Factors Influencing Food Habits in Antigua, W.I

A cross-cultural food habit model conscructed aft~r review of available literature was used as a guide for inves.tigating food habit determinants. A food habit survey was conducted in Antigua, w. I.. in December, 1981 and January, 1982 to test the model. Data collected by questionnaire from 305 households included household food purchase and production, food consumption via 24-hour recall and food frequency, food-related beliefs and socioeco~omic characteristics of respondents and household members. Results showed that .the Antiguan diet is starc.h-based and that, despite crop production by about t.lio-thirds of the households sampled, many diets appeared to be def~c~nt in selected fruit and vegetable groups. Diets appeared to be adequate in protein, .with reliance ori fish, egg, chicken and pork. on a product by product basis, diets appeared to .be a mixture of locally-produced items or traditional foods, suc.h as saltfish and local vegetables, and of imported i.tems such as Ovaltine, canned milk, can.ned fish and Cream of ihea~. Noon and evening meal pat:terns resembled traditional breakfast (noon) patterns of English-speaking south Caribbean islanders (Jerome, 1975) • .Investigation into one striking food pattern, the ~idespread consump~ion of a fortified food beverage powder, revealed that the powders supplemented most of the diets with one or more of the selected food groups found to be limiting in respondents• diets and that respondents from a variety of socio.economic groups were consuming the powders. Analysis of cultural factors influe.nci.ng food habits revealed that the major food avoidance was pork, for religious reasons. were asked which foods they would select When respondents .with extra food money and why, major health and preference. variables representing influence food hahits reasons for choices were r~lated to An extensive array of socioeconomic fac±ors in ±he model hypothesized to in Antigua were tested for a significant correlation (Kendall's tau b coefficient,) with a Dietary Adequacy Score reflecting the presence of six selected food groups i.n a respo.nde.nt•s one-day intake. As a result of m·ulticollinearity among the ind.ependen.t variables significantly (p<0.05) related to the Dietary Adeguacy Score, factor analysis of the ind.ependent variables was run. Two of the four resulting factors, re pre sen ting household wealth and the exposure to new ideas, were significantly related to the Dietary Ade~uacy Score in a multiple linear regression model.

fortified food beverage powder, revealed that the powders supplemented most of the diets with one or more of the selected food groups found to be limiting in respondents• diets and that respondents from a variety of socio.economic groups were consuming the powders.
Analysis of cultural factors influe.nci.ng food habits revealed that the major food avoidance was pork, for religious reasons.
were asked which foods they would select Re s po n den t s r a i s i· n g a n i ma 1 prod u c ts for consumption . . . . . Mary's [CFN.I,198'T).
Information collected included the frequency of consumption of selected foods, household production of food, household composition, food preferences a.nd avoidances and opinions about food-C'elated ideas or concepts. Data on food preferences and avoidances were collected by as~ing each respondent, "What would you buy if you had extra money fqr food only, and why" and "What foods, if any, do you or your family members avoid and why." Also included w.ith the gue stionnaire was a 24-hour recall. T.he recall fo.rma t reguired the respondent to name each food or dish consumed over the previous 24 hour period, its ingredients and ±he 3 method of preparation of each individual i~em.
From the 24-hour recall data, a m-easure of dietary ad.eguacy was derived. Due to the iniJ:ial emphasis on gualitative rather t.ha.n quantitative recall data, a Dietary Adeguacy {DA) Score was defined as th.e numb~r of selected food groups in an individual's one-day intake, calculated as follows.
The number of different dishes or foods listed by all respondents was counted. Foods and dishes were listed as given by the respondent: for example, "fisJi and rice" was considered as one dish, where as "fis.h" and "rice" .lisl:ed separately constitu.ted two separate dishes,. These foods were then placed under one or more of the following six food categories according to the nutrient compos~tion of .eac.h food (CFNI, 1974 for labeling certain foods as pregnancy and lactation were "tradi tio.na.l" (eg, unspecific reasons such as "is good for") , · "e xamp.le of o.thers," 11 infl ue12ce of a nurse or doc:tor" and •1nutri tiona.l reason, 11 which included a nutrition term whether correct or incorrect. "Traditional" reasons listed in Table 1 might be "for body s<trength and health," "builds the .body," "very subs tan ti al," "good for the body" and 11 good for t..he .blood. " A number of respondents indicated that t..hey were guided .by a doctor's advice. The reasons which included a nutrition term or concept were "nutritious," "contains lots of 7 erroneous nutrition informacion. For example, carrots ~ere correctly perceived by two respondents as "good for sight ... " However, carrots and bananas were perceived to be iron sources by 8 and 5 respondents, respectively. In addition, lettuce and 11 vegetables" (unspecified) .were identified as protein sources by two respondents.

2,.
Heal th-related food concepts from food avoidance data Table 3 presents the f ood-rela:ted health concepts associated with foods the respondent and/or respondent's family members avoided. Forty-seven percent of t.he respondents reported the avoidance of one Dr more foods; howev _ er, health-related reasons accounted for only 9.% of t.he total foods listed. Of ths xeasons listed in Table 3, "burns the s;tomach, 11 11 o~her all_ ergy or medical problem" and "heal th reasons" appear to represent a few .i:andom ideosyncrasies. However, the use of the term 11 unc.lean 11 is widespread and heal.th-related in that the term denotes local animals that wander anywhere, eat anything and a r e there.fore 11 fil thy" and unfit foe consumption [Parker,l.982}. I.t has not been unusual for pigs, goats and to a less.er .extent cattle to roam the villages and streets of St. John's in search of grazing land or: other sources of food despite laws and regulations prohibiting such fre.edom [Livingston, 1981) .• 8 The term "unclean" is imported meat since it imported meat has been not generally is assumed that used .to descr i.be t.he gualitJ of more ca ref ul.ly con trolled (Bird, i984) • Three respondents linked 11 unclean" .with "religious" reasons for avoiding pork, a possible refer~nc.e to either the Rastafarian dietary influence (Taylor, 1981) or the practice of Seventh Day Adventis~s on the island (Parker,i9821 ,.

DISCUSS.ION
In contrast, Dewalt and Pelto (i977) argued that the people .in the Mexican comunity they studied were already aware of t.he idea of "good nutrition" and simply re~uired the economic means to implement their knowledge.              c. and A. G.
Van Veen (1972 TJle perceived advantage of one food over another can con tribute to competition between foo· ds .• relative advantage is to consider one way to measure both ±he emotional importance of foods and their f reguency of consump±ion. Passim and Bennett (1943) pioneered this approach in whic.h uni~ers~lly eaten, staple items and the mos~ emotionally important foods are labeled "core" foods, "secondary core" foods differ from "core" foods by being wid.espread rat.her than universal in use, and "peripheral" f.oods ar.e infreguently consumed and have the least emotion attached to 69 them. Resistance to change is hypothesized to increas~ vit:h progression from peripheral to s.econdary core to core f .oods (Sanjur, 1982) , .  iO.*-= i68 TYPES** = 17    (Robson, 1976'. introduction of potatoes, mai~e and coffee into Europe took generations (Gordon, 1974); however, one impact of urbanization in the South Pacific has been to rapidly replace some traditional foods and child feeding practices with European vegetables, canned products and bottle-feeding, resulting in significant health changes in just one decade (Lambert, 1982). Another short-t2rm change is the "end of t.he month" syndrome, or the drastic changes in food purchases during the period of time after a household has run out of money and be~ore more income is available, and the nutritional conseguences of t~is change (Bode et al, 1974). In adjuscing to short-term crises such as illness or unemployment, i:eople of ten cu.t back on foqd first because it is more flexible than rent or utili~ies [McKenzie, 1974). Omololu {1971) has observed that these "fast days" coincide with times when food is scarce. Thus, a prac~ice .b.elieved to bring man closer to the Divine (Knutsson and Selinus, 1970) may also be an adaptation to a variable food supply.
Other negative food beliefs may not be linked with severe penalites, but may ne ve rthel.ess directly affect food choice. For example, the traditional principles of "hot 11 versus "cold" and "weak" versus "strong" foods in Latin America (Cosminsky, 1975;Maz.ess, 1968)  Malaysia consume an enormous variety of jungle animals taboos are associat.ed with the consumption of some o1 those animals by certain groups, such as pregnant and lactating women and children (Bolton, 1972). However, if a person consumes a meat and has no negative after effects, .he can thereafter eat that meat.
Where negative food beliefs and values discoura~~ t~e consumption of certain foods, ;there are many values associated with foods which encourage consumption, . For example, eating patterns may be reinforced by association with memories or f . eelings or familiarity and security (Lowenberg, 1968 Pa~ific (Lambert, 1982) and among Japanese in Hawaii (Wenkam and Wolff, 1970). In addition, food habits associated with the d~sire for prestigious foods may create demands for products which are difficult or impossible to ma~e or get locally (Pirie, 1972 Wenkam and Wolff ('1970j ,. In the o.s., personal health is a strong factor infiuencing motivations to diet, and it appeared to increase wit..b age in one group studied (Berman,i975 de Chavez, 1972) to food faddists [Wilson and Lamb, 1968). But nutrition education mus:t fun:: ti on within the fcamework of existing attitudes toward food (Sipple, 1971) or i.t will JJ.ot he successful in motivating the individual to fully integrate the dietary change with his lifestyle, understanding and health concepts [Knutson and Newton, 1960) (Graham, 1973 Morocco after a bakillg specialist demonstrated its use in bread, a major food of the country, and in · Britis..h Honduras and Ceylon, where it also was substituted in certain recipes in school feeding programs (Horan, 1972). DURYEA, a .weaning food developed in Colombia by CPC International, Inc •• was designed to be easily prepared and served in a bottle or as a gruel, in accordance with existing infant feeding practice (Robinson and Weisz, 1972). This product also re~uired :;areful guality control measures for .raw materials which reduced insect and microbial damage and allowed cheaper packaging. It appears that the effort to make DU~YEA a local product -in terms of ingredients, packaging materials and management -was successful.
One of the most widely publiciz.ed failures 0£ a produc:t  (Jelliffe, 1976). In addition, Jelliffe (1975) has point.ed out the populations using indigenous raw materials so as ±o reduce costs and engage local economies (Bauman, 1972;.Robinson and Weisz, 1972;Caruso and Moore, 1976 ,,, Eide et al (1977) have poi.n.ted ouit that some A£rican .women's low attendance at clinics and appare.n-C reluctance to carry out nutrition instruction may be a reflection of ~he limit~d time they have outside of their obli~ations as food producers, housek. eepers, mothers and wives.
has also argued that tiue involved i.n Graham (1973) introducing or demonstrating an innovation (eg, ne.w farm machinery versus a new seed) is critical to its acceptance.
Closely associated with the time factor in food procurement is labor and its division within the house.hol.d.
The resources available to labor affect its productivity. indust.ry to sell higJa protein products (Horan, 1972;Kracht, 1972;Malaspina, 1972;Caruso and Moore, 197.6 Mos. t importantly, the definit.iolls of 11 cor.e," "seco.ndary =ore" and "peripheral" imply that attempts to make cha..nges in food consumption £ace increasing resisitance progressing from peripheral to secondary core to core foods (Sanjur, 1982). Thus, the relative importance of a food to the die± gives predictive inf or ma ti on a.l'J out the relative ease o..f. a food change.
There is not only competition between foods on t~e market, but also between family members for availa.Dle food.
Customs of intrafamily food distribution may limit the food choices of certain family members. In Africa and the males i.n Caribbean, for example, it is .not unusual for the the household to be given first choice at especial! y where animal foods are concerned meal.times, (Sai, i960;Marchione, 1981).

D. Availability
Foods which become core or important i±ems in the diet must hav.e been a va ilab le to the comm uni.ty over a sufficient period of time. Availability (production and supply) of f 00 a. is of primary consider at ion in det.erlllining the diel:a.ry pattern of any group. Factors such as soil, climate and level of economic and agricultural developmsnt limit local food production (Brozek, 1958) and affect what may be produced in exchang. e for food gro.wn ability -or inability -to produce food Gurney (1975) stated that staple foods of the caribbean, which provide :the greatest amount of total energy in the diet, take up significant working time and land area.
With scarce resources, competition ~etween these crops used as food and other crops used for cash income also determines food availability.
The natural occurrence of a food or the proximity to a food source can influence food acceptability. For example, Cassel (1957)  The addition of imported we stern foods to traditional iiets has been regarded as a symptom of "weste.rnizat.io.n" (Pirie, 1972)  islanders (Lambert, 1982) (Jerome, 1975 (Ennever and Standard, 1975;Simmons and Gurney. 1982 (Sinha, 1979) , arrowroot or diluted proprietary baby foods (Lees, 1964). Greiner and Lat.ham ( 1980)  Ant.iguan agriculture is characterized by wet season gluts of produce that are not ade g ua te.l y s~ored or preserved locally, and expensive and limit.ed dry season production due to the lack of irrigation .technology and cosc cf water (Michael, 1981 damage to crops and other property (Livingston, 1981) .• However, there is a great deal of political ~rcssure agai.ns± enforcing these laws (Livingston, 1981 (Benjamin, 1981 Wom.e.n in food production, food handling a.nd nutrition.

1975,.
An anal1sis of health problems of Commonwealth Caribbean coun±ries and the priorities.
L. These two groups appro.x:iua te a "before" and "after" sampling of a population which is hypothesized to be c.hangi.llg its food habits as a result of the exposure .to tourism.. Two parishes, or regional divisions on the is.land, were c.   family likes food

its good for you
cost o.f the food , 77 pa g.e 9 28. How would you best describe the fol.lowing foods in terms of:   (Cook, 1971) in which abou.t .half of the 167 homes examined were using kerosene stoves, JO.I were using wood fires or a coalpot and atout 18ji were using gas. It appears from this survey t.hat the gas stove is t.he major cooking applia.oce that has replaced the coalpo.t.
The second factor is com posed of 11 1 uxur y" items in t.he sense that, with the possible exception of the sewing machine, they could provide an extra or secondary degr.ee of convenience or comfort. This factor includes a fan, color TV, sewing machine, phone and electri.c .was.her and dryer.

Crop Production
About 643 of the respondents indicated that t..hey grc.w :rops (Table 25). Of these, 70% had a backyard gardc.o; t.he remaining .res pond.en ts had access to a "gro.und" (usually a larger plot and located at some di s.tance from the house), or both a backyard garden and ground.
Th.e vasit majority cf respondents with a ground were living in However, 45% of urban respondents .were engaged agriculture.
in .backyard The average garden or ground was planted ~ith five :rops. Crops raised by households were p.laced into t.l!e seven categories shown in Talllle 26, . These categories were chosen to reflect nutritional differences among crop ;types,.
The numbers of individuals producing · crops in more than ore category are given. Root crops such as cassava and s~eet potatoes, fruits such as bananas, papaws (papaya), coconut., mangoes, limes, soursops and sugarapples, vege±ables such as "an trobers" (eggplant) and peas {legumes) and sugarcane lie re the most frequently listed crops in home pro duct ion , . N.early 85% of the respondents growing crops were raising at leas± one fruit crop and consuming part or all of the frui~.

Animal-Raising Practices
Respondents were asked to identi£y w.h.ich of t~€ animals listed, if any, were being raised for household consumption.
The proportion of those raised that was being consumed was not specified. A total of 29% of .the responden.ts were raising chickens (Ta.ble 28) ; of these, 55i were raising them only for eggs and the ..re st for bocth eggs and meat [ were raising cattle. Host households raising cattle did sc for the milk production; very few used only the meaL (Table   29).
Twenty-six households (8%) wer€ rsportedly raising pigs and 25 were raising sheep, while 10% indicated that they had goats. Respondents were also asked i~ t~ey ..raise 'T95 rabbits; how~ver, none did. Only a small percenLage of households in the city were raising animals.
About 243 of t~e respondents were raising animals to sell, exchange oc give away (and also possibly to consume) (Table 30). Of these individuals, mosi: (773) were s~lling some or all of their animal(s) and few were givi.ng animals away to others ris3), selling and giving away their animals (73) or selling, excha.nging and giving away animals (c.ne in di vi dual}.

Food Resources Scores
The Also, respondents from non-fishing households were mere likely to eat sa l.tfish several times a llleek.

Fruits and Vegetables
Frequency data on the consumption of selected fruits and vegetables revealed that fresh fruits, ground provisions {potatoes, yams and other starchy root crops,) and dark gee.en leafy vegetaDles were consumed freguently by masc respondents (see Table 34 However, it appears that close to one-fourth of the individuals surveyed ate pickled mackerel at least once a wee.k.. These smo.ked and pickled fish products were available at the supexmarkEts and to a more limited extent at local stores. When asked if they would eat these seafood products 2DO more often if given the cha~ce, most respondents said yes for fresh fish and saltfish (833 and 633, respectiveln (Table 36). Forty percen.t seemed in.terested in consuming more canned fish a .nd pickled mackerel.
0.11l y about 30 % of the respondents would eat more frozen fish or herring. . X he demand for fish products, particularly fresh .fish alld saltfis.h, is higJ.i.. . In the case of fresh fisA, it appears that the amount of catch is a major factor limitiny ~.he availability of the product (Brownell, 1978).
Since fresh fish and saltfish appeared to be prominent items in the diet of Antigua~s, it might be hypothesized that consumption of these foods is high across all groups qf respondents, where those groups may be diffeL€ntiated by any socioeconomic characteristic of in.teresrt.. A step-wise multiple regression was used to analyze the extent to which selected economic variables (.number of rooms in the house, "Basic" and "Luxury") account for the variance in f.resb fish and saltfish consumption (Table 37)  Plaintain is sliced and fried in oil as a side dish.
Fungee (or dumplings) is a traditional accompaniment to saltfish; fungee may also be served with pepperpo~, a stewed vegetable dish.
Various kinds of fruits and vegetables appeared to be frequently eaten. Twenty-six out of 28 respondents reported ea ting yellow vegetables (probably carrot see the following section) at least once a week. Antigua relies on fresh fruit imports from local islands such as Dominica and 204 Montserrat {Bird, 1984); fresh fruits were freguently consumed by more than half of the respondents. Canned fruit juice is often a substitute for fresh fruit (Bird, 1984) and was apparently consumed daily by a~out one-fiith of the respondents. Canned fruit is considered by some to be a dessert for special occasions (Bird, 1984 well as that of the government school system's home economics program, which is beginning to stress traditional food use (Hurst, 1982).
Two imported products have seemingly become "core" items according to apparent frequency of use in this data: canned milk and fortified food beverage powders such as Oval tine, iii lo and Complan. Canned milk is mostly added to leaf tea or diluted with water and mixed with a fortified food beverage powder which, with sugar or other ingredients optional, is often called "tea" (Bird,. 198£ The individual foods listed on all the 24-hour recalls were grouped into 13 selected food categories which were hypothesized to be culturally, as well as nutritionally, appropriate {Table 39) • The largest group [by the total frequency of items listed) was the meat group, which constituted 203 of the foods listed. This group included fish and chicken, the two most freguentl y lisj:ed meats, as ~ell as turkey, pork, egg, beef and mutton. Saltfish was the most frequently listed fish product, followed by fresh fish. Even though pork was freguently listed elsew~ere in the questionnaire as a food that was avoided (see section D 2.), it was listed more of ten than beef in the 24-hour recalls. Eggs were the next most fre~uently consumed protein after fish and chicken. "Mixed protein dishes," including meat mized with vegetables or rice, soup, pizza and fabricated meats such as chicken franks, were considered as a separate kind of protein dish, inasmuch as the protein was usually diluted (and also supplemented) by other ingredients. This type of dish constituted about 53 of the items listed. More than one key informant stated that Antiguans love bread this was supported by the finding that bread, with or without cheese or meat, was the second most frequently consumed food group after the meat group .• Other starches, such as starchy root crops, pasta and rice,

207
were also important items in the diets, constituting about 1~3 of the foods listed. Vegetables figured prominently, as might be expected during the growing season. By contrasc, fruits constituted only a small proportion of foods eaten.
"Jui:::e 11 or various fruit drinks constituted the largest group of beverages, outnumbering sodas by more than two to one. Tea and milk were popular individual beverages.
Breakfast-type cereals, legume dishes (beans or peas with rice) and dessert/snacks constituted the remaining food groups.
Vitamin supplements, taken exclusively as ~owders mixed with milk or water, were listed freguently and appear to be one indication that many Antiguans are health-conscious.

b. Pattern analysis of foods consumed
Although frequency counts indicate foods freguently consumed, they do not give any information about food patterns, or foods which are consumed together. In order to obtain this kind of infocmation, two analyses were made: exploratory factor analysis for extraction of overall food clusters or factors and a meal pattern analysis based on t.he 1 1 selected categories fcom section 4 a.
Exploratory factor analysis was chosen as a device to reduce the vast array of recall items to clus~ers 0£ foods.

208
These clusters, composed of foods which covary in the data, may or may not represent a full day's intake, but reflect food patterning, or foods consumed with other foods in the data. When foods listed less than 2% of the time were It may be assumed that, over the entire sample, the patterns detected and the proportion of r espondents associated with these patterns is con s tant, at least for the time of year in which the survey was conducted.

2'Ti
Although factor analysis is useful for e~tracting food patterns from 24-hour recall data, it is not known what time periods and therEfore which meal structures these clusters may encompass. Therefore, an additional pattern analysis was conducted on the 24-hour recall data. In this analysis,  (Passim and Bennett, 1943;Jeromer 1975 andSanjur, 1982) who wish to denote other characl:eris±ics such as the emotional importance as well as the freguency of consumption of foods.
3. food patterns respond. ents eating at accounting for the majority cf eaclt time of day (Table 41b) (in all three time periods, a certain proportion of respondents• diets were too variable to conform to any overall pattern).

c. Dietary Complexity Scores
Dietary complexity, as a reflection of t~e variety of foods in the diet, has been associated with dietary status Marchione, 1981). Dietary complexity, as measured by a Guttman scale, was also demonstrated to increase with greater socioeconomic complexity of the individual and household (Chassy et al, 1967;Sanjur et al, 1970 andBeaudry-Daris.me et al, 1972).
The implication is that increased dietary comple~ity reflects an increased variety of foods in the diet, which gives greater chances for good nutrition or good "dietary guality. 11 Since one of the major sour~es of food behavior data in the survey, the 24-hour recall, emf~asized the variety of foods consumed and food preparation methods used over food quantity, the use of this data as an estimation cf dietary quality via dietary compleiity is appropria£e, within the limitations of one-day recall data. Therefore, a Complexity Score was calculated as the sum of al~ the dishes consumed (eg fish, yams, beef and rice, etc as listed according to the 24-hour recall format) by a given individual during the 24-hour period, and ranged from O to 1 8 (Figure 1 ) . An individual's score also reflects any dish or dishes eaten more than once during the 24-hour period~  Focusing exclusively on the subsample of respondents 223 who raised their own food and examining the effect of resource consumption, sale or both on dietary complexity might be a different way to discover a relationship between food resources and diet. To do this, respondents raising both crops and animals were placed into a data subset. Out of this subset, respondents who stated that they sold at least part of their crops and at least part of their animals ("sell" group) were separated from respondents who were only consuming their crops and animals {"eat" group). Either group may include a small proportion oi respondents who were also giving food resources away or exchanging them for other things. Figure 2 shows that the peak complexity score for the group that sold part or all of its resources is shifted slightly higher than the peak complexity score for the group "contains vitamins," "builds the body," "rich in protein," "nutritious," etc.). Fowl -mostly chicken -was the next most fre1uently listed protein group ior preference reasons ("like it," "prefer it") as well as· health-related ones ("nutritious," "good for the body," particularly sweet potatoes, white etc. as "is good "tradi tiona.l" (eg, unspecific for") , "example of others," "influence of a nurse or doctor" or "nutritional. reason," which included a nutrition term -whether correct or incorrect. "Traditional" reasons listed in Tab.le 28 migll t be "for body strength and health," "builds the body," "very substantial," "good for the body" and "good for the blood." A number of respondents indicated that they were guided by a doctor's advice. The reasons which included a nutrition term or concept were "nutritious," "contains lots of vitamins," "rich in protein," "protects the body £.rem starch," "for iron," "to give energy," "for a balanced diet" 227 and "good for si g.ht." The phrase "protects the l:lody fr cm starch" refers to a non-starchy food which is an alternative to more s±archy meal items and is deliberately eaten in place of s±arc.h (Parker,i982).
"Nutritious" was the health-related term used most f re guen tl y as a reason for choosing a food. T h.e protein sources, milk, fish, beef, "meat," poultry and eggs, were also generally associated with "rich in protein, 11 "contains lots of vitamins," "good for the body" and "for body strength and health." Milk, cheese, fresh fish and fOUltry were foods characterized as "substantial." The conceft 11 pcotects the body from starch" was exclusively associated in this data with carrot, cabbage and other non-s~archy vegetables. Also associated with these foods were the concepts "for balanced diet" and "doctor's advice/health reasons." This implies that increased consumption of Finally, although saltfish was liswd as frequently consumed (Table 35) and is available in most food stores, it was listed as an extra food purchase by only three respondents.
By contrast, fresh fish, also fre~uent.ly consumed durinq the 24 hour period, was an item many respondents wished to purchase with extra money. This implies that, among fish consumers, fresh fish is in greater demand than saltfish.
While the data implies an overall preference for fresh fish over o.ther selected fish products (see section c 2.), there is also a general shortage of fresh fish in Antigua (Brownell, 1978 (Bird, 1984).
Fifty-two respondents listed a number of ite ms whi ch are grouped together in  (Parker, 1982 (Table 47).
However, it appeared that household fisni.ng activity was significantly (p < Q.05) related to lower rankings for tDe expense of fresh fish ( probably reflects the fact that the kind of milk product refered to was unspecified.

Perceived Importance of Selected Concepts When
Purchasing Foods Respondents were asked to rank the importance of several concepts hypothesized to be considerations in deciding which foods to purchase for their families. Fer example, they were asked how important it is to purchase 240 food that is "good for you." The end points of a 7-point scale were defined as "unimportant" (i) and "very important" (7) an1 each respondent identified the point along the scale corresponding to the perceived importance of t~e concept.
Results are shown in Figure 5. All respondsnts felt strongly that the selected concepts were important when purchasing food for their families. When freguencies f cr ranks of 6 and 7 were totalled, "family likes t.he food" appeared to be the most important concept, followed by "its good for you" and the cosi: of the food. Ease of preparation appeared to be the least important concept of the four. As indicated previously, Antiguans who have had a government school education are likely to have received some nutrition education. Also, simple nutrition instruction was appearing periodically on local TV and radio programs at the time of the survey. Thus, exposure to food· and health principles was probably wide-spread.

Perceived Importance of Selected Concepts When
Raising Food Respondents were also reguested to rank the imfortance of selected concepts hypothesized to be consideratiQns in deciding which crops to grow. Thus, the responses appear to represent at least those people who were interested in farming, if not those who were farming at the time. Raising a crop that is "easy to grow" and "family likes the food" were ranked as very important (rank = 7) concepts by 64% and 62%, respectively, of those individuals responding to the guestion ( Figure 6). consumption of milk and fruits. It is noteworthy that, of the i6 respondents who reportedly did not eat breakfast, only one consumed milk or fruit later in the day. The afternoon patterns showed many more types of combinations cf food groups. They also included two new groups, legumes and ground provisions, and excluded fruits as a major category.
The primary noon-time meat was chicken, rice replaced bread as the primary starch and a different group of non-starchy vegetables was consumed. The evening meal patterns were extremely similar to the noon ones. These patterns largely coincide with the "traditional breakfast (noon) patterns" of English-speaking south Caribbean islanders as described by Jerome (1975), with local adaptations such as the decreased consumption 0£ fruit and increased consumption of bread and When requested to rank, also on a 7-point scale, the importance of selected concepts when buying or growing food, most respondents felt that "family likes the food" was an important concept. "Its good for you, 11 "the cost of the food" and "easy to prepare" were of next greatest importance when purchasing food as judged by the number of respondents giving the concepts a high rank (= 6 or 7) • Most respondents also felt that a crop that was "easy to grow," the "cost of growing food compared to the market price f cr the food" and the 11 price you get for the food when you sell it" were important concepts when growing or raising food.
Thus, family likes and dislikes seem to te as important as economic and health considerations when buying or growing food.                             * Once to several times a month ** Several times a year, not often, rarely, seldom, seasonally + "Fish" represents respondents from households that caught their own fish; "No Fish" represents respondents from households that did not catch their own fish 2 79 FREQ CJ ENCY OF  2.0 1.0 * Several times a week ** Once to several times a month *** Several times a year, not often, rarely, seldom, seasonally 2 81 * ** *** (2)    In div idu al factor scores on these dimensions were used as variable measures.       (8) TOTAL (9) TOTAL ( 10) TOTAL ( 11) TOTAL ( 12) TOTAL (