2. Background

Evaluating recent perceptions of shrimp products can disclose current and future trends in consumption, as market research is important in assessing consumer preferences. Although the literature pertaining to seafood is not as extensive as other food products, research devoted to consumer perceptions and attitudes has progressed in the past 25 years (e.g., Anderson and Bettencourt, Reference Anderson and Bettencourt1993; Charles and Boude, Reference Charles and Boude2001; Davidson et al., Reference Davidson, Pan, Hu and Poerwanto2012; Hanson, Herrmann, and Dunn, Reference Hanson, Herrmann and Dunn1995; Holland and Wessells, Reference Holland and Wessells1998; Jaffry et al., Reference Jaffry, Pickering, Ghulam, Whitmarsh and Wattage2004; Johnston et al., Reference Johnston, Wessells, Donath and Asche2001; O’Dierno et al., Reference O’Dierno, Govindasamy, Puduri, Myers and Islam2006; Quagrainie, Hart, and Brown, Reference Quagrainie, Hart and Brown2008; Ratliff, Vassalos, and Hu, Reference Ratliff, Vassalos and Hu2018; Roheim, Sudhakaran, and Durham, Reference Roheim, Sudhakaran and Durham2012; Rudd, Pelletier, and Tyedmers, Reference Rudd, Pelletier and Tyedmers2011; Wessells, Kline, and Anderson, Reference Wessells, Kline and Anderson1996; Whitmarsh and Palmieri, Reference Whitmarsh and Palmieri2009). Carlucci et al. (Reference Carlucci, Nocella, Devitiis, Viscecchia, Bimbo and Nardone2015) provide a detailed literature review of recent seafood preference studies in the United States and Europe. The research highlights that quality, safety, and price are important factors for consumers’ seafood purchasing decisions. However, when compared with consumers of other animal proteins, seafood consumers may have unique preferences. For instance, although research indicates that consumers will pay premiums for U.S. meat products (e.g., Lim et al., Reference Lim, Hu, Maynard and Goddard2013; Loureiro and Umberger, Reference Loureiro and Umberger2003), this has not always been the case for seafood. Jaffry et al. (Reference Jaffry, Pickering, Ghulam, Whitmarsh and Wattage2004) found that consumers prefer domestically caught to imported seafood for all the products they examined. However, Kuchler, Krissoff, and Harvey (Reference Kuchler, Krissoff and Harvey2010) found no statistically significant impact of country-of-origin labeling on U.S. household seafood consumption. Furthermore, Poerwanto (Reference Poerwanto2011) found strong consumer heterogeneity in participants’ preference for “previously frozen” and “never frozen” seafood, and Davidson et al. (Reference Davidson, Pan, Hu and Poerwanto2012) also supported that view. The aforementioned discussion indicates the need for further research related to consumer preferences for seafood products, assessing both “traditional” (i.e., fresh vs. frozen) and emerging credence attributes.

Following the release of the National Ocean and Atmospheric Administration’s Aquaculture Policy in June 2011, sustainable aquaculture has been strongly and globally advocated (Aquaculture Stewardship Council, 2016). The BAP is one organization certifying aquaculture production on an international scale. BAP defines the following elements as most important to responsible aquaculture: environmental responsibility, social responsibility, food safety, animal welfare, and traceability (Best Aquaculture Practices Certification, 2016). Through certifying finfish, crustaceans, and mussels, the third-party certification system outlines the elements of responsible aquaculture. BAP’s ecolabel appears on frozen and prepared seafood packages. Keeping in mind that aquaculture is pivotal to increasing seafood supply for food security, the Global Aquaculture Alliance (GAA) advocates responsible and sustainable aquaculture by working with nongovernmental organizations, industry, governments, and academia to meet these challenges (GAA, 2016). As of 2016, BAP certified 40 operations in the United States, of which 12 processed shrimp products (BAP, 2016). However, none originated specifically from Kentucky or South Carolina.

Food systems focusing on direct marketing (e.g., farmers’ markets, community-supported agriculture [CSA], etc.) have sprung up across the United States as states’ agriculture departments promote “local” producers’ products (i.e., Kentucky State Proud, Certified South Carolina Seafood, etc.; Low et al., Reference Low, Adalja, Beaulieu, Key, Martinez, Melton and Perez2015). “Local” food products are intended to represent transparency, traceability, and short supply chains (Marsden, Banks, and Bristow, Reference Marsden, Banks and Bristow2000). Similarly, preference for origin labeling, especially “local,” has made recent strides in the seafood sector. However, there is a lack of research in consumers’ perception of “local” seafood, and the definition is somewhat obscure (Adams and Adams, Reference Adams and Adams2008; Fonner and Sylvia, Reference Fonner and Sylvia2014; Smith and MacKinnon, Reference Smith and MacKinnon2007).

For seafood, consumers may define “local” by port, region, product traveling 175 miles inland, or even country (Brinson, Lee, and Rountree, Reference Brinson, Lee and Rountree2011; Merritt et al., Reference Merritt, DeLong, Griffith and Jensen2018; Nganje, Hughner, and Lee, Reference Nganje, Hughner and Lee2011). For example, the Port Clyde CSF in Maine distributes product to New York and promotes itself as “local” (Brinson, Lee, and Rountree, Reference Brinson, Lee and Rountree2011). For the present study, we utilize the “Kentucky State Proud” and “Certified South Carolina Seafood” labels to represent the term. Both labels allow products grown/harvested or processed in the state to be considered local. As a result, “local” is defined as a product sourced from the survey participants’ state of residence. This gives the term a more tangible depiction and examines how consumers value state-sourced shrimp.

A popular marketing model also placing emphasis on origin and traceability is the CSF, which is designed after the CSA, a food system that has gained extensive support across the United States (Andreatta, Nash, and Martin, Reference Andreatta, Nash and Martin2011). Just as CSA “shares” are allocated, CSF consumers provide up-front payments in exchange for scheduled seafood deliveries, and both consumers and producers share the risk of production (Brinson, Lee, and Rountree, Reference Brinson, Lee and Rountree2011). Although this analysis examines only farmed shrimp products, the “Product of CSF” label is utilized as an attribute. When considering demand markets outside of the CSF itself, CSF producers may opt to sell excess product (leftover from allocated shares) to indirect markets such as retail. Within these retail markets, the CSF label may be applied to represent the business model.

Furthermore, as a supplement to operations, a group of producers could pool their products into one entity (like a cooperative) and sell to an array of indirect outlets. For example, the Yankee Fishermen’s Cooperative in Seabrook, New Hampshire, operated and sold product outside its own shrimp CSF (Brinson, Lee, and Rountree, Reference Brinson, Lee and Rountree2011). In theory, producers may decrease risk by expanding and diversifying their markets. However, in comparison with other agricultural enterprises, it may be difficult to assimilate seafood products. Nevertheless, there have been no studies examining how consumers would evaluate such a label.

Lastly, another label not examined in the literature is HBH. The HBH program was founded by the Kentucky Department of Agriculture (KDA) and distributed nationally through the Farmer Veteran Coalition (KDA, 2015). The KDA was determined to add value to veterans’ farm products (KDA, 2015). Kentucky’s former commissioner of agriculture, James Comer, launched the HBH program in January 2013. The program now includes more than 250 members in 43 states, as the label serves to inform consumers that products were produced by veterans. The label is currently available to farmers, ranchers, fishers, and value-added producers of all branches and eras of military service (Farmer Veteran Coalition, 2015). Thus far, the label has been applied to a diverse array of food and beverage products, reiterating its versatility.

The HBH program was established under the belief that agriculture is a sound occupation for veterans. When returning from deployment overseas to uncertain job prospects, agricultural programs can exist as a viable option for employment. Moreover, veterans have the opportunity to register for mentoring programs on farms and the choice to enroll as a “Beginning Farmer” (eligible for privileges in USDA programs and loans).

The label quickly gained support from other states. South Carolina Agriculture Commissioner Hugh Weathers stirringly endorsed the label’s potential (radio interview) by discussing veterans’ natural fit in such occupations (South Carolina Department of Agriculture, 2015). These discussions are consistent with the establishment of the USDA’s Military Veterans Agricultural (MVA) Liaison, which was instituted in the 2014 Agricultural Act. The MVA Liaison “coordinates USDA leadership across the Department to provide information, resources, and support for active duty military and veterans interested in agriculture” (USDA, 2014). The MVA Liaison also facilitates relations between the USDA, other government agencies, and nonprofits to expand opportunities for veteran employment. In short, the HBH label provides consumers a mechanism to signify support for veterans’ service. Yet, no study, to the best of our knowledge, has investigated consumer preference for such a label. Our study fills this void.

4. Data

Data were collected from a total of 1,011 respondents (505 from Kentucky and 506 from South Carolina). Qualtrics, a large online panel of general consumers, was contracted to distribute the survey. Specific information regarding the sampling process, such as response rate, is proprietary. Nevertheless, Table 2 provides selected sociodemographic statistics for the two states from the 2012 American Community Survey, as well as the sample demographics. Female respondents made up the majority of survey participants (about 69%), but this makes intuitive sense when considering the female role in shopping behavior. For example, females resulted in 60% of the Fonner and Sylvia (Reference Fonner and Sylvia2014) sample when analyzing WTP for seafood attributes. In order to proceed with the survey, participants had to respond “yes” to whether they classified themselves as the primary grocery shopper. Most survey respondents were between the ages of 35 and 54 (43%). When asked about education, “some college, technical school, or associate’s degree” was the majority response (37%), and most respondents earned $50,000 to $74,000 (17%) annually.

Table 2. Sample and population sociodemographic statistics

^a State population is based on 25 years and over.

^b In 2014 inflation-adjusted dollars.

Note: State population statistics are based on the 2012 American Community Survey 1-year estimates.

5. Model

Lancaster’s (Reference Lancaster1966) formative study creates the framework to evaluate consumers’ utility for products with an array of attributes. By recognizing that a collection of traits is significant, Lancaster (Reference Lancaster1966) proposed the theory of demand for products having one or more attributes.

When considering a number of n-choice situations and evaluating consumer i’s selection of a product, McFadden’s (Reference McFadden and Zarembka1974) random utility theory can be applied. Consumer i’s indirect utility (U _ijn ) from selecting the jth product in a group of J products in the nth choice condition (n = 1, 2, 3, …) is described as a linear function of product attributes (X _ijn ) by the following equation:

(1) $${U_{ijn}} = {\boldsymbol X_{ijn}}{\boldsymbol\beta} + {\varepsilon _{ijn}},$$

where β symbolizes a vector of indefinite marginal utilities from product attributes X _ijn of the alternate j in choice situation n, and ε _ijn denotes the random error term of the computed utilities. Assuming that consumers act rationally, utility is maximized through selecting alternatives j in the nth choice framework (McFadden, Reference McFadden and Zarembka1974).

Owing to the extensive application of the conditional logit (CL) choice model for inference in discrete choice experiments, this econometric technique is applied as a baseline model in this research as well. By accepting that the independent and identical distribution (iid) of the error term (ε _ijn ) and independence of irrelevant alternatives (IIA) assumptions hold, the probability of the jth option being selected can be modeled as follows:

(2) $$P({Y_{in}} = j) = {{{\rm{exp}}(\boldsymbol{X_{ijn}}\boldsymbol\beta )} \over {\sum _{j = 1}^J{\rm{exp}}(\boldsymbol{X_{ijn}}\boldsymbol\beta )\;}}\;\;\;\;{\rm{For}}\;j = 1,2,\;...\;J,$$

where Y _in is an indicator variable representing the selection by consumer i in the nth choice situation. Considering a closed-form probability function, the CL method can be assessed using maximum likelihood estimation (McFadden, Reference McFadden and Zarembka1974). Though straightforward to estimate, the IIA assumption is restrictive. Train (Reference Train1998) developed an alternative to CL known as the mixed logit model. The mixed logit offers greater flexibility by relaxing the IIA assumption and incorporating preference heterogeneity, as well as accounting for correlations between multiple choice observations made by each respondent (Bliemer and Rose, Reference Bliemer and Rose2010). This is also referred to as accounting for correlation in unobserved utility over repeated choices made by each respondent, or that the parameter estimates of the marginal utilities vary across respondents. The choice probability identified by the mixed logit is modeled as

(3) $$P({Y_{in}} = j) = \int {{{{\rm{exp}}(\boldsymbol{X_{ijn}} \boldsymbol\beta )} \over {\sum _{(j = 1)}^J{\rm{exp}}(\boldsymbol{X_{ijn}}\boldsymbol\beta )}}} f(\boldsymbol\beta )d\boldsymbol\beta ,$$

where the coefficients in vector β are defined as random variables following density function f as

$${\beta _i} \sim f\left( {{\beta _{0,}}\;\boldsymbol G} \right),$$

with β ₀ as the means of β _i and G as the variance matrix. With the probability evaluated over a range of possible values of β _i and the absence of a closed-form solution, the approach of approximating the likelihood function with simulated maximum likelihood is applied to the model (Train, Reference Train2009).

Following the estimation of β in either the conditional or mixed logit model, marginal WTP measures for an attribute k is approximated as the coefficient estimate for the attribute divided by the negative marginal utility of price (Louviere, Hensher, and Swait, Reference Louviere, Hensher and Swait2000):

(4) $$WT{P_k} = - {{{\beta _k} + {\boldsymbol\beta _{\boldsymbol kD}} \times \boldsymbol D} \over {{\rm{\beta _{\rm price}}}}}$$

where D are covariates entered as interactions with the attributes in the logit model to further decompose the main effect of attribute k, and β_kd are the coefficients associated with the interaction effects. Thus, WTP measures the change in price associated with a unit increase in the respective attribute and approximates the monetary values of product attributes.

7. Conclusion and implications

By considering inland Kentucky and coastal South Carolina, we investigated consumer preferences for shrimp with multiple attributes. With a strong focus toward developing marketing strategies, the attributes examined were deemed possible candidates for producers and policy makers to adopt. By including unfamiliar, emerging, and hypothetical labels, we also contribute to the literature by establishing a foreground for further research.

Survey results had implications for both existing and developing attributes for the shrimp market. Consistent with previous studies, product form (i.e., “Fresh [Never Frozen]” or “Previously Frozen”) produced the highest premium, with both states garnering this result. Criteria referring to product form may infer how customers evaluate quality such as taste, sight, and smell of the product, with results indicating a fresh form consisting of a higher quality versus a previously frozen product. Results showing some South Carolina residents willing to pay higher premiums could suggest living in a coastal state, with closer proximity and access to fresh products, may generate greater preference for nonfrozen products.

Concerning the popularity of the local food movement and support for producers operating within participants’ state of residence, a “local” label allowed for analysis on preference for regional seafood. Although the literature is extensive, this study provided additional context of seafood labeling under which the “local” term was compared with other shrimp attributes. In both states, the “local” state labels generated the second highest premium behind product form, implying that support for shrimp sourced within the state is highly valued. Producers of aquaculture systems could use these results to justify labeling schemes indicating state origin, which ultimately may be more important than attributes such as environmental certification.

With the BAP label signifying environmental standards and sustainable practices, significant premiums resulted in both states, though not as strong as the product form and local attributes. Considering the relative magnitude of results, consumers may not fully understand nor value environmental stewardship as strongly in the case of aquaculture products. A possible explanation for this finding is that within aquaculture production systems, issues such as bycatch and stock (or supply) of certain seafood species is not as relevant as in ocean capture, so the value of certification may be limited. The only negative though insignificant product feature was the “Product of CSF” attribute. This could be the result of consumers’ concerns regarding the quality of farm-raised seafood in a CSF context.

Consumer perception of veteran-sourced products has not been studied within the food literature. Despite the recent emergence of Homegrown by Heroes (as of 2013) and existence in today’s food markets, the label was shown to be significant and produce a premium that was marginally less than “local” labels. Similar results between both states add to the importance of results. Policy makers may be encouraged to expand promotion and work with the Farmer Veteran Coalition to further develop and promote marketing programs for veterans in food production. Last, it is important to highlight that the WTP estimates are addable to each other. Thus, seafood producers can realize substantial premiums if their products include all the positive attributes.

For this study, we would like to mention limitations to better understand how future research can progress. First, the study includes two U.S. states located in the southeastern region; thus, it is hard to justify national implications. Although limiting our study to two states enables us to provide a more focused discussion on similarities and differences, future projects may survey a broader audience with greater sampling. Results can then be assessed from a national perspective, and more robust conclusions can be drawn on.

National scale is not the only market targeted when considering the global nature of the seafood market. A study of multiple consumers across various countries can support the understanding of international trade, as disputes considering the inflow of imported seafood products (e.g., shrimp) continue to affect domestic markets. An additional concern is the impact of wild-caught fisheries and environmental issues from a global scale, so consumer research in multiple countries could help understand preference in these areas.

Second, this study focuses only on the demand side of the market. Although it has been shown there is positive consumer support for many of the attributes considered, one must assess the feasibility of implementing these various labels by producers. With both mislabeling and transparency being prominent issues affecting the seafood supply chain, the supply side (or processing and distribution) must be considered, especially in strategizing marketing programs successfully. Future research may include developing producer surveys to analyze whether participation in certain programs would occur and if labels justify the economic investment to follow certain standards.

Third, the current analysis provides a snapshot of seafood consumption focusing on farm-raised shrimp. Certain participants who do not prefer shrimp or assess seafood attributes for other seafood species from a different perspective may affect conclusions made about certain labels. Studying other popular forms of seafood (e.g., salmon, tuna, etc.) could make conclusions more robust. A broader understanding of the overall consumption and, more important, long-term consumption trends is an interesting area for future research. Though many attributes included are emerging within the marketplace, one must also assess the sustainability of demand for future implementation.