In June 2019, a paper titled: The characterization of seafood mislabeling: A global meta-analysis by Luque & Donlan, was published in Biological Conservation. The paper cited our story from last November, The irony of Oceana’s seafood fraud campaign, validating our remarks that Oceana, and many seafood fraud studies to date, employed flawed study design that inflated claims of “widespread” seafood mislabeling in the US and abroad.
Luque & Donlan conducted a global meta-analysis of existing seafood mislabeling literature to examine true rates of fraud, but also to attempt to characterize seafood mislabeling results in a more constructive light. The researchers built a model from over 27,000 seafood samples tested in hundreds of mislabeling studies then filtered out species without enough data to estimate a global mislabeling rate of 8%.
They defined mislabeled seafood as, “seafood product being physically substituted for another (i.e., species substitution).” Labeling that wasn’t specific enough (i.e. “salmon” but not “Atlantic salmon”) was not considered. If the label reads “salmon”, and the product is in the Salmonidae family, it is labeled correctly. Only misinformation equaled mislabeling, missing details did not—this is the right way to measure seafood fraud.
In our post we expressed concern with mislabeling study methodologies (particularly Oceana) whose methods were skewed to produce specific results. Luque & Donlan came to the same conclusions, citing two major issues: (1) pooled sampling across multiple studies without taking into account sampling effort, and (2) convenience sampling with little explanation of how or why seafood was chosen for sampling. They also felt that a study-level mean was too often utilized to present results in seafood mislabeling studies which, “is almost certainly overestimating mislabeling in many cases, and may be misleading the public even when qualifications are included (e.g., 30% of the seafood samples tested were mislabeled), particularly given the current sampling practices and that uncertainty in estimates is rarely reported.”
For reference, the 2018 Oceana Canada seafood mislabeling study that inspired our first post reported 44% mislabeling in the samples they tested.
Luque & Donlan are specifically referencing Oceana when they mention qualifiers being used to characterize the results. Indeed, that was Oceana’s response to us via email after we posted our November story – they claimed their press releases were not misleading because results were of “samples tested” and were not supposed to represent all seafood. Oceana’s press release can be found here.
The latest science estimates 8% of global seafood is mislabeled.
Another important point in our November post shared by Luque & Donlan is for seafood mislabeling researchers and policy makers to consider production scale when discussing fraud. For example, swordfish was chosen as one of a select few seafood species under the US Seafood Import Monitoring Program because of its risk for fraud. Yet Luque & Dolan found swordfish mislabeling to be less than 5%. Similarly, Oceana targeted hogfish as a primary species in one of its recent mislabeling studies, despite less than 15 tons of hogfish being landed on average annually. “A product with relatively low mislabeling rate but high production could be precipitating greater impacts compared to a species with a higher mislabeling rate, but low production,” explained Luque & Donlan. Policies should consider the impact of seafood fraud based on scale, and future seafood mislabeling studies would be more impactful if they focused on the most consumed species rather than obscure species like hogfish just to inflate mislabeling rates.
This paper does fall short of examining exactly where in the supply chain seafood mislabeling begins. Luque & Donlan acknowledged this need for future studies and that it will prove tricky to measure. To date there has been little research at the port level to examine mislabeling. If mislabeling occurs at the dock then it will have widespread effects across the supply chain, making it unjust to blame restaurants and retailers without further .
Luque & Donlan conclude with suggestions for future seafood mislabeling studies in order to produce results that can better characterize true seafood mislabeling rates:
- Study-level and naive means without measures of uncertainty, particularly from studies that include multiple species, are of limited utility and will often overestimate true mislabeling rates. More focus on products of concern, particularly in specific settings, are likely to be more insightful.
- Convenience sampling should be avoided if possible. Probabilistic sampling strategies that take into account products and venues will result in estimates with more utility and generalizability.
- Tools such as power analyses can inform sampling strategies, as well as any monitoring effort interested in detecting changes in mislabeling through time and space. Attention a priori to how seafood is sampled can help avoid under sampling, increase the ability to detect differences, and maximize the value of mislabeling studies.
- High-effort studies that focus on specific products in specific countries across the entire supply chain should be a priority and could reveal important insights into seafood fraud. This is particularly true for invertebrates and wholesale venues, which are both major information gaps for seafood mislabeling.
- While important, product mislabeling estimates alone cannot inform solutions to reduce seafood fraud. Rather, estimates must be combined with other data (e.g., production and prices) in order to understand the extent of mislabeling, as well as the potential causes and consequences.
This paper will be a useful stepping stone into more targeted, representative seafood mislabeling studies in the future. It supports our basic critique of Oceana’s seafood fraud campaign and many similar studies and reports: Study design and sampling methodology is important and should not be skewed to find fraud, but rather to examine the true extent of seafood mislabeling in our supply chains. Also, we do not have enough information to understand where in the supply chain mislabeling occurs, and thus cannot point the finger at restaurants, retailers, and even wholesalers, who may not be aware of mislabeled product in their possession.
Editor’s note: The authors of the paper have just launched a site for consumers to explore the mislabeling data from their meta-analysis: https://www.seafoodethics.org/
Jack Cheney
Jack has sourced, sold, cooked, and sustainably certified seafood over the past 10 years. In addition to his contributions to Sustainable Fisheries UW, he is working to increase traceability into supply chains and educate consumers, chefs and retailers on the value of environmentally sustainable seafood. He earned a Master's in Marine Affairs from the University of Washington in 2015.
Check out our other coverage of seafood mislabeling:
The Irony of Oceana’s Seafood Fraud Campaign
Oceana’s seafood fraud campaign is based on poor science that misleads the public, stokes consumer fear, and hurts fishermen, mongers, and chefs.
Oceana’s response to Seafood Fraud Campaign Criticism
Dr. Kimberly Warner, a senior scientist at Oceana, responds to previous criticism of Oceana’s Seafood Fraud campaign.
Latest Oceana seafood fraud report takes aim at SIMP, but misses mark
Oceana’s latest seafood fraud report takes aim at reforming SIMP, however the science behind their advocacy is misleading, once again.
