July 27, 2012 · 0 Comments
Source: Beat the Press
By Dean Baker:
Elaine K. Hill, a doctoral candidate in Cornell University’s department of applied economics and management, found evidence that fracking is associated with the frequency of low birth weight babies. The findings of her study implied that for mothers living close to a fracking site, the probability of a low birth weight baby increased by 25 percent.
While this might be important information for government officials and the general public to have when considering restrictions on fracking, New York Times reporter Andrew Revkin is outraged that an unpublished study is being widely circulated and could impact public policy. From his blogpost, it sounds like Revkin gave Hill a really serious grilling about the ethics of allowing her unpublished study to influence debate on a major national issue. (Don’t you wish reporters would just once give the same sort of grilling to Jamie Dimon or some other corporate honcho?)
There are two problems with Revkin’s outrage. First, while he wants to be a real tough guy and show that this study should not be taken seriously, absolutely nothing in his piece calls into question the main findings of the research. Revkin presents at some length the views of David Ropeik, who Wikepedia identifies as “an independent consultant to government, business, trade associations, consumer groups, and educational institutions.”
While Ropeik appears angered about Hill’s speculation on how fracking might affect the number of low birth weight children, he gives no reason whatsoever to question the main finding. Specifically, Ropeik does not in any way question the statistical relationship that Hill found between fracking and low birth weight children. If the study was as bad as he seems to think it is, it should have been easy to find at least some potential flaws in its statistical analysis. Apparently he didn’t.
Another source raises some very tentative questions about the statistical method, but does not really draw them out. As a practical matter, Revkin has given readers no substantive reason for questioning the basic finding, that being near a fracking site substantially increases the risk for pregnant women of having a low birth weight baby.
Moving beyond the substance, Revkin seems to find it outrageous that an unpublished study would have an impact on public debate. Surely no serious news outlet would ever take the findings of an unpublished study seriously.
Is that true? Five years ago, the Washington Post gave a column to an economist affiliated with the conservative Manhattan Institute to tout the benefits of new drugs. The economist, Frank Lichtenberg, argued that new drugs increased life expectancy and lowered overall health care costs based on the findings of a then unpublished study. The quite explicit moral of the story was that the government should be willing to pay for expensive new drugs through Medicare, Medicaid and other public health care programs. (Here’s one of my pieces making fun of Lichtenberg’s study.)
In economics at least, unpublished studies find their way into print all the time. That would concern me, except I have seen all sorts of dreck also find its way into highly respected journals. It would be nice if we could view the academic review process as providing efficient and effective quality control, but we don’t live in that world.
Hill has uncovered an important finding. If there is some fundamental error in her methodology then the more senior people in the field who are condemning her, should be able to quickly identify it. Revkin found people with plenty of bad things to say about Hill, but he was apparently unable to find anyone with fundamental questions about her methodology or who could suggest an alternative explanation for her findings.
Given the importance of these findings, it would have been irresponsible for Hill not to make them public. It’s unfortunate she has to deal with people who are more concerned about credentials than science.