forrest decided that fighting slavery was his philanthropic cause, he went for advice to Bill Gates, who told him to find a way to quantify the problem, because “if you can’t measure it, it doesn’t exist.”

Thus we have the Walk Free Foundation’s Global Slavery Index, which confidently says that there were 35.8 million “modern slaves” in the world in 2014. The foundation gathered some original information by adding questions to the Gallup World Poll, a giant annual survey Gallup does for a variety of international clients, including the UN and the World Bank. (The survey can be found here) The foundation hired Gallup to ask 1000 people in six countries (and 2000 in Russia) such questions as if anyone in the person polled’s family had done forced labor or had a job where his or her passport was taken way.

Using the polling data, the Walk Free analysts calculate that Indonesia has somewhere between 214,296 and 2,962,791 modern slaves, a ridiculously wide confidence interval.

But even with their polls, Walk Free covered only seven countries. They came up with estimates for the remaining countries by taking data from some countries and then applying them to others. “Thus data for the United States is considered relevant to calculate Italy’s total of 11,400 slaves,” Kessler writes. “South Africa’s number of slaves—supposedly 100,000—was derived from the fact that GSI (Global Slavery Index) researchers decided the country is 70 percent “Western Europe” and 30 percent “African” (specifically, an amalgam of Ethiopia, Nigeria, Niger, and Namibia).”

George Washington University sociologist Roland Weitzer, writing in the Annals of the American Academy of Political and Social Sciences, only thinks the Global Slavery Index is questionable. He argues that any international estimates of the size and scope of human trafficking cannot be trusted. “The claim that human trafficking involves a massive number of people is unsubstantiated,” he writes, “it simply cannot be established at the macrointernational level.” He calls for more “microlevel”—or what we would call local—research.

Anti-trafficking analyst Helen Gallagher agrees that the Walk Free Foundation report is questionable. She notes in this piece for The Guardian that “After noting that little reliable information exists on slavery in China, the index’s authors declare that they are comfortable with China being considered pretty much the same as other east Asian nations like South Korea, Taiwan and Japan.”

What implications do the problems with the Global Slavery Index have for donors? Donors interested in problems of poverty, immigration, and prostitution should deal with local problems in their neighborhoods rather than trusting dubiously precise international reports. Large foundations are well-equipped to deal with international medical problems (such as fighting polio and malaria). They are far less trustworthy when calculating issues of human behavior, such as human trafficking, that almost certainly cannot be accurately measured.