Certification of Shrimp and Salmon for Best Aquaculture Practices: Assessing Consumer Preferences in Rhode Island

Abstract This study uses a conjoint experiment to evaluate seafood consumers' preferences for wild versus farmed seafood in Rhode Island, while providing an option for farmed products to be certified for best aquaculture practices, focusing upon salmon and shrimp. The definition for best aquaculture practices provided to respondents in the survey is broadly based upon standards currently in use by aquaculture certification groups, highlighting sustainability of fish feed, and control of antibiotic use, water quality and stocking density. Using data from an in-person intercept survey, a conditional logit model shows that a sample of 250 seafood consumers in Rhode Island choose wild products over farmed even when farmed products are certified, and by an entity preferred by the consumer. Results warrant both further study of consumer preferences for certified aquaculture products across a broader population, and study of the effect of different explanations of 'best aquaculture practices' upon preferences. Keywords: aquaculturebest aquaculture practicescertificationconjoint analysisconsumer preferences ACKNOWLEDGMENTS We would like to thank David Beutel of URI for his helpful suggestions, and the Rhode Island Sea Grant and the RI Agricultural Experiment Station for financial assistance. We would also like to thank Sam Grimley and Jonathon Peros for their valuable assistance in data collection. Notes Standard errors in parentheses. ***1% significance level; **5% significance level; *10% significance level. See http://www.gaalliance.org/cmsAdmin/uploads/BAP-SalmonF-611.pdf. We use the term "best aquaculture practices" to adapt from usage by organizations such as the Global Aquaculture Alliance (GAA). We are not, however, attempting to utilize the exact definition of the GAA in our research or in any way endorse or pass judgment upon this definition. A complete list of all 48 conjoint questions, 24 salmon and 24 shrimp, is available upon request from the authors. In contrast to some other studies (e.g., Olynk, Tonsor, & Wolf, Citation2010) the effects coding is specified such that the attribute takes on a value of 0 when applicable, 1 if not applicable, and −1 otherwise. This specification affects the WTP calculation as well, in that the numerator in (6) need not be multiplied by 2 (Lusk, Roosen, & Fox, Citation2003). A series of nested tests were conducted to arrive at a final model of main effects with interactive effects for each species.