Characteristics of Fish Buyers Likely to Purchase Farm-Raised Tilapia
in Honduras and Nicaragua 10PDVR1
Characteristics of Fish Buyers Likely to Purchase Farm-Raised Tilapia
in Honduras and Nicaragua
Product Diversification Research 1 (10PDVR1)/Study/Honduras and Nicaragua
Escuela Agrícola Panamericana, Zamorano, Honduras
Universidad Centroamericana, Nicaragua
University of Arkansas at Pine Bluff
Carole R. Engle
1) Identify those characteristics and preferences of restaurants, supermarkets and open-air fish
market vendors in Honduras and in Nicaragua that sell tilapia, those that used to sell tilapia, and those
that never sold tilapia.
2) Identify those characteristics and preferences of restaurants, supermarkets and open-air fish
market vendors in Honduras and in Nicaragua that are likely to add tilapia to their product lines.
Tilapia growers in Honduras and in Nicaragua have requested guidance identifying the most
promising markets and in developing efficient marketing strategies. Small and medium-scale producers, in
particular, do not have adequate resources to conduct market research and fewer resources to conduct
advertising campaigns and to develop markets. Identification of very specific types of market outlets that are most
likely to either sell more tilapia or to begin to sell tilapia would enable tilapia growers to minimize their
marketing expenses by targeting those types of markets likeliest to purchase their products.
Discrete choice analysis using qualitative dependent variables has been shown to be an effective
analytical tool in the analysis of perceptions and attitudes. While several studies have used these techniques in
U.S. markets for aquaculture products, they have not been applied to Latin American markets for
aquaculture products. Foltz et al. (1999) used discrete choice (both binary and multi-choice) techniques to show
that individuals with an urban background and desiring nutritious, easy-to-prepare foods preferred fillets
to whole-dressed trout. Engle and Kouka (1995) used discrete choice logit techniques to evaluate
potential consumer acceptance of canned bighead carp in Arkansas. Probabilities estimated showed that
canned bighead competed more favorably with canned tuna than with canned salmon. Income, region, and
gender significantly affected perceptions. Gempesaw et al. (1995) used logit analysis to show that most decisions
to purchase fresh seafood products for consumption were based on perceptions that products tasted good
and added variety to the diet.
Quantified Anticipated Benefits
The primary direct beneficiaries of this study will be tilapia growers in Honduras and in Nicaragua.
Small-scale growers, in particular, have struggled to develop markets without the benefit of comprehensive
market information and of characteristics of particular market segments in Honduras and in Nicaragua.
Small-scale growers have greater difficulty meeting the volume and sizing requirements of export markets and,
thus, domestic markets are their primary market targets. Yet without guidance based on quantitative analyses,
they will be required to expend considerable resources to identify market opportunities by trial and error.
The results of this study will assist them to concentrate their limited marketing time and resources on those
types of market channels that are likeliest to result in sales.
The Peruvian graduate student who will work on this project as part of his thesis research will benefit
from acquiring advanced quantitative analytical skills. His training, in turn, will benefit Peru by increasing
its human capital and building capacity in aquaculture marketing and analysis.
The PD/A CRSP will be a secondary beneficiary because improvements in market development will result
in industry expansion that will allow for greater adoption and need for PD/A CRSP research results. The
overall impact of the PD/A CRSP will be greater as the industry grows in Central America.
Location of Work: This work will be done at the University of Arkansas at Pine Bluff. This study will utilize
the existing database to conduct the analyses and publish the results.
Separate surveys were conducted of restaurants, supermarkets, and open-air market vendors
in both Honduras and in Nicaragua in the Ninth Work Plan. Direct personal interviews were conducted
in Honduras in 1999 and in Nicaragua in 2000. Complete censuses were taken of supermarkets and
open-air fish market vendors in both countries. In Honduras, random samples were taken of restaurants
in Tegucigalpa and San Pedro Sula (the two main urban population centers in Honduras) and in
selected small rural towns. Small rural towns were selected along the primary route from north to south
through the country to collect data along a possible gradient of preferences between the Pacific and
Atlantic coasts. Honduras is the only Central American country with good access between the two coasts
where this might be possible. Additional towns that were large enough to be included on maps and located
to the east and west of the Tegucigalpa-San Pedro Sula highway were included. In all, the following
small rural towns were included in the survey: Catascamas, Siguatepeque, Santa Barbara, Comaguela, Lago
de Yojoa, Choluteca, Puerto Cortes, Juticalpa, Comayagua, La Paz, Santa Maria del Real, and Campamento
La Lima. In Nicaragua, interviews were completed in Managua, Los Pueblos, Masaya, Granada, Boaco,
Jinotega, Matagalpa, Esteli, Chinandega, Leon, and Rivas. Sampled restaurants were selected from a list
of restaurant telephone listings. Fast-food eating establishments, bars, cafes, and Chinese restaurants
were excluded from the restaurant survey; only full-service restaurants were represented.
Supermarkets represented in the survey excluded convenience stores. In the fish markets, only those vendors with
a market stand within the market were interviewed. The response rates were very high (99 to 100%) for
the surveys. This is likely due to the novelty of marketing surveys in Central America. In all, there were
73 and 118 completed questionnaires for restaurants, 54 and 53 completed questionnaires for
supermarkets, and 66 and 106 completed questionnaires for open-air fish market vendors in Honduras and
The database developed in the Ninth Work Plan contains data on sales, supply, attitudes, and
socio-demographic and economic characteristics. Sales data include information on quantities of fish
and seafood species and product forms sold per week, peak seasonal demand periods, quantities of
tilapia sold by product form, prices of fish and seafood products, prices of tilapia and competing freshwater
fish species., and for those not selling tilapia, information on the reasons that they do not sell tilapia and
the likelihood that they will begin to sell tilapia the next year. On supply issues, the database includes
the primary sources of supplies of fish and seafood in general and of tilapia in particular, the principal
means of transporting fish and seafood and associated costs. Additional questions were asked related to
the consistency of supply of tilapia and of other similar freshwater cichlid fishes in Central America.
Respondents to the survey were presented with statements related to a variety of attributes of
tilapia. They were asked to assign a value from 1 to 10 in response to statements concerning each attribute.
A score of one represented complete disagreement with the sentence and a score of 10 represented
complete agreement. In that way, we measured the perceptions of buyers towards tilapia and
seafood characteristics. The statements included responses to the following attributes: flavor, odor, size,
price, nutritional value, quality, customer preference, reliability of supply, availability of tilapia, variety,
Socio-demographic and economic characteristics data include the following: customer income,
weekly sales, clientele type, age, location, size of store, education, years in business, and seating capacity
and cuisine type for restaurants.
Discrete choice (logit/probit) regression models will be developed to quantify
the effects of socio-economic and demographic variables on the decision to sell or not to sell tilapia and
the likelihood of the restaurant, grocery store, and market stands to sell tilapia the next year, respectively.
Development of regression models depend on the type of dependent variable. Since questions in
the survey related to fish sales decisions are consistently associated with discrete-valued answers,
we expected all dependent variables to be discrete values. For example, questions related to the decision
to sell (or not to sell) tilapia involve a one or zero answer, where one (zero) represents that the seller
had chosen to sell (or not sell) tilapia. Questions related to a restaurant's willingness to put tilapia on
the menu have answers coded in multi-valued discrete numbers, i.e., not willing (code = 0), willing (code
= 1), very willing (code = 2), and a multi-choice ordered dependent variable will be used.
Logit models explain the variation of y and assume the existence of an underlying latent
(or unobservable) variable which exhibits its value through an observed dependent variable. These
models explain the variation of y, using the variation of independent variables which are postulated to have
a causal linkage with the dependent variable. The types of independent variables used will
include: location of market (region/size of town or city), market outlet descriptors (size, clientele
characteristics, cuisine type, experience selling seafood, and attitudes and preferences). Logit model regressions are
run by maximizing the likelihood of the respective models with respect to the choice of coefficients in such
a manner that the probability of observing the given values of the dependent variable are as high
as possible. A likelihood function can be developed from the product of different probabilities for
the dependent variable.
The regional plan for Central America refers specifically to planning and implementing marketing
research activities in the region and for cooperation with the proposal PI. Research needs for tilapia culture in
Central America refer specifically to market development.
7/1/01 Initiate project.
4/30/02 Complete discrete choice analyses of restaurant data, Honduras and Nicaragua.
6/30/02 Complete discrete choice analyses of supermarket data, Honduras and Nicaragua.
12/30/02 Complete discrete choice analyses of open-air fish market vendors, Honduras and Nicaragua.
4/30/03 Final report submitted.
Engle, C.R. and P.J. Kouka, 1995. Potential consumer acceptance for canned bighead carp: a structural
model analysis. Marine Resource Economics, 10:101116.
Gempesaw, C.M., J.R. Bacon, C.R. Wessels, and A. Manalo, 1995. Consumer perceptions of
aquaculture products. American Journal of Agricultural Economics, 77:130612.
Foltz, J., S. Dasgupta, and S. Devadoss, 2000. Consumer perceptions of trout as a food item.
International Food and Agribusiness Management Review, 2(1):83101.