<%BANNER%>

An Evaluation of factors influencing away-from-home consumption of crawfish in the Gulf region

University of Florida Institutional Repository

PAGE 1

AN EVALUATION OF FACTORS IN FLUENCING AWAY-FROM-HOME CONSUMPTION OF CRAWFISH IN THE GULF REGION By XUMIN ZHANG A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2004

PAGE 2

Copyright 2004 by Xumin Zhang

PAGE 3

To Miao, our kid, and my parents

PAGE 4

ACKNOWLEDGMENTS I would like to deeply thank Dr. Lisa House, chair of my supervisory committee, for her advice and timeless help on this thesis. Also, I would like to thank Dr. Richard Kilmer for giving me the chance to do research with him during my masters program. Those experiences have broadened my view and enriched my knowledge base. Dr. House and Dr. Kilmer have given me the most help and advice during my masters program, and I would like to express my appreciation again to them. I also want to thank the Food and Resource Economics Department for those who have given me support and encouragement, like Dr. Robert (Jeff) Burkhardt, Dr. Donna Lee, and Dr. Allen Wysocki, and particularly Jessica Herman, our program assistant. Finally, my deepest gratitude goes to Miao and my parents for their support throughout these years. iv

PAGE 5

TABLE OF CONTENTS Page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES............................................................................................................vii LIST OF FIGURES.........................................................................................................viii ABSTRACT.......................................................................................................................ix CHAPTER 1 INTRODUCTION........................................................................................................1 Preamble.......................................................................................................................1 Problematic Situation....................................................................................................1 Researchable Problem..................................................................................................7 Objectives.....................................................................................................................8 Hypotheses....................................................................................................................9 2 LITERATURE REVIEW...........................................................................................11 3 SURVEY INSTRUMENT..........................................................................................17 Survey Instrument.......................................................................................................17 Survey Contents..........................................................................................................19 4 DATA.........................................................................................................................22 5 THEORETICAL MODEL AND MODEL SPECIFICATION..................................30 Theoretical Model.......................................................................................................30 Model Specification....................................................................................................33 Truncated-at-Zero Count Data Double Hurdle Model...............................................38 6 EMPIRICAL RESULTS............................................................................................40 Issues of Crawfish Consumption................................................................................40 Regional Consumption........................................................................................40 Farm-Raised Crawfish.........................................................................................40 Domestic Versus Imported Crawfish..................................................................41 v

PAGE 6

Product Form.......................................................................................................42 Reasons for Consuming and Not-Consuming.....................................................44 Reasons for consuming................................................................................44 Reasons for not consuming..........................................................................44 Away-From-Home Crawfish Consumption...............................................................47 7 CONCLUSIONS AND IMPLICATIONS.................................................................62 Summary.....................................................................................................................62 Conclusions.................................................................................................................67 Implications................................................................................................................68 APPENDIX 2004 MARKET SURVEY OF FOOD CONSUMPTION...........................71 LIST OF REFERENCES...................................................................................................76 BIOGRAPHICAL SKETCH.............................................................................................80 vi

PAGE 7

LIST OF TABLES Table page 4-1. Residence region of survey respondents....................................................................23 4-2. Summary of demographic information.......................................................................27 4-2. Continued....................................................................................................................28 4-3. Descriptive statistics on other factors included in the Double-Hurdle model............29 6-1. Sample frequency distribution of the dependent variable (n=733)............................48 6-2. Description of variables included in the double-hurdle model...................................50 6-2. Continued....................................................................................................................51 6-3. Empirical results of single-decision and double-hurdle: maximum-likelihood estimates and marginal effects.................................................................................53 6-3. Continued....................................................................................................................54 vii

PAGE 8

LIST OF FIGURES Figure page 1-1. United States per capita total, fresh & frozen fish and shellfish consumption.............2 1-2. Percentage of away-from-home food expenditure at 1988 prices................................3 1-3. Percentages of seafood meals eaten away-from-home.................................................4 4-1. Comparison of U.S. population in the Southeast and survey respondents by age.....26 4-2. Comparison of incomes of all the U.S. population and incomes of the survey respondents...............................................................................................................26 6-1. Regional percentage of crawfish consumption...........................................................41 6-2. Comparison of preferences over domestic and imported crawfish products..............42 6-3. Categories of reasons given for consuming crawfish (crawfish consumers vs. away-from-home consumers)............................................................................................45 6-4. Categories of reasons given for consuming crawfish (away-from-home consumers only, at-home consumers only, and all crawfish consumers)..................................45 6-5. Categories of reasons given for not consuming or not consuming more crawfish (consumers vs. non-consumers)...............................................................................46 6-6. Categories of reasons given for not consuming or not consuming more crawfish (away-from-home only, at-home only, and non crawfish consumers)....................48 A-1. Crawfish pictures.......................................................................................................71 viii

PAGE 9

Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science AN EVALUATION OF FACTORS INFLUENCING AWAY-FROM-HOME CONSUMPTION OF CRAWFISH IN THE GULF REGION By Xumin Zhang May 2004 Chair: Lisa A. House Major Department: Food and Resource Economics U.S. consumers are traditionally known as away-from-home consumers of seafood products. Over the past 20 years, consumption of ethnic foods has steadily increased. For example, approximately 25 percent of Americans say they enjoy Cajun cooking. In the United States, consumption of crawfish is usually associated with special events or Cajun cooking, which differs from consumption of crawfish in other countries. Growth in consumption has provided opportunities as well as challenges to Chinese exporters and U.S. importers as well as domestic farmers, processors, and marketers of crawfish. This thesis focuses on developing an understanding of factors influencing away-from-home consumption of crawfish in the Gulf region. A web-based consumer survey was conducted to obtain information about consumer preferences. Demographics as well as stated consumer preferences have significant effects on both participation and consumption decisions, but in different scales or directions. For example, enjoying the flavor, preferring Cajun cooking, and adding variety in the diet were the main reasons for ix

PAGE 10

consuming crawfish. The results also indicated that crawfish is an inferior good and relatively price sensitive. Caucasians tend to be less likely to consume crawfish away-from-home and to consume less frequently. Additionally, the lowest income group has the highest likelihood and frequency of crawfish consumption. x

PAGE 11

CHAPTER 1 INTRODUCTION Preamble Crawfish belongs to the scientific class Crustaceans. Crawfish has a hard external shell, which provides protection to its body. These small crustaceans are related to lobsters and closely resemble them. The Red Swamp Crawfish (Procambarus clarkia) is the species that is the most acceptable for cooking, and is commercially produced and consumed. The shell of the adult Red Swamp Crawfish is dark red to nearly black ( Ladewig and Schaer, 1993 ). In the Southeastern United States and the Gulf region, particularly in Louisiana, crawfish farming and consumption have been a part of the culture ( Avery and Lorio, 1999 ). Catches from the wild are usually seasonal and unpredictable. Increases in product demand beyond the amount traditionally caught in the wild led to crawfish farming ( DAbramo et al., 2002 ). In the United States, consumption of crawfish is usually associated with special events or Cajun cooking, which differs from consumption of crawfish in other countries. This thesis will focus on developing an understanding of factors that influence consumption of crawfish in the Gulf region. Problematic Situation Seafood includes aquaculture products such as crawfish and harvests from salt and fresh water sources. In the United States, the consumption of seafood has become an important part of the diet. According to the United States Department of Agriculture (USDA), the consumption of seafood has increased in recent decades. The estimated per 1

PAGE 12

2 capita seafood consumption increased by three pounds from 1970 to 1992, to 14.7 pounds, and it increased further in the 1990s. In 2000, the estimated per capita consumption of seafood was 15.6 pounds (Figure 1-1). 02468101214161819701973197619791982198519881991199419972000lb/person Total Fresh & Frozen Source: USDA-ERS, 2002a Figure 1-1. United States per capita total, fresh & frozen fish and shellfish consumption. Nearly two-thirds of the consumption of seafood was in the form of fresh and frozen products, the remainder was in processed products such as canned and cured. This compares to 5760 percent in the 1970s (Figure 1-1). In value terms, according to the National Fisheries Institute (NFI), American consumers spend almost $50 billion each year on a wide variety of fish and shellfish products. The thousands of firms that produce, process and distribute fish and shellfish are located throughout the United States, and annually contribute more than $25 billion to the U.S. gross national product ( NFI, 2003 ). Along with eating more seafood, U.S. consumers are dining out more often than ever before ( USDA-ERS, 1999a ). Away-from-home food expenditures increased from 36

PAGE 13

3 percent of total food expenditures in 1970 to 47.5 percent in 1996, and 48 percent in 2001 (Figure 1-2). Reasons for this trend include smaller household size, more affordable and convenient fast food services, a growing number of women working outside the home, and higher household incomes ( USDA-ERS, 1999a ). 0%10%20%30%40%50%60%19701973197619791982198519881991199419972000 Source: USDA-ERS, 2002b ; USDA-ERS, 2002c Figure 1-2. Percentage of away-from-home food expenditure at 1988 prices. US consumers are traditionally known as away-from-home consumers of seafood products. Although no precise data are available, one estimate by Keithly ( 1985 ) suggested that the quantity of away-from-home consumption of seafood products ranged from one-third to two-thirds of all seafood consumed. A recent study by Selassie, House, and Sureshwaran ( 2002 ) found 57, 62, and 58 percent of meals of shrimp, oysters, and catfish, respectively, were consumed away-from-home (Figure 1-3). This set of figures compares to general food consumption, where 16 percent of the meals were eaten away-from-home in 1978, a figure that increased to 29 percent by 1995 ( USDA-ERS, 1999a ).

PAGE 14

4 Crawfish can be seen as a healthy food, which is high in minerals and protein, and low in calories and saturated fat ( Ladewig and Schaer, 1993 ). Also, according to the Florida Department of Agriculture and Consumer Services (FDACS), crawfish is known for its delicious meat, and is consumed all over the United States ( FDACS, 1992 ). 0%10%20%30%40%50%60%70%Meals*Shrimp Oyster CatfishSeafood Note: All meals data were from 1995, other categories were from 2001 (more recent data not available for all meals). Source: Selassie, House, and Sureshwaran, 2002 ; USDA-ERS, 1999a Figure 1-3. Percentages of seafood meals eaten away-from-home. When talking about the consumption of crawfish, people used terms such as Louisiana parties, family gatherings, and Cajun-style boiled crawfish ( Ladewig and Schaer, 1993 ). Traditionally, most crawfish in the United States was consumed in Louisiana ( Dellenbarger et al., 1996 ). Over the past 20 years, consumption of ethnic foods in the United States has steadily increased, with approximately 25 percent of Americans indicating they enjoy Cajun cooking ( RoperASW, 2002 ). Particularly, a large proportion of young consumers indicate a preference for ethnic foods ( FMI, 2003 ). Since younger consumers are usually seen to set future trends, their preference of ethnic foods, including Cajun-cooking might hint at an increasing demand for crawfish. As a result, the

PAGE 15

5 demand for Cajun dishes that include crawfish is steadily growing, and the potential for growth in the crawfish industry is high. Due to the increases in year-round demand for crawfish and the seasonality and unpredictability of crawfish caught from the wild, crawfish aquaculture has arisen. The majority of crawfish produced in the United States is used for food and originates in Louisiana ( McCullough et al., 2001 ), and the price generally is considered expensive, particularly in other regions. An increased interest in finding alternative sources of income has led farmers to transform their fields into crawfish farms, and has expanded production along the Gulf coastal regions of the United States in the last 30 years ( Ladewig and Schaer, 1993 ). A suitable crawfish pond needs relatively flat land, soil with a high clay content, and an adequate water source. The states in the Southeast and Gulf region have geographic and climatic advantages in crawfish production ( FDACS, 1992 ). Nevertheless, although the growing demand for crawfish has heightened the interest of agricultural producers, the industry in states other than Louisiana is still comparatively weak. For example, due to the state regulations on crawfish aquaculture in Florida, there were only 21 facilities certified as crawfish producers in the year of 2003. Among those 21 farms, less than 15 farmers produce crawfish for the food market, and the acreage for crawfish producing is relatively small ( FDACS, 2003 ). In 1990, there were about 160,000 acres of managed crawfish ponds in the United States, and total harvest of crawfish was around 760 million pounds a year ( Ladewig and Schaer, 1993 ). However, due to the rapid increase of imports since 1994, crawfish production decreased to about 36 million pounds in 1998 ( USDA-ERS, 1999b ). In the

PAGE 16

6 year 2000, total harvest of crawfish in the US was around 100 million pounds ( FDACS, 2001 ), partly as a result of tariffs on imported crawfish. Crawfish production is mainly in the lower Gulf Coast regions, such as Texas, Mississippi, Florida, and Louisiana, with most of the acreage found in Louisiana ( Avery and Lorio, 1999 ). It is worth mentioning that since 1994, an increasing proportion of the market volume has been imported from foreign countries such as China and Spain. The market share of imported crawfish increased to 87 percent in 1996 ( Roberts, 2000 ), and imports continuously increased 173 percent in 1998. The total quantity was over six million pounds in 1998, mainly in the form of tail meat ( McCullough et al., 2001 ). The International Trade Commission ( ITC, 1997 ) has found that the increasing crawfish tail meat imports from China have caused material injury to the domestic crawfish processing industry, and has imposed Less Than Fair Value (LTFV) duties on crawfish since 2000 ( Roberts, 2000 ). The average import tariff is as high as 123 percent, and there also are levies on past shipments. Over 90 percent of crawfish imports come from China even with restrictive duties ( McCullough et al., 2001 ). Because of the limited quantities of crawfish available in the U.S. market, prices have been higher in recent years. Given the deficit in domestic supply, it appears there will be opportunities for crawfish farming in domestic production. Although consumption data for crawfish are not available, according to several studies on crawfish consumption ( Ladewig and Schaer, 1993 ; Yen et al., 1995 Dellenbarger et al., 1996 ; McCullough et al., 2001 ), the demand for crawfish has increased, and the growth in consumption has provided both opportunities and challenges to Chinese exporters and US importers as well as domestic farmers, processors, and

PAGE 17

7 marketers of crawfish. However, domestic market feasibility, market growth potential, and market size and trends are still unknown, and there are concerns about the viability of crawfish production and marketing. Knowledge about consumers preferences and perceptions in terms of crawfish consumption becomes important in identifying the viability of crawfish production and marketing. The lack of information about consumer preferences is a serious barrier to answering questions about the viability of crawfish production and marketing. Therefore, for the growth of the industry, especially in the Gulf region, new information is needed on the factors that influence crawfish consumption patterns. For crawfish, factors that might influence the consumption decision and frequency of consumption are unknown. Further, it is not known if these factors have the same influence on at-home consumption as compared to away-from-home consumption decisions. Developing an understanding of those factors that influence consumption of crawfish becomes comparatively important. Researchable Problem Growth in consumption has offered opportunities and challenges to producers, processors, and marketers of crawfish. Knowledge about consumers preferences and perceptions in terms of crawfish consumption is important for identifying the viability of crawfish production and marketing. While potential producers and importers of crawfish may question the profitability of new crawfish ventures, processors, and marketers may question the viability and profitability of crawfish marketing. The lack of information about consumers preferences and perceptions is a serious barrier to assessing the viability of crawfish production and marketing. Therefore, new information is needed about factors, such as demographic, socioeconomic, and consumers preferences and

PAGE 18

8 perceptions about crawfish, which might influence the probability and the frequency of consumption. Due to the fact that Americans are traditionally away-from-home seafood consumers and most crawfish are consumed in the Southeastern United States, and away-from-home consumption is the main means of introducing seafood to new consumers, this study will focus on developing an understanding of factors that might influence away-from-home consumption of crawfish in the Southeastern United States, particularly in the Gulf region. Objectives The overall objective of this study is to develop an understanding of the factors that influence away-from-home consumption of crawfish. We will collect information from existing and potential crawfish consumers in the Gulf region, and study the effects of geographical regions and household characteristics for away-from-home crawfish consumption that are not available in aggregate time series. Also, stated consumer preferences and perceptions regarding taste, flavor, nutrition, safety, appearance, and availability will be examined to see whether they affect the consumption decision and consumption frequency. Specific objectives are: To inform producers, processors, and marketers about what attributes of crawfish, such as taste, flavor, nutrition, safety, and appearance, and which form of crawfish product (whole boiled or tail meat) are preferred by current and potential consumers in the Gulf region. To inform importers of crawfish about the factors that will influence the demand and market trends of crawfish in the near future. Due to the domestic popularity of crawfish in China, Chinese exporters target the United States as one of the biggest importing countries of crawfish. Information about factors that influence consumption might help Chinese exporters better understand the different preferences and perceptions regarding crawfish.

PAGE 19

9 Hypotheses Patterns of away-from-home consumption are likely to differ from at-home consumption. Moreover, for crawfish, it is likely to differ from other seafood consumption and differ from consumption in other countries. Many factors might influence participation and consumption decisions. Those factors can be separated as demographics and preferences or perceptions. Demographics might include geographical regions, ethnicity, household income, and educational levels. In addition to price, consumer perception regarding taste, flavor, nutrition, safety, appearance, and availability might influence crawfish consumption. Given these considerations, the following hypotheses are proposed: Crawfish consumption will vary by region of residence. Consumers from the main production regions will be more likely to consume crawfish. Farm raised and/or domestically produced crawfish will be preferred by consumers over wild caught and/or imported crawfish. Fresh crawfish and tail meat will be preferred by consumers in accordance with the market trends toward convenience and freshness. Education and household income will negatively affect the probability of away-from-home crawfish consumption. Household size, education, and household income will positively affect the frequency of away-from-home consumption among households that do consume crawfish. Other demographics, such as age, race, and ethnicity will influence both the probability and frequency of away-from-home crawfish consumption. For example, older groups will be more likely to consume crawfish. On the contrary, Caucasians are less likely to consume crawfish and consume it less frequently. Consumer preferences, including taste, nutrition, and availability issues, will affect the participation decision for away-from-home crawfish consumption. Consumer preferences, including taste, nutrition, and availability issues, will affect the frequency of away-from-home crawfish consumption.

PAGE 20

10 Consumer perceptions regarding seafood safety and appearance will negatively affect participation and consumption decisions.

PAGE 21

CHAPTER 2 LITERATURE REVIEW Not many empirical studies focusing on crawfish consumption patterns exist in the literature. Few studies have focused on crawfish consumption, and no study has examined away-from-home consumption of crawfish in the Gulf region. Studies on crawfish consumption patterns were mostly conducted in the early to middle 1990s. As a result, new information is needed. Two previous studies, Yen et al. ( 1995 ) and Dellenbarger et al. ( 1996 ), examined crawfish consumption in South Louisiana and Louisiana, respectively. Yen et al. ( 1995 ) investigated the determinants of crawfish consumption in South Louisiana using a generalized limited dependent variable model that is similar to Craggs double hurdle model ( Cragg, 1971 ). The study included socio-economics and demographics as the independent variables, and no price or expenditure data were collected. The independent variables included income, household size, and dummy variables indicating professional types, employment status, education, religion, and race. The quantities (in pounds) of crawfish consumed by the responding households during a five-day period previous to the survey were collected as the dependent variable. It is worth mentioning here, among its 915 responding households, only 200 households, or 21.9 percent, reported consumption of crawfish. The study found that demographics, such as income, household size, and skilled labor, increased the likelihood of crawfish consumption but not the conditional level of 11

PAGE 22

12 consumption. For example, households characterized by the attributes of higher income, larger household size, skilled labor, Catholic, and white are more likely to consume crawfish than others. In terms of the conditional level of crawfish consumption, they found that education and employment status are among the household characteristics that determine the level of consumption. Households with skilled laborers or unemployed workers consume more crawfish than other groups. In addition, the results suggest that consumption of crawfish is income inelastic, although insignificant. However, the study only centered attention on the most recent five-day period of crawfish consumption, and thus ignored the seasonality of the crawfish harvest. Moreover, the study only included socio-economics and demographics that might influence consumption; however, it did not illustrate other factors regarding consumer preferences and perceptions, which might also influence crawfish consumption. A shortcoming of the study is the method in which the dependant variable was obtained. Respondents were asked to recall how many pounds of crawfish they had consumed in the five-day period. Respondents were then converted to live-product equivalents. The calculation of the dependent variable caused discrepancies from realistic quantities. However, it is not unexpected as respondents may find it hard to estimate the quantity consumed in pounds. Dellenbarger et al. ( 1996 ) examined the consumption of boiled crawfish in Louisiana by using a mail survey. A Logit model was used to estimate the probability of household consumption of boiled crawfish. The study surveyed four rural and four urban parishes of Louisiana to identify their crawfish and seafood consumption patterns.

PAGE 23

13 As a result of the small survey region, interpretation of the study should be conducted with caution. Of 858 effective observations, 203 households, or 23.6 percent, reported consumption. The study only focused on participation decision, and households were asked if they had consumed boiled crawfish in the last five days. As a result, a Logit analysis was used. A value of one represented households consuming boiled crawfish, and a value of zero represented non-consumption. The dependent variables included income, household size, urban resident, race, religion, and education. The study found that religion had an influence on the probability of crawfish consumption. Catholic and Protestant respondents were more likely to consume crawfish. Also, white households were more likely to consume boiled crawfish than non-white households. However, unlike Yen et al.s findings, education was not a statistically significant variable in determining the likelihood of boiled crawfish consumption. The study suggested households with incomes below $25,000 should be targeted for boiled crawfish consumption, since crawfish consumption declined with income. Additionally, the results showed that urban households had a lower probability of boiled crawfish consumption compared to rural households. Schupp et al. ( 1991 ) studied U.S. food stores handling of crawfish, finding that many consumers were not knowledgeable about aquaculture products, especially in areas of limited or no local production ( Schupp et al., 1991 ). In addition, market expansion for crawfish outside the South Central region was found to be largely dependent on obtaining the support of grocery stores. The study found that many existing markets for crawfish are not being met by the current domestic supply and many

PAGE 24

14 people do not know how to prepare crawfish at home. Availability is limited and crawfish is likely not to be found in local supermarkets. Other empirical studies focusing on seafood consumption patterns include Keithly ( 1985 ), Cheng and Capps ( 1988 ), Yen and Huang ( 1996 ), and Drammeh et al. ( 2002 ). Keithly ( 1985 ), using food consumption survey data, focused on a set of socio-economic and demographic factors that affect at-home consumption of total seafood and five specific products. He found that region, urbanization, race, household size, money value of meals consumed away-from-home, and income were all contributing factors that helped to explain at-home seafood consumption patterns. Cheng and Capps ( 1988 ) investigated the key socio-demographic determinants of at-home demand for several fresh and frozen finfish and shellfish species. They found factors explaining the variation of expenditures on seafood were own price, household income, household size, coupon value, geographic region, urbanization, race, and seasonality. Drammeh et al. ( 2003 ) found source of seafood for consumption, enjoyment of flavor, availability, price, allergies, gender, and geographic reasons to be significant in determining probability of participation in oyster consumption. They found the double-hurdle model was a significantly better fit than the tobit model. Variables significant in the level of consumption of oysters included source of seafood for consumption, enjoyment of flavor, tradition, price, product safety, geographic region, income, and age. Yen and Huang ( 1996 ) performed a detailed study on household demand for finfish. The study estimated household demand for finfish in the United States using a limited

PAGE 25

15 dependent variable model similar to Craggs double hurdle model, which accounted for both participation and consumption decisions. The study found that the price of finfish, shopping frequency, geographic region, race, and life-cycle variable were the key factors that significantly affect both the probability of participation and the level of household finfish consumption in the United States. Furthermore, they found a variable might exert opposite effects on the probability and level of seafood consumption. Also, the study concluded that the double hurdle model was particularly relevant for studying seafood consumption behavior, because the participation and consumption decision were likely to differ. Results of the analysis were useful for seafood marketers in planning and developing marketing strategies. Based upon the results, seafood marketers can differentiate between the factors that influence participation decisions and the factors that influence consumption decisions. In conclusion, there have not been many empirical studies of crawfish consumption in the Southeast and the Gulf region, particularly in recent years. As a result, research conducted to study the consumption of crawfish, including regions beyond Louisiana, would be useful. In addition to demographics, factors regarding consumer preferences also influence the consumption of crawfish either at-home or away-from-home. It has been suggested to include those factors in the analysis of crawfish consumption patterns. Those factors include information from current and potential crawfish consumers about their preferences and perceptions of crawfish to determine which of these might influence decisions and the frequency of crawfish consumption. Attention to the consumption

PAGE 26

16 patterns and the identification of those factors could be helpful in developing marketing strategies for the industry targeting these markets.

PAGE 27

CHAPTER 3 SURVEY INSTRUMENT This study will examine away-from-home consumption of crawfish in the lower Gulf regions using data collected from an Internet-based survey. The web-based survey was conducted in late January of 2004, which was just before the main crawfish harvest season. However, the survey asked the frequency of away-from-home crawfish consumption in the previous 12-month period to accommodate the seasonality of crawfish consumption. Although it is expected that recalling the frequency of consumption over 12 months may be difficult, in the case of crawfish the low overall level of consumption during one year reduces this concern. For example, people can remember one time without difficulty, but cannot easily remember 100 times. Survey Instrument Web-based surveys are increasing in popularity with the growth of the Internet. Some advantages of a web-based survey include the ability to use color graphics, higher response rate and higher completion rate, and lower cost compared to other survey instruments ( Larkin et al., 2002 ). Color graphics can be very expensive for mail surveys and impossible for telephone surveys. In our food market survey for crawfish consumption, color graphics were included to illustrate crawfish, following a verbal description. Since crawfish resemble lobster, it is extremely important to avoid confusion by providing color graphics. Participants were recruited through Survey Sampling, Inc.s web-survey service. Similar to telephone survey centers, this service charged per completed survey, thus 17

PAGE 28

18 allowing the researcher to guarantee the number of responses. This differs from mail surveys where the response rate is highly dependent on factors like survey length. By surveying on the Internet, respondents might be more likely to complete the entire survey. Additionally, in each question page, blank responses can be disallowed in some critical questions by using reminder dialogue boxes. Compared to mail surveys and focus groups, the web-based survey lowers the cost and increases the completion rate. With the growth of the Internet and Information technology, the web-based survey is one of the most effective survey instruments. However, there are still some disadvantages of conducting a web-based survey such as non-representation of consumers without Internet access ( Larkin et al., 2002 ). Uncompleted surveys still exist. Also, if a respondent completes the survey several times, it would be difficult to pick only one as the final response. Steps can be taken to address the disadvantages. 1 Another important issue included in the web-based survey instrument is how to obtain potential respondents. Our method involved contracting a survey sampling company, which can provide the stratified survey sample with the potential respondents fitting our specific requests such as region of residence, demographic compositions comparable to the Census data, etc. An additional advantage of using an Internet survey service is that participants have agreed to participate in web surveys, leading to a higher completion rate. The procedures of how respondents completed the Internet survey are shown below: Once the respondents had read the welcome page and started the survey, a time stamp 1 See Larkin et al. (2001) for details of developing an Internet survey instrument.

PAGE 29

19 would automatically record when the respondent started the survey. Responses were 'submitted' when the respondent completed one page and wanted to move to the next web page. To receive the data in different stages, the survey questions were organized into a series of five pages. The resulting files for each page were retrieved only by the survey authors, not by the survey sampling company. Survey Contents The purpose of the web-based survey is to collect information about crawfish consumption patterns. Once the respondents began the survey, responses were submitted when the respondent completed each page of the five-page survey. Interruption during completion had smaller impacts on completion rate than an Internet survey with only one long page. For example, a survey was not considered completed for the survey sampling company until all five pages were completed. However, we obtained data from respondents who completed less than all five pages of the survey. The designing of survey questions was the critical part for the success of the web-based survey. Our web-based survey can be accessed at http://www.agsurveys.org/food and the contents are included in the Appendix. After the respondents read the welcome page and gave their informed consent, they were asked to click the start button to begin the survey if they were the member of the household that usually decides what food to purchase. As we mentioned before, a time and date stamp recorded the beginning time to monitor how long the respondents took to complete the survey. Basic information, including the description, scientific names, and color graphics of crawfish, was provided on the second page. The first two questions asked respondents whether they could distinguish between crawfish, lobster, and langostino. Because crawfish are related to lobsters and closely resemble to them, it is

PAGE 30

20 believed that respondents might not perceive the difference. One reason to ask these questions was that we wanted to understand how knowledgeable people were about crawfish. Another reason was we wanted to make sure that, after reading the description, respondents could tell that, although related, crawfish are different from lobster and langostino, and to ascertain whether they were answering the survey for crawfish and not other products. The following two questions asked the respondent if they had ever eaten crawfish and, if so, if they had eaten crawfish in the last 12 months. A simple redirect was used to lead respondents who had not eaten crawfish in the last 12 months to the final page, while others proceeded to answer questions about crawfish consumption. The dependent variable in this study is the frequency of away-from-home consumption of crawfish. For each respondent who had consumed crawfish, we first questioned them on how many times they had eaten crawfish in the last 12 months. To obtain the dependent variable for this study, we then questioned them on how many times they had eaten crawfish both at-home and away-from-home in the last 12 months. For the crawfish consumers, questions were designed to obtain respondents perceptions and preferences on attributes such as farm-raised vs. wild caught crawfish and domestic vs. imported crawfish, respectively. Preferences on attributes regarding product forms were also obtained for away-from-home and at-home consumption. Main product forms of crawfish in this study include whole-boiled crawfish and crawfish tail meat, which are the dominant forms in the existing market. Whole-live crawfish were not covered because we were only concerned about away-from-home consumption and people seemed unknowledgeable

PAGE 31

21 about cooking crawfish at home according to our pre-survey study. Other attributes included frozen and fresh crawfish. In order to obtain price preferences on crawfish, we asked questions about how much they would pay for one pound whole-boiled crawfish and one pound crawfish tail. Typical prices were obtained through a pre-survey study, mainly from the local market and the web-based crawfish trader. Typical prices included whole-boiled crawfish (medium, three to five dollars per pound, and large, six to seven dollars per pound) and tail meat, 10 to 14 dollars per pound. The preferences on price might give insight on how much consumers would be willing to pay for one pound whole-boiled crawfish and one pound tail meat. Also, respondents were asked to select the top three reasons why they did not eat crawfish, including price. Other variables for eating crawfish included enjoyment of flavor and taste, enjoyment of texture, health and nutrition benefits, tradition and habit, prices or cost, availability, variety in diet, and prefer Cajun style. The final page of the survey (completed by both those who had consumed crawfish in the last 12 months and those who had not) included a question asking respondents to identify the top three reasons why crawfish was not eaten, or was not eaten more frequently. Variables for the top three reasons included prices, do not like the product form, no custom, do not like texture and taste, do not like the smell, no local farm-raised available, safety concerns, bad impression, health reasons (allergy), and vegetarian. The final part of the survey included socio-economic and demographic questions such as household income, household size, employment status, marital status, religions, and race. The information became extremely important in analysis of away-from-home consumption of crawfish, based on previous empirical studies.

PAGE 32

CHAPTER 4 DATA The data for this study were obtained through a web-based consumer survey. Targeting the Gulf region, the web-based survey was conducted in late January of 2004. Respondents were asked about the frequency of away-from-home crawfish consumption during the previous 12 months. Recall bias for the crawfish consumption is minimized because of the nature of the web-based survey. Crawfish consumption is expected to be seasonal. We asked for a 12-month recall period to accommodate the seasonality of crawfish consumption, given the low overall level of crawfish consumption in one year. Survey Sampling, Inc. provided a stratified survey sample for the survey. The chosen stratified sample was confined to the Southeast and the Gulf region. Different census regions were expected to be a significant determinant of both the participation and consumption decision on away-from-home consumption of crawfish. The survey was published on the website ( agsurveys.org/food ) in late January of 2004, with a total of 765 responses to the questions related to crawfish consumption. As mentioned before, compared to the mail survey instrument, the web-based survey is more effective and efficient in data collecting. Although the response rate for our survey was about 14.1 percent, we have a very high completion rate, 90 percent, which means, once the respondents started the survey, 90 percent would finish it. In this study, we used 733 effective records, which provided full information on crawfish consumption for this study. Cases with incomplete socio-economic and demographic information, such as income and gender, were dropped from the sample. In addition, responses with uncompleted 22

PAGE 33

23 answers and irrational answers were eliminated as well. The information obtained from these 733 responses is summarized below. We considered three census regions. For comparison, we divided the large Southeast Atlantic region into Northern and Southern areas (Table 4-1). The demographic data collected in this study indicated that the response rate per region in the Gulf region was relatively comparable, ranging from 129 effective responses from the East South Central region to 208 responses from the Southern Southeast Atlantic region (Table 4-1). The comparison to the percentage of population in these regions is included in the table. Table 4-1. Residence region of survey respondents. Region of Residence States Included in Region Number of Responses % of Survey Responses % of the Population in the region Northern Southeast Atlantic Delaware, Washington D.C., Maryland, North Carolina, Virginia, West Virginia 190 26.0 28.1 Southern Southeast Atlantic Florida, Georgia, South Carolina 208 28.4 23.5 East South Central Alabama, Kentucky, Mississippi, Tennessee, 129 17.6 17.0 West South Central Arkansas, Louisiana, Oklahoma, Texas 206 28.1 31.4 Compared with U.S. Census data ( US Census Bureau, 2000 ), the responses did appear biased towards Caucasians. A large percentage of the respondents in the survey, 88.2 percent, indicated they were Caucasian, followed by 8.0 percent Black or African-American, 1.5 percent Asian, 2.3 percent Indian, and 3.7percent other. These compare to the 2000 U.S. Census data, with approximately 75 percent of the U.S. population Caucasian, 12.3 percent Black or African-American, and 4.2 percent Asian. The results also indicated a bias towards non-Hispanic, with only 3.4 percent of the respondents indicating they were Hispanic or Latino. This compares to the Census data, with 12.5

PAGE 34

24 percent of Hispanic origin ( U.S. Census Bureau, 2000 ). However, this is not uncommon for survey responses on seafood consumption. Figure 4-1 shows the percentage of U.S. population and survey respondents in different age groups. Although discrepancies exist, survey respondents were somewhat comparable to the population average, with 46.4 percent of the U.S. population over the age of 45. Also, the survey respondents tended to have comparable household incomes to those from the U.S. Census data (Figure 4-2). The median income of the survey respondents fell in the $40,000 $59,999 category, compared to a U.S. median income of $41,994. However, our survey respondents were more educated than average, with approximately 40 percent of the sample having a four year college degree compared to 26 percent in the general population, according to the 2000 Census data. As a result, interpretation of the results should be conducted recognizing the bias, and future studies should make an attempt to focus on the underrepresented population (e.g. Hispanic origin). Overall, among the 733 responses, 58.4 percent (428 responses) indicated they had consumed crawfish, 30.0 percent (220 responses) indicated they had consumed crawfish in the last 12 months, and 25.8 percent (189 responses) indicated they had consumed crawfish away-from-home during the last 12-month period. Only 11 percent (82 responses) indicated at-home consumption, and 4.2 percent (31 responses) indicated consumed crawfish only at-home. Tables 4-2 and 4-3 provide descriptive statistics for the survey, including the respondents who had indicated crawfish consumption in the last 12-month period. Respondents were asked to indicate the times they have consumed crawfish at-home and

PAGE 35

25 away-from-home in the last 12 months. The frequency of away-from-home consumption of crawfish is the sum of responses to the two questions about at-restaurant and at-other-away-from-home consumption. Respondents who ate crawfish away-from-home consumed crawfish, on average, 2.97 times (standard deviation 2.682) in the last 12-month period. Additionally, other factors included socio-economic and demographic variables (age, ethnicity, household income, education, etc.), preferences for consuming or not consuming, and knowledge about crawfish. Descriptive statistics for the survey sample are shown in Tables 4-2 and 4-3.

PAGE 36

26 05101520253018-2425 to 3435 to 4445 to 5455 &abovePercent U.S. Census Survey Figure 4-1. Comparison of U.S. population in the Southeast and survey respondents by age. 05101520253035$19,999 orless$20,000 -$39,999$40,000 -$59,999$60,000 -$79,999$80,000 orgreaterPercent U.S. Census Survey Figure 4-2. Comparison of incomes of all the U.S. population and incomes of the survey respondents.

PAGE 37

27 Table 4-2. Summary of demographic information. Non-Consumers Consumers Non-Away-From-Home Consumers a Away-From-Home Consumers b Overall Sample Number of Observations 513 220 544 189 733 Age of Respondents % % % % % Greater than 55 19.9 19.4 19.9 19.5 19.7 Between 45 and 55 26.7 25.2 26.5 25.9 26.3 Between 35 and 45 23.2 27.9 24.3 25.9 24.7 Between 25 and 35 22.2 19.4 21.7 20.6 21.4 Between 18 and 25 8.0 8.1 7.2 7.9 7.7 Gender Percent Female 82.7 69.5 82.2 68.8 78.7 Household Income Less than $19,999 15.4 11.7 15.1 12.2 14.3 Between $20,000 and $39,999 29.8 28.2 30.3 26.4 29.4 Between $40,000 and $59,999 25.3 20.0 25.0 20.1 23.7 Between $60,000 and $79,999 15.0 14.5 15.1 14.3 14.8 $80,000 or greater 14.4 25.4 14.5 27.0 17.8 Region of Residence Northern Southeast Atlantic 32.0 11.8 30.9 11.6 26.0 Southern Southeast Atlantic 28.3 28.6 27.6 30.7 28.3 East South Central 18.1 16.4 18.8 14.3 17.6 West North Central 21.6 43.2 22.8 43.4 28.1 Education High School or less 45.8 37.3 46.0 35.4 43.2 Some College (2 Yr) 17.9 14.5 17.5 15.3 17.0 College Degree (4 Yr) 27.5 30.0 27.6 30.2 28.2 Above College Degree 8.8 18.2 9.0 19.0 11.6 Number of Children (under 16 living at home) No Children 61.8 57.7 60.8 59.8 60.6 1 or 2 Children 29.2 30.9 29.2 31.2 29.7 Above 3 Children 9.0 11.4 9.9 9.0 9.7 Household Size 1 Only 15.0 9.5 14.5 10.0 13.4 2 People 36.3 35.9 35.7 37.6 36.1 3 People 21.4 18.6 20.8 20.1 20.6 4 People 15.0 21.0 15.6 20.1 16.8 5 and above 12.3 15.0 13.4 12.2 13.1

PAGE 38

28 Table 4-2. Continued. Non-Consumers Consumers Non-Away-From-Home Consumers a Away-From-Home Consumers b Overall Sample % % % % % Race (races are alone and in combination) White 88.9 87.3 89.3 85.7 88.4 Black 8.8 5.9 8.5 6.3 7.9 Asian 1.2 1.8 1.1 2.1 1.4 Indian 2.0 3.2 2.0 3.2 2.3 Other Race 2.5 6.4 2.4 7.4 3.7 Religion Christian c 51.9 53.2 52.8 50.8 52.2 Catholic 12.7 11.4 12.3 12.2 12.3 Jewish 2.1 1.8 2.0 2.1 2.0 Other Religions 4.5 4.1 4.4 4.2 4.4 No Religion 28.9 29.5 28.5 30.7 29.1 Hispanic Descendent Hispanic 2.9 4.5 2.9 4.8 3.4 a. Non-Away-From-Home consumers included non-consumers and only at-home consumers. b. Away-From-Home consumers included only away-from-consumers and away-from-home consumers who also were at-home consumers of crawfish. c. Christian including Protestant.

PAGE 39

29 Table 4-3. Descriptive statistics on other factors included in the Double-Hurdle model. Mean, Non Away-From-Home (544) Mean, Away-From-Home (189) Overall Mean (733) Frequency of Consumption (times in the last 12 months): Away-From-Home 0 2.97 (2.682) a 0.77 (1.882) At-Home 0.18 (1.32) 0.72 (1.611) 0.32 (1.423) Knowledge of Crawfish % % % Knowing the difference between crawfish and lobster 73.5 89.5 77.7 Knowing the difference between crawfish and langostino 25.6 45.0 30.6 Indicated the following was one of the top three reasons for consuming Enjoy flavor 5.2 86.4 26.3 Health/Nutrition 0.9 14.1 4.4 Tradition/Habit 1.7 21.5 6.8 Price is attractive 1.1 14.1 4.5 Availability 1.3 23.6 7.1 Variety in diet 2.0 37.7 11.3 Prefer Cajun style 3.1 44.0 13.7 Indicated the following was one of the top three reasons for not consuming: Price too high 24.8 58.2 33.3 No preferred product form available 32.4 12.0 27.1 Not part of custom 20.8 20.0 20.5 Dislike taste 27.4 5.8 21.8 Dislike smell 31.1 16.9 27.5 Lack of domestically produced product 26.5 22.0 25.3 Product safety concerns 9.6 20.1 12.3 Bad impression 31.4 12.6 26.5 Health concerns 6.1 5.8 6.0 a Standard deviations were reported in the parenthesis.

PAGE 40

CHAPTER 5 THEORETICAL MODEL AND MODEL SPECIFICATION Theoretical Model Neoclassical demand theory suggests own price, income, the prices of substitutes and complements, socio-economic and demographic variables are determinants of demand. Eulers theorem for homogeneous demand functions shows ( Nicholson, 2002 ): 0,,,, IxPzxPyxPxxeeee where e x,Px is the own-price elasticity of demand, e x,Py and e x,Pz and all cross-price elasticities in-between are cross-price elasticities of demand with respect to all other goods, and e x,I is the income elasticity of demand. This property of the demand function will give insight into crawfish consumption analysis, although we did not calculate those elasticities in quantitative terms. In this study, the sensitivity of x (crawfish) to price (P x P y P z ) and income levels will be evaluated in general qualitative terms. However, consumption of substitute products (such as other seafood) was not available. With information about the sensitivity of x (crawfish) to own-price and income levels, we can infer a potential relationship between crawfish and its substitutes. In this study of crawfish consumption patterns, the main forms of crawfish consumed away-from-home are whole-boiled crawfish and tail meat. Different product forms may have different prices, which are difficult to aggregate. Moreover, it is hard to get aggregate prices because of the seasonality of the crawfish harvest and variety of consumption locations (at-home vs. away-from-home). The main purpose of this study is to develop an understanding of the factors influencing away-from-home consumption of 30

PAGE 41

31 crawfish, not to derive the demand function for crawfish. As a result, we did not include prices as a quantitative independent variable. However, in order to examine the impact price has on away-from-home consumption of crawfish and to examine the own-price elasticity in comparative terms, we include price as a binary variable. If respondents think prices are the main reason for their participation and/or consumption decision, the value of the variable would be one; otherwise, it would be zero. This study will examine the factors influencing away-from-home consumption of crawfish in the Gulf region using data from a web-based consumer survey conducted in late January of 2004. The survey instrument was discussed in the previous chapter. The frequency of crawfish consumed away-from-home in the last 12-month period is used as the dependent variable. Respondents were asked to indicate how many times they consumed crawfish away-from-home in the last 12 months. As mentioned before, the reasons for a 12-month period includes seasonality of crawfish consumption and overall low levels of consumption in one year. In this study, consumers, on average, ate crawfish away-from-home three times during the last 12 months. If asking for one month, the average times could be greater than, equal to or less than the annual average of three times. The frequency of at-home consumption of crawfish is included to check the relationship between at-home and away-from-home consumption and to infer if increases in away-from-home consumption would occur at the expense of at-home consumption of crawfish. Factors, including stated consumer preferences and perceptions regarding crawfish, also are included as explanatory variables. For example, the survey includes questions about the top three reasons for consuming or not consuming crawfish

PAGE 42

32 (respondents were asked to select the top three reasons from a list, including prices, availability, tradition, variety, flavor and taste, safety concerns, and health reasons). Moreover, socio-economic and demographic data, such as age, education level, income, gender, and ethnicity, will be examined to determine their impact on participation and consumption decisions. Censoring and truncation of the dependent variable is a very common problem in survey data. In previous seafood consumption studies ( Keithly, 1985 ; Lin and Milon, 1993 ; Yen and Huang, 1996 ; House et al, 2003 ), researchers found significant proportions of households with zero observations. Conventional regression methods fail to account for the qualitative difference between limit (zero) observations and non-limit (positive) observations. Early studies with limited dependent variables used the Tobit model ( Tobin, 1958 ). For example, Keithly ( 1985 ) used a Tobit model to accommodate the problem of zero consumption. The Tobit model specifies that y i =' x i + i i ~ N(0, 2 ) y i = 0 if y i 0, y i = y i if y i >0 where y i is the latent variable, and y i is the dependent variable. The model implies that y i will only be positive given a value of y i greater than zero. The Tobit model assumes the factors that affect the level of consumption are the same as those that determine the probability of consumption. See Amemiya ( 1984 ) for a detailed review. Although the Tobit model might be consistent with consumer behavior and might be useful in studying the consumption patterns of households, it is restrictive in parameter

PAGE 43

33 estimation because it is assumed that both the probability and the level of consumption are affected identically by the same factors. The Tobit model assumes that the variables and estimated parameters determine the consumption level as well as the participation probability. This might not be true. Respondents may have the potential to participate; however, other reasons might curb consumption. For example, safety issues might not inhibit shellfish consumption, but reports of contaminated shellfish products might do so temporarily. This implies that participation and consumption decisions might depend upon different explanatory variables and parameters, which might have opposite effects ( Lin and Milon, 1993 ). Model Specification The restriction of the Tobit model has been recognized in demand analysis for food and seafood, including finfish and shellfish, and has consistently been rejected in recent seafood studies. Cheng and Capps ( 1988 ), Lin and Milon ( 1993 ), Yen and Huang ( 1996 ), Drammeh et al. ( 2002 ), and House et al ( 2003 ) all realized the restrictions of using a Tobit model in demand analysis of seafood. They found that the decision to consume a seafood product has determinants that are independent of the level of consumption. For instance, income is expected to have opposite effects on participation and the consumption of crawfish. Based upon empirical results of previous studies, lower income households tended to have a higher probability to consume crawfish; however, lower income households tended to have lower levels of consumption than other crawfish consumers. To accommodate the restrictions of using a Tobit model in analysis for seafood consumption, a double-hurdle model proposed by Cragg ( 1971 ) has been used extensively in recent studies. In the Tobit model, a variable that increases the probability

PAGE 44

34 of an observation also increases the quantity of consumption simultaneously. In the model proposed by Cragg, the probability of an observation is independent of the regression model of the quantity consumed. Cheng and Capps ( 1988 ) used Heckmans two-step procedure, and Yen and Huang ( 1996 ) used a generalized double hurdle model to analyze household demand for finfish. Lin and Milon ( 1993 ) used a count-data double-hurdle model to examine the impacts of attributes and food safety perceptions of seafood consumption. Drammeh et al. ( 2002 ) and House et al. ( 2003 ) also used a double hurdle model similar to Yen and Huang (1996) to analyze U.S. oyster and catfish consumption, respectively. There are several different but related models in the mentioned literature; however, all are specified from Craggs double-hurdle model. For crawfish consumption, Yen et al. ( 1995 ) also found the complicating feature of household survey data, which is the significant proportion of zero observations. A large quantity of zero observations cannot be treated as true non-consumption because zero observations may be due to, besides non-consumption, infrequency of consumption or conscientious abstention, which could have other behavioral explanations ( Yen et al., 1995 ). As a result, the Tobit model also was rejected. The authors separated the sets of parameters in the probability of consumption and level of consumption equations. To allow for violation of the distribution assumption of Craggs double hurdle model, they used the Box-Cox transformation on the dependent variable, which is an extension of the double-hurdle model. We show the benchmark model below using the double-hurdle model framework specified by Cragg ( 1971 ):

PAGE 45

35 Individual i's participation equation can be expressed as d i = z i + v i with d i = 0 or 1 Individual i's consumption equation can be expressed as y i = x i + u i where y i represents the latent consumption decision, and d i is a latent variable describing participation. z i and x i are vectors of exogenous variables, and and are parameter vectors. Random errors u i and v i are normally distributed as N(0, 1) and N(0, 2 ), respectively. It also is assumed that u i and v i are independent. The double-hurdle model has separate participation and consumption equations that are related as: y i = y i if y i >0 and d i *>0 = 0 otherwise. where observed consumption y i relates to latent consumption y i *, only if y i >0 and d i *>0. It also is the conditional decision to consume the product (y i >0). The double-hurdle model includes two equations: the consumption equation y i and the participation equation d i *. Thus, the probability of consumption and level of consumption are determined by separate sets of parameters ( and ). The use of the same sets of variables, z i and x i in both equations is not restrictive because these variables affect consumption and participation differently through the different parameter coefficient ( and ), and might have opposite effects. The combination of the above model can be estimated in two parts. A Probit model can be used to estimate the participation equation, and the parameters of the consumption equation can be estimated independently using the truncated regression model.

PAGE 46

36 The maximum likelihood estimation of a Probit model is used to evaluate the censoring rule (z i ), given a normal distribution. Since we impose the assumption of a normal distribution, if y i *>0, then a truncated regression (i.e., truncated Tobit) applies. Thus, another maximum-likelihood estimation that accounts for a truncated normal distribution is used for the truncated regression. The likelihood function for the double hurdle can be derived as 0),|(),|1(Pr),|0(PrxzziiiiiiyfdobdobL where the 0 under the product sign indicates products over those observations which are zero, and the + signs under the product sign indicates products over those observations which are positive, y i >0. For simplicity, Prob denotes the probability, and f(y i |) is the conditional (truncated) density of y i given y i >0. and are vectors of parameters. As we indicated, the assumption of independence is imposed, with the participation and consumptions decisions as a binomial probability distribution and truncated at zero conditional distribution, respectively. The structure of the data determines the specific forms of the distributions. Because it is beyond the scope of this study, we will not cover the discussion in this study. As indicated above, the double hurdle model allows the parameters in the Probit equation to differ from those in the Tobit model, so that the complete model is a Probit model for d and a separate truncated regression model for the positive values of y. It is also assumed in this study that the participation and frequency of away-from-home crawfish consumption are independent of each other, and the probability and frequency of crawfish consumption are determined by separate sets of parameters. Since the Tobit

PAGE 47

37 log likelihood is simply the sum of the Probit and truncated regression log likelihood ( Greene, 1995 ), a test of the Tobit model as a restriction is implemented. The chi-squared specification test ( Greene, 1995 ) 2 = -2(f Tobit f Probit f Truncated ) is used to determine whether the double hurdle model is a better fit than the Tobit specification, where f i is the log-likelihood function value. It is testable with the likelihood ratio test by estimating the three implied models, with the specific degrees of freedom, which is equal to the number of variables in x. If the null hypothesis is rejected, we can conclude that the double hurdle model is a better fit than a Tobit model. The model was estimated by maximizing the logarithm of the likelihood function using a modified double hurdle model ( Cragg, 1971 ), with the frequency of away-from-home consumption of crawfish as the dependent variable. In practical terms, the Probit model and the truncated Tobit model as well as the compound Tobit model have been incorporated in many software packages (i.e., LIMDEP, SAS, and TSP). As a result, the estimation is now essentially on the level of ordinary regression, and will not be covered in this study. The marginal effects of the independent variables (similar to the Tobit model), on the probability of participation, and the conditional level of consumption can be derived by differentiating the probability and conditional mean, F(z)/x i and E(y*|y*>0)/x i where F(z) and E(y*|y*>0) are the probability of participation and the conditional mean of consumption, respectively, and x is the specific independent variable. The marginal effects on the unconditional mean of consumption, which can be associated with all observations, can be derived by E(y)/x, where E(y)=F(z)E(y*|y*>0). The marginal

PAGE 48

38 effects on the unconditional level evaluate what contributes to the consumption level by increasing or decreasing either the probability or conditional level. The marginal effects are computed by LIMDEP by computing the derivatives for each observation and averaging the derivatives ( Greene, 1995 ). Although the approach ignores the fact that some variables are binary, the results are almost exactly the same as the results of computation for binary variables using numerical methods ( Greene, 1999 ). In this study, the dependent variable is the frequency of away-from-home consumption of crawfish in the last 12-month period. Based upon the discrete nature of the dependent variable, it is a count data variable. Consequently, instead of using a continuous double hurdle model, a count data double hurdle model should be considered. In this study, a count data double hurdle model developed by Mullahy ( 1986 ), which is similar to Lin and Milons ( 1993 ) and Bilgic and Florkowskis ( 2003 ), is used. Truncated-at-Zero Count Data Double Hurdle Model Because the dependent variable, the frequency of away-from-home consumption of crawfish, is a count, our analysis adopted the count data double-hurdle model similar to Lin and Milons ( 1993 ) and Bilgic and Florkowskis ( 2003 ). A good review of the modified count data model for the hurdle model was developed by Mullahy ( 1986 ). According to Bilgic and Florkowski ( 2003 ), because of the possible overdispersion in the observations, the assumption of the familiar Poisson distribution truncated at-zero might not always fit the sample. However, in this study, we tested the overdispersion and compared the Poisson and negative binomial distributions of both zero and positive counts.

PAGE 49

39 Compared to the Poisson model, the negative binomial model is not encumbered by the Poissons mean=variance nature ( YEY var ) ( Mullahy, 1986 ). The variance of a negative binomial variant can exceed its mean, and the relationship between population mean and variance of the negative binomial distribution is defined as )1(varYEYEY For both the Poisson and negative binomial models, a hurdle (truncated-at-zero) specification is derived. See Mullahy ( 1986 ) for a detailed discussion. As the result of Mullahys modification, the binomial probabilities (the participation decision) are identical to those of a standard binomial Logit model. Thus, it is appropriate to test both Probit and Logit models for the binary probabilities. The consumption decision distributes as either a truncated-at-zero Poisson or a truncated-at-zero negative binomial model, depending on the overdispersion. Again, the specified model is estimated using standard maximum likelihood techniques, and ML estimates can be obtained by maximizing the log likelihood function. The compound model (single decision) and truncated-at-zero model have been incorporated in several software packages (i.e., LIMDEP). It is worth mentioning here that we included all the same determinants regarding specific perceptions and preferences, socio-economics, and demographics, since we have no basis to exclude any variable in either decision.

PAGE 50

CHAPTER 6 EMPIRICAL RESULTS Issues of Crawfish Consumption Regional Consumption Crawfish consumption tended to vary by region of residence in the targeted Southeast and the Gulf region (Figure 6-1). Visual inspection shows that consumers in the West South Central region (AR, LA, OK, TX) are most likely to consume crawfish both away-from-home and at-home (Figure 6-1). Overall, 39.8 percent of the respondents from the West South Central region consumed crawfish away-from-home, compared to 11.6 percent in the northern South Atlantic region (DE, DC, MD, NC, VA, WV). Chi-squared tests on significance are included below the table. Chi-square probabilities below 0.05 indicate a significant difference in the variables. In Figure 6-1, the chi-square probability of 0.001 indicates crawfish consumption is significantly different in the different regions. The result is consistent with our expectation, since the majority of crawfish produced in the United States is from the West South Central region where the consumption rate is expected to be higher than in other regions. The southern part of the South Atlantic region, including Florida, Georgia, and South Carolina, ranked second in crawfish consumption. Farm-Raised Crawfish Farm-raised crawfish production is mainly in the lower Gulf Coast region, such as Texas, Mississippi, Florida, and Louisiana, with most of the acreage found in Louisiana ( Avery and Lorio, 1999 ). In this study, crawfish consumers were asked if they were 40

PAGE 51

41 aware of whether the crawfish they had eaten was farm-raised or not. Only 29 percent of the crawfish consumers answered they were aware, but among those respondents, over 80 percent had eaten farm-raised crawfish in the previous 12 months. Among those respondents who had eaten farm-raised crawfish, over 94 percent indicated they would eat farm-raised crawfish again. Among the remaining crawfish consumers who had not eaten, or were not aware they had eaten, farm-raised crawfish, over 67 percent indicated they would be willing to consider consuming farm-raised crawfish. This implies that it might be appropriate to use educational approaches to promote the farm-raised crawfish and let the consumers or potential consumers have more knowledge about farm-raised crawfish. 05101520253035404550Northern SouthAtlanticSouthern SouthAtlanticEast South CentralWest South CentralPercent Away-From-Home Consumption Overall Consumption Note: Chi-square probability < 0.001. Figure 6-1. Regional percentage of crawfish consumption. Domestic Versus Imported Crawfish Wal-Mart in Louisiana carries crawfish produced in China, which even the most dedicated economists espousing comparative advantage had not forecast. Do consumers

PAGE 52

42 really like imported crawfish over the domestic crawfish? As a special interest of this study, crawfish consumers were also asked to indicate whether they prefer domestically produced or imported crawfish. The survey indicates that although 58 percent of the crawfish consumers did not know where the crawfish they had consumed came from, about 52 percent of the consumers indicated they preferred locally produced and only six percent indicated they preferred imported crawfish. A relatively large percentage of consumers had no opinion, meaning they either do not know or do not care (Figure 6-2). This implies that properly promoted domestically produced crawfish, targeting those crawfish consumers, might win back the market share lost to the imported product. 0102030405060Preferred Not PreferredDo Not Care Do Not KnowPercent Domestic Crawfish Imported Crawfish Figure 6-2. Comparison of preferences over domestic and imported crawfish products. Product Form Since most of the crawfish imported are in the form of tail meat, and domestically produced tail meat is usually high in price, the product form might explain the reason why imported crawfish have an increasing market share. In the survey, crawfish

PAGE 53

43 consumers were asked to identify what form of crawfish they prefer: whole crawfish or tail meat? Live whole crawfish and boiled whole crawfish as well as tail meat are the main forms of crawfish products normally found on the market. However, in order to simplify the question and focus on away-from-home consumption, we included only whole and tail meat product form in this study. The majority of existing crawfish consumers (72%) indicated that they preferred tail meat to whole crawfish in away-from-home consumption. Interestingly, although the price of tail meat is comparatively higher than the price of whole crawfish, a large percentage of consumers (71%) preferred tail meat to whole crawfish in at-home consumption. This might be caused by the trend of consumer preferences towards convenience. Also, this implies that the introduction of new crawfish products targeting consumer trends, such as convenience and health concerns, could encourage consumption or more consumption to some extent. Domestic crawfish producers and processors can design marketing strategies accordingly to compete with imported products. Additionally, existing crawfish consumers were asked if they preferred fresh or frozen crawfish. The majority (75%) indicated that they preferred fresh crawfish products, with only one percent preferring the frozen form and 24 percent having no preference. This is consistent with the major consumer trend of consuming fresh products. This implies that domestically produced fresh crawfish can be competitive with imported frozen products if prices are reasonable. However, it is believed that existing fresh crawfish products are far more expensive compared to those imported products.

PAGE 54

44 Reasons for Consuming and Not-Consuming Reasons for consuming In addition to the frequency of away-from-home and at-home consumption and the demographic variables, respondents were asked to identify their main reasons for consuming crawfish. Results from the 220 crawfish consumers (including the 189 away-from-home crawfish consumers) who responded to this question are illustrated in Figure 6-3. As indicated by about 87 percent of the crawfish consumers and the away-from-home crawfish consumers, the principal reason for consuming crawfish was enjoyment of flavor. Next was preference for Cajun style, followed by adding variety to diet (Figure 6-3). It is worth mentioning that there was no statistically significant difference between away-from-home consumers and overall crawfish consumers (Chi-square probabilities ranging from 0.2817 to 0.8813). Among the 220 crawfish consumers, 138 respondents indicated they only consumed crawfish away-from-home, 31 indicated they only consumed at-home, and 51 indicated they consumed both away-from-home and at-home. Results from these three types of consumers who responded to this question are illustrated in Figure 6-4. As indicated, although the percentage levels are somewhat different for all three types of consumers, the principal reason for consuming crawfish was enjoyment of flavor. Next was preference for Cajun style, followed by adding variety to diet (Figure 6-4). Reasons for not consuming Both consumers and non-consumers of crawfish were asked to identify the top three reasons for either not consuming or not consuming more crawfish (Figure 6-5). For non-consumers, lack of availability of the preferred product form, bad impression, and smell were the top three reasons for not consuming; however, taste also was ranked high

PAGE 55

45 0102030405060708090100Enjoy FlavorPrefer CajunStyleVariety in DietAvailabilityTradition/HabitEnjoy TexturePriceHealth/NutritionPercent Away-From-Home Consumption Overall Consumption Note: Chi-square probabilities ranging from 0.2817 to 0.8813 Figure 6-3. Categories of reasons given for consuming crawfish (crawfish consumers vs. away-from-home consumers). 0102030405060708090100Enjoy FlavorPrefer CajunStyleVariety in DietAvailabilityTradition/HabitEnjoy TexturePriceHealth/NutritionPercent Away-From-Home Only At-Home Only All Crawfish Consumers Note: Chi-square probabilities between AFH-Only and AH-Only ranging from 0.1707 to 0.9787 Figure 6-4. Categories of reasons given for consuming crawfish (away-from-home consumers only, at-home consumers only, and all crawfish consumers).

PAGE 56

46 in the reasons for not consuming crawfish. In contrast, crawfish consumers gave significantly different responses to the questions (Chi-square probabilities < 0.0001). Price is too high, followed by lack of availability of domestic products and custom were the top three reasons for not consuming among existing crawfish consumers. 010203040506070PriceDomestic notAvailableNo CustomProduct Safety SmellBad ImpressionPreferred Formnot AvailableTasteHealth/NutritionPercent Crawfish Consumers Non-Consumers Note: Chi-square probabilities < 0.0001 Figure 6-5. Categories of reasons given for not consuming or not consuming more crawfish (consumers vs. non-consumers). We also illustrate the top three reasons for not consuming for away-from-home consumers and non-away-from-home consumers separately. For non-away-from-home crawfish consumers, unavailability of product form, bad impression, and smell also were the top three reasons for not consuming. Taste also was ranked high in the reasons for not consuming crawfish by those non-away-from-home consumers. Away-from-home consumers gave different responses towards the question. Price is too high, followed by unavailability of domestic products and safety concerns were the top three reasons for not

PAGE 57

47 consuming by existing away-from-home consumers. No custom or tradition became the fourth important reason for not consuming crawfish. It is worth testing the relationship of reasons for not consuming among non-crawfish consumers, only away-from-home consumers, and only at-home consumers (Figure 6-6). For non-crawfish consumers, unavailability of product form, bad impression, and smell were the top three reasons for not consuming. Taste also was ranked high in the reasons of not consuming crawfish. Consumers who only consumed crawfish away-from-home gave different responses towards the question. Price is too high, followed by unavailability of domestic products and no custom were the top three reasons. Safety concerns became the fourth important reason for not consuming crawfish among away-from-home consumers. Interestingly, consumers who only consumed crawfish at-home gave somewhat different responses to the question. Price is too high and unavailability of domestic products ranked much higher than the other reasons. No custom and smell ranked third and fourth. However, safety concerns was not as important for at-home consumers, who might have been influenced by the fact that they prepare the crawfish themselves. Away-From-Home Crawfish Consumption Consumers were asked to identify how many times they had consumed crawfish both at-home and away-from-home in the previous 12-month period. However, to focus on away-from-home consumption, we treat the frequency of away-from-home consumption as the dependent variable in the count data double hurdle model. Because of its discrete nature, we reported the sample frequency distribution of the dependent variable (FREAFH) in Table 6-1. The table shows the number of times a consumer consumed crawfish away-from-home in the previous 12-month period. Average

PAGE 58

48 consumption of the overall sample (733 responses) was 0.8 times; average consumption of the 189 away-from-home crawfish consumers was 3.0 times. 010203040506070PriceDomestic notAvailableNo CustomProduct Safety SmellBad ImpressionPreferred Formnot AvailableTasteHealth/NutritionPercent AFH-Only AH-Only Non-Consumers Note: Chi-square probabilities between AFH-Only and AH-Only ranging from 0.1223 to 0.9924. Figure 6-6. Categories of reasons given for not consuming or not consuming more crawfish (away-from-home only, at-home only, and non crawfish consumers). Table 6-1. Sample frequency distribution of the dependent variable (n=733). Away-From-Home Consumption (times) Number of Consumers (n=733) % 0 1 2 3 4 5 6 7 8 10 12 15 544 54 63 24 14 18 3 1 1 5 3 3 74.2 7.4 8.6 3.3 1.9 2.5 0.4 0.1 0.1 0.7 0.4 0.4 Note: Overall 733 respondents, values indicate the times of away-from-home crawfish consumption in the previous 12 months period.

PAGE 59

49 The following model was estimated using a count data double hurdle model developed by Mullahy ( 1986 ), with the frequency of away-from-home consumption of crawfish as the dependent variable. The dependent and independent variables are described in Table 6-2. As mentioned in the previous chapter, an interesting and distinctive feature of the hurdle model is that the probability of participation is assumed to be independent of the level of consumption. The count data double hurdle model consists of two equations describing participation and consumption decisions separately. In this study, the participation equation depicts the decision of whether to be an away-from-home crawfish consumer, and the consumption equation refers to how many times to consume as an away-from-home crawfish consumer. For the participation decision stage, we used a Logit model to estimate the probabilities of being an away-from-home crawfish consumer. The coefficients from the Logit model (participation decision) as well as the marginal effects (calculated at mean value) are reported in Table 6-3. The Logit model correctly predicts a consumers likelihood of being or not being an away-from-home crawfish consumers 93.7 percent of the time (incorrectly predicted being a consumer 3.4 percent of the time and not being a consumer 2.9 percent of the time). This can be compared to a nave prediction, which would result in correctly predicting away-from-home crawfish consumption 74 percent of the time. The prediction rate supports the appropriateness of using a Logit model (with PseudoR 2 =0.67). In the second stage, the dependent variable is a count variable for the number of times crawfish was consumed away-from-home in the previous 12 months. We have

PAGE 60

50 Table 6-2. Description of variables included in the double-hurdle model. Variants VariableName Description FREAFH Frequency of crawfish consumption away-from-home Consumption of Crawfish FREAH Frequency of crawfish consumption at-home Knowledge of Crawfish KNLAN 1 if respondent indicated being aware of the difference between crawfish and langostino Reasons for consuming crawfish The following variables are 1 if this reason was listed as one of the top three reasons for consuming crawfish: FLAVOR Enjoy flavor HEALTH Health/nutrition TRAHABIT Tradition/habit YESPRICE Price AVAILTY Availability FARMRSD Farm raised crawfish VDIET Variety in diet CAJUN Prefer Cajun Style Reasons for not consuming, or not consuming crawfish more frequently The following variables are 1 if this reason was listed as one of the top three reasons for NOT consuming, or not consuming MORE crawfish: NOPRICE Price NOFORM Lack of preferred product form NOCUSTOM Custom NOTASTE Disliketaste NOSMELL Dislikesmell NOLOCAL Lack of availability of domestically produced products NOSAFE Product safety concerns NOIMPR Badimpression NOHEALTH Health concerns (allergy) Region of residence (Gulf region) REGION1 Northern Southeast Atlantic (DE, DC, MD, NC, VA, WV) REGION2 Southern Southeast Atlantic (FL, GA, SC) REGION3 East South Central (AL, KY, MS, TN) REGION4 West South Central (AR, LA, OK, TX) (omitted category)

PAGE 61

Table 6-2. Continued. 51 Variants VariableName Description Race/Ethnicity WHITE 1 if Caucasian, 0 otherwise BLACK 1 if Black, 0 otherwise (omitted category) ASIAN 1 if Asian, 0 otherwise (omitted category) INDIAN 1 if Indian, 0 otherwise (omitted category) OTHRACE Other race/ethnicity (omitted category) INCOME1 <$19,999 (omitted category) INCOME2 $20,000 $39,999 INCOME3 $40,000 $59,999 INCOME4 $60,000 $79,999 Income INCOME5 $80,000 or above EDUCAT1 High School degree or less (omitted category) EDUCAT2 2 year college degrees EDUCAT3 Degree from college (4 year) Education EDUCAT4 Above college degree Age AGE1 Ages 18-24 (omitted category) AGE2 Ages25-35 AGE3 Ages35-45 AGE4 Ages45-55 AGE5 Age 55 or older Employment Status EMPLOY1 1 if currently full time employed EMPLOY2 1 if currently part time employed (omitted category) EMPLOY3 1 if currently unemployed (omitted category) EMPLOY4 1 if student (omitted category) EMPLOY5 1 if house worker without pay (omitted category) EMPLOY6 1 if retired (omitted category)

PAGE 62

52 considered both truncated-at-zero Poisson and truncated-at-zero Negative Binomial models for estimation. However, after we tested the Poisson process, the overdispersion test indicated that the assumption of the Poisson distribution truncated-at-zero could not be rejected and might fit the sample. 1 Therefore, we report the results from the truncated-at-zero Poisson model. Both coefficient estimates and the marginal effects are reported in Table 6-3 as well as the estimates from a Logit model. To test the compound (single-decision) count data model (standard Poisson) against the hurdle count model, we use Vuongs t-test. The standard normal statistics were calculated. For Vuongs t-test, a value greater than 1.96 favors the altered model. In this case, the value of 5.66 (given by LIMDEP) was obtained for the null hypothesis that the standard Poisson and its hurdle Poisson modification are the same. In other words, Vuongs statistic favors the hurdle Poisson model. The result shows that splitting the count model is preferred to the standard count model. We also included the single decision (standard Poisson) estimates in Table 6-3 for comparison. As expected, results indicated that variables affected the decision to consume crawfish away-from-home and the frequency of consumption differently. If consumers ate crawfish at-home more frequently, they were significantly less likely to be away-from-home crawfish consumers, but their frequency of away-from-home consumption increased. In other words, as the frequency of at-home consumption increases, the likelihood to consume crawfish away-from-home decreases. A possible reason for the 1 Overdispersion test: H0: var[yi]=i ; H1: var[yi]=i+g(i) with g(i)=i. t-statistic=1.453<1.96; the result could not reject the null hypothesis at 5% level. See LIMDEP Manual ( Greene, 1995 ) for the overdispersion test in Poisson regression.

PAGE 63

53 Table 6-3. Empirical results of single-decision and double-hurdle: maximum-likelihood estimates and marginal effects. Single Decision Double Hurdle Count Data Variable Standard Poisson Logit Coefficient Marginal Effects Truncated Poisson Coefficient Marginal Effects Frequency of Consumption: Crawfish at-home -0.032*** (0.020) a -0.412* b (0.125) -0.040* 0.052** (0.028) 0.084** Knowledge about crawfish: Crawfish and langostino 0.322* (0.094) 0.254 (0.357) 0.025 0.339* (0.116) 0.547* Indicated the following was one of the top three reasons for consuming: Enjoy flavor 2.874* (0.180) 3.352* (0.457) 0.327* 0.559* (0.198) 0.902* Health/Nutrition 0.234*** (0.134) 1.241*** (0.736) 0.121*** 0.189 (0.152) 0.305 Tradition/Habit 0.224** (0.115) 0.851 (0.594) 0.083 0.131 (0.129) 0.212 Price is attractive 0.184 (0.137) 1.522** (0.787) 0.149** -0.158 (0.164) -0.255 Availability 0.201*** (0.114) 1.552* (0.607) 0.152* -0.049 (0.135) -0.079 Farm-raised available -0.232 (0.210) 0.195 (0.915) 0.019 -0.123 (0.240) -0.020 Variety in diet -0.033 (0.105) 2.135* (0.507) 0.208* -0.306* (0.124) -0.494* Prefer Cajun cooking 0.249* (0.097) 2.078* (0.498) 0.202* 0.055 (0.109) 0.088 Indicated the following was one of the top three reasons for not consuming: Price too high -0.241* (0.096) 0.169 (0.371) 0.017 -0.356* (0.111) -0.574* No preferred product form available -0.159 (0.151) -0.603 (0.466) -0.059 -0.142 (0.186) -0.230 Not part of custom 0.092 (0.125) -0.287 (0.428) -0.028 0.091 (0.150) 0.147 Taste 0.023 (0.225) -0.246 (0.566) -0.024 0.276 (0.274) 0.446 Smell -0.198 (0.141) -1.145* (0.478) -0.111* -0.114 (0.176) -0.185 No domestic products available -0.137 (0.116) -1.141* (0.408) -0.111* -0.067 (0.135) -0.108 Product safety concerns 0.089 (0.127) 0.243 (0.510) 0.024 0.069 (0.149) 0.111 Bad impression -0.365* (0.149) -0.910** (0.484) -0.089** -0.289*** (0.184) -0.467*** Health concerns 0.563* (0.180) 0.790 (0.589) 0.077 0.344*** (0.207) 0.556***

PAGE 64

54 Table 6-3. Continued. Single Decision Double Hurdle Count Data Variable Standard Poisson Logit Coefficient Marginal Effects Truncated Poisson Coefficient Marginal Effects Demographics Caucasian -0.423* (0.131) -1.090* (0.431) -0.106* -0.489* (0.145) -0.789* 2 year College degree 0.056 (0.147) 0.350 (0.470) 0.034 -0.129 (0.175) -0.208 4 year College degree -0.008 (0.118) -0.067 (0.430) -0.007 -0.017 (0.143) -0.027 Postgraduate 0.203 (0.137) 0.499 (0.608) 0.049 0.114 (0.162) 0.184 Between 18 and 24 -0.142 (0.226) -0.163 (0.651) -0.016 -0.402 (0.279) -0.649 Between 25 and 34 -0.335** (0.163) -0.741 (0.531) -0.072 -0.511* (0.193) -0.824* Between 35 and 44 0.055 (0.143) -1.289* (0.518) -0.126* 0.175 (0.162) 0.282 Between 45 and 54 0.056 (0.138) -1.022** (0.511) -0.100** 0.146 (0.157) 0.236 Northern Southeast Atlantic -0.572* (0.180) -0.936** (0.493) -0.091** -0.545* (0.229) -0.879* Southern Southeast Atlantic 0.131 (0.112) -0.218 (0.394) -0.021 0.063 (0.128) 0.102 East South Central 0.201 (0.127) -0.596 (0.523) -0.058 0.248*** (0.150) 0.340*** $20,000 $39,999 -0.540* (0.161) -1.332* (0.495) -0.130* -0.648* (0.192) -1.045* $40,000 $59,999 -0.643* (0.178) -1.125** (0.551) -0.110** -0.675* (0.209) -1.089* $60,000 $79,999 -0.456* (0.173) -1.256** (0.607) -0.123** -0.349*** (0.202) -0.564*** $80,000 and above -0.703* (0.173) -0.665 (0.619) -0.065 -0.856* (0.209) -1.382* Full time employed 0.411* (0.105) 0.468 (0.374) 0.046 0.425* (0.125) 0.687* Constant -1.511* (0.268) ----1.264* (0.332) 2.040* Log-Likelihood -599.79 -137.95 -340.45 Correct prediction in Logit (%) 93.7 Vuongs statistic 5.66 a Standard errors of the coefficients are reported in parentheses. b One, two, and three asterisks indicate significance at the 0.01, 0.05 and 0.10 levels, respectively.

PAGE 65

55 result is that at-home crawfish consumers might be dedicated crawfish consumers, and they might have preparation knowledge. Compared to eating out, at-home consumption could be cheaper, and thus made them prefer at-home consumption. However, due to the fact that they might be dedicated and more frequent crawfish consumers than those who usually ate crawfish at restaurants, their frequency of away-from-home consumption increases as their frequency of at-home consumption increases. Our results indicated that for each one-unit increase in at-home consumption of crawfish (one unit equals one time in the last 12 months), the respondents were four percent less likely to be away-from-home crawfish consumers. The average crawfish consumer ate crawfish away-from-home 2.97 times in the 12-month period. For each one-unit increase in at-home consumption, respondents increased away-from-home consumption significantly by 0.08 to 3.05 times. We did not include the test of the relationship between crawfish consumption and other seafood consumption (consumption of substitutes) in this study. Further research could be done to see whether an increase in crawfish consumption will occur at the expense of other seafood consumption and whether an increase in away-from-home crawfish consumption will occur at the expense of at-home consumption of other seafood products. However, in this study we can infer a potential relationship between crawfish and its substitutes from the information about the sensitivities of crawfish to own-price and income levels, and the discussion will be covered in the next chapter. Consumers who know the difference between crawfish and langostino (more knowledgeable about crawfish) consumed crawfish away-from-home significantly more frequently. Nevertheless, it did not affect the likelihood of being a crawfish consumer.

PAGE 66

56 Knowledgeable respondents consumed crawfish 0.55 times more often away-from-home (3.52 times/year). This implies that using educational programs to promote crawfish could increase away-from-home consumption, but not persuade non-consumers to consume crawfish. Variables representing the respondents top three reasons for consuming crawfish as well as the top three reasons for not consuming were included in the model and provided insight for preferences on away-from-home consumption of crawfish. The results indicated that those reasons significantly affect both the likelihood to be an away-from-home consumer and the frequency of away-from-home consumption. Our results indicated that consumers were more likely to consume crawfish away-from-home (15 percent more likely) if they selected price attractiveness or availability as reasons for consumption. Price does matter in crawfish consumption, and this result infers that crawfish consumption could be comparatively price sensitive. Consumers were 20 percent more likely to consume crawfish away-from-home if they selected prefer Cajun cooking as the reason for consuming, but it was insignificant in determining how frequently a consumer ate crawfish. If respondents indicated they consumed crawfish because they enjoyed the flavor or for health/nutritional reasons, as expected, they were 33 percent and 12 percent respectively more likely to be an away-from-home crawfish consumer. Frequency of consumption away-from-home was significantly affected by different variables or in different directions. For example, consumers indicating they consumed crawfish to add variety to their diet were likely to consume crawfish away-from-home less frequently. It is interesting to note that if consumers chose crawfish to add variety to

PAGE 67

57 their diet, although they were more likely (20 percent) to be away-from-home consumer, it decreased the frequency of consumption (0.5 times less than average). This suggests that someone interested in adding variety to the diet may eat crawfish, but on a less frequent basis. Frequency of away-from-home crawfish consumption also was influenced by the variable representing flavor, implying that consumers who selected crawfish because they really enjoyed the flavor were more likely to consume crawfish more frequently than consumers who rated other reasons as important. In numeric terms, respondents who chose flavor as the reason for consuming were 33 percent more likely to consume crawfish away-from-home, and were likely to consume 0.90 times more. Additionally, consumers were more likely to consume crawfish away-from-home more frequently if they indicated health/nutrition as the reason for consuming (0.3 times more). Variables representing farm-raised crawfish and tradition/habit were not significant in determining either participation or consumption decisions for away-from-home crawfish consumption. Respondents also were asked to identify the top three reasons why they did not consume or did not consume more crawfish. Probability of away-from-home consumption of crawfish decreased if the respondents indicated they did not consume crawfish due to smell or bad impression. However, those two reasons did not influence the frequency of consumption for existing consumers. Respondents who believed domestic product unavailability was one of the reasons for not consuming or not consuming more crawfish were less likely to consume crawfish away-from-home. The magnitude of these impacts was similar; respondents were 11 percent less likely to

PAGE 68

58 consume crawfish away-from-home because of smell, and 11 percent less likely to consume because of bad impression. Similarly, consumers were nine percent less likely to eat crawfish away-from-home because domestically produced crawfish were not available. Frequency of away-from-home crawfish consumption was not significantly impacted by most variables in this category. However, away-from-home crawfish consumption decreased by 0.57 times if respondents believed price was too high and 0.56 times if respondents had health concerns. The relationship between perceived price unattractiveness and the frequency of crawfish consumption infers again that consumption of crawfish is relatively price sensitive. When respondents believe price is attractive, they are more likely to consume crawfish. On the contrary, when consumers believe the price is high, they dramatically decrease the frequency of away-from-home consumption by 0.57 times, a 19.2 percent drop. As a result, we assert that crawfish consumption is relatively price sensitive. Additionally, respondents who had a bad impression of crawfish affected both the probability and the level of away-from-home crawfish consumption significantly. If the respondents indicated not consuming due to bad impression, they were nine percent less likely to consume crawfish away-from-home, and consumed 0.47 times less away-from-home. Perhaps the most interesting result is that product safety concerns did not influence either the decision to consume crawfish or the frequency of away-from-home consumption significantly. This implies that although crawfish scored a bad impression, it was not due to safety concerns by existing and potential consumers. This result differs

PAGE 69

59 from the pattern of crawfish consumption in other countries such as China. Generally, Chinese people are more concerned with safety issues due to environmental pollution in the crawfish production areas. Demographics had an effect on both the decision and the frequency of away-from-home consumption. Region significantly impacted the probability and the frequency of away-from-home crawfish consumption. In general, consumers in the northern part of the Southeast Atlantic region were significantly less likely to consume crawfish compared to consumers in the West South Central region. Although not significantly, consumers in the southern part of the Southeast Atlantic region and the East South Central region seemed less likely to consume crawfish compared to the consumers in the West South Central region. This result is consistent with our expectation. Since the West South Central region is the main crawfish production area in the United States, consumers in this region tend to have a higher likelihood of consuming crawfish. Additionally, consumers in the northern part of the Southeast Atlantic region were significantly more likely to consume crawfish away-from-home on a less frequent basis (0.88 times less than the West South Central region). However, it is interesting to note that consumers in the East South Central region (Alabama, Kentucky, Mississippi, and Tennessee) consumed crawfish away-from-home significantly more frequently than those in the West South Central region. This implies that although consumers in the East South Central were less likely to consume crawfish, once they became crawfish consumers, they consumed more than consumers in the West South Central region. This could indicate a new target market for the crawfish industry.

PAGE 70

60 One disturbing piece of information delivered by the results is that Caucasian consumers were significantly less likely to consume crawfish and to consume crawfish away-from-home less frequently. It is worth noting that education levels did not have a significant impact on either the participation or the consumption decisions in this study. However, education levels did have a significant impact on crawfish consumption according to several previous crawfish consumption studies, and lower educated groups seemed more likely to consume crawfish and to consume it more frequently. Our survey had quite a representative sample in terms of education level, and probably was better than those in other similar seafood and crawfish consumption studies. Further research with more representative samples might explain the impact of education on crawfish consumption. In our sample, we found that the correlation between education and income was 0.32, but not significant. Also, we found income levels had significant impacts on both participation and consumption decisions, which partly explains the lack of impact by education levels. In general, our result indicated that most income groups above the base group, $20,000-or-less, were about 12 percent less likely to consume crawfish away-from-home. All income groups above the base group consumed crawfish away-from-home less frequently, ranging from 0.56 times less (a 18.9 percent drop) to 1.38 time less (a 46.5 percent drop). This result infers that crawfish consumption is sensitive to income levels and that crawfish is an inferior good. The lowest income groups have the highest likelihood to consume crawfish away-from-home and to consume more frequently than other income groups. People with higher incomes tend to consume less crawfish away-from-home and to consume less frequently.

PAGE 71

61 Age did significantly influence away-from-home crawfish consumption to some extent. Compared to the oldest age group (age 55 and above), middle age groups (between 35 and 44, between 45 and 54) were significantly less likely to consume crawfish away-from-home. The younger age group (between 25 and 34) consumed crawfish away-from-home less frequently, and their probability of consuming also decreased. Given that away-from-home consumption might be the main means of introducing crawfish to potential consumers and that at-home consumption, consumers need more preparation knowledge, it is important to understand the factors influencing away-from-home crawfish consumption.

PAGE 72

CHAPTER 7 CONCLUSIONS AND IMPLICATIONS Summary This thesis focuses on developing an understanding of the factors that influence away-from-home consumption of crawfish in the Gulf region. The main objective was to analyze the significant factors that influence the probability and frequency of crawfish consumption, particularly for away-from-home consumption. The information then could be used by crawfish producers, processors, and marketers to develop market strategies for growth in the crawfish industry. Also, market trends could be evaluated, and the crawfish industry could follow the trends found in this study to make marketing decisions accordingly. Results showed that the factors that influenced the decision to consume crawfish away-from-home differed from the factors that influenced the decision of how often to consume crawfish away-from-home. This is consistent with the results of previous studies of crawfish and other seafood consumption. The crawfish industry can use this information to target potential consumers and existing consumers separately. Keep in mind that one marketing strategy might not work on both groups. One potential concern about increasing away-from-home consumption is that it might occur at the expense of at -home consumption. This study found that as the frequency of at-home consumption increases, the likelihood to consume crawfish away-from-home decreases. However, as consumers frequency of at-home crawfish 62

PAGE 73

63 consumption increases, their frequency of away-from-home crawfish consumption increases as well. Price is an important determinant in crawfish consumption. Consumers regarded price attractiveness as a reason for consuming crawfish, and those consumers were more likely to be away-from-home consumers of crawfish. Respondents also indicated that price unattractiveness is the main reason for not consuming or not consuming more crawfish. Consumers who indicated this dramatically decreased their frequency of away-from-home crawfish consumption, by 0.57 times, or a decrease of 19.2 percent of the overall level of consumption in one year. Although quantitative price elasticity was not available, these results infer that crawfish consumption is relatively price sensitive and that increasing the price of crawfish will decrease the frequency of away-from-home crawfish consumption. On the other hand, decreasing the price slightly might encourage consumption and increase the total revenue for the industry. The differences in consumer perceptions of crawfish can also be seen through the relationships of variables representing reasons for consuming and not consuming the product. Interestingly, product safety concerns did not impact consumers decisions or the frequency of consumption. However, having a bad impression of crawfish (i.e., mud bug) did decrease both the likelihood of being a consumer and the frequency of away-from-home consumption. This result implies, rather than working on improving crawfish safety or perception of crawfish safety issues, the industry should work on promoting crawfish as a farm-raised seafood product. For example, instead of calling it mud bug as it is traditionally called, the industry could call it miniature lobster in their promotion campaign. There is a lack of knowledge about the difference between

PAGE 74

64 crawfish, lobster, and langostino. For instance, of the survey respondents, 70 percent indicated they do not know the difference between crawfish and langostino, and 22 percent indicated they do not know the difference between crawfish and lobster. Therefore it might be easy to reposition the image of crawfish in potential consumers mind. Our results also indicated that if consumers are more familiar with crawfish, they would consume more frequently. As a result, work by the industry to increase knowledge about the product, such as combined demonstrations and educational advertising, could be successful in increasing the demand for away-from-home consumption. Compared to average consumers, consumers who ate crawfish to add variety to their diet were more likely to consume crawfish away-from-home, and they were more likely to consume crawfish away from home less frequently (0.5 times less per year). In other words, if consumers choose crawfish to add variety to their diet, they eat crawfish, but less frequently. Of particular use to the industry are the results surrounding the variables reflecting availability. Consumers who knew that crawfish were available were more likely to consume crawfish. On the contrary, consumers who did not know domestic products were available were less likely to consume crawfish. Because of the perceived small size of the crawfish market, not many supermarket chains carry crawfish, and live crawfish have to be ordered online from Louisiana processors. This implies that the small market size is not only caused by the lack of demand, but also by the distribution channel, and the latter one might play a even more important role in the consumption of crawfish. It

PAGE 75

65 appears that a broader distribution channel and more availability might encourage consumption. Additionally, smell was an issue for crawfish consumption, which decreased the likelihood to consume crawfish by 11 percent. Similar to other seafood products, a fishy smell is always a big concern. Innovation in the industry would include pre-cleaned, pre-processed crawfish products such as ready-to-cook tail meat. In general, strategies to overcome the perceived smell issues could increase demand. This might work, particularly for at-home consumption. As with other seafood products, the impact of smell should not be ignored, although the magnitude of the impact on away-from-home consumption is less than the impact on at-home consumption. Persistent smell issues are related to bad impression, which could seriously impact both participation and consumption decisions. Relationships seen in the demographics are generally consistent with a priori expectations with the exception of education level. The relationships between income levels and away-from-home consumption tell an important story. The lowest income group, $20,000 and under, has the highest likelihood and frequency of crawfish consumption. As income increases, people tend to be less likely to consume crawfish away-from-home and to consume it less frequently. For example, the highest income group, $80,000 and above, consumed crawfish away-from-home 1.38 times less frequently than the lowest income group, a decrease of 46.5 percent. This result infers that crawfish is an inferior good, although a quantitative income elasticity is not available. This is not the result the industry wants to see, but the information can help the industry to further stratify consumer segments and to work on segments effectively. How to

PAGE 76

66 transfer crawfish from an inferior good to a normal good becomes a must for the industry if they want to encourage consumption of crawfish among higher income groups. Strategies such as differentiating by value adding characteristics to processed crawfish products could be successful. Additionally, the information on inferior goods and the sensitivity of price and income levels on crawfish can infer a potential relationship between crawfish and its substitutes using Eulers theorem (e x,Px + e x,Py ++ e x,Pz + e x,I =0). Although the price and income elasticities are not calculated, the empirical result makes an assertion that crawfish tend to have many substitutes and/or is sensitive to the price changes of substitutes, given the relatively large hypothetically positive values of e x,Py e x,Pz and in-between cross price elasticities implied by the theorem. The industry should design appropriate marketing-specific strategies to compete with other seafood products. For instance, differentiation and adding value characteristics to crawfish products, (i.e., flavored, ready to cook crawfish tail meat) to reduce the number of substitutes could win market share. Not much can be said on age groups. The oldest age groups were more likely to consume the product away-from-home and to consume frequently away-from-home. One possible explanation is because of the lifestyle of the older segment. This finding is not a trend that an industry would like to see since younger consumers often set future trends. Crawfish consumption often is seen as regional because the regions where crawfish is mainly produced are significantly more likely to consume crawfish. Interesting though, was the finding that the East South Central region was significantly likely to consume crawfish away-from-home most frequently.

PAGE 77

67 Another finding in this study is that Caucasians are either less likely to consume or to consume crawfish less frequently away-from-home. Crawfish consumption is traditionally associated with ethnic culture and Cajun cooking in this country. This might be the reason why Caucasians still have lower participation and consumption rates. However, in recent years, ethnic food has become increasingly more popular, with more and more people enjoying Cajun cooking, including crawfish. Our study found that the dislike of taste of crawfish was not the significant reason for not consuming crawfish, and preferring Cajun cooking was the second reason for consuming crawfish. However, due to the fact that this study had a larger percentage of Caucasians responding than what is representative, further study into this issue would be useful. Conclusions In this study, we found that patterns of away-from-home crawfish consumption differ from at-home consumption and other seafood consumption. Crawfish consumption tends to vary by region of residence, and consumers from the main production regions are more likely to consume crawfish. Consumers prefer farm-raised and domestically produced crawfish. Fresh crawfish and tail meat are preferred because of convenience and freshness. In addition, factors regarding consumer preferences and perceptions influence participation and consumption decisions differently. Crawfish consumption is relatively price sensitive, as respondents indicated price is one of the reasons for consuming and the main reason for not consuming. Crawfish is perceived as an inferior good, and away-from-home crawfish consumption is relatively sensitive to income level. Higher income groups are less likely to consume crawfish away-from-home and to consume it less frequently. In addition, crawfish may have many substitutes and is relatively sensitive to consumers income levels.

PAGE 78

68 We also found that dislike of taste is not a significant factor influencing away-from-home consumption and that nutrition issues tend to positively affect both the probability and the frequency of away-from-home crawfish consumption. Consumers do not feel product safety is an issue for crawfish. However, the bad impression of crawfish decreases both the probability and frequency of away-from-home consumption. There is a lack of knowledge about the differences between crawfish, lobster, and langostino among consumers. If the industry can improve the image of crawfish and increase consumers familiarity with crawfish, away-from-home, as well as at-home, crawfish consumption would increase as a result. Our results found availability is a critical factor to increase crawfish consumption. The industry should increase the availability of crawfish through different distribution channels in order to encourage consumption. In addition, smell is a concern of consumers. Product innovation in the industry can be used to reduce this concern. In terms of demographic factors that influence away-from-home crawfish consumption, we found education level and household size do not impact either the probability or frequency of away-from-home crawfish consumption. Other demographics, such as age, race, and ethnicity, also influence both probability and frequency of away-from-home crawfish consumption. Older age groups are more likely to consume crawfish, while Caucasians are less likely to consume crawfish and to consume it less frequently. Implications Since the International Trade Commission imposed Less Than Fair Value (LTFV) tariffs on imported crawfish in 2000, it appears there are opportunities for domestically produced crawfish to increase its market share. The findings of this study could be used

PAGE 79

69 by the industry to target potential and existing crawfish consumers and to increase crawfish consumption, specifically away-from-home consumption. For existing consumers, product development, which focuses on developing new products for present customers, and market penetration, which focuses on improving existing products fro present customers, can be used to increase crawfish consumption. Increasing consumption among existing consumers could be less costly than attracting new consumers. The crawfish industry could use the findings of this study to target potential consumers to encourage consumption of crawfish, particularly away-from-home consumption. Although it is a little costly, it is necessary for the long-term development of the industry. The industry should target the regions surrounding the main production region and promote crawfish as domestically farm-raised seafood product. The study found that once consumers in other regions start to eat crawfish, they consume it more frequently. In addition, the industry should increase the availability of crawfish products, particularly tail meat, through different distribution channels to reach consumers. Since crawfish may have many substitutes and is perceived as an inferior good, it is appropriate for the industry to develop new products with value adding characteristics to differentiate crawfish from its substitutes and to market it as a normal good. Product innovation (i.e., ready to cook tail meat products) can be successful in marketing crawfish. Our study also found crawfish to be relatively price sensitive. As a result, carefully pricing crawfish products, particularly tail meat, is very important in increasing crawfish consumption and in competing with imports of crawfish and other seafood products. Appropriate promotions, such as demonstrations, educational advertising, and

PAGE 80

70 repositioning the image of crawfish in consumers mind, will also be successful in increasing the demand for crawfish. As a result of this study, some areas identified that need further research include quantitative price and income elasticities, the relationship between crawfish and other seafood products (would increases in crawfish consumption occur at the expense of other seafood products), and consumption patterns of different races and ethnicities. In addition, more balanced sample data are needed for further study. It is believed that an appropriate interpretation of the research findings and the development of related marketing strategies could win back the domestic market share lost to Chinese imports, and that the opportunity still exists for the domestic crawfish industry.

PAGE 81

APPENDIX 2004 MARKET SURVEY OF FOOD CONSUMPTION Notice: Any information reported below is strictly confidential. This data will be used only by persons engaged in this survey, and will not be disclosed or released to others for any purpose. Directions: Please have the member of the household that usually decides what food you purchase fill out this survey. Thank you in advance for taking the time to fill out this survey. Crawfish are crustaceans with a hard outer shell, which provides some protection and gives rigidity to their bodies. The shell of adults is dark red to nearly black with a wedge-shaped stripe on the abdomen. These small crustaceans are related to lobsters and closely resemble them. Common Names: Red Swamp Crayfish, White River Crayfish, Crawfish. Scientific Names: Procambarus clarkii Procambarus zonangulus, Procambarus paeninsulanus, P. alleni. Figure A-1. Crawfish pictures. 71

PAGE 82

72 1. Do you feel that crawfish & lobster are different? Yes (different) No (same) Do not Know 2. Do you feel that crawfish & langostino are different? Yes (different) No (same) Do not Know 3. Have you ever eaten crawfish? Yes No 4. Have you eaten crawfish in the last 12 months? Yes No 5. How many times have you eaten crawfish in the last 12 months? (Please write the number) ________ times. 6. Thinking of the occasions you ate crawfish over the past 12 months, how many times did you eat crawfish in each of the following locations? _______ times at home _______ times at restaurants _______ times at other away-from-home locations 7. Are you aware of whether the crawfish you have eaten was farm-raised or wild-caught? Yes (Know) No (Dont know) If YES, have you ever consumed farm-raised crawfish? Yes No Do not know If YES, would you consume it again? Yes No Do not care If NO, would you consider consuming farm-raised crawfish? Yes No Do not care 8. Are you aware of where the crawfish you have consumed came from? (Check any of the following options that apply) Domestic Imported Do not know 9. Would you be more likely to consume crawfish that was produced locally or in the region? (Check one) Yes No Do not care Do not know 10. Would you be more likely to consume crawfish that was imported? (Check One)

PAGE 83

73 Yes No Do not care Do not know 11. What product forms of crawfish do you normally consume Away-From-Home? (Check all that apply) Whole with head Tail meat only If you could choose, which form would you prefer? (Check one) Whole with head Tail meat only 12. What product forms of crawfish do you normally consume At-Home? (Check all that apply) Whole with head Tail meat only If you could choose, which form would you prefer? (Check one) Whole with head Tail meat only 13. Do you prefer fresh or frozen crawfish? (Check one) Fresh Frozen Do not care 14. What price would you pay for one-pound of whole crawfish? (See below for typical prices) _______ $/pound. 15. What price would you pay for one-pound of tail meat? (See below for typical prices) _______ $/pound. (Typical prices are: Boiled Whole Crawfish, Medium $3.00 to $5.00 per pound, Large $6.00 to $7.00 per pound; Tail meat $10.00 to $14.00 per pound). 16. For Away-From-Home consumption, check the top three reasons you EAT crawfish: Enjoy Flavor and Taste Enjoy Texture Health/Nutrition Benefits Tradition/Habit Price (Cost compare to other seafood) Availability Farm-raised Religion Variety in Diet Convenience Prefer Cajun Style

PAGE 84

74 17. Please check the top three reasons you DO NOT EAT crawfish (or do not eat it more often): Price Don't like the Product Form Health/Nutrition Benefits No custom Religion Don't like texture and taste Don't like smell No local farm-raised available Concern about safety Bad impression (i.e., mud bug) Health reasons Vegetarian 18. Gender: Male Female 19. Your Age? _________ 20. Marital status: Never Married Married Separated/Divorced Widowed 21. What is the highest level of education you have completed? No formal education Less than high school diploma High school diploma University undergraduate degree University postgraduate degree Other, please clarify _________ 22. Current Employment: Full time Part time Current not working Student Unpaid family worker Retired 23. Please indicate your approximate household income before taxes: Less than $20,000 $20,000 to $39,999 $40,000 to $59,999 $60,000 to $79,999 $80,000 to $99,999 $100,000 to $119,999 $120,000 to $139,999 $120,000 to $139,999 More than $140,000 24. Number of children under 16 living at home (If none, put 0): _______ 25. Number of people currently living in your household (including yourself)? ______ 26. Please indicate your race (Check all that apply):

PAGE 85

75 White Black and African American Asian American Indian Other, please clarify________ 27. Are you of Hispanic descent? Yes No 28. Please list your Religion: __________ (If Not Applicable, please write down None) 29. What is your zip code? _______ Thank you for your time and effort completing this survey. Please be assured that all answers will be kept strictly confidential and used only for the purposes of this research. If you have any questions about the survey, please contact us at (352) 392-1826 or send email to xzhang@ifas.ufl.edu

PAGE 86

LIST OF REFERENCES Amemiya, T. (1984). Tobit Models: A Survey. Journal of Econometrics, 24, 3-61. Avery, J. and Lorio, W. (1999). Crawfish Production Manual. Pub. 2637. Louisiana State University Agricultural Center, Baton Rouge, LA (July). Bilgic, A., and Florkowski, W. (2003). Truncated-at-Zero Count Data Models with Partial Observability: An Application to the Freshwater Fishing Demand in the Southeastern U.S. Selected Paper presented at the Annual Meetings of the Southern Agricultural Economics Association. Mobile, AL (February). Retrieved February 8, 2004 from http://agecon.lib.umn.edu/ Cheng, H., and Capps, O. Jr. (1988). Demand Analysis for Fresh and Frozen Finfish and Shellfish in the United States. American Journal of Agricultural Economics, 70, 533-542. Cragg, J. (1971). Some Statistical Models for Limited Dependent Variables with Application to the Demand for Durable Goods. Econometrica, 39, 829-844. DAbramo, L. R., Ohs, C. L., Hanson, T. R., and Montanez, J. L. (2002). Production of the Red Swamp Crayfish in Earthen Ponds without Planted Forage: Management Practices and Economics. Bulletin 1115. Mississippi Agricultural & Forestry Experiment Station, Mississippi State University, Mississippi State, MS (February). Dellenbarger, L. E., August, P., and Schupp, A. (1996). Boiled Crawfish Consumption in Louisiana. Journal of Food Distribution Research, (February), 99-101. Drammeh, L., House, L., Sureshwaran, S., and Selassie, H. (2002). Analysis of Factors Influencing the Frequency of Catfish Consumption in the United States. Selected Paper presented at the Annual Meetings of the American Agricultural Economics Association, Long Beach, CA (July). Retrieved November 19, 2003 from http://agecon.lib.umn.edu/ Florida Department of Agriculture and Consumer Services [FDACS]. (1992). Aquaculture Species Resource Series: Crawfish. FDACS, Tallahassee, FL. Florida Department of Agriculture and Consumer Services [FDACS]. (2001). Agriculture Press Release. FDACS, Tallahassee, FL (September). Retrieved April 19, 2003 from http://www.fl-ag.com/ 76

PAGE 87

77 Florida Department of Agriculture and Consumer Services [FDACS]. (2003). Aquaculture Producer Database. Division of Aquaculture. FDACS, Tallahassee, FL. Food Marketing Institute [FMI]. (2003). Consumer Trends in the United States. Food Marketing Institute, Washington, D.C. Retrieved December 23, 2003 from http://www.fmi.org/ Greene, W. (1995). Limdep Version 7.0 Users Manual. Plainview, NY: Econometric Software, Inc. Greene, W. (1999). Econometrics Analysis 4 th Edition, A LIMDEP Guide to the Examples. Upper Saddle River, NJ: Prentice Hall. House, L., Hanson, T., and Sureshwaran, S. (2003). U.S. Consumers: Examining the Decision to Consume Oysters and the Decision of How Frequently to Consume Oysters. Journal of Shellfish Research, 22(1), 51-59. International Trade Commission [ITC]. (1997). Crawfish Tail Meat from China. Investigation No. 731-TA-752, Publication 3057. International Trade Commission, Washington, D.C. Keithly, M. R. (1985). Socioeconomic Determinants of At-Home Seafood Consumption: A Limited Dependent Variable Analysis of Existing and Latent Consumers. Unpublished Ph.D. dissertation. University of Florida, Gainesville, FL. Ladewig, F. K., and Schaer, L. S. (1993). Crawfish-A Healthy Choice. SRAC Publication No. 243. Southern Regional Aquaculture Center, Stoneville, MS (September). Larkin, S., Degner, R. L., and Rubinstein, W. (2002). "Market Preferences, Wholesale Demand & Breakeven Prices for Marine Ornamental Fish Cultured and Collected in Florida." Staff PaperSP-02-5, Department of Food and Resource Economics, University of Florida, Gainesville, FL (73pp.). Retrieved December 23, 2003 from http://agsurveys.org/RLRA29finalreport.pdf Larkin, S., Tucker, J.E., and Degner, R.L. (2001). "Developing an Internet Survey Instrument: Application for Florida Sea Grant Marketing Study." Staff Paper SP-01-8, Department of Food and Resource Economics, University of Florida, Gainesville, FL (20pp.). Lin, C. T., and Milon, J.W. (1993). Attribute and Safety Perceptions in a Double-Hurdle Model of Shellfish Consumption. American Journal of Agricultural Economics, 75, 724-729. McCullough, A., Gempesaw. C. M. II, Daniels, W. H., and Bacon, J. R. (2001). Simulating the Economic Viability of Crawfish Production: A Two-stage Modeling Approach. Aquaculture Economics and Management, 5(1/2) 2001: 69-79.

PAGE 88

78 Mullahy, J. (1986). Specification and Testing of Some Modified Count Data Models. Journal of Econometrics, 33 (December), 341-365. National Fisheries Institute [NFI]. (2003). About Our Industry. The Fish and Seafood Trade Association, NFI, Arlington, VA (March). Retrieved April 19, 2003 from http://www.nfi.org/ Nicholson, W. (2002). Microeconomic Theory: Basic Principle and Extensions (8th edition). Mason, OH: South-Western Educational Publishing. Roberts, J. K. (2000). Imported and Consumer Impacts of U.S. Antidumping Tariffs: Freshwater Crawfish from China. Research Summary, Department of Agricultural Economics and Agribusiness, Louisiana State University Agricultural Center, Baton Rouge, LA. RoperASW (2002). Food Trends in America Presentation. RoperASW, New York, NY. Retrieved December 23, 2003 from http://www.roperasw.com/ Schupp, A., Pomeroy, R., and Dellenbarger, L. (1991). U.S. Food Store Experience in Handling Crawfish. Journal of Food Distribution Research, (June), 1-9. Selassie, H., House, L., and Sureshwaran, S. (2002). Characteristics and Consumer Preferences Affecting Marketing of Farm Raised Fish. Research Monograph. South Carolina State University, Orangeburg, SC (December). Tobin, J. (1958). Estimation of Relationships for Limited Dependent Variables. Econometrica, 26, 24-36. U.S. Census Bureau (2000). U.S. Census 2000, Washington, D.C. Retrieved February 24, 2004 from http://www.census.gov/main/www/cen2000.html U.S. Department of Agriculture, Economic Research Service [USDA-ERS]. (1999a). Contribution of Away-From-Home Foods to American Diet Quality. Research Summaries, Family Economics and Nutrition Review, 12 (3, 4). USDA-ERS, Washington, D.C. U.S. Department of Agriculture, Economic Research Service [USDA-ERS]. (1999b). Aquaculture Outlook. USDA-ERS, Washington, D.C. U.S. Department of Agriculture, Economic Research Service [USDA-ERS]. (2002a). Food Consumption Trends: Food Consumption (per capita) Data System. USDA-ERS, Washington, D.C. Retrieved May 19, 2003 from http://www.ers.usda.gov/ U.S. Department of Agriculture, Economic Research Service [USDA-ERS]. (2002b). Food Away From Home. Food Market Indicators, Food CPI, Prices and Expenditure: ERS Food Expenditure Tables. USDA-ERS, Washington, D.C. Retrieved May 19, 2003 from http://www.ers.usda.gov/Briefing/CPIFoodAndExpenditures/Data

PAGE 89

79 U.S. Department of Agriculture, Economic Research Service [USDA-ERS]. (2002c). Food Consumption (per capita) Data System. USDA-ERS, Washington, D.C. Retrieved May 19, 2003 from http://www.ers.usda.gov/ Yen, T. S., and Huang, L. C. (1996). Household Demand for Finfish: A Generalized Double-Hurdle Model. Journal of Agricultural and Resource Economics, 21, 220-234. Yen, T. S., Dellenbarger, L. E., and Schupp, A. R. (1995). Determination of Participation and Consumption: The Case of Crawfish in South Louisiana. Journal of Agriculture and Applied Economics, 27(1), 253-262.

PAGE 90

BIOGRAPHICAL SKETCH Xumin Zhang was born on November 19, 1972, in China. After graduating from high school, he enrolled in the Food Science and Technology Department at the Shanghai Fisheries University, Shanghai, China, in 1991, and graduated with a bachelors degree in science in June 1995. In September 2000, after working several years with Metro, a German supermarket chain in China, Mr. Zhang immigrated to the United States and continued his education. After studying at Santa Fe Community College, Gainesville, Florida, for one and one-half years, Mr. Zhang was accepted in the Master of Science program of the Food and Resource Economics Department at the University of Florida. He began his masters study at the Food and Resource Economics Department in August of 2002, with fields of specialization in food marketing and management. 80


Permanent Link: http://ufdc.ufl.edu/UFE0004460/00001

Material Information

Title: An Evaluation of factors influencing away-from-home consumption of crawfish in the Gulf region
Physical Description: Mixed Material
Language: English
Creator: Zhang, Xumin ( Dissertant )
House, Lisa A. ( Thesis advisor )
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2004
Copyright Date: 2004

Subjects

Subjects / Keywords: Food and Resource Economics thesis, M.S
Dissertations, Academic -- UF -- Food and Resource Economics
Spatial Coverage: United States

Notes

Abstract: U.S. consumers are traditionally known as away-from-home consumers of seafood products. Over the past 20 years, consumption of ethnic foods has steadily increased. For example, approximately 25 percent of Americans say they enjoy Cajun cooking. In the United States, consumption of crawfish is usually associated with special events or Cajun cooking, which differs from consumption of crawfish in other countries. Growth in consumption has provided opportunities as well as challenges to Chinese exporters and U.S. importers as well as domestic farmers, processors, and marketers of crawfish. This thesis focuses on developing an understanding of factors influencing away-from-home consumption of crawfish in the Gulf region. A web-based consumer survey was conducted to obtain information about consumer preferences. Demographics as well as stated consumer preferences have significant effects on both participation and consumption decisions, but in different scales or directions. For example, enjoying the flavor, preferring Cajun cooking, and adding variety in the diet were the main reasons for consuming crawfish. The results also indicated that crawfish is an inferior good and relatively price sensitive. Caucasians tend to be less likely to consume crawfish away-from-home and to consume less frequently. Additionally, the lowest income group has the highest likelihood and frequency of crawfish consumption.
Subject: consumer, consumption, crawfish, double, hurdle, marketing, model, preference
General Note: Title from title page of source document.
General Note: Document formatted into pages; contains 90 pages.
General Note: Includes vita.
Thesis: Thesis (M.S.)--University of Florida, 2004.
Bibliography: Includes bibliographical references.
General Note: Text (Electronic thesis) in PDF format.

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0004460:00001

Permanent Link: http://ufdc.ufl.edu/UFE0004460/00001

Material Information

Title: An Evaluation of factors influencing away-from-home consumption of crawfish in the Gulf region
Physical Description: Mixed Material
Language: English
Creator: Zhang, Xumin ( Dissertant )
House, Lisa A. ( Thesis advisor )
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2004
Copyright Date: 2004

Subjects

Subjects / Keywords: Food and Resource Economics thesis, M.S
Dissertations, Academic -- UF -- Food and Resource Economics
Spatial Coverage: United States

Notes

Abstract: U.S. consumers are traditionally known as away-from-home consumers of seafood products. Over the past 20 years, consumption of ethnic foods has steadily increased. For example, approximately 25 percent of Americans say they enjoy Cajun cooking. In the United States, consumption of crawfish is usually associated with special events or Cajun cooking, which differs from consumption of crawfish in other countries. Growth in consumption has provided opportunities as well as challenges to Chinese exporters and U.S. importers as well as domestic farmers, processors, and marketers of crawfish. This thesis focuses on developing an understanding of factors influencing away-from-home consumption of crawfish in the Gulf region. A web-based consumer survey was conducted to obtain information about consumer preferences. Demographics as well as stated consumer preferences have significant effects on both participation and consumption decisions, but in different scales or directions. For example, enjoying the flavor, preferring Cajun cooking, and adding variety in the diet were the main reasons for consuming crawfish. The results also indicated that crawfish is an inferior good and relatively price sensitive. Caucasians tend to be less likely to consume crawfish away-from-home and to consume less frequently. Additionally, the lowest income group has the highest likelihood and frequency of crawfish consumption.
Subject: consumer, consumption, crawfish, double, hurdle, marketing, model, preference
General Note: Title from title page of source document.
General Note: Document formatted into pages; contains 90 pages.
General Note: Includes vita.
Thesis: Thesis (M.S.)--University of Florida, 2004.
Bibliography: Includes bibliographical references.
General Note: Text (Electronic thesis) in PDF format.

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0004460:00001


This item has the following downloads:


Full Text












AN EVALUATION OF FACTORS INFLUENCING AWAY-FROM-HOME
CONSUMPTION OF CRAWFISH IN THE GULF REGION















By

XUMIN ZHANG


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE

UNIVERSITY OF FLORIDA


2004


































Copyright 2004

by

Xumin Zhang

































To Miao, our kid, and my parents















ACKNOWLEDGMENTS

I would like to deeply thank Dr. Lisa House, chair of my supervisory committee,

for her advice and timeless help on this thesis. Also, I would like to thank Dr. Richard

Kilmer for giving me the chance to do research with him during my master's program.

Those experiences have broadened my view and enriched my knowledge base. Dr. House

and Dr. Kilmer have given me the most help and advice during my master's program, and

I would like to express my appreciation again to them.

I also want to thank the Food and Resource Economics Department for those who

have given me support and encouragement, like Dr. Robert (Jeff) Burkhardt, Dr. Donna

Lee, and Dr. Allen Wysocki, and particularly Jessica Herman, our program assistant.

Finally, my deepest gratitude goes to Miao and my parents for their support

throughout these years.
















TABLE OF CONTENTS
Page

ACKNOW LEDGM ENTS ........................................ iv

LIST OF TABLES ............. ........ ............. ........... ........ vii

LIST OF FIGURES .......... ............................. viii

ABSTRACT.................................. .............. ix

CHAPTER

1 INTRODUCTION ................... .................. .............. .... ......... .......

Pream ble .................................................... ...............
Problematic Situation................ ...... ...............
Researchable Problem .............................................. ..... .. .7
Obj ectives ................... ...................................... ............... ......... 8
H ypotheses................................................................ 9

2 L IT E R A TU R E R E V IE W ...................................................................................... 11

3 SURVEY IN STRUM EN T................................................................................. .. ......17

Survey Instrument.......................... ...............17
Survey Contents................................ ....... .........19

4 D A T A ....................................................... 22

5 THEORETICAL MODEL AND MODEL SPECIFICATION............... ...............30

T h eoretical M odel ................... .................................... ...........................30
M odel Specification....................... ............... .. .............33
Truncated-at-Zero Count Data Double Hurdle Model ........................................38

6 EM PIRICAL RESULTS ................................................ ............... 40

Issues of Craw fish Consum ption .......................................................................... ........40
R regional Consum ption ................................................ ............... 40
Farm-Raised Crawfish............................. ......... ......... 40
D om estic Versus Im ported Crawfish ...................................... ........... ....41










Product Form ............................... .................. ....... 42
Reasons for Consuming and Not-Consuming ...................................44
R seasons for consum ing ........................................ ................. 44
Reasons for not consuming .............................. ............... 44
Away-From-Home Crawfish Consumption ........................................47

7 CONCLUSIONS AND IMPLICATIONS ..........................................62

Sum m ary ............... ....... .. .. ......... .. .. ........................ 62
Conclusions............................................. .........67
Implications ................................................68

APPENDIX 2004 MARKET SURVEY OF FOOD CONSUMPTION..........................71

L IST O F R E F E R E N C E S ............................................................................................. 76

BIOGRAPHICAL SKETCH .................................................. ............... 80
















LIST OF TABLES


Table page

4-1. Residence region of survey respondents. .............................................. .....23

4-2. Summary of demographic information.............................. ...............27

4-2. Continued......................................... ......... 28

4-3. Descriptive statistics on other factors included in the Double-Hurdle model............29

6-1. Sample frequency distribution of the dependent variable (n=733). ........................48

6-2. Description of variables included in the double-hurdle model............... ...............50

6-2. Continued......................................... ......... 51

6-3. Empirical results of single-decision and double-hurdle: maximum-likelihood
estim ates and m arginal effects. ......................................................... 53

6 -3 C ontinu ed................ ........................................................................ .................. ...... 54
















LIST OF FIGURES


Figure page

1-1. United States per capital total, fresh & frozen fish and shellfish consumption............2

1-2. Percentage of away-from-home food expenditure at 1988 prices................................3

1-3. Percentages of seafood meals eaten away-from-home..............................................4

4-1. Comparison of U.S. population in the Southeast and survey respondents by age. ....26

4-2. Comparison of incomes of all the U.S. population and incomes of the survey
respondents...................................... ............................... ......... 26

6-1. Regional percentage of crawfish consumption........................ ............... 41

6-2. Comparison of preferences over domestic and imported crawfish products..............42

6-3. Categories of reasons given for consuming crawfish (crawfish consumers vs. away-
from -hom e consum ers). .............................................................. 45

6-4. Categories of reasons given for consuming crawfish (away-from-home consumers
only, at-home consumers only, and all crawfish consumers). ..............................45

6-5. Categories of reasons given for not consuming or not consuming more crawfish
(consum ers vs. non-consumers). ................................ ............... 46

6-6. Categories of reasons given for not consuming or not consuming more crawfish
(away-from-home only, at-home only, and non crawfish consumers). ................48

A-1. Crawfish pictures. ............................................... ........ 71
















Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science

AN EVALUATION OF FACTORS INFLUENCING AWAY-FROM-HOME
CONSUMPTION OF CRAWFISH IN THE GULF REGION

By

Xumin Zhang

May 2004

Chair: Lisa A. House
Major Department: Food and Resource Economics

U.S. consumers are traditionally known as away-from-home consumers of seafood

products. Over the past 20 years, consumption of ethnic foods has steadily increased. For

example, approximately 25 percent of Americans say they enjoy Cajun cooking. In the

United States, consumption of crawfish is usually associated with special events or Cajun

cooking, which differs from consumption of crawfish in other countries. Growth in

consumption has provided opportunities as well as challenges to Chinese exporters and

U.S. importers as well as domestic farmers, processors, and marketers of crawfish.

This thesis focuses on developing an understanding of factors influencing away-

from-home consumption of crawfish in the Gulf region. A web-based consumer survey

was conducted to obtain information about consumer preferences. Demographics as well

as stated consumer preferences have significant effects on both participation and

consumption decisions, but in different scales or directions. For example, enjoying the

flavor, preferring Cajun cooking, and adding variety in the diet were the main reasons for









consuming crawfish. The results also indicated that crawfish is an inferior good and

relatively price sensitive. Caucasians tend to be less likely to consume crawfish away-

from-home and to consume less frequently. Additionally, the lowest income group has

the highest likelihood and frequency of crawfish consumption.














CHAPTER 1
INTRODUCTION

Preamble

Crawfish belongs to the scientific class Crustaceans. Crawfish has a hard external

shell, which provides protection to its body. These small crustaceans are related to

lobsters and closely resemble them. The Red Swamp Crawfish (Procambarus clarkia) is

the species that is the most acceptable for cooking, and is commercially produced and

consumed. The shell of the adult Red Swamp Crawfish is dark red to nearly black

(Ladewig and Schaer, 1993).

In the Southeastern United States and the Gulf region, particularly in Louisiana,

crawfish farming and consumption have been a part of the culture (Avery and Lorio,

1999). Catches from the wild are usually seasonal and unpredictable. Increases in product

demand beyond the amount traditionally caught in the wild led to crawfish farming

(D'Abramo et al., 2002). In the United States, consumption of crawfish is usually

associated with special events or Cajun cooking, which differs from consumption of

crawfish in other countries. This thesis will focus on developing an understanding of

factors that influence consumption of crawfish in the Gulf region.

Problematic Situation

Seafood includes aquaculture products such as crawfish and harvests from salt and

fresh water sources. In the United States, the consumption of seafood has become an

important part of the diet. According to the United States Department of Agriculture

(USDA), the consumption of seafood has increased in recent decades. The estimated per











capital seafood consumption increased by three pounds from 1970 to 1992, to 14.7 pounds,

and it increased further in the 1990s. In 2000, the estimated per capital consumption of

seafood was 15.6 pounds (Figure 1-1).


18

16

14

12

o 10

8 8

6

4

2

0




Total Fresh & Frozen

Source: USDA-ERS, 2002a.
Figure 1-1. United States per capital total, fresh & frozen fish and shellfish consumption.

Nearly two-thirds of the consumption of seafood was in the form of fresh and

frozen products, the remainder was in processed products such as canned and cured. This


compares to 57- 60 percent in the 1970s (Figure 1-1). In value terms, according to the

National Fisheries Institute (NFI), American consumers spend almost $50 billion each


year on a wide variety of fish and shellfish products. The thousands of firms that produce,

process and distribute fish and shellfish are located throughout the United States, and

annually contribute more than $25 billion to the U.S. gross national product (NFI, 2003).


Along with eating more seafood, U.S. consumers are dining out more often than

ever before (USDA-ERS, 1999a). Away-from-home food expenditures increased from 36










percent of total food expenditures in 1970 to 47.5 percent in 1996, and 48 percent in 2001

(Figure 1-2). "Reasons for this trend include smaller household size, more affordable and

convenient fast food services, a growing number of women working outside the home,

and higher household incomes" (USDA-ERS, 1999a).


60%

50%

40%

30%

20%

10%

0%



Source: USDA-ERS, 2002b; USDA-ERS, 2002c.
Figure 1-2. Percentage of away-from-home food expenditure at 1988 prices.

US consumers are traditionally known as away-from-home consumers of seafood

products. Although no precise data are available, one estimate by Keithly (1985)

suggested that the quantity of away-from-home consumption of seafood products ranged

from one-third to two-thirds of all seafood consumed. A recent study by Selassie, House,

and Sureshwaran (2002) found 57, 62, and 58 percent of meals of shrimp, oysters, and

catfish, respectively, were consumed away-from-home (Figure 1-3). This set of figures

compares to general food consumption, where 16 percent of the meals were eaten away-

from-home in 1978, a figure that increased to 29 percent by 1995 (USDA-ERS, 1999a).










Crawfish can be seen as a healthy food, which is "high in minerals and protein, and

low in calories and saturated fat" (Ladewig and Schaer, 1993). Also, according to the

Florida Department of Agriculture and Consumer Services (FDACS), crawfish is known

for its delicious meat, and is consumed all over the United States (FDACS, 1992).


70%

60% -

50% -

40% -

30% -

20% -

10% -

0%
Meals* Shrimp Oyster Catfish Seafood

Note: All meals data were from 1995, other categories were from 2001 (more recent data not
available for all meals).
Source: Selassie, House, and Sureshwaran, 2002; USDA-ERS, 1999a.
Figure 1-3. Percentages of seafood meals eaten away-from-home.

When talking about the consumption of crawfish, people used terms such as

"Louisiana parties, family gatherings, and Cajun-style boiled crawfish" (Ladewig and

Schaer, 1993). Traditionally, most crawfish in the United States was consumed in

Louisiana (Dellenbarger et al., 1996). Over the past 20 years, consumption of ethnic

foods in the United States has steadily increased, with approximately 25 percent of

Americans indicating they enjoy Cajun cooking (RoperASW, 2002). Particularly, a large

proportion of young consumers indicate a preference for ethnic foods (FMI, 2003). Since

younger consumers are usually seen to set future trends, their preference of ethnic foods,

including Cajun-cooking might hint at an increasing demand for crawfish. As a result, the









demand for Cajun dishes that include crawfish is steadily growing, and the potential for

growth in the crawfish industry is high.

Due to the increases in year-round demand for crawfish and the seasonality and

unpredictability of crawfish caught from the wild, crawfish aquaculture has arisen. "The

majority of crawfish produced in the United States is used for food and originates in

Louisiana" (McCullough et al., 2001), and the price generally is considered expensive,

particularly in other regions. An increased interest in finding alternative sources of

income has led "farmers to transform their fields into crawfish farms, and has expanded

production along the Gulf coastal regions of the United States" in the last 30 years

(Ladewig and Schaer, 1993).

A suitable crawfish pond needs relatively flat land, soil with a high clay content,

and an adequate water source. The states in the Southeast and Gulf region have

geographic and climatic advantages in crawfish production (FDACS, 1992). Nevertheless,

although the growing demand for crawfish has heightened the interest of agricultural

producers, the industry in states other than Louisiana is still comparatively weak. For

example, due to the state regulations on crawfish aquaculture in Florida, there were only

21 facilities certified as crawfish producers in the year of 2003. Among those 21 farms,

less than 15 farmers produce crawfish for the food market, and the acreage for crawfish

producing is relatively small (FDACS, 2003).

In 1990, there were about 160,000 acres of managed crawfish ponds in the United

States, and total harvest of crawfish was around 760 million pounds a year (Ladewig and

Schaer, 1993). However, due to the rapid increase of imports since 1994, crawfish

production decreased to about 36 million pounds in 1998 (USDA-ERS, 1999b). In the









year 2000, total harvest of crawfish in the US was around 100 million pounds (FDACS,

2001), partly as a result of tariffs on imported crawfish. Crawfish production is mainly in

the lower Gulf Coast regions, such as Texas, Mississippi, Florida, and Louisiana, with

most of the acreage found in Louisiana (Avery and Lorio, 1999).

It is worth mentioning that since 1994, an increasing proportion of the market

volume has been imported from foreign countries such as China and Spain. The market

share of imported crawfish increased to 87 percent in 1996 (Roberts, 2000), and imports

continuously increased 173 percent in 1998. The total quantity was over six million

pounds in 1998, mainly in the form of tail meat (McCullough et al., 2001). The

International Trade Commission (ITC, 1997) has found that the increasing crawfish tail

meat imports from China have caused material injury to the domestic crawfish processing

industry, and has imposed Less Than Fair Value (LTFV) duties on crawfish since 2000

(Roberts, 2000). The average import tariff is as high as 123 percent, and there also are

levies on past shipments. Over 90 percent of crawfish imports come from China even

with restrictive duties (McCullough et al., 2001). Because of the limited quantities of

crawfish available in the U.S. market, prices have been higher in recent years. Given the

deficit in domestic supply, it appears there will be opportunities for crawfish farming in

domestic production.

Although consumption data for crawfish are not available, according to several

studies on crawfish consumption (Ladewig and Schaer, 1993; Yen et al., 1995,

Dellenbarger et al., 1996; McCullough et al., 2001), the demand for crawfish has

increased, and the growth in consumption has provided both opportunities and challenges

to Chinese exporters and US importers as well as domestic farmers, processors, and









marketers of crawfish. However, domestic market feasibility, market growth potential,

and market size and trends are still unknown, and there are concerns about the viability of

crawfish production and marketing.

Knowledge about consumers' preferences and perceptions in terms of crawfish

consumption becomes important in identifying the viability of crawfish production and

marketing. The lack of information about consumer preferences is a serious barrier to

answering questions about the viability of crawfish production and marketing. Therefore,

for the growth of the industry, especially in the Gulf region, new information is needed

on the factors that influence crawfish consumption patterns.

For crawfish, factors that might influence the consumption decision and frequency

of consumption are unknown. Further, it is not known if these factors have the same

influence on at-home consumption as compared to away-from-home consumption

decisions. Developing an understanding of those factors that influence consumption of

crawfish becomes comparatively important.

Researchable Problem

Growth in consumption has offered opportunities and challenges to producers,

processors, and marketers of crawfish. Knowledge about consumers' preferences and

perceptions in terms of crawfish consumption is important for identifying the viability of

crawfish production and marketing. While potential producers and importers of crawfish

may question the profitability of new crawfish ventures, processors, and marketers may

question the viability and profitability of crawfish marketing. The lack of information

about consumers' preferences and perceptions is a serious barrier to assessing the

viability of crawfish production and marketing. Therefore, new information is needed

about factors, such as demographic, socioeconomic, and consumers' preferences and









perceptions about crawfish, which might influence the probability and the frequency of

consumption.

Due to the fact that Americans are traditionally away-from-home seafood

consumers and most crawfish are consumed in the Southeastern United States, and away-

from-home consumption is the main means of introducing seafood to new consumers,

this study will focus on developing an understanding of factors that might influence

away-from-home consumption of crawfish in the Southeastern United States, particularly

in the Gulf region.

Objectives

The overall objective of this study is to develop an understanding of the factors that

influence away-from-home consumption of crawfish. We will collect information from

existing and potential crawfish consumers in the Gulf region, and study the effects of

geographical regions and household characteristics for away-from-home crawfish

consumption that are not available in aggregate time series. Also, stated consumer

preferences and perceptions regarding taste, flavor, nutrition, safety, appearance, and

availability will be examined to see whether they affect the consumption decision and

consumption frequency. Specific objectives are:

* To inform producers, processors, and marketers about what attributes of crawfish,
such as taste, flavor, nutrition, safety, and appearance, and which form of crawfish
product (whole boiled or tail meat) are preferred by current and potential
consumers in the Gulf region.

* To inform importers of crawfish about the factors that will influence the demand
and market trends of crawfish in the near future.

* Due to the domestic popularity of crawfish in China, Chinese exporters target the
United States as one of the biggest importing countries of crawfish. Information
about factors that influence consumption might help Chinese exporters better
understand the different preferences and perceptions regarding crawfish.









Hypotheses

Patterns of away-from-home consumption are likely to differ from at-home

consumption. Moreover, for crawfish, it is likely to differ from other seafood

consumption and differ from consumption in other countries. Many factors might

influence participation and consumption decisions. Those factors can be separated as

demographics and preferences or perceptions. Demographics might include geographical

regions, ethnicity, household income, and educational levels. In addition to price,

consumer perception regarding taste, flavor, nutrition, safety, appearance, and availability

might influence crawfish consumption. Given these considerations, the following

hypotheses are proposed:

* Crawfish consumption will vary by region of residence. Consumers from the main
production regions will be more likely to consume crawfish.

* Farm raised and/or domestically produced crawfish will be preferred by consumers
over wild caught and/or imported crawfish.

* Fresh crawfish and tail meat will be preferred by consumers in accordance with the
market trends toward convenience and freshness.

* Education and household income will negatively affect the probability of away-
from-home crawfish consumption.

* Household size, education, and household income will positively affect the
frequency of away-from-home consumption among households that do consume
crawfish.

* Other demographics, such as age, race, and ethnicity will influence both the
probability and frequency of away-from-home crawfish consumption. For example,
older groups will be more likely to consume crawfish. On the contrary, Caucasians
are less likely to consume crawfish and consume it less frequently.

* Consumer preferences, including taste, nutrition, and availability issues, will affect
the participation decision for away-from-home crawfish consumption.

* Consumer preferences, including taste, nutrition, and availability issues, will affect
the frequency of away-from-home crawfish consumption.






10


* Consumer perceptions regarding seafood safety and appearance will negatively
affect participation and consumption decisions.














CHAPTER 2
LITERATURE REVIEW

Not many empirical studies focusing on crawfish consumption patterns exist in the

literature. Few studies have focused on crawfish consumption, and no study has

examined away-from-home consumption of crawfish in the Gulf region. Studies on

crawfish consumption patterns were mostly conducted in the early to middle 1990s. As a

result, new information is needed.

Two previous studies, Yen et al. (1995) and Dellenbarger et al. (1996), examined

crawfish consumption in South Louisiana and Louisiana, respectively. Yen et al. (1995)

investigated the determinants of crawfish consumption in South Louisiana using a

generalized limited dependent variable model that is similar to Cragg's double hurdle

model (Cragg, 1971).

The study included socio-economics and demographics as the independent

variables, and no price or expenditure data were collected. The independent variables

included income, household size, and dummy variables indicating professional types,

employment status, education, religion, and race. The quantities (in pounds) of crawfish

consumed by the responding households during a five-day period previous to the survey

were collected as the dependent variable. It is worth mentioning here, among its 915

responding households, only 200 households, or 21.9 percent, reported consumption of

crawfish.

The study found that demographics, such as income, household size, and skilled

labor, increased the likelihood of crawfish consumption but not the conditional level of









consumption. For example, households characterized by the attributes of higher income,

larger household size, skilled labor, Catholic, and white are more likely to consume

crawfish than others. In terms of the conditional level of crawfish consumption, they

found that education and employment status are among the household characteristics that

determine the level of consumption. Households with skilled laborers or unemployed

workers consume more crawfish than other groups. In addition, the results suggest that

consumption of crawfish is income inelastic, although insignificant.

However, the study only centered attention on the most recent five-day period of

crawfish consumption, and thus ignored the seasonality of the crawfish harvest.

Moreover, the study only included socio-economics and demographics that might

influence consumption; however, it did not illustrate other factors regarding consumer

preferences and perceptions, which might also influence crawfish consumption.

A shortcoming of the study is the method in which the dependant variable was

obtained. Respondents were asked to recall how many pounds of crawfish they had

consumed in the five-day period. Respondents were then converted to live-product

equivalents. The calculation of the dependent variable caused discrepancies from realistic

quantities. However, it is not unexpected as respondents may find it hard to estimate the

quantity consumed in pounds.

Dellenbarger et al. (1996) examined the consumption of boiled crawfish in

Louisiana by using a mail survey. A Logit model was used to estimate the probability of

household consumption of boiled crawfish. The study surveyed "four rural and four

urban parishes of Louisiana to identify their crawfish and seafood consumption patterns".









As a result of the small survey region, interpretation of the study should be conducted

with caution.

Of 858 effective observations, 203 households, or 23.6 percent, reported

consumption. The study only focused on participation decision, and households were

asked if they had consumed boiled crawfish in the last five days. As a result, a Logit

analysis was used. A value of one represented households consuming boiled crawfish,

and a value of zero represented non-consumption. The dependent variables included

income, household size, urban resident, race, religion, and education.

The study found that religion had an influence on the probability of crawfish

consumption. Catholic and Protestant respondents were more likely to consume crawfish.

Also, white households were more likely to consume boiled crawfish than non-white

households. However, unlike Yen et al.'s findings, education was not a statistically

significant variable in determining the likelihood of boiled crawfish consumption.

The study suggested "households with incomes below $25,000 should be targeted

for boiled crawfish consumption, since crawfish consumption declined with income".

Additionally, the results showed that urban households had a lower probability of boiled

crawfish consumption compared to rural households.

Schupp et al. (1991) studied U.S. food stores' handling of crawfish, finding that

"many consumers were not knowledgeable about aquaculture products, especially in

areas of limited or no local production" (Schupp et al., 1991). In addition, market

expansion for crawfish outside the South Central region was found to be largely

dependent on obtaining the support of grocery stores. The study found that many

existing markets for crawfish are not being met by the current domestic supply and many









people do not know how to prepare crawfish at home. Availability is limited and crawfish

is likely not to be found in local supermarkets.

Other empirical studies focusing on seafood consumption patterns include Keithly

(1985), Cheng and Capps (1988), Yen and Huang (1996), and Drammeh et al. (2002).

Keithly (1985), using food consumption survey data, focused on a set of socio-

economic and demographic factors that affect at-home consumption of total seafood and

five specific products. He found that region, urbanization, race, household size, money

value of meals consumed away-from-home, and income were all contributing factors that

helped to explain at-home seafood consumption patterns.

Cheng and Capps (1988) investigated the key socio-demographic determinants of

at-home demand for several fresh and frozen finfish and shellfish species. They found

factors explaining the variation of expenditures on seafood were own price, household

income, household size, coupon value, geographic region, urbanization, race, and

seasonality.

Drammeh et al. (2003) found source of seafood for consumption, enjoyment of

flavor, availability, price, allergies, gender, and geographic reasons to be significant in

determining probability of participation in oyster consumption. They found the double-

hurdle model was a significantly better fit than the tobit model. Variables significant in

the level of consumption of oysters included source of seafood for consumption,

enjoyment of flavor, tradition, price, product safety, geographic region, income, and age.

Yen and Huang (1996) performed a detailed study on household demand for finfish.

The study estimated household demand for finfish in the United States using a limited









dependent variable model similar to Cragg's double hurdle model, which accounted for

both participation and consumption decisions.

The study found that the price of finfish, shopping frequency, geographic region,

race, and life-cycle variable were the key factors that significantly affect both the

probability of participation and the level of household finfish consumption in the United

States. Furthermore, they found "a variable might exert opposite effects on the

probability and level of seafood consumption".

Also, the study concluded that the double hurdle model was particularly relevant

for studying seafood consumption behavior, because the participation and consumption

decision were likely to differ. Results of the analysis were useful for seafood marketers in

planning and developing marketing strategies. Based upon the results, seafood marketers

can differentiate between the factors that influence participation decisions and the factors

that influence consumption decisions.

In conclusion, there have not been many empirical studies of crawfish consumption

in the Southeast and the Gulf region, particularly in recent years. As a result, research

conducted to study the consumption of crawfish, including regions beyond Louisiana,

would be useful. In addition to demographics, factors regarding consumer preferences

also influence the consumption of crawfish either at-home or away-from-home. It has

been suggested to include those factors in the analysis of crawfish consumption patterns.

Those factors include information from current and potential crawfish consumers about

their preferences and perceptions of crawfish to determine which of these might influence

decisions and the frequency of crawfish consumption. Attention to the consumption






16


patterns and the identification of those factors could be helpful in developing marketing

strategies for the industry targeting these markets.














CHAPTER 3
SURVEY INSTRUMENT

This study will examine away-from-home consumption of crawfish in the lower

Gulf regions using data collected from an Internet-based survey. The web-based survey

was conducted in late January of 2004, which was just before the main crawfish harvest

season. However, the survey asked the frequency of away-from-home crawfish

consumption in the previous 12-month period to accommodate the seasonality of

crawfish consumption. Although it is expected that recalling the frequency of

consumption over 12 months may be difficult, in the case of crawfish the low overall

level of consumption during one year reduces this concern. For example, people can

remember one time without difficulty, but cannot easily remember 100 times.

Survey Instrument

Web-based surveys are increasing in popularity with the growth of the Internet.

Some advantages of a web-based survey include "the ability to use color graphics, higher

response rate and higher completion rate, and lower cost compared to other survey

instruments" (Larkin et al., 2002).

Color graphics can be very expensive for mail surveys and impossible for

telephone surveys. In our food market survey for crawfish consumption, color graphics

were included to illustrate crawfish, following a verbal description. Since crawfish

resemble lobster, it is extremely important to avoid confusion by providing color graphics.

Participants were recruited through Survey Sampling, Inc.'s web-survey service.

Similar to telephone survey centers, this service charged per completed survey, thus









allowing the researcher to guarantee the number of responses. This differs from mail

surveys where the response rate is highly dependent on factors like survey length. By

surveying on the Internet, respondents might be more likely to complete the entire survey.

Additionally, in each question page, blank responses can be disallowed in some critical

questions by using reminder dialogue boxes.

Compared to mail surveys and focus groups, the web-based survey lowers the cost

and increases the completion rate. With the growth of the Internet and Information

technology, the web-based survey is one of the most effective survey instruments.

However, there are still some disadvantages of conducting a web-based survey such as

"non-representation of consumers without Internet access" (Larkin et al., 2002).

Uncompleted surveys still exist. Also, if a respondent completes the survey several times,

it would be difficult to pick only one as the final response. Steps can be taken to address

the disadvantages.1

Another important issue included in the web-based survey instrument is how to

obtain potential respondents. Our method involved contracting a survey sampling

company, which can provide the stratified survey sample with the potential respondents

fitting our specific requests such as region of residence, demographic compositions

comparable to the Census data, etc. An additional advantage of using an Internet survey

service is that participants have agreed to participate in web surveys, leading to a higher

completion rate.

The procedures of how respondents completed the Internet survey are shown below:

Once the respondents had read the welcome page and started the survey, a time stamp


1 See Larkin et al. (2001) for details of developing an Internet survey instrument.









would automatically record when the respondent started the survey. Responses were

'submitted' when the respondent completed one page and wanted to move to the next web

page. To receive the data in different stages, the survey questions were organized into a

series of five pages. The resulting files for each page were retrieved only by the survey

authors, not by the survey sampling company.

Survey Contents

The purpose of the web-based survey is to collect information about crawfish

consumption patterns. Once the respondents began the survey, responses were submitted

when the respondent completed each page of the five-page survey. Interruption during

completion had smaller impacts on completion rate than an Internet survey with only one

long page. For example, a survey was not considered completed for the survey sampling

company until all five pages were completed. However, we obtained data from

respondents who completed less than all five pages of the survey. The designing of

survey questions was the critical part for the success of the web-based survey. Our web-

based survey can be accessed at http://www.agsurveys.org/food, and the contents are

included in the Appendix.

After the respondents read the welcome page and gave their informed consent, they

were asked to click the 'start' button to begin the survey if they were the member of the

household that usually decides what food to purchase. As we mentioned before, a time

and date stamp recorded the beginning time to monitor how long the respondents took to

complete the survey. Basic information, including the description, scientific names, and

color graphics of crawfish, was provided on the second page. The first two questions

asked respondents whether they could distinguish between crawfish, lobster, and

langostino. Because crawfish are related to lobsters and closely resemble to them, it is









believed that respondents might not perceive the difference. One reason to ask these

questions was that we wanted to understand how knowledgeable people were about

crawfish. Another reason was we wanted to make sure that, after reading the description,

respondents could tell that, although related, crawfish are different from lobster and

langostino, and to ascertain whether they were answering the survey for crawfish and not

other products.

The following two questions asked the respondent if they had ever eaten crawfish

and, if so, if they had eaten crawfish in the last 12 months. A simple redirect was used to

lead respondents who had not eaten crawfish in the last 12 months to the final page, while

others proceeded to answer questions about crawfish consumption.

The dependent variable in this study is the frequency of away-from-home

consumption of crawfish. For each respondent who had consumed crawfish, we first

questioned them on how many times they had eaten crawfish in the last 12 months. To

obtain the dependent variable for this study, we then questioned them on how many times

they had eaten crawfish both at-home and away-from-home in the last 12 months. For

the crawfish consumers, questions were designed to obtain respondents' perceptions and

preferences on attributes such as farm-raised vs. wild caught crawfish and domestic vs.

imported crawfish, respectively.

Preferences on attributes regarding product forms were also obtained for away-

from-home and at-home consumption. Main product forms of crawfish in this study

include whole-boiled crawfish and crawfish tail meat, which are the dominant forms in

the existing market. Whole-live crawfish were not covered because we were only

concerned about away-from-home consumption and people seemed unknowledgeable









about cooking crawfish at home according to our pre-survey study. Other attributes

included frozen and fresh crawfish.

In order to obtain price preferences on crawfish, we asked questions about how

much they would pay for one pound whole-boiled crawfish and one pound crawfish tail.

Typical prices were obtained through a pre-survey study, mainly from the local market

and the web-based crawfish trader. Typical prices included whole-boiled crawfish

(medium, three to five dollars per pound, and large, six to seven dollars per pound) and

tail meat, 10 to 14 dollars per pound. The preferences on price might give insight on how

much consumers would be willing to pay for one pound whole-boiled crawfish and one

pound tail meat. Also, respondents were asked to select the top three reasons why they

did not eat crawfish, including price. Other variables for eating crawfish included

enjoyment of flavor and taste, enjoyment of texture, health and nutrition benefits,

tradition and habit, prices or cost, availability, variety in diet, and prefer Cajun style.

The final page of the survey (completed by both those who had consumed crawfish

in the last 12 months and those who had not) included a question asking respondents to

identify the top three reasons why crawfish was not eaten, or was not eaten more

frequently. Variables for the top three reasons included prices, do not like the product

form, no custom, do not like texture and taste, do not like the smell, no local farm-raised

available, safety concerns, bad impression, health reasons (allergy), and vegetarian.

The final part of the survey included socio-economic and demographic questions

such as household income, household size, employment status, marital status, religions,

and race. The information became extremely important in analysis of away-from-home

consumption of crawfish, based on previous empirical studies.














CHAPTER 4
DATA

The data for this study were obtained through a web-based consumer survey.

Targeting the Gulf region, the web-based survey was conducted in late January of 2004.

Respondents were asked about the frequency of away-from-home crawfish consumption

during the previous 12 months. Recall bias for the crawfish consumption is minimized

because of the nature of the web-based survey. Crawfish consumption is expected to be

seasonal. We asked for a 12-month recall period to accommodate the seasonality of

crawfish consumption, given the low overall level of crawfish consumption in one year.

Survey Sampling, Inc. provided a stratified survey sample for the survey. The

chosen stratified sample was confined to the Southeast and the Gulf region. Different

census regions were expected to be a significant determinant of both the participation and

consumption decision on away-from-home consumption of crawfish. The survey was

published on the website (agsurveys.org/food) in late January of 2004, with a total of 765

responses to the questions related to crawfish consumption. As mentioned before,

compared to the mail survey instrument, the web-based survey is more effective and

efficient in data collecting. Although the response rate for our survey was about 14.1

percent, we have a very high completion rate, 90 percent, which means, once the

respondents started the survey, 90 percent would finish it. In this study, we used 733

effective records, which provided full information on crawfish consumption for this study.

Cases with incomplete socio-economic and demographic information, such as income

and gender, were dropped from the sample. In addition, responses with uncompleted









answers and irrational answers were eliminated as well. The information obtained from

these 733 responses is summarized below.

We considered three census regions. For comparison, we divided the large

Southeast Atlantic region into Northern and Southern areas (Table 4-1). The demographic

data collected in this study indicated that the response rate per region in the Gulf region

was relatively comparable, ranging from 129 effective responses from the East South

Central region to 208 responses from the Southern Southeast Atlantic region (Table 4-1).

The comparison to the percentage of population in these regions is included in the table.

Table 4-1. Residence region of survey respondents.
% of the
Region of Number of % of Survey
S States Included in Region Population in
Residence Responses Responses r
the region
Northern Delaware, Washington D.C., 190 26.0 28.1
Southeast Maryland, North Carolina,
Atlantic Virginia, West Virginia
Southern Florida, Georgia, South 208 28.4 23.5
Southeast Carolina
Atlantic
East South Alabama, Kentucky, 129 17.6 17.0
Central Mississippi, Tennessee,
West South Arkansas, Louisiana, 206 28.1 31.4
Central Oklahoma, Texas

Compared with U.S. Census data (US Census Bureau, 2000), the responses did

appear biased towards Caucasians. A large percentage of the respondents in the survey,

88.2 percent, indicated they were Caucasian, followed by 8.0 percent Black or African-

American, 1.5 percent Asian, 2.3 percent Indian, and 3.7percent other. These compare to

the 2000 U.S. Census data, with approximately 75 percent of the U.S. population

Caucasian, 12.3 percent Black or African-American, and 4.2 percent Asian. The results

also indicated a bias towards non-Hispanic, with only 3.4 percent of the respondents

indicating they were Hispanic or Latino. This compares to the Census data, with 12.5









percent of Hispanic origin (U.S. Census Bureau, 2000). However, this is not uncommon

for survey responses on seafood consumption.

Figure 4-1 shows the percentage of U.S. population and survey respondents in

different age groups. Although discrepancies exist, survey respondents were somewhat

comparable to the population average, with 46.4 percent of the U.S. population over the

age of 45. Also, the survey respondents tended to have comparable household incomes to

those from the U.S. Census data (Figure 4-2). The median income of the survey

respondents fell in the $40,000 $59,999 category, compared to a U.S. median income of

$41,994. However, our survey respondents were more educated than average, with

approximately 40 percent of the sample having a four year college degree compared to 26

percent in the general population, according to the 2000 Census data. As a result,

interpretation of the results should be conducted recognizing the bias, and future studies

should make an attempt to focus on the underrepresented population (e.g. Hispanic

origin).

Overall, among the 733 responses, 58.4 percent (428 responses) indicated they had

consumed crawfish, 30.0 percent (220 responses) indicated they had consumed crawfish

in the last 12 months, and 25.8 percent (189 responses) indicated they had consumed

crawfish away-from-home during the last 12-month period. Only 11 percent (82

responses) indicated at-home consumption, and 4.2 percent (31 responses) indicated

consumed crawfish only at-home.

Tables 4-2 and 4-3 provide descriptive statistics for the survey, including the

respondents who had indicated crawfish consumption in the last 12-month period.

Respondents were asked to indicate the times they have consumed crawfish at-home and









away-from-home in the last 12 months. The frequency of away-from-home consumption

of crawfish is the sum of responses to the two questions about at-restaurant and at-other-

away-from-home consumption. Respondents who ate crawfish away-from-home

consumed crawfish, on average, 2.97 times (standard deviation 2.682) in the last 12-

month period.

Additionally, other factors included socio-economic and demographic variables

(age, ethnicity, household income, education, etc.), preferences for consuming or not

consuming, and knowledge about crawfish. Descriptive statistics for the survey sample

are shown in Tables 4-2 and 4-3.

















*U.S. Census
O Survey


18-24 25 to 34 35 to 44 45 to 54 55 &
above



Figure 4-1. Comparison of U.S. population in the Southeast and survey respondents by
age.


SU.S. Census
O Survey


$19,999 or $20,000 $40,000 $60,000 $80,000 or
less $39,999 $59,999 $79,999 greater


Figure 4-2. Comparison of incomes of all the U.S. population and incomes of the survey
respondents.










Table 4-2. Summary of demographic information.

Non- Non-Away- Away-From- Overall
Consumers From-Home Home
Consumers a b Sample
_Consumers _Consumers Consumersb Sample
Number of Observations 513 220 544 189 733
Age of Respondents % % % % %
Greater than 55 19.9 19.4 19.9 19.5 19.7
Between 45 and 55 26.7 25.2 26.5 25.9 26.3
Between 35 and 45 23.2 27.9 24.3 25.9 24.7
Between 25 and 35 22.2 19.4 21.7 20.6 21.4
Between 18 and 25 8.0 8.1 7.2 7.9 7.7
Gender
Percent Female 82.7 69.5 82.2 68.8 78.7
Household Income
Less than $19,999 15.4 11.7 15.1 12.2 14.3
Between $20,000 and 29.8 28.2 30.3 26.4 29.4
$39,999
Between $40,000 and 25.3 20.0 25.0 20.1 23.7
$59,999
Between $60,000 and 15.0 14.5 15.1 14.3 14.8
$79,999
$80,000 or greater 14.4 25.4 14.5 27.0 17.8
Region of Residence
Northern Southeast 32.0 11.8 30.9 11.6 26.0
Atlantic
Southern Southeast 28.3 28.6 27.6 30.7 28.3
Atlantic
East South Central 18.1 16.4 18.8 14.3 17.6
West North Central 21.6 43.2 22.8 43.4 28.1
Education
High School or less 45.8 37.3 46.0 35.4 43.2
Some College (2 Yr) 17.9 14.5 17.5 15.3 17.0
College Degree (4 Yr) 27.5 30.0 27.6 30.2 28.2
Above College 8.8 18.2 9.0 19.0 11.6
Degree
Number of Children (under 16 living at home)
No Children 61.8 57.7 60.8 59.8 60.6
1 or 2 Children 29.2 30.9 29.2 31.2 29.7
Above 3 Children 9.0 11.4 9.9 9.0 9.7
Household Size
1 Only 15.0 9.5 14.5 10.0 13.4
2 People 36.3 35.9 35.7 37.6 36.1
3 People 21.4 18.6 20.8 20.1 20.6
4 People 15.0 21.0 15.6 20.1 16.8
5 and above 12.3 15.0 13.4 12.2 13.1










Table 4-2. Continued.


Race (races are alone and in combination)
White 88.9 87.3 89.3 85.7 88.4
Black 8.8 5.9 8.5 6.3 7.9
Asian 1.2 1.8 1.1 2.1 1.4
Indian 2.0 3.2 2.0 3.2 2.3
Other Race 2.5 6.4 2.4 7.4 3.7
Religion
Christian" 51.9 53.2 52.8 50.8 52.2
Catholic 12.7 11.4 12.3 12.2 12.3
Jewish 2.1 1.8 2.0 2.1 2.0
Other Religions 4.5 4.1 4.4 4.2 4.4
No Religion 28.9 29.5 28.5 30.7 29.1
Hispanic Descendent
Hispanic 2.9 4.5 2.9 4.8 3.4
a Non-Away-From-Home consumers included non-consumers and only at-home consumers.
b Away-From-Home consumers included only away-from-consumers and away-from-home
consumers who also were at-home consumers of crawfish.
Christian including Protestant.










Table 4-3. Descriptive statistics on other factors included in the Double-Hurdle model.
Mean, Non Away- Mean, Away- Overall
From-Home (544) From-Home (189) Mean (733)
Frequency of Consumption (times in the last 12 months):
Away-From-Home 0 2.97 (2.682)a 0.77 (1.882)
At-Home 0.18(1.32) 0.72 (1.611) 0.32 (1.423)
Knowledge of Crawfish % % %
Knowing the difference between 73.5 89.5 77.7
crawfish and lobster
Knowing the difference between 25.6 45.0 30.6
crawfish and langostino
Indicated the following was one of the top three reasons for consuming
Enjoy flavor 5.2 86.4 26.3
Health/Nutrition 0.9 14.1 4.4
Tradition/Habit 1.7 21.5 6.8
Price is attractive 1.1 14.1 4.5
Availability 1.3 23.6 7.1
Variety in diet 2.0 37.7 11.3
Prefer Cajun style 3.1 44.0 13.7
Indicated the following was one of the top three reasons for not consuming:
Price too high 24.8 58.2 33.3
No preferred product form 32.4 12.0 27.1
available
Not part of custom 20.8 20.0 20.5
Dislike taste 27.4 5.8 21.8
Dislike smell 31.1 16.9 27.5
Lack of domestically produced 26.5 22.0 25.3
product
Product safety concerns 9.6 20.1 12.3
Bad impression 31.4 12.6 26.5
Health concerns 6.1 5.8 6.0
a Standard deviations were reported in the parenthesis.















CHAPTER 5
THEORETICAL MODEL AND MODEL SPECIFICATION

Theoretical Model

Neoclassical demand theory suggests own price, income, the prices of substitutes

and complements, socio-economic and demographic variables are determinants of

demand. Euler's theorem for homogeneous demand functions shows (Nicholson, 2002):

ex, Px + ex, Py +* + ex, P + ex, I 0

where e,P, is the own-price elasticity of demand, e,py and e,p and all cross-price

elasticities in-between are cross-price elasticities of demand with respect to all other

goods, and ex, is the income elasticity of demand. This property of the demand function

will give insight into crawfish consumption analysis, although we did not calculate those

elasticities in quantitative terms. In this study, the sensitivity of x (crawfish) to price (Px,

P,, P,) and income levels will be evaluated in general qualitative terms. However,

consumption of substitute products (such as other seafood) was not available. With

information about the sensitivity ofx (crawfish) to own-price and income levels, we can

infer a potential relationship between crawfish and its substitutes.

In this study of crawfish consumption patterns, the main forms of crawfish

consumed away-from-home are whole-boiled crawfish and tail meat. Different product

forms may have different prices, which are difficult to aggregate. Moreover, it is hard to

get aggregate prices because of the seasonality of the crawfish harvest and variety of

consumption locations (at-home vs. away-from-home). The main purpose of this study is

to develop an understanding of the factors influencing away-from-home consumption of









crawfish, not to derive the demand function for crawfish. As a result, we did not include

prices as a quantitative independent variable. However, in order to examine the impact

price has on away-from-home consumption of crawfish and to examine the own-price

elasticity in comparative terms, we include price as a binary variable. If respondents think

prices are the main reason for their participation and/or consumption decision, the value

of the variable would be one; otherwise, it would be zero.

This study will examine the factors influencing away-from-home consumption of

crawfish in the Gulf region using data from a web-based consumer survey conducted in

late January of 2004. The survey instrument was discussed in the previous chapter. The

frequency of crawfish consumed away-from-home in the last 12-month period is used as

the dependent variable. Respondents were asked to indicate how many times they

consumed crawfish away-from-home in the last 12 months. As mentioned before, the

reasons for a 12-month period includes seasonality of crawfish consumption and overall

low levels of consumption in one year. In this study, consumers, on average, ate crawfish

away-from-home three times during the last 12 months. If asking for one month, the

average times could be greater than, equal to or less than the annual average of three

times.

The frequency of at-home consumption of crawfish is included to check the

relationship between at-home and away-from-home consumption and to infer if increases

in away-from-home consumption would occur at the expense of at-home consumption of

crawfish. Factors, including stated consumer preferences and perceptions regarding

crawfish, also are included as explanatory variables. For example, the survey includes

questions about the top three reasons for consuming or not consuming crawfish









(respondents were asked to select the top three reasons from a list, including prices,

availability, tradition, variety, flavor and taste, safety concerns, and health reasons).

Moreover, socio-economic and demographic data, such as age, education level, income,

gender, and ethnicity, will be examined to determine their impact on participation and

consumption decisions.

Censoring and truncation of the dependent variable is a very common problem in

survey data. In previous seafood consumption studies (Keithly, 1985; Lin and Milon,

1993; Yen and Huang, 1996; House et al, 2003), researchers found significant proportions

of households with zero observations. Conventional regression methods fail to account

for the qualitative difference between limit (zero) observations and non-limit (positive)

observations. Early studies with limited dependent variables used the Tobit model (Tobin,

1958). For example, Keithly (1985) used a Tobit model to accommodate the problem of

zero consumption.

The Tobit model specifies that

P' xi +, N(0, a2)

y, = 0 ify, <0,

y, = y* ify,* >0

where y,* is the latent variable, and y, is the dependent variable. The model implies that y,

will only be positive given a value of y, greater than zero. The Tobit model assumes the

factors that affect the level of consumption are the same as those that determine the

probability of consumption. See Amemiya (1984) for a detailed review.

Although the Tobit model might be consistent with consumer behavior and might

be useful in studying the consumption patterns of households, it is restrictive in parameter









estimation because it is assumed that both the probability and the level of consumption

are affected identically by the same factors. The Tobit model assumes that the variables

and estimated parameters determine the consumption level as well as the participation

probability. This might not be true. Respondents may have the potential to participate;

however, other reasons might curb consumption. For example, safety issues might not

inhibit shellfish consumption, but reports of contaminated shellfish products might do so

temporarily. This implies that participation and consumption decisions might depend

upon different explanatory variables and parameters, which might have opposite effects

(Lin and Milon, 1993).

Model Specification

The restriction of the Tobit model has been recognized in demand analysis for food

and seafood, including finfish and shellfish, and has consistently been rejected in recent

seafood studies. Cheng and Capps (1988), Lin and Milon (1993), Yen and Huang (1996),

Drammeh et al. (2002), and House et al (2003) all realized the restrictions of using a

Tobit model in demand analysis of seafood. They found that the decision to consume a

seafood product has determinants that are independent of the level of consumption. For

instance, income is expected to have opposite effects on participation and the

consumption of crawfish. Based upon empirical results of previous studies, lower income

households tended to have a higher probability to consume crawfish; however, lower

income households tended to have lower levels of consumption than other crawfish

consumers.

To accommodate the restrictions of using a Tobit model in analysis for seafood

consumption, a double-hurdle model proposed by Cragg (1971) has been used

extensively in recent studies. In the Tobit model, a variable that increases the probability









of an observation also increases the quantity of consumption simultaneously. In the

model proposed by Cragg, the probability of an observation is independent of the

regression model of the quantity consumed.

Cheng and Capps (1988) used Heckman's two-step procedure, and Yen and Huang

(1996) used a generalized double hurdle model to analyze household demand for finfish.

Lin and Milon (1993) used a count-data double-hurdle model to examine the impacts of

attributes and food safety perceptions of seafood consumption. Drammeh et al. (2002)

and House et al. (2003) also used a double hurdle model similar to Yen and Huang (1996)

to analyze U.S. oyster and catfish consumption, respectively. There are several different

but related models in the mentioned literature; however, all are specified from Cragg's

double-hurdle model.

For crawfish consumption, Yen et al. (1995) also found the complicating feature of

household survey data, which is the significant proportion of zero observations. A large

quantity of zero observations cannot be treated as true non-consumption because zero

observations may be due to, besides non-consumption, infrequency of consumption or

conscientious abstention, which could have other behavioral explanations (Yen et al.,

1995). As a result, the Tobit model also was rejected. The authors separated the sets of

parameters in the probability of consumption and level of consumption equations. To

allow for violation of the distribution assumption of Cragg's double hurdle model, they

used the Box-Cox transformation on the dependent variable, which is an extension of the

double-hurdle model.

We show the benchmark model below using the double-hurdle model framework

specified by Cragg (1971):









Individual i's participation equation can be expressed as

d,* = zia + v, with d,= 0 or 1

Individual i's consumption equation can be expressed as

y = xiP + u,

where y, represents the latent consumption decision, and d, is a latent variable

describing participation. zi and xi are vectors of exogenous variables, and a and P are

parameter vectors. Random errors u, and v, are normally distributed as N(O, 1) and N(O,

a), respectively. It also is assumed that u, and v, are independent.

The double-hurdle model has separate participation and consumption equations that

are related as:

y, = y,* if y,* >0 and d,*>O

= 0 otherwise.

where observed consumption y, relates to latent consumption y,*, only ify,* >0 and d,*>0.

It also is the conditional decision to consume the product (y,* >0).

The double-hurdle model includes two equations: the consumption equationy* and

the participation equation d,*. Thus, the probability of consumption and level of

consumption are determined by separate sets of parameters (a and P). The use of the

same sets of variables, zi and xi, in both equations is not restrictive because these

variables affect consumption and participation differently through the different parameter

coefficient (a and p), and might have opposite effects.

The combination of the above model can be estimated in two parts. A Probit model

can be used to estimate the participation equation, and the parameters of the consumption

equation can be estimated independently using the truncated regression model.









The maximum likelihood estimation of a Probit model is used to evaluate the

censoring rule (zia), given a normal distribution. Since we impose the assumption of a

normal distribution, if y,*>0, then a truncated regression (i.e., truncated Tobit) applies.

Thus, another maximum-likelihood estimation that accounts for a truncated normal

distribution is used for the truncated regression.

The likelihood function for the double hurdle can be derived as

L = YProb(d = 01 zi,a) Prob(d =1 zi,a) f (y, xi,P)
0
where the 0 under the product sign indicates products over those observations which are

zero, and the + signs under the product sign indicates products over those observations

which are positive, y, >0. For simplicity, Prob denotes the probability, andf(y, -) is the

conditional (truncated) density of y,, given y,>0. a and P are vectors of parameters. As we

indicated, the assumption of independence is imposed, with the participation and

consumption decisions as a binomial probability distribution and truncated at zero

conditional distribution, respectively. The structure of the data determines the specific

forms of the distributions. Because it is beyond the scope of this study, we will not cover

the discussion in this study.

As indicated above, the double hurdle model allows the parameters in the Probit

equation to differ from those in the Tobit model, so that the complete model is a Probit

model for d and a separate truncated regression model for the positive values of y. It is

also assumed in this study that the participation and frequency of away-from-home

crawfish consumption are independent of each other, and the probability and frequency

of crawfish consumption are determined by separate sets of parameters. Since the Tobit









log likelihood is simply the sum of the Probit and truncated regression log likelihood

(Greene, 1995), a test of the Tobit model as a restriction is implemented.

The chi-squared specification test (Greene, 1995)

X = 2(fTobit -fProbit -/Truncated)

is used to determine whether the double hurdle model is a better fit than the Tobit

specification, where is the log-likelihood function value. It is testable with the

likelihood ratio test by estimating the three implied models, with the specific degrees of

freedom, which is equal to the number of variables in x. If the null hypothesis is rejected,

we can conclude that the double hurdle model is a better fit than a Tobit model.

The model was estimated by maximizing the logarithm of the likelihood function

using a modified double hurdle model (Cragg, 1971), with the frequency of away-from-

home consumption of crawfish as the dependent variable. In practical terms, the Probit

model and the truncated Tobit model as well as the compound Tobit model have been

incorporated in many software packages (i.e., LIMDEP, SAS, and TSP). As a result, the

estimation is now essentially on the level of ordinary regression, and will not be covered

in this study.

The marginal effects of the independent variables (similar to the Tobit model), on

the probability of participation, and the conditional level of consumption can be derived

by differentiating the probability and conditional mean, di(z)/l, and ci(y*|y*>0)/a,,

where F(z) and E(y*|y*>0) are the probability of participation and the conditional mean of

consumption, respectively, and x is the specific independent variable. The marginal

effects on the unconditional mean of consumption, which can be associated with all

observations, can be derived by ci(y)/cc, where E(y)=F(z)-E(y*|y*>0). The marginal









effects on the unconditional level evaluate what contributes to the consumption level by

increasing or decreasing either the probability or conditional level. The marginal effects

are computed by LIMDEP "by computing the derivatives for each observation and

averaging the derivatives" (Greene, 1995). Although the approach ignores the fact that

some variables are binary, the results are almost exactly the same as the results of

computation for binary variables using numerical methods (Greene, 1999).

In this study, the dependent variable is the frequency of away-from-home

consumption of crawfish in the last 12-month period. Based upon the discrete nature of

the dependent variable, it is a count data variable. Consequently, instead of using a

continuous double hurdle model, a count data double hurdle model should be considered.

In this study, a count data double hurdle model developed by Mullahy (1986), which is

similar to Lin and Milon's (1993) and Bilgic and Florkowski's (2003), is used.

Truncated-at-Zero Count Data Double Hurdle Model

Because the dependent variable, the frequency of away-from-home consumption of

crawfish, is a count, our analysis adopted the count data double-hurdle model similar to

Lin and Milon's (1993) and Bilgic and Florkowski's (2003). A good review of the

modified count data model for the hurdle model was developed by Mullahy (1986).

According to Bilgic and Florkowski (2003), because of the possible overdispersion

in the observations, the assumption of the familiar Poisson distribution truncated at-zero

might not always fit the sample. However, in this study, we tested the overdispersion and

compared the Poisson and negative binomial distributions of both zero and positive

counts.









Compared to the Poisson model, the negative binomial model is not encumbered by

the Poisson's mean=variance nature (var[Y] = E[Y]) (Mullahy, 1986). The variance of a

negative binomial variant can exceed its mean, and the relationship between population

mean and variance of the negative binomial distribution is defined as

var[Y]= E[Y( + OE[Yb
For both the Poisson and negative binomial models, a hurdle (truncated-at-zero)

specification is derived. See Mullahy (1986) for a detailed discussion.

As the result of Mullahy's modification, the binomial probabilities (the

participation decision) are identical to those of a standard binomial Logit model. Thus, it

is appropriate to test both Probit and Logit models for the binary probabilities.

The consumption decision distributes as either a truncated-at-zero Poisson or a

truncated-at-zero negative binomial model, depending on the overdispersion. Again, the

specified model is estimated using standard maximum likelihood techniques, and ML

estimates can be obtained by maximizing the log likelihood function. The compound

model (single decision) and truncated-at-zero model have been incorporated in several

software packages (i.e., LIMDEP). It is worth mentioning here that we included all the

same determinants regarding specific perceptions and preferences, socio-economics, and

demographics, since we have no basis to exclude any variable in either decision.














CHAPTER 6
EMPIRICAL RESULTS

Issues of Crawfish Consumption

Regional Consumption

Crawfish consumption tended to vary by region of residence in the targeted

Southeast and the Gulf region (Figure 6-1). Visual inspection shows that consumers in

the West South Central region (AR, LA, OK, TX) are most likely to consume crawfish

both away-from-home and at-home (Figure 6-1). Overall, 39.8 percent of the

respondents from the West South Central region consumed crawfish away-from-home,

compared to 11.6 percent in the northern South Atlantic region (DE, DC, MD, NC, VA,

WV). Chi-squared tests on significance are included below the table. Chi-square

probabilities below 0.05 indicate a significant difference in the variables. In Figure 6-1,

the chi-square probability of 0.001 indicates crawfish consumption is significantly

different in the different regions. The result is consistent with our expectation, since the

majority of crawfish produced in the United States is from the West South Central region

where the consumption rate is expected to be higher than in other regions. The southern

part of the South Atlantic region, including Florida, Georgia, and South Carolina, ranked

second in crawfish consumption.

Farm-Raised Crawfish

Farm-raised crawfish production is mainly in the lower Gulf Coast region, such as

Texas, Mississippi, Florida, and Louisiana, with most of the acreage found in Louisiana

(Avery and Lorio, 1999). In this study, crawfish consumers were asked if they were










aware of whether the crawfish they had eaten was farm-raised or not. Only 29 percent of

the crawfish consumers answered they were aware, but among those respondents, over 80

percent had eaten farm-raised crawfish in the previous 12 months. Among those

respondents who had eaten farm-raised crawfish, over 94 percent indicated they would

eat farm-raised crawfish again. Among the remaining crawfish consumers who had not

eaten, or were not aware they had eaten, farm-raised crawfish, over 67 percent indicated

they would be willing to consider consuming farm-raised crawfish. This implies that it

might be appropriate to use educational approaches to promote the farm-raised crawfish

and let the consumers or potential consumers have more knowledge about farm-raised

crawfish.


50
45
40
35 -
30
2 25
20 -
15 -
10


0
Northern South Southern South East South Central West South Central
Atlantic Atlantic

Away-From-Home Consumption 0 Overall Consumption

Note: Chi-square probability < 0.001.
Figure 6-1. Regional percentage of crawfish consumption.

Domestic Versus Imported Crawfish

Wal-Mart in Louisiana carries crawfish produced in China, which even the most

dedicated economists espousing comparative advantage had not forecast. Do consumers










really like imported crawfish over the domestic crawfish? As a special interest of this

study, crawfish consumers were also asked to indicate whether they prefer domestically

produced or imported crawfish. The survey indicates that although 58 percent of the

crawfish consumers did not know where the crawfish they had consumed came from,

about 52 percent of the consumers indicated they preferred locally produced and only six

percent indicated they preferred imported crawfish. A relatively large percentage of

consumers had no opinion, meaning they either do not know or do not care (Figure 6-2).

This implies that properly promoted domestically produced crawfish, targeting those

crawfish consumers, might win back the market share lost to the imported product.


60

50 -

40 -

| 30 -

2 0 ------------

10 -

0
Preferred Not Preferred Do Not Care Do Not Know

Domestic Crawfish 0 Imported Crawfish


Figure 6-2. Comparison of preferences over domestic and imported crawfish products.

Product Form

Since most of the crawfish imported are in the form of tail meat, and domestically

produced tail meat is usually high in price, the product form might explain the reason

why imported crawfish have an increasing market share. In the survey, crawfish









consumers were asked to identify what form of crawfish they prefer: whole crawfish or

tail meat? Live whole crawfish and boiled whole crawfish as well as tail meat are the

main forms of crawfish products normally found on the market. However, in order to

simplify the question and focus on away-from-home consumption, we included only

whole and tail meat product form in this study. The majority of existing crawfish

consumers (72%) indicated that they preferred tail meat to whole crawfish in away-from-

home consumption. Interestingly, although the price of tail meat is comparatively higher

than the price of whole crawfish, a large percentage of consumers (71%) preferred tail

meat to whole crawfish in at-home consumption. This might be caused by the trend of

consumer preferences towards convenience. Also, this implies that the introduction of

new crawfish products targeting consumer trends, such as convenience and health

concerns, could encourage consumption or more consumption to some extent. Domestic

crawfish producers and processors can design marketing strategies accordingly to

compete with imported products.

Additionally, existing crawfish consumers were asked if they preferred fresh or

frozen crawfish. The majority (75%) indicated that they preferred fresh crawfish products,

with only one percent preferring the frozen form and 24 percent having no preference.

This is consistent with the major consumer trend of consuming fresh products. This

implies that domestically produced fresh crawfish can be competitive with imported

frozen products if prices are reasonable. However, it is believed that existing fresh

crawfish products are far more expensive compared to those imported products.









Reasons for Consuming and Not-Consuming

Reasons for consuming

In addition to the frequency of away-from-home and at-home consumption and the

demographic variables, respondents were asked to identify their main reasons for

consuming crawfish. Results from the 220 crawfish consumers (including the 189 away-

from-home crawfish consumers) who responded to this question are illustrated in Figure

6-3. As indicated by about 87 percent of the crawfish consumers and the away-from-

home crawfish consumers, the principal reason for consuming crawfish was enjoyment of

flavor. Next was preference for Cajun style, followed by adding variety to diet (Figure 6-

3). It is worth mentioning that there was no statistically significant difference between

away-from-home consumers and overall crawfish consumers (Chi-square probabilities

ranging from 0.2817 to 0.8813).

Among the 220 crawfish consumers, 138 respondents indicated they only

consumed crawfish away-from-home, 31 indicated they only consumed at-home, and 51

indicated they consumed both away-from-home and at-home. Results from these three

types of consumers who responded to this question are illustrated in Figure 6-4. As

indicated, although the percentage levels are somewhat different for all three types of

consumers, the principal reason for consuming crawfish was enjoyment of flavor. Next

was preference for Cajun style, followed by adding variety to diet (Figure 6-4).

Reasons for not consuming

Both consumers and non-consumers of crawfish were asked to identify the top

three reasons for either not consuming or not consuming more crawfish (Figure 6-5). For

non-consumers, lack of availability of the preferred product form, bad impression, and

smell were the top three reasons for not consuming; however, taste also was ranked high





































SAway-From-Home Consumption 0 Overall Consumption

Note: Chi-square probabilities ranging from 0.2817 to 0.8813
Figure 6-3. Categories of reasons given for consuming crawfish (crawfish consumers vs.
away-from-home consumers).


C) C)-' .
o 4
"0
-e, C)
* ~ e n C) C)-
a e e, C)
W a rH


MAway-From-Home Only OAt-Home Only OAll Crawfish Consumers

Note: Chi-square probabilities between AFH-Only and AH-Only ranging from 0.1707 to 0.9787
Figure 6-4. Categories of reasons given for consuming crawfish (away-from-home
consumers only, at-home consumers only, and all crawfish consumers).










in the reasons for not consuming crawfish. In contrast, crawfish consumers gave

significantly different responses to the questions (Chi-square probabilities < 0.0001).

Price is too high, followed by lack of availability of domestic products and custom were

the top three reasons for not consuming among existing crawfish consumers.


70

60

50

S40

30

20 -

10









Crawfish Consumers O Non-Consumers



crawfsh (consumers vs. non-consumers).
0 .






sh by those nonCrawfish Consumers Non-Consumers

Note: Chi-square probabilities < 0.0001
Figure 6-5. Categories of reasons given for not consuming or not consuming more
crawfish (consumers vs. non-consumers).

We also illustrate the top three reasons for not consuming for away-from-home

consumers and non-away-from-home consumers separately. For non-away-from-home

crawfish consumers, unavailability of product form, bad impression, and smell also were

the top three reasons for not consuming. Taste also was ranked high in the reasons for not

consuming crawfish by those non-away-from-home consumers. Away-from-home

consumers gave different responses towards the question. Price is too high, followed by

unavailability of domestic products and safety concerns were the top three reasons for not









consuming by existing away-from-home consumers. No custom or tradition became the

fourth important reason for not consuming crawfish.

It is worth testing the relationship of reasons for not consuming among non-

crawfish consumers, only away-from-home consumers, and only at-home consumers

(Figure 6-6). For non-crawfish consumers, unavailability of product form, bad

impression, and smell were the top three reasons for not consuming. Taste also was

ranked high in the reasons of not consuming crawfish. Consumers who only consumed

crawfish away-from-home gave different responses towards the question. Price is too

high, followed by unavailability of domestic products and no custom were the top three

reasons. Safety concerns became the fourth important reason for not consuming crawfish

among away-from-home consumers. Interestingly, consumers who only consumed

crawfish at-home gave somewhat different responses to the question. Price is too high

and unavailability of domestic products ranked much higher than the other reasons. No

custom and smell ranked third and fourth. However, safety concerns was not as important

for at-home consumers, who might have been influenced by the fact that they prepare the

crawfish themselves.

Away-From-Home Crawfish Consumption

Consumers were asked to identify how many times they had consumed crawfish

both at-home and away-from-home in the previous 12-month period. However, to focus

on away-from-home consumption, we treat the frequency of away-from-home

consumption as the dependent variable in the count data double hurdle model. Because of

its discrete nature, we reported the sample frequency distribution of the dependent

variable (FREAFH) in Table 6-1. The table shows the number of times a consumer

consumed crawfish away-from-home in the previous 12-month period. Average







48


consumption of the overall sample (733 responses) was 0.8 times; average consumption

of the 189 away-from-home crawfish consumers was 3.0 times.


70

60

50 -

2 40 -

30

20
10



H AF- nI O HOI IO No n-CEsu.
0


B i u! ,a 'S5
-9-



EAFH-Only DAH-Only O Non-Consumers

Note: Chi-square probabilities between AFH-Only and AH-Only ranging from 0.1223 to 0.9924.
Figure 6-6. Categories of reasons given for not consuming or not consuming more
crawfish (away-from-home only, at-home only, and non crawfish consumers).

Table 6-1. Sample frequency distribution of the dependent variable (n=733).
Away-From-Home Number of Consumers
Consumption (times) (n=733)
0 544 74.2
1 54 7.4
2 63 8.6
3 24 3.3
4 14 1.9
5 18 2.5
6 3 0.4
7 1 0.1
8 1 0.1
10 5 0.7
12 3 0.4
15 3 0.4
Note: Overall 733 respondents, values indicate the times of away-from-home crawfish
consumption in the previous 12 months period.









The following model was estimated using a count data double hurdle model

developed by Mullahy (1986), with the frequency of away-from-home consumption of

crawfish as the dependent variable. The dependent and independent variables are

described in Table 6-2.

As mentioned in the previous chapter, an interesting and distinctive feature of the

hurdle model is that the probability of participation is assumed to be independent of the

level of consumption. The count data double hurdle model consists of two equations

describing participation and consumption decisions separately. In this study, the

participation equation depicts the decision of whether to be an away-from-home crawfish

consumer, and the consumption equation refers to how many times to consume as an

away-from-home crawfish consumer.

For the participation decision stage, we used a Logit model to estimate the

probabilities of being an away-from-home crawfish consumer. The coefficients from the

Logit model (participation decision) as well as the marginal effects (calculated at mean

value) are reported in Table 6-3. The Logit model correctly predicts a consumers'

likelihood of being or not being an away-from-home crawfish consumers 93.7 percent of

the time (incorrectly predicted being a consumer 3.4 percent of the time and not being a

consumer 2.9 percent of the time). This can be compared to a naive prediction, which

would result in correctly predicting away-from-home crawfish consumption 74 percent of

the time. The prediction rate supports the appropriateness of using a Logit model (with

Pseudo- R2=0.67).

In the second stage, the dependent variable is a count variable for the number of

times crawfish was consumed away-from-home in the previous 12 months. We have













Table 6-2. Description of variables included in the double-hurdle model.
Variants Variable Name Description
Consumption of Crawfish FREAFH Frequency of crawfish consumption away-from-home
FREAH Frequency of crawfish consumption at-home
Knowledge of Crawfish KNLAN 1 if respondent indicated being aware of the difference between crawfish and
langostino
Reasons for consuming The following variables are 1 if this reason was listed as one of the top three reasons for consuming crawfish:
crawfish


Reasons for not consuming, or
not consuming crawfish more
frequently










Region of residence (Gulf
region)


FLAVOR Enjoy flavor
HEALTH Health/nutrition
TRAHABIT Tradition/habit
YESPRICE Price
AVAILTY Availability
FARMRSD Farm raised crawfish
VDIET Variety in diet
CAJUN Prefer Cajun Style
The following variables are 1 if this reason was listed as one of the top three reasons for NOT consuming, or not
consuming MORE crawfish:


NOPRICE
NOFORM
NOCUSTOM
NOTASTE
NOSMELL
NOLOCAL
NOSAFE
NOIMPR
NOHEALTH
REGION 1

REGION
REGION
REGION


Price
Lack of preferred product form
Custom
Dislike taste
Dislike smell
Lack of availability of domestically produced products
Product safety concerns
Bad impression
Health concerns (allergy)
Northern Southeast Atlantic (DE, DC, MD, NC, VA, WV)

Southern Southeast Atlantic (FL, GA, SC)
East South Central (AL, KY, MS, TN)
West South Central (AR, LA, OK, TX) (omitted category)













Table 6-2. Continued.


Variants
Race/Ethnicity




Income




Education



Age




Employment Status


Variable Name
WHITE
BLACK
ASIAN
INDIAN
OTHRACE
INCOME
INCOME
INCOME
INCOME
INCOME
EDUCAT 1
EDUCAT2
EDUCAT3
EDUCAT4
AGE1
AGE2
AGE3
AGE4
AGE5
EMPLOY 1
EMPLOY2
EMPLOY
EMPLOY4
EMPLOY
EMPLOY6


Description
1 if Caucasian, 0 otherwise
1 if Black, 0 otherwise (omitted category)
1 if Asian, 0 otherwise (omitted category)
1 if Indian, 0 otherwise (omitted category)
Other race/ethnicity (omitted category)
<$19,999 (omitted category)
$20,000 $39,999
$40,000 $59,999
$60,000 $79,999
$80,000 or above
High School degree or less (omitted category)
2 year college degrees
Degree from college (4 year)
Above college degree
Ages 18-24 (omitted category)
Ages 25-35
Ages 35-45
Ages 45-55
Age 55 or older
1 if currently full time employed
1 if currently part time employed (omitted category)
1 if currently unemployed (omitted category)
1 if student (omitted category)
1 if house worker without pay (omitted category)
1 if retired (omitted catesorv)


"









considered both truncated-at-zero Poisson and truncated-at-zero Negative Binomial

models for estimation. However, after we tested the Poisson process, the overdispersion

test indicated that the assumption of the Poisson distribution truncated-at-zero could not

be rejected and might fit the sample.' Therefore, we report the results from the truncated-

at-zero Poisson model. Both coefficient estimates and the marginal effects are reported

in Table 6-3 as well as the estimates from a Logit model.

To test the compound (single-decision) count data model (standard Poisson) against

the hurdle count model, we use Vuong's t-test. The standard normal statistics were

calculated. For Vuong's t-test, a value greater than 1.96 favors the altered model. In this

case, the value of 5.66 (given by LIMDEP) was obtained for the null hypothesis that the

standard Poisson and its hurdle Poisson modification are the same. In other words,

Vuong's statistic favors the hurdle Poisson model. The result shows that splitting the

count model is preferred to the standard count model. We also included the single

decision (standard Poisson) estimates in Table 6-3 for comparison.

As expected, results indicated that variables affected the decision to consume

crawfish away-from-home and the frequency of consumption differently. If consumers

ate crawfish at-home more frequently, they were significantly less likely to be away-

from-home crawfish consumers, but their frequency of away-from-home consumption

increased. In other words, as the frequency of at-home consumption increases, the

likelihood to consume crawfish away-from-home decreases. A possible reason for the





1 Overdispersion test: Ho: var[y,]=[; H, var[y,]=[ +ag([1) with g(p1)=[ t-statistic=1.453<1.96; the result
could not reject the null hypothesis at 5% level. See LIMDEP Manual (Greene, 1995) for the
overdispersion test in Poisson regression.










Table 6-3. Empirical results of single-decision and double-hurdle: maximum-likelihood
estimates and marginal effects.
Single
siongle Double Hurdle Count Data
Decision
Truncated
Standard Logit Marginal Truncated Marginal
Poisson Coefficient Effects Poisson Effects
Coefficient
Frequency of Consumption:
Crawfish at-home -0.032*** -0.412* b 0.052**
(0.020) a (0.125) (0.028)
Knowledge about crawfish:


Crawfish and 0.322*
langostino (0.094)
Indicated the following was one of the


Enjoy flavor


Health/Nutrition

Tradition/Habit

Price is attractive

Availability

Farm-raised available

Variety in diet

Prefer Cajun cooking


2.874*
(0.180)
0.234***
(0.134)
0.224**
(0.115)
0.184
(0.137)
0.201***
(0.114)
-0.232
(0.210)
-0.033
(0.105)
0.249*
(0.097)


0.254 0025
(0.357)
top three reasons for consuming


3.352*
(0.457)
1.241***
(0.736)
0.851
(0.594)
1.522**
(0.787)
1.552*
(0.607)
0.195
(0.915)
2.135*
(0.507)
2.078*
(0.498)


0.327*

0.121***

0.083

0.149**

0.152*

0.019

0.208*

0.202*


0.339*
(0.116)

0.559*
(0.198)
0.189
(0.152)
0.131
(0.129)
-0.158
(0.164)
-0.049
(0.135)
-0.123
(0.240)
-0.306*
(0.124)
0.055
(0.109)


Indicated the following was one of the top three reasons for not consuming:


Price too high


No preferred product
form available
Not part of custom

Taste

Smell

No domestic products
available
Product safety
concerns
Bad impression

Health concerns


-0.241*
(0.096)
-0.159
(0.151)
0.092
(0.125)
0.023
(0.225)
-0.198
(0.141)
-0.137
(0.116)
0.089
(0.127)
-0.365*
(0.149)
0.563*
(0.180)


0.169
(0.371)
-0.603
(0.466)
-0.287
(0.428)
-0.246
(0.566)
-1.145*
(0.478)
-1.141*
(0.408)
0.243
(0.510)
-0.910**
(0.484)
0.790
(0.589)


0.017

-0.059

-0.028

-0.024

-0.111*

-0.111*

0.024

-0.089**

0.077


-0.356*
(0.111)
-0.142
(0.186)
0.091
(0.150)
0.276
(0.274)
-0.114
(0.176)
-0.067
(0.135)
0.069
(0.149)
-0.289***
(0.184)
0.344***
(0.207)


0.547*


0.902*

0.305

0.212

-0.255

-0.079

-0.020

-0.494*

0.088


-0.574*

-0.230

0.147

0.446

-0.185

-0.108

0.111

-0.467***

0.556***










Table 6-3. Continued.


Variable


Demographics
Caucasian

2 year College degree

4 year College degree

Postgraduate

Between 18 and 24

Between 25 and 34

Between 35 and 44

Between 45 and 54

Northern Southeast
Atlantic
Southern Southeast
Atlantic
East South Central

$20,000 $39,999

$40,000 $59,999

$60,000 $79,999

$80,000 and above

Full time employed

Constant

Log-Likelihood
Correct prediction in
Logit (%)
Vuong's statistic


Single
Decision

Standard
Poisson


-0.423*
(0.131)
0.056
(0.147)
-0.008
(0.118)
0.203
(0.137)
-0.142
(0.226)
-0.335**
(0.163)
0.055
(0.143)
0.056
(0.138)
-0.572*
(0.180)
0.131
(0.112)
0.201
(0.127)
-0.540*
(0.161)
-0.643*
(0.178)
-0.456*
(0.173)
-0.703*
(0.173)
0.411*
(0.105)
-1.511*
(0.268)
-599.79


Double Hurdle Count Data

Truncated
Logit Marginal Psn
Co ff ci nt Ef ec s Poisson
Coefficient Effects Pis
Coefficient


-1.090*
(0.431)
0.350
(0.470)
-0.067
(0.430)
0.499
(0.608)
-0.163
(0.651)
-0.741
(0.531)
-1.289*
(0.518)
-1.022**
(0.511)
-0.936**
(0.493)
-0.218
(0.394)
-0.596
(0.523)
-1.332*
(0.495)
-1.125**
(0.551)
-1.256**
(0.607)
-0.665
(0.619)
0.468
(0.374)


-137.95


-0.489*
-0.106* (0.145)
(0.145)
-0.129
(0.175)
-0.017
-0.007 -0.017
(0.143)
0.114
0.049 (0.1
(0.162)
-0.402
-0.016 0.
(0.279)
-0.511*
-0.072 -0.511*
(0.193)
0.175
-0.126* 0.162)
(0.162)
0.146
(0.157)
-0.545*
-0.091** -0.545
(0.229)
0.063
-0.021 0063
(0.128)
0.248***
(0.150)
-0.648*
-0.130* -0
(0.192)
-0110* --0.675*
(0.209)
-0.349***
-0.123** -0.349
(0.202)
-0.856*
-0.065
(0.209)
0.425*
0.046
(0.125)
1.264*
(0.332)
-340.45


Marginal
Effects


-0.789*

-0.208

-0.027

0.184

-0.649

-0.824*

0.282

0.236

-0.879*

0.102

0.340***

-1.045*

-1.089*

-0.564***

-1.382*

0.687*

2.040*


93.7


5.66


a Standard errors of the coefficients are reported in parentheses.
b One, two, and three asterisks indicate significance at the 0.01, 0.05 and 0.10 levels, respectively.









result is that at-home crawfish consumers might be dedicated crawfish consumers, and

they might have preparation knowledge. Compared to eating out, at-home consumption

could be cheaper, and thus made them prefer at-home consumption. However, due to the

fact that they might be dedicated and more frequent crawfish consumers than those who

usually ate crawfish at restaurants, their frequency of away-from-home consumption

increases as their frequency of at-home consumption increases.

Our results indicated that for each one-unit increase in at-home consumption of

crawfish (one unit equals one time in the last 12 months), the respondents were four

percent less likely to be away-from-home crawfish consumers. The average crawfish

consumer ate crawfish away-from-home 2.97 times in the 12-month period. For each

one-unit increase in at-home consumption, respondents increased away-from-home

consumption significantly by 0.08 to 3.05 times.

We did not include the test of the relationship between crawfish consumption and

other seafood consumption (consumption of substitutes) in this study. Further research

could be done to see whether an increase in crawfish consumption will occur at the

expense of other seafood consumption and whether an increase in away-from-home

crawfish consumption will occur at the expense of at-home consumption of other seafood

products. However, in this study we can infer a potential relationship between crawfish

and its substitutes from the information about the sensitivities of crawfish to own-price

and income levels, and the discussion will be covered in the next chapter.

Consumers who know the difference between crawfish and langostino (more

knowledgeable about crawfish) consumed crawfish away-from-home significantly more

frequently. Nevertheless, it did not affect the likelihood of being a crawfish consumer.









Knowledgeable respondents consumed crawfish 0.55 times more often away-from-home

(3.52 times/year). This implies that using educational programs to promote crawfish

could increase away-from-home consumption, but not persuade non-consumers to

consume crawfish.

Variables representing the respondents' top three reasons for consuming crawfish

as well as the top three reasons for not consuming were included in the model and

provided insight for preferences on away-from-home consumption of crawfish. The

results indicated that those reasons significantly affect both the likelihood to be an away-

from-home consumer and the frequency of away-from-home consumption.

Our results indicated that consumers were more likely to consume crawfish away-

from-home (15 percent more likely) if they selected price attractiveness or availability as

reasons for consumption. Price does matter in crawfish consumption, and this result

infers that crawfish consumption could be comparatively price sensitive. Consumers

were 20 percent more likely to consume crawfish away-from-home if they selected prefer

Cajun cooking as the reason for consuming, but it was insignificant in determining how

frequently a consumer ate crawfish. If respondents indicated they consumed crawfish

because they enjoyed the flavor or for health/nutritional reasons, as expected, they were

33 percent and 12 percent respectively more likely to be an away-from-home crawfish

consumer.

Frequency of consumption away-from-home was significantly affected by different

variables or in different directions. For example, consumers indicating they consumed

crawfish to add variety to their diet were likely to consume crawfish away-from-home

less frequently. It is interesting to note that if consumers chose crawfish to add variety to









their diet, although they were more likely (20 percent) to be away-from-home consumer,

it decreased the frequency of consumption (0.5 times less than average). This suggests

that someone interested in adding variety to the diet may eat crawfish, but on a less

frequent basis.

Frequency of away-from-home crawfish consumption also was influenced by the

variable representing flavor, implying that consumers who selected crawfish because they

really enjoyed the flavor were more likely to consume crawfish more frequently than

consumers who rated other reasons as important. In numeric terms, respondents who

chose flavor as the reason for consuming were 33 percent more likely to consume

crawfish away-from-home, and were likely to consume 0.90 times more. Additionally,

consumers were more likely to consume crawfish away-from-home more frequently if

they indicated health/nutrition as the reason for consuming (0.3 times more).

Variables representing farm-raised crawfish and tradition/habit were not significant

in determining either participation or consumption decisions for away-from-home

crawfish consumption.

Respondents also were asked to identify the top three reasons why they did not

consume or did not consume more crawfish. Probability of away-from-home

consumption of crawfish decreased if the respondents indicated they did not consume

crawfish due to smell or bad impression. However, those two reasons did not influence

the frequency of consumption for existing consumers. Respondents who believed

domestic product unavailability was one of the reasons for not consuming or not

consuming more crawfish were less likely to consume crawfish away-from-home. The

magnitude of these impacts was similar; respondents were 11 percent less likely to









consume crawfish away-from-home because of smell, and 11 percent less likely to

consume because of bad impression. Similarly, consumers were nine percent less likely

to eat crawfish away-from-home because domestically produced crawfish were not

available.

Frequency of away-from-home crawfish consumption was not significantly

impacted by most variables in this category. However, away-from-home crawfish

consumption decreased by 0.57 times if respondents believed price was too high and 0.56

times if respondents had health concerns. The relationship between perceived price

unattractiveness and the frequency of crawfish consumption infers again that

consumption of crawfish is relatively price sensitive. When respondents believe price is

attractive, they are more likely to consume crawfish. On the contrary, when consumers

believe the price is high, they dramatically decrease the frequency of away-from-home

consumption by 0.57 times, a 19.2 percent drop. As a result, we assert that crawfish

consumption is relatively price sensitive.

Additionally, respondents who had a bad impression of crawfish affected both the

probability and the level of away-from-home crawfish consumption significantly. If the

respondents indicated not consuming due to bad impression, they were nine percent less

likely to consume crawfish away-from-home, and consumed 0.47 times less away-from-

home.

Perhaps the most interesting result is that product safety concerns did not influence

either the decision to consume crawfish or the frequency of away-from-home

consumption significantly. This implies that although crawfish scored a bad impression,

it was not due to safety concerns by existing and potential consumers. This result differs









from the pattern of crawfish consumption in other countries such as China. Generally,

Chinese people are more concerned with safety issues due to environmental pollution in

the crawfish production areas.

Demographics had an effect on both the decision and the frequency of away-from-

home consumption. Region significantly impacted the probability and the frequency of

away-from-home crawfish consumption. In general, consumers in the northern part of

the Southeast Atlantic region were significantly less likely to consume crawfish

compared to consumers in the West South Central region. Although not significantly,

consumers in the southern part of the Southeast Atlantic region and the East South

Central region seemed less likely to consume crawfish compared to the consumers in the

West South Central region. This result is consistent with our expectation. Since the West

South Central region is the main crawfish production area in the United States,

consumers in this region tend to have a higher likelihood of consuming crawfish.

Additionally, consumers in the northern part of the Southeast Atlantic region were

significantly more likely to consume crawfish away-from-home on a less frequent basis

(0.88 times less than the West South Central region). However, it is interesting to note

that consumers in the East South Central region (Alabama, Kentucky, Mississippi, and

Tennessee) consumed crawfish away-from-home significantly more frequently than those

in the West South Central region. This implies that although consumers in the East South

Central were less likely to consume crawfish, once they became crawfish consumers,

they consumed more than consumers in the West South Central region. This could

indicate a new target market for the crawfish industry.









One disturbing piece of information delivered by the results is that Caucasian

consumers were significantly less likely to consume crawfish and to consume crawfish

away-from-home less frequently. It is worth noting that education levels did not have a

significant impact on either the participation or the consumption decisions in this study.

However, education levels did have a significant impact on crawfish consumption

according to several previous crawfish consumption studies, and lower educated groups

seemed more likely to consume crawfish and to consume it more frequently. Our survey

had quite a representative sample in terms of education level, and probably was better

than those in other similar seafood and crawfish consumption studies. Further research

with more representative samples might explain the impact of education on crawfish

consumption.

In our sample, we found that the correlation between education and income was

0.32, but not significant. Also, we found income levels had significant impacts on both

participation and consumption decisions, which partly explains the lack of impact by

education levels. In general, our result indicated that most income groups above the base

group, $20,000-or-less, were about 12 percent less likely to consume crawfish away-

from-home. All income groups above the base group consumed crawfish away-from-

home less frequently, ranging from 0.56 times less (a 18.9 percent drop) to 1.38 time less

(a 46.5 percent drop). This result infers that crawfish consumption is sensitive to income

levels and that crawfish is an inferior good. The lowest income groups have the highest

likelihood to consume crawfish away-from-home and to consume more frequently than

other income groups. People with higher incomes tend to consume less crawfish away-

from-home and to consume less frequently.









Age did significantly influence away-from-home crawfish consumption to some

extent. Compared to the oldest age group (age 55 and above), middle age groups

(between 35 and 44, between 45 and 54) were significantly less likely to consume

crawfish away-from-home. The younger age group (between 25 and 34) consumed

crawfish away-from-home less frequently, and their probability of consuming also

decreased.

Given that away-from-home consumption might be the main means of introducing

crawfish to potential consumers and that at-home consumption, consumers need more

preparation knowledge, it is important to understand the factors influencing away-from-

home crawfish consumption.














CHAPTER 7
CONCLUSIONS AND IMPLICATIONS

Summary

This thesis focuses on developing an understanding of the factors that influence

away-from-home consumption of crawfish in the Gulf region. The main objective was to

analyze the significant factors that influence the probability and frequency of crawfish

consumption, particularly for away-from-home consumption. The information then

could be used by crawfish producers, processors, and marketers to develop market

strategies for growth in the crawfish industry. Also, market trends could be evaluated,

and the crawfish industry could follow the trends found in this study to make marketing

decisions accordingly.

Results showed that the factors that influenced the decision to consume crawfish

away-from-home differed from the factors that influenced the decision of how often to

consume crawfish away-from-home. This is consistent with the results of previous

studies of crawfish and other seafood consumption. The crawfish industry can use this

information to target potential consumers and existing consumers separately. Keep in

mind that one marketing strategy might not work on both groups.

One potential concern about increasing away-from-home consumption is that it

might occur at the expense of at -home consumption. This study found that as the

frequency of at-home consumption increases, the likelihood to consume crawfish away-

from-home decreases. However, as consumers' frequency of at-home crawfish









consumption increases, their frequency of away-from-home crawfish consumption

increases as well.

Price is an important determinant in crawfish consumption. Consumers regarded

price attractiveness as a reason for consuming crawfish, and those consumers were more

likely to be away-from-home consumers of crawfish. Respondents also indicated that

price unattractiveness is the main reason for not consuming or not consuming more

crawfish. Consumers who indicated this dramatically decreased their frequency of away-

from-home crawfish consumption, by 0.57 times, or a decrease of 19.2 percent of the

overall level of consumption in one year. Although quantitative price elasticity was not

available, these results infer that crawfish consumption is relatively price sensitive and

that increasing the price of crawfish will decrease the frequency of away-from-home

crawfish consumption. On the other hand, decreasing the price slightly might encourage

consumption and increase the total revenue for the industry.

The differences in consumer perceptions of crawfish can also be seen through the

relationships of variables representing reasons for consuming and not consuming the

product. Interestingly, product safety concerns did not impact consumers' decisions or

the frequency of consumption. However, having a bad impression of crawfish (i.e., mud

bug) did decrease both the likelihood of being a consumer and the frequency of away-

from-home consumption. This result implies, rather than working on improving crawfish

safety or perception of crawfish safety issues, the industry should work on promoting

crawfish as a farm-raised seafood product. For example, instead of calling it "mud bug"

as it is traditionally called, the industry could call it "miniature lobster" in their

promotion campaign. There is a lack of knowledge about the difference between









crawfish, lobster, and langostino. For instance, of the survey respondents, 70 percent

indicated they do not know the difference between crawfish and langostino, and 22

percent indicated they do not know the difference between crawfish and lobster.

Therefore it might be easy to reposition the image of crawfish in potential consumers'

mind.

Our results also indicated that if consumers are more familiar with crawfish, they

would consume more frequently. As a result, work by the industry to increase knowledge

about the product, such as combined demonstrations and educational advertising, could

be successful in increasing the demand for away-from-home consumption.

Compared to average consumers, consumers who ate crawfish to add variety to

their diet were more likely to consume crawfish away-from-home, and they were more

likely to consume crawfish away from home less frequently (0.5 times less per year). In

other words, if consumers choose crawfish to add variety to their diet, they eat crawfish,

but less frequently.

Of particular use to the industry are the results surrounding the variables reflecting

availability. Consumers who knew that crawfish were available were more likely to

consume crawfish. On the contrary, consumers who did not know domestic products

were available were less likely to consume crawfish. Because of the perceived small size

of the crawfish market, not many supermarket chains carry crawfish, and live crawfish

have to be ordered online from Louisiana processors. This implies that the small market

size is not only caused by the lack of demand, but also by the distribution channel, and

the latter one might play a even more important role in the consumption of crawfish. It









appears that a broader distribution channel and more availability might encourage

consumption.

Additionally, smell was an issue for crawfish consumption, which decreased the

likelihood to consume crawfish by 11 percent. Similar to other seafood products, a fishy

smell is always a big concern. Innovation in the industry would include pre-cleaned, pre-

processed crawfish products such as ready-to-cook tail meat. In general, strategies to

overcome the perceived smell issues could increase demand. This might work,

particularly for at-home consumption. As with other seafood products, the impact of

smell should not be ignored, although the magnitude of the impact on away-from-home

consumption is less than the impact on at-home consumption. Persistent smell issues are

related to bad impression, which could seriously impact both participation and

consumption decisions.

Relationships seen in the demographics are generally consistent with a priori

expectations with the exception of education level. The relationships between income

levels and away-from-home consumption tell an important story. The lowest income

group, $20,000 and under, has the highest likelihood and frequency of crawfish

consumption. As income increases, people tend to be less likely to consume crawfish

away-from-home and to consume it less frequently. For example, the highest income

group, $80,000 and above, consumed crawfish away-from-home 1.38 times less

frequently than the lowest income group, a decrease of 46.5 percent. This result infers

that crawfish is an inferior good, although a quantitative income elasticity is not available.

This is not the result the industry wants to see, but the information can help the industry

to further stratify consumer segments and to work on segments effectively. How to









transfer crawfish from an inferior good to a normal good becomes a must for the industry

if they want to encourage consumption of crawfish among higher income groups.

Strategies such as differentiating by value adding characteristics to processed crawfish

products could be successful.

Additionally, the information on inferior goods and the sensitivity of price and

income levels on crawfish can infer a potential relationship between crawfish and its

substitutes using Euler's theorem (e,px+ e,py+... + e,p, + ex, =0). Although the price and

income elasticities are not calculated, the empirical result makes an assertion that

crawfish tend to have many substitutes and/or is sensitive to the price changes of

substitutes, given the relatively large hypothetically positive values of ex,py, ex,Pz, and in-

between cross price elasticities implied by the theorem. The industry should design

appropriate marketing-specific strategies to compete with other seafood products. For

instance, differentiation and adding value characteristics to crawfish products, (i.e.,

flavored, ready to cook crawfish tail meat) to reduce the number of substitutes could win

market share.

Not much can be said on age groups. The oldest age groups were more likely to

consume the product away-from-home and to consume frequently away-from-home. One

possible explanation is because of the lifestyle of the older segment. This finding is not a

trend that an industry would like to see since younger consumers often set future trends.

Crawfish consumption often is seen as regional because the regions where crawfish

is mainly produced are significantly more likely to consume crawfish. Interesting though,

was the finding that the East South Central region was significantly likely to consume

crawfish away-from-home most frequently.









Another finding in this study is that Caucasians are either less likely to consume or

to consume crawfish less frequently away-from-home. Crawfish consumption is

traditionally associated with ethnic culture and Cajun cooking in this country. This might

be the reason why Caucasians still have lower participation and consumption rates.

However, in recent years, ethnic food has become increasingly more popular, with more

and more people enjoying Cajun cooking, including crawfish. Our study found that the

dislike of taste of crawfish was not the significant reason for not consuming crawfish, and

preferring Cajun cooking was the second reason for consuming crawfish. However, due

to the fact that this study had a larger percentage of Caucasians responding than what is

representative, further study into this issue would be useful.

Conclusions

In this study, we found that patterns of away-from-home crawfish consumption

differ from at-home consumption and other seafood consumption. Crawfish consumption

tends to vary by region of residence, and consumers from the main production regions are

more likely to consume crawfish. Consumers prefer farm-raised and domestically

produced crawfish. Fresh crawfish and tail meat are preferred because of convenience

and freshness. In addition, factors regarding consumer preferences and perceptions

influence participation and consumption decisions differently.

Crawfish consumption is relatively price sensitive, as respondents indicated price is

one of the reasons for consuming and the main reason for not consuming. Crawfish is

perceived as an inferior good, and away-from-home crawfish consumption is relatively

sensitive to income level. Higher income groups are less likely to consume crawfish

away-from-home and to consume it less frequently. In addition, crawfish may have

many substitutes and is relatively sensitive to consumers' income levels.









We also found that dislike of taste is not a significant factor influencing away-

from-home consumption and that nutrition issues tend to positively affect both the

probability and the frequency of away-from-home crawfish consumption. Consumers do

not feel product safety is an issue for crawfish. However, the bad impression of crawfish

decreases both the probability and frequency of away-from-home consumption. There is

a lack of knowledge about the differences between crawfish, lobster, and langostino

among consumers. If the industry can improve the image of crawfish and increase

consumers' familiarity with crawfish, away-from-home, as well as at-home, crawfish

consumption would increase as a result. Our results found availability is a critical factor

to increase crawfish consumption. The industry should increase the availability of

crawfish through different distribution channels in order to encourage consumption. In

addition, smell is a concern of consumers. Product innovation in the industry can be used

to reduce this concern.

In terms of demographic factors that influence away-from-home crawfish

consumption, we found education level and household size do not impact either the

probability or frequency of away-from-home crawfish consumption. Other demographics,

such as age, race, and ethnicity, also influence both probability and frequency of away-

from-home crawfish consumption. Older age groups are more likely to consume crawfish,

while Caucasians are less likely to consume crawfish and to consume it less frequently.

Implications

Since the International Trade Commission imposed Less Than Fair Value (LTFV)

tariffs on imported crawfish in 2000, it appears there are opportunities for domestically

produced crawfish to increase its market share. The findings of this study could be used









by the industry to target potential and existing crawfish consumers and to increase

crawfish consumption, specifically away-from-home consumption.

For existing consumers, product development, which focuses on developing new

products for present customers, and market penetration, which focuses on improving

existing products fro present customers, can be used to increase crawfish consumption.

Increasing consumption among existing consumers could be less costly than attracting

new consumers. The crawfish industry could use the findings of this study to target

potential consumers to encourage consumption of crawfish, particularly away-from-home

consumption. Although it is a little costly, it is necessary for the long-term development

of the industry.

The industry should target the regions surrounding the main production region and

promote crawfish as domestically farm-raised seafood product. The study found that

once consumers in other regions start to eat crawfish, they consume it more frequently.

In addition, the industry should increase the availability of crawfish products, particularly

tail meat, through different distribution channels to reach consumers.

Since crawfish may have many substitutes and is perceived as an inferior good, it is

appropriate for the industry to develop new products with value adding characteristics to

differentiate crawfish from its substitutes and to market it as a normal good. Product

innovation (i.e., ready to cook tail meat products) can be successful in marketing crawfish.

Our study also found crawfish to be relatively price sensitive. As a result, carefully

pricing crawfish products, particularly tail meat, is very important in increasing crawfish

consumption and in competing with imports of crawfish and other seafood products.

Appropriate promotions, such as demonstrations, educational advertising, and









repositioning the image of crawfish in consumers' mind, will also be successful in

increasing the demand for crawfish.

As a result of this study, some areas identified that need further research include

quantitative price and income elasticities, the relationship between crawfish and other

seafood products (would increases in crawfish consumption occur at the expense of other

seafood products), and consumption patterns of different races and ethnicities. In

addition, more balanced sample data are needed for further study. It is believed that an

appropriate interpretation of the research findings and the development of related

marketing strategies could win back the domestic market share lost to Chinese imports,

and that the opportunity still exists for the domestic crawfish industry.















APPENDIX
2004 MARKET SURVEY OF FOOD CONSUMPTION

Notice: Any information reported below is strictly confidential. This data will be used
only by persons engaged in this survey, and will not be disclosed or released to others for
any purpose.

Directions: Please have the member of the household that usually decides what food you
purchase fill out this survey. Thank you in advance for taking the time to fill out this
survey.

Crawfish are crustaceans with a hard outer shell, which provides some protection and
gives rigidity to their bodies. The shell of adults is dark red to nearly black with a wedge-
shaped stripe on the abdomen. These small crustaceans are related to lobsters and closely
resemble them.
Common Names: Red Swamp Crayfish, White River Crayfish, Crawfish.
Scientific Names: Procambarus clarkii, Procambarus zonangulus, Procambarus
paeninsulanus, P. alleni.


Figure A-1. Crawfish pictures.









1. Do you feel that crawfish & lobster are different?

[ Yes (different) [ No (same) [ Do not Know

2. Do you feel that crawfish & langostino are different?

D Yes (different) D No (same) [ Do not Know

3. Have you ever eaten crawfish?

a Yes a No

4. Have you eaten crawfish in the last 12 months?

a Yes a No

5. How many times have you eaten crawfish in the last 12 months? (Please write the
number) times.

6. Thinking of the occasions you ate crawfish over the past 12 months, how many
times did you eat crawfish in each of the following locations?

times at home times at restaurants times
at other away-from-home locations

7. Are you aware of whether the crawfish you have eaten was farm-raised or wild-
caught?

1 Yes (Know) 1 No (Don't know)

If YES, have you ever consumed farm-raised crawfish?
1 Yes 1 No I Do not know

If YES, would you consume it again? D Yes D No I Do not care
If NO, would you consider consuming farm-raised crawfish?
1 Yes 1 No I Do not care

8. Are you aware of where the crawfish you have consumed came from? (Check any
of the following options that apply)

D Domestic 0 Imported D Do not know

9. Would you be more likely to consume crawfish that was produced locally or in the
region? (Check one)

I Yes D No I Do not care 0 Do not know

10. Would you be more likely to consume crawfish that was imported? (Check One)









0 Yes D No I Do not care 0 Do not know

11. What product forms of crawfish do you normally consume Away-From-Home?
(Check all that apply)

D Whole with head D Tail meat only

If you could choose, which form would you prefer? (Check one)
D Whole with head D Tail meat only

12. What product forms of crawfish do you normally consume At-Home? (Check all
that apply)

D Whole with head D Tail meat only

If you could choose, which form would you prefer? (Check one)
D Whole with head D Tail meat only

13. Do you prefer fresh or frozen crawfish? (Check one)

D Fresh I Frozen 0 Do not care

14. What price would you pay for one-pound of whole crawfish? (See below for typical
prices) $/pound.

15. What price would you pay for one-pound of tail meat? (See below for typical prices)
$/pound.

(Typical prices are: Boiled Whole Crawfish, Medium $3.00 to $5.00 per pound, Large
$6.00 to $7.00 per pound; Tail meat $10.00 to $14.00 per pound).

16. For Away-From-Home consumption, check the top three reasons you EAT
crawfish:

0 Enjoy Flavor and Taste
0 Enjoy Texture
1 Health/Nutrition Benefits
1 Tradition/Habit
0 Price (Cost compare to other seafood)
D Availability
D Farm-raised
I Religion
D Variety in Diet
0 Convenience
0 Prefer Cajun Style









17. Please check the top three reasons you DO NOT EAT crawfish (or do not eat it
more often):

[ Price
[ Don't like the Product Form
I Health/Nutrition Benefits
D No custom
I Religion
[ Don't like texture and taste
[ Don't like smell
[ No local farm-raised available
0 Concern about safety
D Bad impression (i.e., mud bug)
D Health reasons
D Vegetarian

18. Gender: 0 Male 0 Female

19. Your Age?

20. Marital status:

0 Never Married D Married
D Separated/Divorced D Widowed

21. What is the highest level of education you have completed?

D No formal education 0 Less than high school diploma
0 High school diploma 0 University undergraduate degree
D University postgraduate degree 0 Other, please clarify

22. Current Employment:

0 Full time 0 Part time 0 Current not working
D Student I Unpaid family worker I Retired

23. Please indicate your approximate household income before taxes:

0 Less than $20,000 0 $20,000 to $39,999 O $40,000 to $59,999
O $60,000 to $79,999 O $80,000 to $99,999 O $100,000 to $119,999
O $120,000 to $139,999 O $120,000 to $139,999 O More than $140,000

24. Number of children under 16 living at home (If none, put 0):

25. Number of people currently living in your household (including yourself)?


26. Please indicate your race (Check all that apply):






75


D White 0 Black and African American 0 Asian
0 American Indian D Other, please clarify

27. Are you of Hispanic descent? D Yes D No

28. Please list your Religion: (If Not Applicable, please write down None)

29. What is your zip code?

Thank you for your time and effort completing this survey. Please be assured that all
answers will be kept strictly confidential and used only for the purposes of this research.
If you have any questions about the survey, please contact us at (352) 392-1826 or send
email to xzhang@ifas.ufl.edu.
















LIST OF REFERENCES


Amemiya, T. (1984). Tobit Models: A Survey." Journal ofEconometrics, 24, 3-61.

Avery, J. and Lorio, W. (1999). "Crawfish Production Manual." Pub. 2637. Louisiana
State University Agricultural Center, Baton Rouge, LA (July).

Bilgic, A., and Florkowski, W. (2003). Truncated-at-Zero Count Data Models with
Partial Observability: An Application to the Freshwater Fishing Demand in the
Southeastern U.S." Selected Paper presented at the Annual Meetings of the
Southern Agricultural Economics Association. Mobile, AL (February). Retrieved
February 8, 2004 from http://agecon.lib.umn.edu/.

Cheng, H., and Capps, 0. Jr. (1988). "Demand Analysis for Fresh and Frozen Finfish and
Shellfish in the United States." American Journal ofAgricultural Economics, 70,
533-542.

Cragg, J. (1971). "Some Statistical Models for Limited Dependent Variables with
Application to the Demand for Durable Goods." Econometrica, 39, 829-844.

D'Abramo, L. R., Ohs, C. L., Hanson, T. R., and Montanez, J. L. (2002). "Production of
the Red Swamp Crayfish in Earthen Ponds without Planted Forage: Management
Practices and Economics." Bulletin 1115. Mississippi Agricultural & Forestry
Experiment Station, Mississippi State University, Mississippi State, MS (February).

Dellenbarger, L. E., August, P., and Schupp, A. (1996). "Boiled Crawfish Consumption
in Louisiana." Journal ofFood Distribution Research, (February), 99-101.

Drammeh, L., House, L., Sureshwaran, S., and Selassie, H. (2002). "Analysis of Factors
Influencing the Frequency of Catfish Consumption in the United States." Selected
Paper presented at the Annual Meetings of the American Agricultural Economics
Association, Long Beach, CA (July). Retrieved November 19, 2003 from
http://agecon.lib.umn.edu/.

Florida Department of Agriculture and Consumer Services [FDACS]. (1992).
"Aquaculture Species Resource Series: Crawfish." FDACS, Tallahassee, FL.

Florida Department of Agriculture and Consumer Services [FDACS]. (2001). Agriculture
Press Release. FDACS, Tallahassee, FL (September). Retrieved April 19, 2003
from http://www.fl-ag.com/.









Florida Department of Agriculture and Consumer Services [FDACS]. (2003).
Aquaculture Producer Database. Division of Aquaculture. FDACS, Tallahassee, FL.

Food Marketing Institute [FMI]. (2003). Consumer Trends in the United States. Food
Marketing Institute, Washington, D.C. Retrieved December 23, 2003 from
http://www.fmi.org/.

Greene, W. (1995). Limdep Version 7.0 User 's Manual. Plainview, NY: Econometric
Software, Inc.

Greene, W. (1999). Econometrics Analysis 4t Edition, A LIMDEP Guide to the Examples.
Upper Saddle River, NJ: Prentice Hall.

House, L., Hanson, T., and Sureshwaran, S. (2003). "U.S. Consumers: Examining the
Decision to Consume Oysters and the Decision of How Frequently to Consume
Oysters." Journal of.\ l//llip Research, 22(1), 51-59.

International Trade Commission [ITC]. (1997). "Crawfish Tail Meat from China."
Investigation No. 731-TA-752, Publication 3057. International Trade Commission,
Washington, D.C.

Keithly, M. R. (1985). "Socioeconomic Determinants of At-Home Seafood Consumption:
A Limited Dependent Variable Analysis of Existing and Latent Consumers."
Unpublished Ph.D. dissertation. University of Florida, Gainesville, FL.

Ladewig, F. K., and Schaer, L. S. (1993). "Crawfish-A Healthy Choice." SRAC
Publication No. 243. Southern Regional Aquaculture Center, Stoneville, MS
(September).

Larkin, S., Degner, R. L., and Rubinstein, W. (2002). "Market Preferences, Wholesale
Demand & Breakeven Prices for Marine Ornamental Fish Cultured and Collected
in Florida." Staff PaperSP-02-5, Department of Food and Resource Economics,
University of Florida, Gainesville, FL (73pp.). Retrieved December 23, 2003 from
http://agsurveys.org/RLRA29finalreport.pdf

Larkin, S., Tucker, J.E., and Degner, R.L. (2001). "Developing an Internet Survey
Instrument: Application for Florida Sea Grant Marketing Study." Staff Paper SP-
01-8, Department of Food and Resource Economics, University of Florida,
Gainesville, FL (20pp.).

Lin, C. T., and Milon, J.W. (1993). "Attribute and Safety Perceptions in a Double-Hurdle
Model of Shellfish Consumption." American Journal ofAgricultural Economics,
75, 724-729.

McCullough, A., Gempesaw. C. M. II, Daniels, W. H., and Bacon, J. R. (2001).
"Simulating the Economic Viability of Crawfish Production: A Two-stage
Modeling Approach." Aquaculture Economics and Management, 5(1/2) 2001: 69-
79.









Mullahy, J. (1986). "Specification and Testing of Some Modified Count Data Models."
Journal ofEconometrics, 33 (December), 341-365.

National Fisheries Institute [NFI]. (2003). "About Our Industry." The Fish and Seafood
Trade Association, NFI, Arlington, VA (March). Retrieved April 19, 2003 from
http://www.nfi.org/.

Nicholson, W. (2002). Microeconomic Theory: Basic Principle and Extensions (8th
edition). Mason, OH: South-Western Educational Publishing.

Roberts, J. K. (2000). "Imported and Consumer Impacts of U.S. Antidumping Tariffs:
Freshwater Crawfish from China." Research Summary, Department of Agricultural
Economics and Agribusiness, Louisiana State University Agricultural Center,
Baton Rouge, LA.

RoperASW (2002). Food Trends in America Presentation. RoperASW, New York, NY.
Retrieved December 23, 2003 from http://www.roperasw.com/.

Schupp, A., Pomeroy, R., and Dellenbarger, L. (1991). "U.S. Food Store Experience in
Handling Crawfish." Journal ofFood Distribution Research, (June), 1-9.

Selassie, H., House, L., and Sureshwaran, S. (2002). "Characteristics and Consumer
Preferences Affecting Marketing of Farm Raised Fish." Research Monograph.
South Carolina State University, Orangeburg, SC (December).

Tobin, J. (1958). "Estimation of Relationships for Limited Dependent Variables."
Econometrica, 26, 24-36.

U.S. Census Bureau (2000). U.S. Census 2000, Washington, D.C. Retrieved February 24,
2004 from http://www.census.gov/main/www/cen2000.html.

U.S. Department of Agriculture, Economic Research Service [USDA-ERS]. (1999a).
"Contribution of Away-From-Home Foods to American Diet Quality." Research
Summaries, Family Economics and Nutrition Review, 12 (3, 4). USDA-ERS,
Washington, D.C.

U.S. Department of Agriculture, Economic Research Service [USDA-ERS]. (1999b).
"Aquaculture Outlook." USDA-ERS, Washington, D.C.

U.S. Department of Agriculture, Economic Research Service [USDA-ERS]. (2002a).
Food Consumption Trends: Food Consumption (per capital) Data System. USDA-
ERS, Washington, D.C. Retrieved May 19, 2003 from http://www.ers.usda.gov/.

U.S. Department of Agriculture, Economic Research Service [USDA-ERS]. (2002b).
"Food Away From Home." Food Market Indicators, Food CPI, Prices and
Expenditure: ERS Food Expenditure Tables. USDA-ERS, Washington, D.C.
Retrieved May 19, 2003 from
http://www.ers.usda.gov/Briefing/CPIFoodAndExpenditures/Data.






79


U.S. Department of Agriculture, Economic Research Service [USDA-ERS]. (2002c).
Food Consumption (per capital) Data System. USDA-ERS, Washington, D.C.
Retrieved May 19, 2003 from http://www.ers.usda.gov/.

Yen, T. S., and Huang, L. C. (1996). "Household Demand for Finfish: A Generalized
Double-Hurdle Model." Journal ofAgricultural and Resource Economics, 21, 220-
234.

Yen, T. S., Dellenbarger, L. E., and Schupp, A. R. (1995). "Determination of
Participation and Consumption: The Case of Crawfish in South Louisiana." Journal
ofAgriculture and Applied Economics, 27(1), 253-262.















BIOGRAPHICAL SKETCH

Xumin Zhang was born on November 19, 1972, in China. After graduating from

high school, he enrolled in the Food Science and Technology Department at the Shanghai

Fisheries University, Shanghai, China, in 1991, and graduated with a bachelor's degree in

science in June 1995.

In September 2000, after working several years with Metro, a German supermarket

chain in China, Mr. Zhang immigrated to the United States and continued his education.

After studying at Santa Fe Community College, Gainesville, Florida, for one and one-half

years, Mr. Zhang was accepted in the Master of Science program of the Food and

Resource Economics Department at the University of Florida. He began his master's

study at the Food and Resource Economics Department in August of 2002, with fields of

specialization in food marketing and management.