Update, September 24, 2018, 8:57, a.m., EST: Two days after the publication of this story, CDC confirmed to us in an email that it had misclassified farmers as farming, fishing, and forestry, or “Triple-F,” workers. You can read more about the implications of this error in our follow-up story.
Last December, The Guardian ran a feature that claimed farmers were committing suicide at record rates. The author cited a 2016 study by the Centers for Disease Control and Prevention (CDC) that found that “farming, fishing, and forestry” had the highest suicide rate of any occupational group. Over the next few months, the New York Times, Mother Jones, NPR (twice) and numerous other outlets used the same study as the basis for stories about farmer suicides. In response, the House and Senate each released legislation to extend mental health care services to farmers. “Farmers … across America are in the midst of a suicide crisis,” said Representative Tom Emmer, a Republican from Minnesota, in a press release that announced the House bill.
There is one problem with this narrative: it’s not true.
We analyzed the study and found that either the media misinterpreted its results, mistaking a farmworker suicide crisis for a farmer one, or the CDC made a simple calculating error—one with far-reaching results. Regardless of what happened, the media’s main takeaway from the report—that farmers have the highest suicide rate in the country—is not supported by the study’s underlying data. This mistake has already had real-world consequences: despite research showing that agricultural workers have exceptionally high rates of mental distress, legislators have almost completely focused on helping farmers.
The CDC study, which the authors wrote was “not nationally representative,” attempted to calculate suicide rates by occupational group across 17 states. The researchers used a national reporting system to access suicide records from 2012 and associated them with occupational categories determined by the federal government. They also used estimates for the total number of workers in each occupational category from a federal survey. With these two figures, they computed suicide rates by occupational group.
Under the federal occupational guidelines, farmers are classified as having a “management occupation,” not a “farming, fishing, and forestry occupation.” Yet it was the farming, fishing, and forestry, or “Triple-F,” occupational group that had the highest suicide rate in the country: 84.5 per 100,000 people, over 4 times the overall average of 20.3 among people in the workforce. The suicide rate among managers, in contrast, was exactly average.
The Triple-F category is made up of, as the name suggests, workers in agriculture, commercial fishing, hunting, and forestry. The emphasis is on “workers.” Managers, business owners, administrators, advisors, and other white-collar or entrepreneurial positions within these industries are generally categorized separately. Almost no reporters noticed that the definition of Triple-F workers does not include farmers: agricultural workers constitute almost 90 percent of the group. These laborers tend to engage in repetitive manual tasks like planting, picking, and packing crops. They make little money—their median household income is between $20,000 and $25,000 a year, according to the latest National Agricultural Workers Survey—and are 80 percent Hispanic. They are often employed by farmers, whose median family income is around $79,000.
When we first asked the authors, in April, how farmers were categorized, the corresponding author responded that “farmers who were managers [were categorized as managers].” After we asked them in early May to clarify whether any farmers were purposefully categorized in the Triple-F category, the corresponding author told us that they needed to review “the programming and data files and [would] get back in touch [soon].” We asked again on June 18, but they did not respond prior to publication.
We became convinced that one of two hypothetical situations was true: 1) CDC had correctly classified suicides in the Triple-F category, or 2) CDC had incorrectly classified farmer suicides with Triple-F suicides. If 1) were true, then CDC’s study found that a group made up almost entirely of agricultural workers had the highest suicide rate in the country. If 2) were true, then we could not make conclusions about the respective suicide rates of Triple-F workers and farmers. But we did notice, when we attempted to reproduce the study’s results, that the researchers appeared to have divided by Triple-F workers only. Had they divided by farmers as well, they would have halved the reported rate, which would have dropped it to third in the study. The CDC may also have not divided by enough agricultural workers: Philip Martin, an emeritus professor at the University of California, Davis, says the survey CDC relied on undercounts them by half. With those workers included, the reported rate would have fallen to fifth or sixth. So, if CDC did lump farmer suicides in with Triple-F suicides, then the researchers should have reported a rate no higher than third in the study, and as low as sixth, rather than the highest.
On the other hand, if farmer suicides were not grouped together with Triple-F ones, then the suicide rate for Triple-F workers should have been about 50 per 100,000 people—and ranked second or third highest—after taking Martin’s correction into account.
Even without our level of analysis, news outlets should have been skeptical. Had they assumed workers were categorized correctly, they should have concluded that agricultural workers, not farmers, have the highest suicide rate in the country. Had they suspected farmers and agricultural workers were in the same category, they should have concluded that CDC provided no definitive findings for either.
Reporters could have also turned to research on agricultural workers, which, while incomplete, suggests they tend to suffer from extraordinarily high rates of anxiety, depression, and suicidal ideation (there is no research on their suicide rates throughout the United States). But instead, media outlets ran stories on the “farmer suicide crisis,” a narrative repeated by politicians as they began to work on the farm bill.
The current draft of the 2018 Senate Farm Bill funds a mental health support network for anyone who works in agriculture, but does so in a way that would likely restrict it to farmers and ranchers. It funds training programs for “farmers and ranchers” only, and limits the United States Department of Agriculture’s (USDA) long-term strategy and program assessment to the mental health of farmers. We contacted Senator Tammy Baldwin, a Democrat from Wisconsin, who sponsored a marker bill called the FARMERS FIRST Act, to ask if she planned to amend the bill’s language to make it less restrictive. Her office said that they are working to ensure that the language will encompass all agriculture-related occupations.
The CDC study is being used to push for more than mental health services. After China responded to a round of new tariffs by retaliating with restrictions on a number of U.S. agricultural goods, New York magazine and Mother Jones, among others, warned that this could exacerbate the so-called farmer suicide crisis. NPR went even further, taking Congress to task for not acting to increase agricultural subsidies. At the same time, media outlets and policymakers have largely ignored the devastating impact of our labor laws—which exclude agricultural workers from common protections—despite research showing that poor working conditions play a major role in their high rates of mental distress.
Many farmers struggle, and society fails everyone who dies by their own hand. But why journalists and policymakers have decided to focus on farmers, rather than farm workers, is worth examining. Researchers make mistakes, reporters are under extreme deadline pressures, and politicians often rely on outside experts to inform their initiatives. Yet, somehow, the least advantaged group discussed in this article may bear the harshest consequence.
Correction: An earlier version of this article included estimates of the median incomes of farmers and agricultural workers from the Occupational Employment System (OES) survey. Since the OES survey does not capture most farmers, these data have been replaced with the median household incomes of agricultural workers from the 2013-14 National Agricultural Workers Survey and of farmers from the 2016 Agricultural Resource Management Survey.