The energy industry has long insisted that hydraulic fracking—the practice of fracturing rock to extract gas and oil deep beneath the earth’s surface—is safe for people who live nearby. New research suggests this is not true for some of the most vulnerable humans: newborn infants.

In a study presented today at the annual meeting of the American Economic Association in Philadelphia, the researchers—Janet Currie of Princeton University, Katherine Meckel of Columbia University, and John Deutch and Michael Greenstone of the Massachusetts Institute of Technology—looked at Pennsylvania birth records from 2004 to 2011 to assess the health of infants born within a 2.5-kilometer radius of natural-gas fracking sites. They found that proximity to fracking increased the likelihood of low birth weight from about 5.6 percent to more than 9 percent. The chances of a low Apgar score, a summary measure of the health of newborn children, roughly doubled, to more than 5 percent.

The study, which has yet to be peer-reviewed or posted online, comes at a time when state and federal officials are grappling with how to regulate fracking and, in the case of New York State, whether to allow the practice at all. Much of the available research has been sponsored either by the energy industry or by its critics. Independent studies have found evidence of well-water contamination in areas close to fracking activity. Establishing a direct link between fracking and human health, though, has been complicated by a lack of information on the chemical substances used in the process and the difficulty of obtaining health records that include residence data.

Currie and her colleagues, who had financial support from the Environmental Protection Agency and the John D. and Catherine T. MacArthur Foundation, obtained a confidential version of Pennsylvania birth records containing the latitude and longitude of the mothers’ residences, matching them to the locations of fracking sites. In doing so, they built on the work of Elaine Hill, a PhD student at Cornell University who sparked controversy in 2012 with a study showing that infants born near fracked gas wells had more health problems than infants born near sites that had merely been permitted for fracking. One criticism of Hill’s study was that fracking activity might change the demography of an area, attracting more mothers who are likely to give birth to infants with health problems.

The new research addresses such concerns by following a constant group of mothers who had children both before and after the onset of fracking, and by controlling for geographical differences in mothers’ initial health characteristics. It seeks to achieve the rigor of a controlled experiment by focusing on mothers who, due to their locations and the dates of their pregnancies, were effectively selected at random to be exposed to fracking.

While the study strongly indicates that fracking is bad for infant health, more work is needed to understand why. Surprisingly, water contamination does not appear to be the culprit: The researchers found similar results for mothers who had access to regularly monitored public water systems and mothers who relied on the kind of private wells that fracking is most likely to affect. Another possibility is that infants are being harmed by air pollution associated with fracking activity.

The study doesn’t necessarily tell us whether or not fracking is worth doing. There may be offsetting health benefits related to the added jobs fracking creates, to lower energy prices or to the reduced use of coal or other fuels as more natural gas becomes available. “Given how important fracking is for the economy generally, it might make sense to compensate people for the cost of moving away from a site rather than shutting it down,” said Currie.

Still, evidence that our demand for cheap energy could be doing irreversible harm to children should be reason for serious pause.

(Mark Whitehouse is a member of the Bloomberg View editorial board.)