Data Doesn't Justify the USDA's New Catfish Regulation

A new USDA rule is the latest example of a regulatory response to a phantom problem.

On social media, a catfish is someone who prowls the Internet pretending to be someone he or she is not, often for the purposes of luring others into romantic situations. Notre Dame football star Manti Te’o famously claimed to be the victim of a catfish prank, and MTV even named a TV show after the phenomenon. Unfortunately, the U.S. Department of Agriculture is engaging in a different kind of catfish hoax, but this time the victim is the American public.

On March 1, a new USDA food safety regulation took effect that is supposed to reduce food poisonings from Siluriformes, the order of fish including catfish. But a close look at the USDA’s risk assessment accompanying the regulation reveals that something very fishy is going on with respect to this new rule.

According to the USDA’s risk assessment, only one “suspected” outbreak of salmonella has been reported from Siluriformes “in the past 20 years.” The one example the agency cites took place more than 20 years ago – in 1991 – and the USDA says there were only 10 illnesses associated with it. Based on this one data point and the magic of probabilistic modelling, the USDA makes the preposterous claim that Siluriformes cause an average of 2,400 salmonella illnesses each year. Elsewhere in the analysis, when presenting concentration levels of salmonella in catfish, the USDA lacked any data for catfish, so it used concentration levels in chicken instead.

Continue reading