Widespread Seafood Fraud and Some Thoughts on Scientific Sampling
This report in the Washington Post is getting great traction this week. The report builds on about two years of DNA testing by Oceana to understand whether the seafood we buy is, in fact, actually the seafood we wanted to buy. As pointed out in the past, seafood fraud generally results in customers buying cheaper seafood substitutes for fish such as cod, tuna, and snapper. And sometimes it even poses health risks as consumers eat seafood that can cause gastro-intestinal distress or, perhaps, is higher in mercury content.
I think this just goes to show how we need far better traceability and oversight in our seafood distribution systems.
I do have one point of caution, however. The study gives these great percentages, such as those appearing above in the infographic. While I think they point to a serious problem, it would be incorrect to say something like “71% of all tuna sold in sushi restaurants in the cities studied sell fraudulent products.”
Why not? It has to do with the sampling strategy, which was an expedient, but not scientific, approach.
Here’s how Oceana described their sampling study:
From 2010 to 2012, Oceana staff and supporters purchased 1,247 seafood samples from 674 retail
outlets in major metropolitan areas in 21 states. More than 350 Oceana supporters volunteered to
purchase seafood items in select cities. Supporters were encouraged to purchase targeted fish of
interest, such as fish species that had been found to be mislabeled in other studies and those with
regional significance. All samples submitted, including targeted and non-targeted fish, were analyzed
using DNA methods. (bold added)
The 674 retail outlets consisted of restaurants, sushi venues, grocery stores and seafood markets. More
than one seafood sample was obtained from many of the retail outlets visited, resulting in numerical 5
differences in the number of outlets visited and samples collected. Staff selected many of the restaurants
and sushi venues visited by searching menus of retail outlets listed in Zagat and Yelp for the fish types of
interest. Grocery stores and markets were selected mostly at random or based on convenience of staff
Thus no random selection of sales venues and highly inconsistent sampling from venue to venue and city to city. More can be found here in Oceana’s report.
Though the study gives a great picture of a serious problem, the picture it gives is blurry. More research would sharpen things, but is clearly unnecessary to see that regulatory and traceability changes are needed.