We seem to be having a run of splashy, peer-reviewed GMO (genetically modified organism) studies that are of questionable merit. Several weeks ago, a team of French researchers published results that linked cancerous tumors in rats to the GM corn they were fed. But many scientists cast doubt on the study's legitimacy almost immediately, and other odd circumstances surrounding it raised some big red flags for science journalists. I covered the particulars of all this in my recent Slate piece. This week another study on genetically modified crops is generating big news, thanks to a widely propagated Reuters article headlined: "Pesticide use ramping up as GMO crop technology backfires." This is big news because biotech crops are believed to reduce the amount of pesticide used in agriculture. But a just-released study claims the opposite is happening due to the rise of certain herbicide resistant weeds. As I noted on Twitter, the reporting in the Reuters story is rather thin. The only scientist quoted is the lone author of the study, Charles Benbrook of Washington State University. Nor did the Reuters reporter appear to vet the study with any outside experts. So the resulting article reads more like a press release. And in fact, one of Benbrook's quotes is straight from the actual university press release:
Resistant weeds have become a major problem for many farmers reliant on GE crops, and are now driving up the volume of herbicide needed each year by about 25 percent.
Just so we're clear: A story making this kind of claim should not hinge entirely on the word of the scientist whose study is being reported on. It's not good journalism. (Alas, this didn't prevent people that should know better from tweeting and retweeting the story.) I have to think the Reuters reporter knows this and simply rushed the story out to beat the competition. Happens all the time. In this case, Reuters got a huge jump on everyone, as its story was picked up worldwide and spread throughout the U.S. via high traffic outlets like MSNBC and the Huffington Post. I'm sure Benbrook is ecstatic over the huge amount of uncritical publicity his study has received this week. Especially since his purported findings are not new. (I'll get to that in a minute.) Benbrook is also the chief science consultant for the Colorado-based Organic Center, a research-focused organization that aims to raise awareness of "the health and environmental benefits of organic food." In a 2011 interview he was asked: "The biotech industry claims that genetically engineered (GE) foods decrease pesticide use. Is that true?" He responded:
The Organic Center has done four reports on this question and has found that crops like corn, cotton and soybeans genetically engineered to be resistant to herbicides have actually increased herbicide use by hundreds of millions of pounds over what herbicide use would have been had these crops not been commercialized. So when the biotech industry says that today's GE crops have reduced and are reducing pesticide use, they're factually wrong.
In 2009, Anastasia Bodnar, a maize geneticist now at the National Institutes of Health (NIH), critiqued one of these reports at Biofortified, "a group website devoted to providing factual information and fostering discussion about agriculture, especially plant genetics and genetic engineering." She wrote:
I can't help but get the feeling that Dr. Benbrook started with a conclusion and found data to fit rather than starting with a general review then finding significant conclusions.
When I asked around in the plant science/biotech community about the new study by Benbrook, I was referred to Bodnar's detailed 2009 critique. That's because his study, as Andrew Kniss, a University of Wyoming agronomist puts it,
is an updated version of a report that The Organic Center published in 2009.
In her analysis of that iteration, Bodnar writes:
There are a lot of problems with this report, but I'm particualrly concerned with the way Dr. Benbrook fails, for the most part, to distinguish between different biotech traits, fails to distinguish and between different pesticides, and fails to consider non-biotech traits that could increase pesticide use. First, all GMOs are not created equal. The two biotech traits currently on the market are herbicide tolerance and insect resistance (Bt). These traits are obviously very different, but most of the report just lumps them together as "GE crops", even though the report clearly states multiple times that Bt crops have reduced insecticide use.
To my eye, the same lumping and conflation appears to hold true for the study Benbrook published this week. In an email communication, Pamela Ronald, the University of California plant geneticist, agrees and adds,
he [Benbrook] does not take into account that glyphosate has displaced more toxic herbicides, thus there is a net reduction in toxicity.
Glyphosate is the compound used in Monsanto's Round-up Ready crop, which Benbrook cites as the main culprit for the pesticide uptick. Others have pointed out the same thing as Ronald, that the metric Benbrook uses is disingenuous. Additionally, Ronald said to me that
the [Benbrook study] conclusions conflict with virtually all peer reviewed studies, including two recent studies in PNAS and Nature which demonstrate reductions in synthetic insecticide use and enhanced biological diversity in GE cotton fields.
Bodnar, after taking a fresh look at Benbrook's newly minted study, has posted some thoughts, including this:
There are a lot of assertions made by Benbrook that just aren't supported.
Kniss, in his own examination, has raised some questions about the study's methodology. He notes that Benbrook "modeled" data for some of the years data on pesticide spraying wasn't available. Specifically, Kniss found that "70% of his [Benbrook's] estimated increase in herbicide use is based on...extrapolated soybean data, that are not well supported by actual USDA data." I made Benbrook aware of this critique by Kniss and he responded at the latter's website. I don't have the technical knowledge to independently assess the scientific merits of Benbrook's published findings or the statistical critique of it by Kniss. But in the last 24 hours I did communicate with numerous scientists familiar with Benbrook's work and poked around enough to learn the backstory of the study he published this week. In sum, I can safely conclude one thing: The widely disseminated Reuters article on Benbrook's study is a gross disservice to the public.