Fracking and Babies: It’s Complicated

Fracking and Babies: It’s Complicated

by | Jun 10, 2015 | Causality | 4 comments

A new study finds an association between living close to fracking wells and babies who were born small for their gestational age. Small for gestational age (SGA) is a particular diagnosis; if a baby is born below the tenth percentile for weight among babies born with a similar gestational age, he or she is considered to be small for gestational age. Among women and their babies in the study, the 25 percent who lived furthest from the wells had a smaller percentage of small babies, compared to the 25 percent who lived closest to the wells.

The headlines were predictable: “Fracking Associated With Smaller Babies,” said the The New York Times; “’Fracking’ Linked to Low Birth Weight Babies,” said Healthday; “Fracking May Cause Lower Birth Weights in Babies, Study Says,” said Healthline News While the media uses words like “association,” which has a technical meaning in statistics, many people, in this case, will have understood “association”—and, more importantly “linked”— as “fracking causes smaller babies.”

Equally predictable, industry sources dismissed the study. The New York Times quoted an industry representative saying, “The researchers by their own admission rely heavily on two anti-oil and natural gas studies that have been thoroughly debunked. They admit that ‘a number of unknown factors limit the research.’”

The first thing to note is that an observational study like this one cannot prove causality; it can only find that the numbers are skewed from what you’d expect if low birth weight were a random occurrence. And interpreting the study’s findings requires subtlety: babies were not more likely to be born prematurely. In fact, the percentage of babies born with SGA is 10 percent (by definition), yet the percentages in both groups in Pennsylvania was less than this.

Indeed, the authors of the study point to the need to do more research, since environmental influences on fetuses are extremely difficult to pin down. The authors have to contend with a wide range of variables from changing methods of fracking to different demographics in different counties, different income levels among those who live near to, or far from, the wells, and several other known potential confounders. At the same time, the question remains: does fracking put babies at risk for a lower birth weight? Or are there other unmeasured factors that created this difference? Or was this difference a simple random outcome?

The sample used in the study was large enough to obtain “statistically significant” results, meaning that if the people close to wells and those far from wells are equally likely to have SGA babies, it would be unlikely (a less than 5 percent chance) to see the differences observed in the study. The birth weights of the babies born with the fewest wells nearby was higher than that of those born with more wells nearby—by about 20 grams (3344 versus 3323g)—and the percentage with the official diagnosis of SGA is 4.8 percent in the 25 percent of women far from a well, versus 6.5 percent of the 25 percent of the women living close to several wells.

As the authors acknowledge, the clinical significance is unclear. About 90 percent of babies born with SGA reach a normal weight by age two. And both groups have a lower percentage of SGA babies than the population in the United States. The longer run outcomes need to be considered.

As far as what causes this difference between the two groups, new efforts designed to isolate the cause are called for. Observational studies are notorious for spurious and meaningless outcomes without clear reasons. For example, this study also found that the women in the second quartile of well exposure had a statistically significant increase in premature babies compared to those in the first quartile. The effect was not seen in those with even greater exposure to wells, so it’s unlikely to be due to differences in wells; it reflects, the nature of the variance in the population without any impact from fracking.

Certainly, this study has produced a finding with important implications, the biggest one being that further research should be done. Both the news media and the fracking industry would do well, no pun intended, to explore both the need for more information and the strengths and weaknesses of current research trying to protect populations living near gas wells.

Rebecca Goldin is Professor of Mathematical Sciences at George Mason University and Director of She received her undergraduate degree from Harvard University and her Ph.D. from the Massachusetts Institute of Technology. She taught at the University of Maryland as a National Science Foundation postdoctoral fellow before joining George Mason in 2001. Her academic research is in symplectic geometry, group actions and related combinatorics. In 2007, she received the Ruth I. Michler Memorial Prize, presented by the Association for Women in Mathematics. Goldin is supported in part by NSF grant #1201458.
Image adapted from an original by WildEarth Guardians via Creative Commons

Please note that this is a forum for statisticians and mathematicians to critically evaluate the design and statistical methods used in studies. The subjects (products, procedures, treatments, etc.) of the studies being evaluated are neither endorsed nor rejected by Sense About Science USA. We encourage readers to use these articles as a starting point to discuss better study design and statistical analysis. While we strive for factual accuracy in these posts, they should not be considered journalistic works, but rather pieces of academic writing. 


  1. Terence

    I suppose the “implications” of this study could be said about any study no matter how erroneous, “that further research should be done”. Spoken like a true statistician.

  2. Roger S

    You say, “The birth weights of the babies born with the fewest wells nearby was lower than that of those born with more wells nearby—by about 20 grams (3344 versus 3323g)”. I’m sure you meant to say that the weights were HIGHER.

    • statsorg

      Thanks for catching that!

  3. Jack M

    Without having read the study itself, I might posit the following questions:

    What would a study of pre-2005 SGA rates reveal (using the identical geographical parameters)?

    I reside in one of the subject counties. Here the locus of an unconventional well tends to be situated in rural, under-developed areas of the county. Would SGA rates by nature be statistically higher in these areas vis a vis higher income areas of suburbia with no wells immediately present?


Submit a Comment

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Share This