Experts find analysis flawed and based on erroneous assumptions
Washington, DC—In an attempt to justify a controversial proposal to remove all limits on line speeds in hog slaughter plants, the U.S. Department of Agriculture’s Food Safety and Inspection Service (FSIS) relied on a flawed analysis that downplayed the dangers posed to workers on hog processing lines, according to a review of the analysis by public health and statistical experts.
“This proposed new hog slaughter inspection system is a radical change that would jeopardize food safety,” said Debbie Berkowitz, program director for worker safety and health with the National Employment Law Project. “The proposal would decrease and privatize food safety inspections and allow for an unlimited increase in swine slaughter line speeds, endangering workers, consumers, and animal welfare.”
Berkowitz and an academic researcher recently obtained the USDA analysis and the underlying data through a Freedom of Information Act request submitted in February 2018, after the USDA refused repeated requests to make its analysis public, which prevented stakeholders from commenting on the analysis during the rulemaking process.
In its analysis, the USDA compared injury rates in traditional plants with those from five hog slaughter plants that were part of a 1990s pilot program that allowed the plants to run faster lines. The agency claimed that its preliminary analysis showed that the sped-up pilot plants “had lower mean injury rates” than traditional plants.
But two researchers at Texas State University who reviewed the USDA analysis concluded that the USDA’s analysis cannot be used to draw any statistically valid conclusions about worker-injury-rate differences in the pilot plants versus traditional plants.
In a scathing critique, the researchers—Dr. Celeste Monforton, DrPH, MPH, a lecturer in public health who specializes in occupational health and safety, and Phillip Vaughan, PhD, a research scientist who specializes in quantitative methods and statistics—wrote that the USDA’s analysis is replete with data limitations that undermine its assertions. These limitations include a non-random sample of only five pilot hog slaughter plants and 24 traditional plants. Moreover, USDA used injury rates that the Labor Department collected during the years 2002 to 2010—data that the Department had alerted users was not representative and warned against using to draw general conclusions.
The researchers wrote that USDA’s analysis was unsuitable for the available data. The agency’s analysis compared different and varying years of data for each plant. “Statisticians are likely to scratch their heads” at USDA’s claims that its analysis is statistically convincing, stated Monforton and Vaughan.
“In plain language, the USDA’s analysis is a joke. Had their students turned in that kind of analysis, these professors would have given them an F,” said Debbie Berkowitz of NELP. “This is another example of the USDA using flimsy data to push through a proposal that will adversely affect food safety, worker safety, and animal welfare.”
Berkowitz concluded: “USDA is using a faulty data analysis, one that it tried to hide from the public, to justify a proposal that will clearly endanger workers. Before any new proposal is finalized, USDA must conduct a new analysis of the impact of line speed increases on worker safety. The agency should withdraw the current proposal until that is complete.”
###
