Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Determinants of Food Choice and Prospects for Modifying Food Aides and Behavior MANY CHANGES have taken place in the United States during the past century with respect to food selection and attitudes toward diet and health. In this chapter, these changes are reviewed by the committee as are the prospects for future changes in behaviors and attitudes to meet dietary recommendations. The focus is on the general U.S. population rather than specific high-risk groups. Recommendations for changing eating habits are more likely to be adopted if their framers (1) are knowledgeable about the factors known to affect food choices, (2) recognize current trends in food consump- tion and attitudes toward food, (3) base their recommendations on basic theory and research related to changing attitudes and behav- iors, and (4) learn from previous attempts to change diet for health purposes. This chapter is organized to address each of these items in sequence. To select the most appropriate targets of change (e.g., the most critical beliefs or behaviors), one must know the basic determi- nants of food choice and which of these are subject to modification. Likewise, it would not be prudent to recommend specific methods or programs of change without knowledge of basic theory and research on the determinants of behavior change and of the techniques that have already proven successful. For example, it is critical to under- stand why an individual's knowledge alone about the links between diet and health is unlikely to change dietary behavior. It is also important to know what changes are already taking place in the United States and whether these trends are likely to facilitate or hinder implementation of dietary recommendations. 33
34 IMPROVING AMERICA'S DIET AND HEALTH This chapter concludes with a critical review of some intervention programs designed explicitly to improve eating patterns. These in- clude programs instituted at the individual, organizational, and com- munity levels. DETERMINANTS OF FOOD CHOICE Why any group of humans eats what it does is considerably more difficult to explain than is the eating behavior of other species. In humans, appetite is not simply a physiological drive toward food but, rather, a complex set of physical, emotional, and cognitive stimuli compounded from events widely separated in time and space. Because humans from birth through childhood depend for their survival on nurturance provided by other members of their species, they are uniquely vulnerable to developing affective relationships with food and feeders. Thus, nourishment for humans almost inevi- tably becomes associated with powerful emotional attachments. These, overlaid with beliefs and feelings that continue to accumulate around food consumption as individuals mature, combine with immediate environmental stimuli to direct food choices. Beneath their socialization, however, humans remain animals, en- dowed with sets of sensors that underlie all their subsequent encoun- ters with food. Taste is one of these. The evidence is overwhelming, for example, that humans are innately programmed to like sweet tastes at birth and even in utero (Montagu, 1962; Weiffenbach, 1977), and there are very tentative indications that they are also born with some sort of attraction to meat, to its fat, or to both (Beauchamp and Moran, 1982; Drewnowski et al., 1985; Farb and Armelagos, 1980; Harris, 19851. Since many poisons are bitter and fruits become sweeter as they ripen toward greater nutritiousness, a taste for sweetness may well have given its possessor a selective advantage. A prefer- ence for fat would also have favored survival under conditions of calorie deprivation. Humans also like salt, although it is not an innately Preferred taste in infancy (Davis, 1928; Steiner, 1977~. J 1 People who change to low- salt foods come to consider formerly acceptable foods too salty. Thus, the preferred level of saltiness appears to be strongly affected by experience. Such is not the case for sweetness, however; preferences for different levels of sweetness appear to be inborn (Desor et al., 1977~. The ability to taste certain isolated flavor chemicals (e.g., phe- nylthiocarbamide) varies with individuals and is genetically controlled. Although there appears to be little genetic control over the liking for particular foods (Fabsitz et al., 1980), early research on food choice
DETERMINANTS OF FOOD CHOICE 35 suggested that infants protected from poisonous foods and exposed only to uncombined foods whose indigenous components have been neither concentrated nor diluted with added fat, salt, or sweeteners are able to select fully nutritious diets instinctively (Davis, 1928~. Humans also appear to be born with automatic regulators of energy need (of a complexity not yet deciphered) that are set differently at or before birth (Ravussin et al., 1988; Roberts et al., 1988~. This physiological base on which eating behavior is built appears to explain very little about the food choices people actually make. The experiences of individuals as members of specific families in a particular culture tend almost inevitably to overwhelm many (if not all) the signals coming from what lean Mayer long ago called the "animal within" (Mayer, 1968~. Young babies universally find chili aversive. Yet, as Rozin and Schiller (1980) have demonstrated, this innate aversion to chili, along with the innate preference for sweets, is spontaneously overcome in Mexican children, almost half of whom at age 6 or 7 years will select a spicy hot snack over a sweet one when given a choice. In an earlier classic study, Moskowitz et al. (1975) found that chronic exposure to tamarind among a group of Asian Indians overrode the relative dislike of the sour taste charac- teristic among humans. As indicated by the prevalence of overweight people in the United States, culture can also override in humans the bodyweight-regulating mechanism that operates effectively in all other species (except when they are domesticated). Factors other than physiological ones that affect food choice can all be attributed to either nurture or culture. However, since culture heavily affects the ways in which a society nurtures, even this divi- sion is somewhat artificial. One such factor is early feeding experi- ences, which involve, in addition to tastes and smells, the sounds, sights, textures, and emotions associated with feeders. Thus, they deeply affect infants who are entirely dependent on their feeders for survival. These individual feelings about eating, implanted in infancy, may be difficult to modify. There is evidence that food aversions result- ing from even a single, powerful negative experience with a food can be very long lasting (Garb and Stunkard, 1974~. Therefore, if patterns es- tablished in early infancy need modification, recommendations would need to be directed to food providers, usually parents. There is no direct evidence, however, that food preferences learned in infancy are permanent, but there has been little systematic study of early feeding interactions and their effects on later eating behav- iors. "There are currently no prospective or longitudinal data with human subjects to provide support for Ethel assumption that early food acceptance patterns are . . . reflected in food acceptance patterns
36 IMPROVING AMERICA 'S DIET AND HEALTH later in life" (Birch, 1987, p. 127). Nor is it known what makes cer- tain children like certain foods. Davis (1928) found that infants pre- viously unexposed to any food but mother's milk, and protected from outside influences on food choice, expressed a wide range of preferences when they were allowed to choose from a variety of simply prepared, unsalted, unsweetened foods. Birch (1987), who investigated food acceptance by young children, concluded that "sweetness" and "fa- miliarity" were two characteristics of food that seemed largely to account for children's food choices in the United States. In this country, preschoolers are almost unavoidably often exposed to highly sweetened foods; thus, the sweetness factor that drives food acceptability in this and similar cultures may arise in large part from familiarity, as may the preference for spicy hot foods in Mexico. There is evidence that continuous exposure to sweetness sustains the neonate's preference for sweetness (Beauchamp and Moran, 19821. Davis (1934) could not determine whether the initial food choices of infants in her study were random or whether they were based on color, odor, or both. It is known that children's acceptance of food can be influenced by the choices of their eating companions (Birch, 1987~. Since parents or other caretakers normally select the foods to be made available to very young children and their eating compan- ions, adults have a strong influence over children's food choices. In the United States, however, children are exposed from early infancy to adults other than members of their own households who tell them what to eat. Many of these adults are seen on television, advertising edible products consisting largely of sugared cereals, candy, and fast- service foods. A recent study in Quebec indicated that the parents of children who watch child-oriented television that carries commer- cials purchase more brands of breakfast cereal directed specifically toward children than do the parents of children who watch commer- cial-free children's programming (Goldberg and Hartwick, 1990~. Thus, advertising demonstrably influences parents as they select food products for their children. In a study of influences on the food choices of elementary school children and adolescents, Contento and Michela found the two most important variables in both groups to be "parents serve it" and "tastes good" (Contento et al., 1988; Michela and Contento, 1986~. Taste predominated among the adolescents; serving by parents took first place among the younger children. Although the sweetness variable was not examined directly in either study, the same authors found in an earlier investigation that sweetness was a highly salient dimen- sion in children's spontaneous classification of foods into groups (Michela and Contento, 1984~.
DETERMINANTS OF FOOD CHOICE 37 Examining adolescent food choices more closely, Contento et al. (1988) found that their subjects could be divided into subgroups "with different motivations for food choice irrespective of ethnicity and gender" (p. 2971. Subjects at one extreme were "hedonistic" choos- ing foods even if they could identify those foods as causing heart disease or containing sugar or fat. Subjects at the other extreme were "health oriented" in that they avoided these same foods and ate foods they perceived as "healthful." Although the food choices of peers were reported to be an important influence on the "hedonistic" group and on others, it was not an important factor in all subgroups. The evidence thus suggests that children, some of them even into adoles- cence, are heavily influenced by their parents' choices of food to serve. As they grow older, however, some children choose foods they perceive as healthful or unhealthful; others are more strongly influenced by other social and environmental factors and by taste. Many of the factors known to affect food choice beyond adoles- cence cannot readily be modified by educational or other populationwide interventions. These include individuals' positive or aversive food or eating experiences that may make certain foods especially palatable or nauseatingly unacceptable, as well as simple familiarity, which probably plays an important role in determining food choice in adulthood just as it does among children. But even though education cannot change an individual's historic relationship with certain foods, food likes and dislikes can be modified with further experience; a new and wholesome food, once tried, may become both familiar and liked. Other variables often identified as determining food choices in adults include age, sex, race, place of birth, time of day, season of year, marital status, children's ages, household size, employment status, income, and perhaps less obviously media events affecting the public's perception of the safety or wholesomeness of the food supply. A1- though none of these determinants of food selection can be intention- ally altered by policymakers, many of them can change over relatively short or long periods. For example, women's increasing participa- tion in the work force encourages more frequent eating outside the home. People at different ages or at different stages of their life cycles will also respond differently to messages about food. For example, the population as a whole is aging (DHHS, 1988), and an aging popu- lation is likely to be more aware of and concerned about health and may therefore be more disposed to seek out certain food components (e.g., fiber and calcium) or avoid others (e.g., fat). Much of the research concerned with modifiable determinants of food selection has been conducted either by marketers attempting to determine which appeals will be most effective in selling products or
38 IMPROVING AMERICA'S DIET AND HEALTH by researchers interested in the factors that promote overeating. Yankelovich, Skelly and White, Inc. (1985) identified "convenience," "price," "nutrition," "variety," "quality," and "good taste" as the variables that will establish "competitive parameters for those who will serve tomorrow's consumers." Rodin (1980), listing "social and immediate environmental influences on food selection," identified "time of day," "accessibility/availability," "expedience," "variety," "media effects," "conditioned stimuli," and "emotions." Many of the factors on both lists are not directly relevant to the question in this chapter: what factors can be manipulated to affect food choice in a healthful direction? The only common factor on the lists is variety, which often represents to a marketer a way of getting a larger share of the market. In that sense, variety is related to newness, which appears to be an inducement to consumers to at least try a product. Evidence indicates that a monotonous diet leads to de- creased food consumption and the availability of a variety of tasty foods leads to increased calorie intake, even among animals who are normally very good at self-regulation (Sclafani and Springer, 1976~. Humans may have room for a dessert even when they are entirely sati- ated from previous courses. Thus, increased variety is unlikely to be helpful in a situation where overconsumption is part of the problem. Price is often mentioned as influencing food choice, and it has played an important role in at least two major health-related dietary changes: the shift from butter to margarine that began during World War II (Green, 1975) and the shift from red meat to chicken that began in earnest in 1976. Between 1976 and 1987, chicken consump- tion increased by 48% while the average retail price of chicken as a percentage of the price of beef decreased from 40 to 32% (Putnam, 1989~. The importance of price as a factor affecting food choice obvi- ously varies, however, with the proportion of the family budget spent on food. Although increased income does not necessarily lead to an improvement in dietary quality, inadequate funds may limit consump- tion of costly fish and (at certain seasons) certain fresh fruits and vegetables. Many other health-promoting foods (e.g., breads and other grain products, starchy tubers, and dry beans) are relatively cheap, and many less desirable foods (e.g., sweet and salty snacks, rich desserts, and heavily marbled beef) are relatively expensive. Thus, price, combined with appropriate education, is a variable that could in some cases favor adoption of dietary recommendations. Very little is known about how individuals (or populations) ac- quire taste preferences. Familiarity appears to be important among adults as well as among children; however, what tastes good or ap- propriate at a given time to any one person undoubtedly relates to
DETERMINANTS OF FOOD CHOICE 39 some of the factors identified by Rodin (1980), for example, "condi- tioned stimuli" (a cocktail with the evening news, popcorn at the movies, hot cocoa at bedtime), "emotions" (candy during times of sadness), or "time of day" (ham and eggs for breakfast). Food pref- erences governed by such factors would be amenable to change if education leads to changed social norms. In discussing factors that affect food choice, a distinction must be made between what is and what is perceived to be reality by the po- tential consumer. In that sense, quality (like nutritiousness or health- fulness) is a belief factor. Quality can mean very different things to different consumers all the way from the fact that a food bears a well-known brand name to the fact that it bears no brand name at all and is purchased fresh from the farmer who produced it. The char- acteristics of foods that groups of people associate with quality can obviously change over time (e.g., among certain groups, marbling in beef has been replaced as a quality factor by beef raised without hormones). Nutritiousness has recently become identifiable as one characteris- tic of a quality product, although no more than 15 years ago, manu- facturers resisted nutritional marketing appeals on the grounds that people were simply not interested (Belasco, 1989~. At present, consumers will, at least some of the time, select food they believe to be nutritious. Perceived nutritiousness, especially in a food already highly desired, is a selling tool. Low calorie is another quality appeal in a culture in which at any given time 33% of women and a smaller percentage of men claim to be dieting (Calorie Control Council, 1989~. The astonishing success of the marketing of diet soft drinks, whose consumption shot up from approximately 1 to 8 gallons per capita between 1954 and 1987 (USDA, 1989), is a clear indicator that identifying a product as diet or low calorie will increase the likelihood that it will be selected by a substantial segment of the population. The selling of the potato as a low-calorie food (Dugas, 1985; Ketchum Communications, 1989) is an example of a marketing approach that might be used to some advan- tage in implementing dietary recommendations. The fact that $3.7 billion is spent annually on food advertising (Advertis- ing Age, 1989) has led to a popular conviction that advertising, especially on television, is a major influence on food choice. Although advertising agencies survive by convincing clients that this is true, the direct effects of the media on food selection are difficult to isolate from all the other promotional factors to which an individual is exposed. Ad- vertising induces people to try new products that might otherwise g unnoticed; it has encouraged the belief that all thirst must be quenched
40 IMPROVING AMERICA'S DIET AND HEALTH from a bottle not from the water tap; it can lead consumers to switch from one brand of soda or tuna to another; and the repeated picturing on television of highly palatable food may induce snacking (Falciglia and Gussow, 1980~. Of all the factors affecting food selection, two availability of foods and knowledge of and beliefs about foods and health are perhaps the most powerful of those amenable to modification. Availability is a much less obvious concept than it seems, incorporating such notions as convenience and technological progress. Real availability-the presence of enough varied food to eat is not an issue in the United States, where variety and quantity abound. However, the sheer number of choices does sometimes constrain availability, since food stores tend to feature the products that sell fastest, so that, for example, refined flour products have been more readily available than those made of whole grains. Nevertheless, certain desirable products such as lower-fat meat and a greater vari- ety of fresh fruits and vegetables are now becoming increasingly avail- able (Duewer, 1989; Greene, 1988~. Perceived availability is a different sort of factor. It changes over time among different groups with different skills and expectations. For many people, a food is now considered to be available only when it can be acquired in a few minutes or is ready to eat at any time of the day or night at a nearby location. This definition of availability restricts many people's food choices; for example, what is available for immediate consumption in many settings may be limited to a variety of bottled liquids and a collection of small packaged snacks. Increasingly, especially in urban areas, there are specialty shops that sell foods with highly concentrated energy components premium ice cream and freshly baked cookies, for example. With regard to these prepacked or proportioned street foods, it is usually more "ex- pedient" (to use Rodin's word) to eat the whole thing the whole cookie (however enormous), the whole package of crackers or nuts, or the whole bottle of beer or soda regardless of actual appetite, since that is what is available. Because snacks and fast-service foods of all kinds are ubiquitous, they are seen as choices, even though they provide a limited variety and less control over fat, sodium, and sugar intake than people might want. Convenience a term applied to something that promises to save work or time-is a subset of availability. To someone who feels time-constrained, a food that requires extended preparation is not perceived as available. To someone without cooking skills, a raw chicken is not available. Technology's impact on food choices results partly from its ability to continually redefine perceived availability.
DETERMINANTS OF FOOD CHOICE 41 Microwave ovens, for example, lead to increased availability, and thus consumption, of microwavable snacks (Erickson, 1989~. Thus, instant heating makes hot snack foods more available. Cultural availability is important, because one's culture determines what constitutes food; all cultures reject some edible parts of their envi- ronments. In the United States, dogs, cats, and horses are seldom eaten, although they are readily available, and Americans do not think of hunting birds and squirrels in parks. In many cultures, milk is considered to be an inappropriate food for adults, and in many others, bread is not spread with butter or margarine as it is in the United States. Because this country is relatively young and culturally diverse, it has no traditional national cuisine no foods that most of its citizens have eaten for generations. Regional foods (e.g., baked beans and brown bread or grits and red-eye gravy) have tended to be displaced by the cuisine offered at franchise restaurants. This lack of a long-standing, strong food tradition may prove to be an advantage to those attempting to produce dietary changes directed toward health. Cultural factors determine not only what but also when, where, how much, and how quickly food is to be eaten. Since the U.S. population has traditionally bolted its food (Fletcher, 1899), fast-service food is nothing really new. People in the United States also spend more time alone than people in many other countries do (Szalai, 1972) and they often eat alone in cars, at their desks, by the refrigerator, or close to vending machines (Lantis, 19621. Foods are increasingly available in quantities designed to be eaten alone, which means that any attempt to alter eating patterns must be directed at different population segments, not merely at adults, since they are no longer the "gatekeepers" identified by Lewin (1943~. Efforts to affect eating patterns need to be attentive to these ambiguous cultural messages since what will be eaten is so often dependent on where, when, and how quickly it is to be eaten. 1 1 ~ . . . . As documented later in this chapter, consumers have become more concerned about the relationship between diet and health and report that they are trying to change their diets accordingly. This interest is confirmed by the increasing emphasis on healthfulness as an important food marketing tool. During the past few decades, there have been substantial changes in overall food consumption patterns (Putnam, 1989~. Survey data show a widespread verbal commitment to eating for health, but consumption data show that declining consumption of beef, eggs, butter, whole milk, and other traditional contributors of saturated fat and cholesterol has been countered somewhat by a rising consumption of cheese, premium ice cream, and other rich sources of saturated fat (Popkin et al., 1989; Putnam, 1989~. In providing con
42 IMPROVING AMERICA 'S DIET AND HEALTH sumers with information that will allow or induce them to act on their stated health concerns, attention will need to be paid to helping them place the confusing bits of information they encounter into a coherent overall picture of the association of diet with health. CHANGES IN FOOD SELECTION It is evident from the preceding discussion that learning why people eat what they do is a complicated undertaking. Finding out exactly what individuals eat is only marginally easier (see, for example, NRC, 1989 and Woteki, 1986). It is possible, however, to obtain reasonably good data on overall changes in the U.S. food supply over time. These are useful for tracking trends in food demand and can be examined to learn whether food consumption patterns are changing in a direc- tion consistent with dietary recommendations. Changes in foods available to the public from 1909 to the present can be identified by examining U.S. Department of Agriculture (USDA) data on the disappearance of foods into wholesale and retail markets. The amounts of foods available to the public in a given year are estimated by subtracting data on exports, year-end inventories, and nonfood uses from data on total production, imports, and inventories at the beginning of the year. These quantities are larger than those actually consumed, since they do not include losses from processing, marketing, and home use (NRC, 1989; Putnam, 19891. USDA has also surveyed food use of households and dietary intakes and patterns of individuals in the Nationwide Food Consumption Surveys (NFCS) and the Continuing Surveys of Food Intake by Individuals (CSFII). Since the overall pattern of changes in both the NFCS and CSFII are generally consistent with the patterns shown in the disappearance data (Popkin et al., 1989), only the latter are presented here. 1 1 Table 3-1 presents the quantities of food available for consumption per person from periods extending from 1909 to 1987 (the latest data available when this report was prepared). Since data on some foods, especially processed vegetables and fruits, were not collected in the earlier years, comparisons of the consumption of these products over time are difficult to make. Changes from 1909 to 1987 Since the first settlers arrived in a New World that was teeming with game, meat has had a dominant position in the diets of its in- habitants. Although beef consumption in 1987 was the lowest since the 1960s, it remained approximately 40 to 50% higher than that dur
DETERMINANTS OF FOOD CHOICE 43 ing World War II and the preceding years back to 1909. Much of the apparent decrease noted in 1987 may be misleading, however, since retail cuts were much more closely trimmed of fat in that year than they were in the past (Putnam, 1989~. Thus, past disappearance data probably included some weight that was trimmed before consumption. Although year-to-year fluctuations in pork consumption have often been quite high, the long-term average weight of pork available per person has varied little during the past eight decades. The-most remarkable change has occurred in poultry consumption, which now averages 78 lb per person nearly five times higher than pre-World War II levels. This increased intake of poultry and a much smaller increase in fish consumption have more than made up for the decrease in beef, veal, and lamb. The annual consumption of total red meat, poultry, and fish in 1987 was not only the highest ever in the United States but it also exceeded that of the traditional leaders Australia and New Zealand. Consumption of dairy products peaked in 1945. Per-capita consumption decreased until the 1960s and 1970s, when consumption of low-fat milk, cheese, and frozen dairy products such as ice cream began to increase. Low-fat milk (1 to 2% fat, skim, buttermilk, and some flavored milk) consumption almost doubled between 1971 and 1987, when consumption of low-fat milk overtook that of whole milk. Egg consumption has decreased to 67% of its World War II high, but is only 15% less than its prewar level and has remained fairly constant in the 1980s. Although butter and margarine use combined has changed relatively little since the 1940s, margarine use has increased at the expense of butter. Consumption of fats and oils has steadily increased, reaching a point in 1987 that was approximately 50% higher than that recorded in the period from 1909 to 1913. Per-capita use of salad and cooking oils has increased markedly in the past two decades. Data on average vegetable consumption are less clear than those for other food groups because the sources of these data have changed. For example, current data are no longer available on several veg- etables. Putnam (1989) reported, however, that per-capita consump- tion of nine major fresh vegetables asparagus, broccoli, carrots, cau- liflower, celery, corn, lettuce, onions, and tomatoes reached a record high in 1987. In the past decade, per-capita consumption of frozen vegetables has increased while consumption of canned vegetables has decreased. Based on disappearance data, fresh fruit consump- tion has increased dramatically in the past two decades. However, food consumption survey data give a somewhat different picture. According to USDA, consumption of fruits and vegetables among women ages 19 to 50 actually declined by an average of 7% between
44 Cat oo A: Con o o UG ·_4 o ._ ~ i 1 1 ~ 1 0 1 $ - o a; ,9 ·_1 ¢ o o o V) . - ·_t ¢ Em 1 _ 1 1 ~ 1 o ._ 1 1 I ~ 1 ~ o ._ a; ! ~ - ._ ¢ ._ o ~ Go ~ Cal o ~ Cat Cr) 00 co en ~ (~- ~ t~ cat AD Cat Go Go Go . . . . . . Cal LO ~ ~ Go Lr) Cal ~ O ~ ~ ~ O ~ ~ 00 ~ . . .. . . .. . . . . O ~ ~ON CO ~ AN~ O ~ Cal Us C~o ~ C~ CN~ CM C~ 00 ~ ~ ~ O ~O ~d~ ~ ~ ~ L~ .. . . . . .. . . .. . . . . C~ L~ ~00 d ~ ~ 00 ~ C~ 00 ~D ~ O ~ ~ ~ ~ ~ ~ O ~ ~ . . . . . . . . . . . . . M cN csX o Oo c~ ~ ~ ~ 00 ~D ~ ~ ~ CO cN ~ cN ~ ~ ~o o ~ oo ~ . . . .. . . . . ~ ~ cN Oo~ ~ ~ cO o Lc) ~ cN cN L~ CO C~ 00 ~ ~ d~ CO ~ 00 0 oO ~ 00 ~ O ~ Lr) CO oo ~ CO ~ CN ~ CO ~ ~ D ~ ~ ~D CN O ~ C~ CN ~ C~ ~ ~ CN ~ di ~ ~ ~ ~ O LC) C~ ~ ~ 00 0 ~ ~D LC) CN CO LC) C~ O L~ ~ ~ ~ CO CO . . . . . . . . . . . . . . . . ~ C~ co 00 LO CO U~ ~ ~ ~ ~ LC) d~ oo O ~ ~ ~ C~ CO ~ C~ ~D ~ ~ CO 00 ~D CN ~ ~ C~ O ~ ~ ~ oo O ~ O C~ c~ ~ ct) ~ O ~ ~ d~ ~ ~ ~LC) ~D ~1 ~ ~cr) ~ O0 CO CO C~ ~ U~ . . . . . C~ ~ 00 ~ ~ ~n ._ _ ~ ~ 3 ~ _ ~ ~ ~ - ~ E Y ~D ~ D ~ ~ ~ e ~ s _4 ~=, ~cIO ~ pL~ ~ C) ~ cn 0 ~ ~
45 Cat ~ LC) ~ ~car CO ~ O ~ ~ Cal didoCM . . . .. .. . . . . . . ... oo ~ ~ C~00 ~ di ~ oo ~ ~ LC) COCMO c~ ~ ~r~ ~0 c ~ c~ 00 ~ ~ ~ O~ CN ~ CO ~ ~ ~LO~ O ~ ~ ~ ~N~ ~00 ~ U~ O CN ~CO O0 ~ ~ ~D~ ~O ~ ~ ~ ~ O C~ oO ~d~ Ocr) oo CM O C~ ~ O ~LC)CO . . . .. .. . . . . . . ... oo ~C~ ~ ~ ~ ~ oo ~ ~o ~ CN ~ LC)~ ~o C ~ ~ ~o d~ ~ ccc ~O O ¢ ¢ r~ ~)t~ ~i C,) 1` ~ 7 7 ON ~00 ~ ~ ~ ~00C~ COO O . . .. . ~ U~ O CM CO ~00 CN ) CN~ O ~ C~) ~S)Ct) C~ ~ ~Lr) ~D O O ¢ ¢ ~O CN 00 C~) 7 ~7 [f) ~00 ~ ~ C~ ~C~ CN ~ O O O =¢ ¢ O ~ O Cix 7 O0 CN CN aJ ~ C ~ . . i - 3 ' 5 A I _ C _ O ~ ~ ~ ~ ~ ~ O .Q ~ V .~ ~ ~ ~ ~ .= o O i~ O O O ~ - .> C~. O O · - .5 ~ O C~ ~ O ~ U 00 ~ C~ crx ~ o . - o - P~ O P~ ~0 . X ~ O ·- .> ~ V O ~ Cl) ~a. ~C.0 -° ~ O O ~ ~ =~ O o~ crs ~ 5 0 ~ ~ O ~ ~ ~ .= X O u ~ ,, ~ ~ ~ 2 ~ z s O c~ ~ `,U, `~ =° a ~ Z ~ ,,, ~ ~ ~ ~\ C, u: ~ u' u .= 5 - o . u . 4 U) U) ~ ~ - ~ .> UD U ,~ ~ oo u - o r, _ o u UO ' C~ ~_ b0 b4 ' -1 ~ l r, o~ U ~ O U ~ ~ UO ~4 u
46 IMPROVING AMERICA'S DIET AND HEALTH 1977 and 1985 (USDA, 1985~. The decline was most pronounced among low-income women (15% for fruits and 21% for vegetables). Consumption of flour and grains has increased in recent years, following a dramatic drop during the first part of this century (Putnam, 1989~. Average consumption of white flour has increased 15% in the past two decades, largely as a result of the greater demand for pasta. Breakfast cereal consumption was also up from 11 lb per capita in 1970 to 1974 to 15.2 lb per capita in 1987. Some of the greatest changes in the U.S. diet during the past two decades have been seen in beverage use (Putnam, 1989~. Between 1966 and 1987, per-capita coffee consumption decreased from 35.8 to 26.5 gal. In contrast, consumption of tea, both hot and iced, increased from 6.5 to 7.0 gal and per-capita consumption of soft drinks went from 17.9 gal in 1966 to 30.3 gal in 1987. The reported adult per- capita consumption of alcoholic beverages also grew from 32.1 to 40.1 gal between 1966 and 1987; declines in use of distilled spirits were countered by increases in the use of beer and wine. Sales of snack foods increased by more than 8% in 1985, reaching a total of nearly $7.5 billion (Supermarket Business, 1986~. Within this category, the greatest change since 1984 has been seen for fruit rolls and fruit bars, which increased nearly 23%. Potato chip sales have grown by more than 9% since 1984, totaling more than $1.8 billion- twice that of any other snack item. In 1985, candy and gum sales reached a total of nearly $9 billion; sales of diet, low-calorie, and sugarless candy and gum accounted for less than 1% of this total. High-calorie, high-fat gourmet foods are selling as well as some of the newer low-calorie, low-fat products. For example, sales of super- premium (higher in butterfat) ice cream increased by 20% in 1985 (Progressive Grocer, 19861. The great growth in the number of fast-service food outlets has had a substantial effect on food consumption patterns (Capps, 1986~. Potato products, cheese, tomatoes, and chicken have benefited from increased consumption away from home. Economic and Demographic Influences on Change From 1970 to 1985, poultry became a lower-priced alternative to red meat. Similarly, consumers switched to vegetable-based fats and oils, having been attracted to their lower prices compared with the more expensive animal-based fats and oils. In contrast, fish con- sumption increased, despite increases in fish prices during the same period. Price, however, probably became less of a factor in the 1980s. The trends of the 1970s and early 1980s may have been fixed in place as a result of increasing concerns about diet and health.
DETERMINANTS OF FOOD CHOICE 47 Historically, as real income rose, demand increased for some rela- tively expensive foods such as beef, poultry, shellfish, processed milk products, and vegetables (Smallwood and Blaylock, 1981~. But large increases in real income are necessary to generate substantial increases in food consumption (Capps, 1986~. At present, income is no longer a primary determinant of food consumption or of nutrient intake, except for those at poverty levels (Senauer, 1986~. Rising real income is associated with greater demand for food products that are conve- nient to prepare and for meals served outside the home (Kinsey, 1983). Demographic changes have also affected food consumption pat- terns. The percentage of single-person households increased from 10.9% in 1950 to 22.5% in 1980, and the proportion of households with more than two people decreased from 60.3 to 46.2% (U.S. De- partment of Commerce, 1983~. Single-person and two-person house- holds use more convenience foods per person than do larger households (Capps, 1986~. Age-related factors are also associated with food selection. For ex- ample, older people are more likely to eat breakfast, whereas younger people are more likely to eat meals away from home (Schoenborn and Cohen, 1986~. CHANGES IN CONSUMER ATTITUDES AND KNOWLEDGE Not surprisingly, just as food choices are changing, so too are con- sumers' attitudes and beliefs about food. An emerging consensus that diet is a risk factor for major chronic diseases such as cardiovas- cular disease, cancer, and hypertension is influencing consumer behavior and the choices available in the marketplace. One useful source of information on consumer attitudes toward nutrition and their food purchasing behaviors is a national survey conducted annually since 1974 by the Food Marketing Institute (FMI) entitled Trends- Consumer Attitudes ~ the Supermarket. For example, in the 1990 survey (FMI, 1990), 95% of respondents reported being "very concerned" or "somewhat concerned" about the "nutritional content" of the food they eat. The proportion of shoppers who re- ported being "very concerned" has held steady since 1987. People age 50 and older were more concerned about nutrition than were people between the ages of 18 and 39, and although they ranked nutrition as less important than taste or product safety, they consid- ered it more important than price or ease of preparation. Of particu- lar note is the increase in the proportion of shoppers (from 9°/O in 1983 to 46% in 1990) who indicated fat content as their primary con- cern about the nutritional composition of foods (FMI, 1990~.
48 IMPROVING AMERICA'S DIET AND HEALTH Changes in the ways that consumers prepare food have tended to reflect nutrition concerns to some extent. The 1989 Trends survey (FMI, 1989) revealed that among the 55% of consumers who reported that they cooked or prepared foods differently since 1984 to 1986, 37% of them were frying less, 24% were using less salt, and 20% were add- ing less fat to foods. Consumers who either had not changed their food preparation practices or had already incorporated these behaviors into their food preparation patterns before 1984 were not surveyed on this question. The percentage of surveyed consumers reporting that they cooked or prepared foods differently since 1985 grew to 61% in the 1990 Trends survey (FMI, 1990~. The Food and Drug Administration's (FDA's) Health and Diet Sur- vey, conducted every 2 years since 1979, assesses changes in public beliefs and knowledge related to diet and health issues. In 1988, 71% of the respondents reported that they had changed their diet recently to prevent heart disease up from 61% in 1986 (Levy et al., 1988~. These data suggest that people who changed their diets did so in accordance with the principles of dietary recommendations. USDA per-capita food consumption data confirm these trends (Putnam, 1989~. As noted earlier, there has been a shift toward consumption of food perceived as low in fat or calories or with other positive characteris- tics (e.g., increase in consumption of poultry, fish, grains and other cereal products, fresh fruits and vegetables, and low-fat milk and yogurt and a decrease in consumption of red meat and whole milk). However, the improvements consumers are making in their eating habits are very selective, as indicated by a 1987 New York Times poll (Burros, 1988) based on telephone interviews with 1,870 adults na- tionwide. For example, consumers who decrease their intake of fat by eating smaller servings of a fatty food, such as a fatty cut of meat, may substitute a salad with 5 to 6 tablespoons of salad dressing, resulting in consumption of more fat and fewer nutrients. Food dis- appearance data support this conclusion (see Table 3-1), showing in- creased consumption of fats and oils (primarily from vegetable sources) between 1970 and 1987. According to a survey by American Health magazine (Mothner, 1987), eating and drinking patterns in the United States are changing to meet the needs of different situations, and specific foods believed to confer unique benefits are being selected. Is the U.S. public selective, or is it confused? The FDA's Health and Diet Surveys (Levy et al., 1988) include several questions related to consumer knowledge. In 1988, for example, 35% of those surveyed considered dietary fat as a risk factor for high blood pressure as compared with only 6% in 1979. As of 1988, 25% considered dietary fat as a risk factor for cancer. From 1979 to 1982, when the National Heart, Lung,
DETERMINANTS OF FOOD CHOICE 49 and Blood Institute (NHLBI) and FDA launched an initiative to inform the public about the relationship between sodium intake and hypertension, there was a 300% increase in respondents who mentioned sodium as a risk factor for hypertension. The Health and Diet Surveys also indicated that during the highly visible advertising campaign for Kellogg All-Bran cereal (1984 to 1986), there was a 350% increase in respondents who mentioned fiber as a factor that might help to prevent cancer. The percentage of people who perceive dietary fats as a risk factor for heart disease rose from 29% in 1983 to 55% in 1988. The mention of cholesterol as a risk factor for heart disease increased similarly. On the other hand, these surveys indicated that respondents did not understand which kinds of foods contain cholesterol, what the term cholesterol-free means, whether vegetable oil contains saturated or unsaturated fat, and whether some fats are higher in calories than others. Seventy percent of respondents to the 1988 Health and Diet Survey were either not sure or believed that all foods containing fat had cholesterol. Forty percent of the respondents thought that a food containing vegetable oil or labeled cholesterol-free would also be low in saturated fat; another 20% of the respondents were not sure what cholesterol-free implied about the saturated fat content of a food. Only 20% of the population realized that all fats have essentially the same number of calories. In short, it appears that much of the public equates cholesterol with saturated fat, which is "bad," while anything free of cholesterol is equated with low saturated fat, which is "good" (Levy et al., 1988~. Yet less than 20% of the respondents knew that hydrogenation made fat more saturated, and only 27% had ever heard of monounsaturated fat. Moreover, 65% of the population questioned by the FDA in 1988 did not know that polyunsaturated fats are more likely than saturated fats to be liquid. These data suggest that the public is highly concerned about diet and health but is lacking the detailed knowledge needed to act effective- ly on these concerns. Another interpretation of the data is that the public, however concerned, does not wish to become knowledgeable in biochemistry in order to eat well. The challenge then becomes not to try to teach them more than they want to know but to give infor- mation that will allow them to make health-promoting food choices based on some easily comprehended rules. This information should be packaged in ways that fit into the busy lives of people. Finally, people experience the pull of many interests and motivations that compete with eating healthfully, and most people are reluctant to make drastic changes in life-style (Light et al., 1989~. The relation- ships between such attitudes and motivations and dietary behavior are explored in greater detail in the following section.
50 IMPROVING AMERICA'S DIET AND HEALTH THEORIES OF ATTITUDE AND BEHAVIOR CHANGE The ultimate goal in implementing dietary recommendations is to produce desirable behavior changes. Such changes can result from a variety of causes both environmental and internal. For example, a person may shift to a lower-fat diet because of a change in the food supply (an environmental cause) or because of new beliefs about the role of fat in the diet or the acquisition of new food preparation skills (internal causes). Since it is unlikely that implementation strategies can rely on environmental changes alone, it is important to examine the role of internal factors in influencing human behavior. Two major approaches to behavior change that rely on internal factors are the communication/persuasion model and the social learning model. Some understanding of the findings of basic research on influ- encing human behavior may help to guard against either overly optimistic or overly pessimistic assessments of the prospects for changing eat- ing habits to promote health. Communication/Persuasion Model The communication/persuasion model focuses on modifying attitudes as a means of changing behavior. Attitudes are people's general pre- dispositions to evaluate other people, objects, and issues either favorably or unfavorably. The attitudes relevant to implementing dietary rec- ommendations range from general attitudes about changing diets for health purposes (i.e., is it perceived to be a worthwhile idea or not?) to attitudes toward specific foods (e.g., do I like premium ice cream?. This construct has achieved a preeminent position because of the assumption that attitude change is an important mediating variable between the acquisition of new knowledge and behavioral change (see reviews by Chaiken and Stangor, 1987; Cialdini et al., 1981; and Cooper and Croyle, 1984~. That is, new knowledge (e.g., saturated fat raises cholesterol levels) is believed to produce new attitudes (e.g., saturated fat is bad), which in turn produces new behavior (e.g., avoidance of foods high in saturated fat). Attitude Change Theories Early theories of persuasion were based on the assumption that effective influence required a sequence of steps (see, for example, McGuire, 1985 and Strong, 1925~. A first step typically was exposure of a person to some new information, from a single channel of com- munication or through multiple channels such as face-to-face con- frontations, the mass media, programs at work sites and churches,
DETERMINANTS OF FOOD CHOICE 51 and in-store food displays. Second, people must attend to the infor- mation presented. Because literally hundreds of messages compete for people's attention each day, relatively few are successful in at- tracting it (Bogart, 1967~. A third step is reception, which involves the storage of selected information segments in long-term memory. Just because a person is consciously aware of an informational presenta- tion, there is no guarantee that any aspect of what has been seen and heard will create more than a fleeting impression. Likewise, just because some new information is learned as a result of an educational campaign, there is no guarantee that this knowledge will lead to attitude or behavior change. Current research strongly indicates that attitude change depends upon the manner in which the information is interpreted, evaluated, and elaborated so that it makes some sense to the person. The more favorable the cognitive or affective response to the information, the more likely that attitudes will change in a positive direction. Once the information received has elicited various thoughts or feelings, these responses must be integrated into an overall evaluation or attitude capable of guiding subsequent action (Petty and Cacioppo, 1984~. A change early in this proposed sequence will not inevitably lead to a change later on, because each step in the sequence may be viewed as a conditional probability. Thus, even if the likelihood of achieving each step is 60%, the probability of achieving all six steps (exposure, attention, reception, elaboration, integration, and action) would be .606 or only 5% (McGuire, 1989~. Another reason why a change early in the sequence may not lead to a change in a later stage is because some steps in the sequence may be independent of each other. For example, although a person's ability to learn and recall new information (e.g., facts about nutrition or diet and health) was often believed to be an important causal determinant of, and prerequisite to, attitude and behavior change, little empirical evidence has accumulated to support this view (McGuire, 1985; Petty and Cacioppo, 19811. Rather, evidence shows that message learning can occur in the absence of attitude change and that people's attitudes may change in the absence of learning the specific information presented. For example, consider two people who hear that the con- sumption of oat bran can reduce one's cholesterol level and reduce the risk of heart disease. One person, who is overweight and constantly dieting, responds to this information with the thought that oat bran must be good for one's health. Later that week in the supermarket, the person passes the cereal aisle and selects an oat bran product because it is perceived as a diet food that will help the person to lose weight. In this instance, the original information about the link between
52 IMPROVING ANIERICA'S DIET AND HEALTH the product and heart disease was lost, but the self-generated (and probably mistaken) elaborations or translation of that information (from disease prevention to healthful to diet) guides the behavior. Another person hears the same information and responds with the thought that there is no need to be concerned since there is no family history of heart disease. Next week in the supermarket, this person passes over the oat bran products, even though there is a perfect recall of the original information that was presented. Since the person does not feel personally vulnerable, learning the new information has no effect on that person's behavior. Current psychological theories of influence focus on how and why various features of the persuasive communication (e.g., the message source, its content, and method of presentation) affect each of the steps in the communication sequence. The most work by far, however, addresses the ways that variables affect the elaboration stage of in- formation processing. This stage is sometimes viewed as the most critical, since it is during this stage that the presented information achieves meaning, is evaluated favorably or unfavorably, and is ac- cepted or rejected. Models of the processes that occur during the elaboration stage emphasize one of two relatively distinct routes to persuasion (Chaiken, 1987; Petty and Cacioppo, 1981, 1986; Sherman, 1987~. The first, or central route, involves cognitive activity whereby the person draws upon experience and knowledge to scrutinize and evaluate carefully the issue-relevant information presented in the communication. For this to occur, the person must possess sufficient motivation, ability, ~~-~~ ~~ ~~~~~~~ ~ Rae 1 and opportunity to think about the perceived merits of the information provided. The end result of this processing is an attitude that is well articulated and integrated into the person's belief structure. Attitudes changed in this way are relatively persistent, predictive of behavior, and resistant to change until they are challenged by cogent contrary information (Petty and Cacioppo, 19861. Using a biological analogy, McGuire (1964) suggests that just as people can be made more resis- tant to disease by giving them a mild form of a germ, people can be made more resistant to attacks on their attitudes by inoculating their new opinions. The inoculation treatment consists of exposing people to a few pieces of attacking information and showing them how to refute it. People whose attitudes are bolstered with inoculation treat- ments become less vulnerable to subsequent attacks on their attitudes than do people whose attitudes are bolstered with supportive infor- mation alone (McGuire and Papageorgis, 1961~. Attitudes may also be changed by a peripheral route in which simple cues in the persuasion context either elicit an affective state (e.g.,
DETERMINANTS OF FOOD CHOICE 53 happiness) that becomes associated with the advocated position (as in classical conditioning) (Staats and Staats, 1958) or trigger a relatively simple inference or rule that the person can use to judge the validity of the message (Chaiken, 1987~. Advertisers attempt to use this strategy when they associate a food product with good times and fun or when they invoke a simple inference such as, "A doctor said it, so it must be true." Changes induced by this peripheral route have been found to be less persistent and predictive of behavior than changes based on more extensive thought about the merits of the arguments in the message. Thus, a person who develops a negative attitude toward salt simply because his or her doctor said it was bad is likely to be less in compliance with the appropriate dietary regimen over the long term than is a person who developed a negative attitude toward salt after careful reflection upon the personal consequences of, and reasons for, the doctor's recommendation. Links Between Attitudes and Behaviors Once a person's attitude has changed, it is important that one's new attitude rather than the old habits guide behavior. A considerable amount of research has addressed the links between attitudes and behavior, and many situational and dispositional factors have been shown to enhance the consistency between them. Attitudes have been found to have a greater impact on behavior when, for example, (1) people are of a certain personality type (e.g., those who tend not to follow the opinions of others), (2) the attitudes are consistent with the person's underlying beliefs, (3) the attitudes are based on extensive amounts of issue-relevant knowledge or personal experience, (4) the attitudes were formed as a result of issue-relevant thinking, and (5) cues indicate that the person's attitude is relevant to the behavior (see Ajzen, 1989 for a comprehensive review). Two general types of theories regarding the process by which atti- tudes guide health-related behavior have achieved widespread ac- ceptance. One of them focuses on thoughtful reasoning processes, whereas the other type focuses on more automatic processes. A good example of the first type of theory is Ajzen and Fishbein's (1980) theory of reasoned action, which assumes that "people consider the im- plications of their actions before they decide to engage or not engage in a given behavior" (p. 5~. According to this theory, people form intentions to perform or not to perform behaviors, and these inten- tions are based on the person's attitude toward the behaviors as well as perceptions of the opinions of others (norms). This theory focuses on the perceived likelihood that certain benefits will accrue or costs
54 IMPROVING AMERICA'S DIET AND HEALTH will be avoided and on the desirability (or aversiveness) of those benefits (or costs). The specific beliefs that are relevant to taking recommended health actions have been outlined in the health belief model (Rosenstock, 1974~. These include beliefs about (1) one's personal susceptibility to some disorder, (2) the perceived severity of the disorder, (3) the subjective benefits of engaging in a recommended action, and (4) the costs (e.g., financial and psychological) of engaging in the behavior. That is, people are assumed to engage in health-related actions to the extent that they believe that some health concern is relevant to them and that the likely effectiveness and other benefits of the recommended action outweigh its costs. The reasoned-action theories have proven remarkably successful in accounting for a wide variety of behaviors (Janz and Becker, 1984; Sheppard et al., 1988~. In contrast to these theories of reasoned action, Fazio (1990) has proposed an accessibility theory, which suggests that much behavior is spontaneous and that attitudes guide behavior by a relatively auto- matic process. Specifically, Fazio argues that attitudes can guide behavior without any deliberate reflection or reasoning if the attitude (1) is highly accessible (i.e., comes to mind spontaneously by the mere presence of the attitude object) and (2) influences the perception of the object i.e., if the attitude is favorable (or unfavorable), the qualities of the object appear favorable (or unfavorable). For example, when confronted with a dish of ice cream, positive feelings may come to mind automatically, thus causing the ice cream to appear more desirable and leading the person to consume the food. The various costs and benefits of eating the ice cream may not be considered at all or may be weighed only after the food is eaten. The theory proposed by Fazio suggests some conditions under which the reasoned or the more spontaneous attitude processes occur. He notes that factors related to motivation, ability, and opportunity will be important in determining the means by which attitudes guide be- havior. Thus, for behavioral decisions that are perceived to have serious personal consequences, attitudes are likely to guide behavior by a deliberate process of reflection, but when consequences are perceived to be less serious, spontaneous attitude activation should be more important. Similarly, as the time allowed for decision making is reduced, the importance of spontaneous attitude activation processes should be increased over more deliberative processes. A typical shopper in a supermarket is confronted with many choices for every category of food and has a limited time to reach a decision. This environment is not likely to foster much cogitation. Rather, simple cues (e.g., packaging), old habits, and well-ingrained attitudes (rather than relatively
DETERMINANTS OF FOOD CHOICE 55 new and less accessible ones) are likely to guide choices. Much can be done, however, to simplify the decision-making environment (e.g., simple shelf labeling) and to provoke thought before selection (e.g., in-store health displays) so that one's new attitudes rather than one's old habits are more likely to guide behavior. Social Learning Model Models of attitude change focus on persuading people to adopt new beliefs and evaluations in order to change behavior. In some areas, the implications of a new attitude for behavior are relatively straightforward and simple to implement (e.g., preference and voting for a particular candidate in an upcoming election). Changes in atti- tude related to diet and health are an important first step but may be insufficient to produce the desired behavioral responses. People may also need to acquire new skills and self-perceptions that allow newly acquired attitudes and intentions to be translated into action. Fur- thermore, once an attitude has yielded a new behavior, this new be- havior may not persist in the absence of incentives. Bandura's (1977, 1986) social (cognitive) learning theory provides a framework for understanding these processes. This theory is based on evidence of a reciprocal linkage of a person's cognitions with both the person's behavior and the environment. Importantly, from the perspective of the social learning theory, the power of vicariously experienced and self-generated consequences in controlling action is recognized. As in the theory of reasoned action and the health belief model, described above, voluntary behavior in the social learning model is determined by the anticipated consequences of various courses of action (Rosenstock et al., 19881. These anticipations of rewards or punishments may be based on personal experience, on the observed experiences of others, or on cognitive reasoning processes. An individ- ual's beliefs about the consequences of behavior can be more influen- tial than the actual consequences in determining what actions are selected, and people are viewed as being capable of providing their own rewards and punishments for their actions (self-regulation). Importantly, before behavior can be changed, it may first be neces- sary for the person to learn new actions (skills) or new sequences of already acquired actions. For example, a person may have devel- oped a negative attitude toward saturated fat but does not have the food preparation skills required to eat less of it. New skills may be acquired through direct experience or through observations of the behavior patterns of others (modeling). The most effective models are people who are admired or people with whom individuals iden
56 IMPROVING AMERICA'S DIET AND HEALTH tify. However, people do not always behave in ways consonant with known behaviors. That is, learning (or knowledge) does not guaran- tee that the person will engage in the behavior that has been learned, because various incentives (both external and intrinsic) may be nec- essary to translate knowledge into action. One particularly important cognitive determinant of whether knowledge and attitudes have behavioral implications concerns people's assessments of their own capabilities (their judgments of self-efficacy) (Bandura, 1982~. People assess their capabilities in a variety of ways, e.g., by assessing their own accomplishments and performances or their current physiological state, by observing and comparing their behaviors with those of other people, and by considering the expectations and norms set for themselves by other people in their lives. Judgments of self-efficacy are important because extensive research indicates that the higher the level of perceived efficacy, the more likely people are to persist in a new, learned behavior (e.g., preparing and eating low- fat foods). Of the various ways to influence self-efficacy, provision of guided practice and specific skills training (e.g., teaching people how to improve their supermarket shopping skills or practicing spe- cific cooking methods) has proven to be especially powerful (Meyer and Henderson, 1974~. Thus, an important procedure for instilling self-efficacy is to set relatively small and easily accomplished goals initially (e.g., change one's breakfast menu for the next week), and as self-efficacy increases, provide more challenging goals (e.g., change one's dinner habits). This procedure is based on the theory that the satisfaction derived from subgoal attainment can build intrinsic interest in the task at hand, thereby leading to persisting change. Implications of Theoretical Perspectives for Changing Diet for Health Although evidence has shown that it is possible to improve people's knowledge of food and nutrition, this new knowledge does not invariably result in attitude and behavior changes (Axelson et al., 1985~. Infor- mation will be successful in producing relatively enduring changes in attitudes and behavior only if people are motivated and able to process the information and if this processing results in favorable cognitive and affective reactions. Furthermore, once attitudes have changed, implementation of change may require learning new behav- ioral skills and developing feelings of self-efficacy. Although some attitudes are based on a careful reasoning process, others are formed as a result of relatively simple cues in the persua- sive message or the surrounding context. There are important conse
DETERMINANTS OF FOOD CHOICE 57 quences of the manner of attitude change, such as whether or not the attitude change will last. For example, it is possible to change attitudes without providing an extensive informational campaign, but the resulting attitudes are likely to be less stable and directive of behavior. Finally, just as some attitudes are thoughtfully based whereas others are not, some behaviors are the product of deliberate reflection on costs and benefits, but others are much more spontaneous. Since implementation of dietary recommendations necessitates long- lasting changes in attitudes with behavioral consequences, the central route to persuasion appears to be the preferred influence strategy. Unfortunately, this is not simple. The recipient of the new information must have the motivation, ability, and opportunity to process the new information. One of the most important determinants of motivation to think about a message is the perceived personal relevance of that message. When personal relevance is great, people are motivated to scrutinize the information presented and to integrate it with their beliefs, but when there is little perceived relevance, messages may be ignored or processed for peripheral cues, such as whether or not the source of the communication is attractive. Perceived personal vulner- ability to some threat has been found to be a particularly important determinant of preventive health beliefs and behaviors (Tanz and Becker, 1984~. Yet many young people in the U.S. population, for example, may believe that health messages about food are not aimed at them or have few consequences for them. An important goal of any implementation strategy should be to increase people's motivation to process messages about diet and health by increasing the perceived personal relevance of these messages. That is, the issue of diet and health must become rele- vant to more people than just those in high-risk populations. Even people motivated to think about diet and health messages must respond to these messages with favorable cognitive and affec- tive reactions before there can be long-lasting changes in attitudes with behavioral consequences. Different types of information will elicit different responses in various segments of the population. Since attitudes toward food are based on a variety of factors (e.g., social norms and perceived healthfulness), different types of appeals and strategies will need to be developed for the diverse targets of influence. Messages concerning diet and health are likely to be relatively complex and therefore difficult to assimilate and implement. Unlike some preventive health messages focused on only one attitude that is to be changed (e.g., cigarettes are bad, so don't smoke), messages on diet and health require considerable sophistication to process. Mul- tiple attitudes are involved, and it is generally inappropriate to label certain foods as invariably bad. Much research is needed to deter
58 IMPROVING AMERICA'S DIET AND HEALTH mine the level of complexity at which messages can be presented to different audiences and the type of information that would elicit posi- tive reactions. Even if the relevant attitudes are changed, the new attitudes can- not influence behavior if they do not come to mind before the oppor- tunity for action is presented or if people lack the necessary skills or confidence to implement them. People will need to be encouraged to think before they eat so that the new attitudes will come to mind. They must also acquire the behavioral skills needed to implement their new attitudes. People may form positive attitudes toward low-fat entrees as the result of an educational campaign, but if the first low-fat meal they try in their workplace cafeteria is unpleasant, two different evalua- tions are formed "low-fat food is good for you" and "low-fat food tastes bad." Since beliefs and attitudes based on direct experiences come to mind more readily than do attitudes that are based solely on externally provided information, the effectiveness of the information favorable to low-fat food (even if it comes to mind eventually) is severely attenuated (Fazio and Zanna, 1981~. Consistent with social learning theory, various procedures for enhancing self-efficacy such as skills training, self-regulation, and reinforcement may be needed in order to maintain behaviors that reflect a person's new attitudes. COMMUNICATION THROUGH THE MEDIA The two models described in the previous section the communi- cation/persuasion model and the social learning model rely heavily on the media to influence behavior. It is therefore appropriate at this point to describe briefly the components and functions of the media. The following section provides examples of the media being used in schools, work sites, and communities to improve dietary patterns. Several of the committee's recommended strategies and actions to implement dietary recommendations, described in Chapters 5 through 8, involve the use of the media. The media are used frequently to inform, educate, and motivate health-seeking individuals and members of communities to improve their dietary behaviors. This may occur in relatively informal ways (e.g., health promotion messages in magazines or television shows) or in planned campaigns as described in the next section. The media are a principal source of information on food and nutrition for many people. Because the success of the media in influencing consumer knowledge and behavior is widely acknowledged, it will continue to play an important role in any comprehensive efforts to improve the
DETERMINANTS OF FOOD CHOICE 59 diets and health of the U.S. population. The committee uses the term media to refer to broadcast media such as radio and television; print media such as newspapers, magazines, and brochures; and newer forms of communication such as videos and computer programs. Two major functions of the mass media are to inform and per- suade. Recent incidents have demonstrated clear links between cov- erage of an issue and consumer response. Extensive media coverage of a report claiming that children were being excessively exposed to pesticides and other agricultural chemicals, particularly from Alar on apples, led to a farce dron in angle consumption (Smith 1ssn~ and 1 l ~ ~ r ~ -rr~~ -a rat A ~ ~ , ~^L~ ~ ~ ~ . . _ _ _ _ _ helped to reduce consumer confidence in the safety of the food supply (FMI, 1989, 1990~. In another example, increased sales of Kellogg high- fiber All-Bran cereal resulted from advertising the product with the National Cancer Institute's message that high-fiber diets may reduce the risk of colon cancer (Freimuth et al., 1988; Levy and Stokes, 1987~. The FDA, reporting the results of its 1979 to 1988 Health and Diet Surveys, noted that increased public knowledge of diet and disease relationships "seems to occur during periods when the diet/disease messages gain access to the mass media" (Levy et al., 1988~. It added that the increased media coverage that effected these gains in public awareness was primarily due to "intensified and coordinated health promotion efforts by industry, government and the scientific community." The four broad sectors of society addressed in this report public, private, health-care professions, and education (Chapters 5 to 8) use the media as a tool to communicate their messages on diet and health because of the media's ability to reach many people in a cost-effective manner. Media access is often purchased, but it may be available at no charge if the media decide that a particular activity or initiative is newswor- thy or that it should be promoted as a public service. To interest the media in promoting dietary recommendations, implementors must find ways to make the messages continually appealing and newsworthy, since the media present material that they believe will attract and interest consumers and thereby enhance their ability to attract funds and advertis- ers. Therefore, implementors should learn, for example, to write effec- tive press releases, provide accurate and useful background material, and offer spokespersons for interviews. The media can also provide instruction and illustrations on how to make and maintain desirable behavior changes. This function of the media needs to be marshaled by implementors in communities across the United States. Media attention to dietary recommendations and their implementation is likely to be most effective in improving eat- ing behaviors when the messages can be integrated and coordinated with other community health promotion efforts such as face-to-face
60 IMPROVING AMERICA'S DIET AND HEALTH instruction in classes, incentive programs at work sites, and wide- spread availability of health-promoting foods at restaurants. One particularly challenging task for implementors will be to use the media in more innovative ways to improve the health-related behaviors of the difficult-to-reach segments of the population, including low-income, poorly educated, low-literacy, and some minority groups. To reach these groups, implementors will need to adapt their messages to make them culturally specific and personally relevant (Freimuth, 1990; Freimuth and Mettger, 1990; Nickens, 1990; White and Maloney, 1990~. Difficult- to-reach people are most likely to understand and accept messages that are simple, concrete, and "adapted sensitively to the target audience's cultural beliefs" (Freimuth, 1990, p. 181~. The messages should probably emphasize immediate rather than long-term benefits from adopting healthy behaviors and be constructed so as to stimulate interest in the subject. Dissemination of information through printed brochures and pamphlets is not likely to be very effective with groups that tend to be infrequent readers and have poor reading skills. REVIEW OF EVIDENCE ON CHANGING DIET TO BENEFIT HEALTH Because of the many complexities involved, public policy initiated to modify diet to benefit health should call for the use of the commu- nication/persuasion and social learning models. Information regard- ing diet and health (communication inputs) can and should come from a variety of sources, including the mass media, face-to-face in- teractions, community events (e.g., health fairs and contests to pro- vide incentives to change), and from alterations in the environment (e.g., provision and clear display of health-promoting food choices in schools and retail food establishments). As shown in Table 3-2, communication roles differ for the sender (the educator) and the recipient (the learner). The sender begins by determining the receiver's needs so that the appropriate information can be presented. The sender must next set a new agenda by gaining the attention of the receiver, a step analogous to the town crier's "Now hear ye." In modern times, the social marketing approach can awaken interest by a variety of means and messages that create a sense that the issue (e.g., nutrition) is personally relevant. The sender's next two steps provide the factual information needed, coupled with incentives for the receiver to make changes. The first three steps shown for the receiver are the essence of the communication/persuasion model. The receiver first becomes aware of the information and may end up with changed beliefs and atti
DETERMINANTS OF FOOD CHOICE TABLE 3-2 The Communication/Persuasion-Social Learning Formulation 61 Communication Functions for the Sender (listed in desired sequence) Determine receiver's needs Set agenda (gain attention) Provide information Provide incentives Provide training Provide cues to action, including environmental change Provide support and self-management skills 1 1 Target Objectives for the Receiver (listed in desired sequence) Exposure to and awareness of communication Reception of message and knowledge change Change in interest, thinking, motives, and attitudes Learn and practice new skills Take action and assess outcomes Maintain action, practice self-management skills Become an opinion leader Give feedback to sender tudes if the persuasion is successful. These changed attitudes and beliefs render the person more open to behavior modification attempts. The next three steps are the essence of the social learning model. Once attitudes are changed, people are motivated to learn new skills to achieve and maintain behavior change. In the last two steps, the receiver becomes a transmitter and advocate. Although there is an extensive literature on efforts to produce di- etary change for the purpose of weight reduction, there are fewer studies of educational interventions designed to produce qualitative dietary change, i.e., changes in the types of foods consumed; and only some of these describe precisely the type and amount of educa- tion provided. This section reviews more than 50 studies reported in the past two decades that have evaluated the effectiveness of nutri- tion education in producing qualitative dietary changes in either spe- cial subgroups or the general population. The results of these stud- ies provide insights into what succeeds (and what fails) and suggest the amount, type, and duration of education needed to achieve im- portant beneficial changes in eating habits. The following types of studies are reviewed: small-group class- room studies; programs that activate social support systems; school 1 ~
62 IMPROVING AMERICA'S DIET AND HEALTH based programs for adolescents; health communication at points of purchase; studies conducted at work sites; regional and national mass media campaigns; and comprehensive, integrated community-based multifactor risk-reduction programs. The first six types of studies are potential components of community-based interventions, which are reviewed later in the chapter. Small-Group Classroom Studies In small-group classroom studies, face-to-face communication has generally been combined with supplementary printed materials to teach selected groups of adults, often those at high risk for develop- ing health problems because of their life-styles or the presence of other risk factors of interest to the investigators (Meyer and Henderson, 1974; MRFIT Research Group, 1982~. These studies have shown that education based on social learning and self-directed change is more effective than the more traditional method of imparting knowledge alone (Bandura, 1986; Meyer and Henderson, 1974~. Programs that provide not only the information needed to change knowledge and attitudes but also instruction in methods of monitoring change and guidance on how to achieve gradual incremental change in dietary habits have been successful, for example, in lowering blood cholesterol levels (Arntzenius et al., 1985; Bruno et al., 1983; Carmody et al., 1986; Glanz, 1985; Hjermann et al., 1981; Meyer and Henderson, 1974; MRFIT Research Group, 1982; Puska, 1985; Wilhelmsen et al., 1986~. Educa- tional programs relying largely on self-help printed instructions that include the social learning features mentioned above have also been successful in lowering cholesterol levels by dietary means for adult participants identified as being at high risk for cardiovascular disease (Crouch et al., 1986~. An increasing number of small-group studies in such high-risk adults have been conducted in different cultural groups in various countries and have reported effecting lasting changes in dietary habits. A limitation of these studies is that their subjects were motivated to participate in them through screening tests that identified such factors as high blood cholesterol levels that placed subjects at high risk for cardiovascular disease. Some effective methods of reaching the general public, including those who will not volunteer for repeated classroom sessions, have been developed and are reported later in this chapter. Programs That Activate Social Support Systems Direct or indirect social support can help bring about dietary change, especially when information and role models are provided to the
DETERMINANTS OF FOOD CHOICE 63 many people who are not strongly motivated to seek either persona counseling or group classroom instruction. Such support can be provided, for example, by television, which is an increasing force in the dissemination of new commercial, political, or religious ideas to individuals linked to the media source but not usually to each other. Nevertheless, traditional networks of peers and family members that provide two-way communication are still the strongest influence on the individual, especially when the safety and good sense of a recommended novel approach to life-style and health are questioned, as reviewed by Rogers (1983~. These tradi- tional networks can be regarded as guardians of existing social norms that can clearly act as barriers to behavior change; however, they can also be activated to promote the adoption of new and beneficial health practices and are, thus, a potentially important resource. Through health communication programs, potential leaders from the general population can be identified to serve as informal health educators within their peer networks. In the Finnish North Karelia Project, for example, such individuals were recruited and, after a brief 4-hour training session, became a useful part of the program. Using face-to-face channels, they helped to communicate the project's goals, including a wide array of desirable dietary changes (Puska et al., 1986). Groups recruited from the general population on the basis of shared attributes, such as having a high risk of a particular disease, may also become effective communicators of health messages. In the Stanford Three-Community Study, for example, health educators gave a group of high-risk adults approximately 20 hours of intensive instruction in self-directed change (see more detailed discussion of this study later in this chapter). In comparison with other members of the adult population, those subjects were found to converse with a larger num- ber of their acquaintances more frequently on health topics (Meyer et al., 1977~. Those at equivalent high risk who were not exposed to this special education were not activated to spread the message. Sub- jects in this study exposed to a mass media health education cam- paign were more likely to become opinion leaders (or unofficial health educators, so to speak) if they felt personally involved by virtue of completing a lengthy survey in which they identified their personal cardiovascular disease-risk status (Meyer et al., 1977~. Given this evidence for the spontaneous creation of opinion lead- ers as a function of the degree of exposure to health education and of perceived personal relevance, one can see why a popularly conceived and implemented health communication program applied to any part of a system (such as a school, a work site, or a community) can lead
64 IMPROVING AMERICA'S DIET AND HEALTH to at least some dissemination of information and potential for be- havior change in individuals within that system. Any individual exposed to such an education program is therefore a potential par- ticipant in the spread of a health message. Results from family-centered programs buttress the view that so- cial support enhances individual instruction and elicits dietary change (Arntzenius et al., 1985; Bruno et al., 1983; Carmody et al., 1982; Ehnholm et al., 1982; Glanz, 1985; Hjermann et al., 1981; Meyer and Henderson, 1974; Nader et al., 1986; Perry et al., 1989; Puska, 1985; Wilhelmsen et al., 1986~. Social support networks are varied and numerous. To provide optimal health communication, one must first use the most effective natural networks (Rogers, 1983~. For example, birth control practices and agricultural innovations diffuse through different networks in the same community (Marshall, 1971~. Nutrition education will have its own natural network that can be identified through social marketing methods, such as the use of focus groups that bring 10 to 20 individuals together to learn who they talk to and rely on for sources of opinion in matters of diet and health (Kotler and Zaltmann, 1971~. School-Based Programs for Adolescents Through school-based programs, information can be channeled to peer networks of young children and adolescents. They have there- fore been used in numerous health communication studies, most no- tably in highly successful attempts to prevent smoking among adolescents (Best et al., 1988; Killen, 1985; Telch et al., 1982~. School-based programs administered to groups of tenth graders in two different areas of northern California have also resulted in self-reported decreases in saturated fat and cholesterol intake from snack foods (King et al., 1988) and statements from students that they would choose more "heart-healthy" snacks (Killer et al., 1988~. The approximately 7 hours of diet and nutrition instruction in this broadly focused health course was divided almost equally among general nutrition information, skills training for change, and skills training to prevent a return to previ- ous habits. Special attention was paid to making the information personally relevant (fostering the central route of attitude change) and to providing guided practice in menu planning, identifying high- risk foods, countering nutrition myths, and resisting peer pressures (following the guidelines of social learning theory). In this same successful diet change study, training in resisting peer pressure included countering erroneous beliefs about food, provid- ing means of identifying external sources of pressure (e.g., advertise
DETERMINANTS OF FOOD CHOICE 65 meets and friends) to make nutritionally poor choices, and guiding the students in the practice of counterarguments. Methods common to successful smoking prevention and diet change programs incorporate elements of attitude change research (e.g., enhancing personal relevance and use of McGuire's inoculation methods) and elements of Bandura's social learning theory (e.g., increasing self-confidence and mastery of skills through stepwise guided practice in new behaviors) (Bandura, 1986; McGuire and Papageorgis, 1961~. These methods were- also used to produce changes in dietary habits in the small-group classes for adults. Therefore, this consistency in results among both young and older people suggests that such methods are generally effective. Reviewing 15 federally funded school-based cardiovascular risk- reduction programs, Stone (1985) concluded that the next urgent step is to provide guidance to the investigators on how to disseminate school curricula focusing on dietary change. Methods should be developed (1) to recruit or persuade the school system of the wisdom of adopting tested curricula, (2) to train the teachers in teaching from these cur- ricula, and (3) to maintain the quality of the curricula (Best, 1989; Stone et al., 1989~. Kreuter et al. (1984) had earlier advocated a similar plan. Effective school-based curricula are available, but the many barriers to their widespread adoption must be overcome by extensive planning, training, and monitoring. Health Communication at Points of Purchase Modification of environmental cues in schools, work sites, restau- rants, and grocery stores in a manner consistent with the social learning theory of Bandura (1977) has produced some observable but transitory changes toward health-promoting food choices (Glanz and Mullis, 1988~. Tactics include provision of more such choices accompanied by easily visible labeling, altered shelf displays, and nutrition infor- mation through posters and brochures. The most extensive point-of-purchase study in health education was conducted by investigators in the Minnesota Heart Health Pro- gram (Mullis et al., 1987~. The meat industry cooperated in this ef- fort by increasing supplies of lean meats and by training meat de- partment personnel to aid consumers in identifying lean meat products (Mullis and Pirie, 1988~. Overall, modest gains in knowledge and slight changes in behavior were observed as a result of the program, but both knowledge and behavior changes were often transitory. Be- cause of the short-term effect of this program, Glanz and Mullis (1988) concluded that point-of-purchase education should be part of a larger campaign to be most effective.
66 IMPROVING AMERICA'S DIET AND HEALTH Recently, Ellison et al. (1989) reported a 20% decline in saturated fat intake among students in boarding high schools following a train- ing program for food service workers to modify the food served in the schools' dining halls. The main purpose of this point-of-con- sumption study was to show that such an environmental change can produce dietary change in the absence of a student educational com- ponent. Despite this lack, student satisfaction with the food pro- vided was unchanged. Studies Conducted at Work Sites Work sites offer excellent opportunities for nutrition education. They provide a convenient locale for small-group classes for adults and allow for activation of a support system through the social networks of the work site. As in schools, there are also opportunities for changing cafeteria policies (e.g., providing point-of-purchase education) and for fostering changes to build a healthy work force. Sallis et al. (1986) reported that approximately 50 work-site pro- grams have been evaluated in the United States in the past decade. Many of these seem to have been effective in achieving at least short- term improvements in various life-styles and risk factor profiles, es- pecially among high-risk individuals in programs designed to foster change in more than one risk factor, including diet. Unfortunately, many work-site interventions are incompletely de- scribed. Successful programs have usually incorporated the self-directed change aspects of social learning theory to bring about dietary change, weight control, or smoking cessation (Blair et al., 1986; Bruno et al., 1983; Glanz, 1985; Klesges et al., 1986; Meyer and Henderson, 1974; Sallis et al., 1986; Stunkard et al., 1985; Wilbur, 19831. Reviewing such work-site programs, Glanz (1985) concluded that those which were most effective also contained changes in regulations (such as smoking policies) and the work environment (such as the inclusion of lower-fat menus in company cafeterias) (see also Glanz and See- wald-Klein, 1986~. Fostering of competition among employees has increased the success of programs designed to achieve smoking cessation (Klesges et al., 1986), but there are no reports of contests used in dietary change programs. An important challenge is whether the sequenced pattern of in- struction on self-directed change can be incorporated into nutrition education programs that rely largely on use of the media, thereby decreasing costs. One study has shown considerable reductions in blood cholesterol levels and in self-reported saturated fat intake among adults in widely dispersed work sites (Miller et al., 19881. This study
DETERMINANTS OF FOOD CHOICE 67 provided a single videotaped instruction session to groups of em- ployees at many widely separated locations. These individuals were linked by telephone to a central computer, which provided guidance in stepwise changes to improve eating behaviors. Printed instructions were mailed to each participant following each telephone contact. These instructions were made personally relevant by linking the results of a self-administered test of dietary practices with the risk of heart attack and stroke. Regional and National Mass Media Campaigns Many investigators have concluded that mass media campaigns (generally based on brief radio and television announcements or spots), although useful for increasing knowledge and changing attitudes, are not usually effective in producing large and lasting behavior change unless they are linked to more comprehensive campaigns (Atkin, 1979; Farquhar et al., 1977, 1985a; Glanz, 1985; Hewitt and Blaine, 1984; McGuire, 1964; Meyer et al., 1977; Puska, 1985; Puska et al., 1981; Roberts and Maccoby, 1985; Rootman, 1985; WHO, 1986~. If they are comprehensive and well conceived, however, mass media efforts may have considerable impact. The 3-year Stanford Three-Community Study was the first com- munity-based cardiovascular disease prevention program conducted in the United States (Farquhar et al., 1977~. The communities were three semirural towns in northern California (economies centered largely around agriculture and related business) with populations of ap- proximately 15,000 each, in which random samples of 500 people from each town were surveyed to determine their knowledge of car- diovascular disease and risk factor levels (smoking rates, body weight, blood cholesterol, and blood pressure). One community, the control, received only the yearly surveys. The two others were exposed to a mass media campaign involving frequent television and radio spots (five times a day, one of approximately 60 different public service announcements were delivered on radio and television), two weekly newspaper columns, and large quantities of mailed booklets contain- ing heart-healthy recipes and specific information on nutrition, exer- cise, and weight control. The study included several unique features: (1) Educational materials were disseminated through multiple com- munication channels (e.g., radio, television, newspapers, mass mail- ings of brief notices and of detailed self-help manuals, billboards, and bus posters); (2) all educational materials were based on the attitude change and social learning elements described earlier; and (3) special means of achieving personal relevance were included (e.g.,
68 IMPROVING AMERICA'S DIET AND HEALTH an hour-long television heart health test provided viewers with an assessment of risks derived from their personal habits of diet, exer- cise, and smoking, and yearly surveys provided a subset of the popu- lation with extra information on their risk factors-blood pressure and blood cholesterol levels). One of the two treatment communities was provided with a 3-month-long series of 10 lessons given to small classes of high-risk adults (Meyer and Henderson, 1974; Meyer et al., 1977~. The study was evaluated through repeated yearly surveys of the 1,500 people surveyed at the baseline. In the two cities exposed to the mass media campaign, there were approximately 30% decreases in self- reported saturated fat and cholesterol intake and also significant but modest decreases of about 3°/O in blood cholesterol levels. Greater changes in dietary patterns and blood cholesterol levels were observed in those who received classroom instruction in addition to the mass media exposure (Farquhar et al., 1977; Fortmann et al., 1981; Maccoby et al., 1977~. These findings reinforce the conclusion that behavior change is more likely in people given evidence of personal relevance. Comprehensive, Integrated Community-Based Multifactor Risk-Reduction Programs Several community-based health communication studies that in- clude dietary change as a goal have been reported or are under way. The two best-known projects are the Stanford Three-Community Study discussed in the preceding section and the Finnish North Karelia Project (Puska, 1985; Puska et al., 1979, 1981,1986; Tuomilehto et al., 1986), which both began in 1972. The Finnish study was conducted in two adjoining counties, one of which received education through multiple communication channels in ways similar to that of the Stanford Three-Community Study. The changes in cholesterol, blood pressure, and smoking achieved after 2 and 3 years of education in the Stanford study were comparable to those achieved after the first 5 years of education in the North Karelia Project. The investigators predicted that an approximately 20 to 25% reduction in coronary events would occur in the future if the risk factor changes persisted for more than 5 years. The risk factor de- creases that were achieved in the Stanford and Finnish studies for blood cholesterol, blood pressure, and smoking rates were approxi- mately 3, 5, and 10%, respectively. In a 10-year follow-up of the Finnish study, mortality from ischemic heart disease was reduced in North Karelia (annual decrease of 2.9% among males and 4.9% among females) to a significantly greater extent compared with that in the remainder of Finland (1.9 and 3.0°/O annual decreases, respectively) (Tuomilehto et al., 1986~.
DETERMINANTS OF FOOD CHOICE 69 A few other community-based studies have been conducted with educational approaches less intensive than, but similar to, those in the Stanford study and the North Karelia project. Significant changes in blood cholesterol levels, blood pressure, and smoking were found in only one of these studies, which was conducted in three small South African towns (Rossouw et al., 1981~. In two others, a four- town study in Switzerland (Gutzwiller et al., 1985) and a three-town study in Australia (Eager et al., 1983), no changes in blood choles- terol levels were observed, but 6 and 9% decreases in smoking rates, respectively, were reported. The reasons for the smaller sizes of these decreases compared with those of the Stanford and Finnish studies are unclear, but it appears that the campaigns were not as extensive nor were the campaigns based on the communication/per- suasion and social learning models. Interim results of another large (total population, 350,000) and complex long-term study, the Stanford Five-City Project (Farquhar et al., 1985b, 1990), show favorable changes in smoking, blood pressure, diet, exercise, and blood cholesterol levels comparable to those achieved in the Stanford Three-Community Study (Farquhar et al., 1990~. The consistency of findings of these two studies in the United States suggests that cost- effective communitywide programs (macrointerventions) are effective in achieving dietary changes in the general population, just as confi- dence exists that small-group classes can bring about change in certain population subgroups (microinterventions). Two additional extensive and well-evaluated studies are now in progress in the United States: the Minnesota Heart Health Program, involving six communities with a total population of 356,000 (Blackburn et al., 1984), and the Pawtucket Heart Health Study in Rhode Island, involving two cities with a total population of 173,000 (Lasater et al., 1984~. Effective Educational Methods The effectiveness of community-based studies in decreasing the risks and incidence of cardiovascular disease is enhanced by ensur- ing the implementation of comprehensive programs that involve the mass media, schools, work sites, and local health organizations (Farquhar, 1985a; Lefebvre and Flora, 1988~. Procedures include analysis of com- munity resources and community needs (including needs of various subgroups within the population) as well as pretesting of various educational methods. The term social marketing has been applied to these methods by Kotler and Zaltman (1971), Manoff (1985), and Lefebvre and Flora (1988~. It is helpful to conduct formal tests of effectiveness in small groups
70 IMPROVING AMERICA'S DIET AND HEALTH before applying methods in large field studies. For example, the Stanford Three-Community Study investigators developed some in- tensive instruction methods that followed Bandura's methods of self directed change (Bandura, 1977) and applied them to 36 high-risk adults at a work site near Stanford University (Meyer and Henderson, 1974~. Approximately 30 hours of group instruction provided by dietitians and health educators over a 12-week period produced sig- nificant changes in the desired directions in dietary patterns (fewer fatty and cholesterol-rich foods consumed), blood cholesterol levels (lowered by an average of at least 17 mg/dl, depending on the treat- ment group), smoking (cut by 50°/0 or more in all cases), and body weight (reduced by an average of at least 6.2 lb). About one-half of the sessions were devoted to nutrition education, including identifi- cation of health-promoting foods; guided practice in successive steps in new eating behaviors; reinforcement through peer approval; and progress monitoring. These methods were subsequently used in the Stanford Three-Community Study (Farquhar et al., 1977~. Similarly, a simplified self-help kit to promote dietary change was pretested in families by the Stanford study investigators. Since the kit was successful in changing eating behaviors, the investigators then used the kit in the field in the subsequent Stanford Five-City Project (Elmore et al., 1982~. The kit contained six lessons based on successive steps, beginning with a personal nutrition risk appraisal and then moving to specific shifts toward health-promoting food choices. This kit contained many other elements that have been effective in face-to-face group sessions, including methods to convey personal relevance, to monitor progress, and to obtain the involvement of family members. To ensure that education is effective, its progress must be tracked through a process evaluation. Any activity such as health fairs, classes, contests, lectures, and point-of-purchase programs must be tracked and measures of its success must be obtained. Education provided through the mass media, including all forms of print, must similarly be tracked to ensure that the target audience is reached and that one's goals for education are reached. Process evaluations principally measure short- term success in achieving the conditions (such as attendance, satisfac- tion, attitudes, and knowledge) leading to behavior change (Farquhar, 1985a,b). Although these data will contribute to evaluating the overall outcome of the health education program, a complete outcome evalua- tion would require longer-term, more complete, and more rigorous methods. Nevertheless, process evaluation provides valuable information to the implementor on ways that educational programs and messages might be altered to improve their effectiveness in future applications.
DETERMINANTS OF FOOD CHOICE Components and Amount of Nutrition Education Needed in Community-Based Programs 71 In the Stanford Three-Community Study and in the ongoing Stanford Five-City Project, investigators developed a general measure of the amount of education needed. Exposure to approximately 5 hours of education throughout the year provided both through the media and through occasional face-to-face encounters and continued for 2 or more years produced favorable change in all cardiovascular risk fac- tors (e.g., blood cholesterol, blood pressure, smoking rate, exercise habits, and resting pulse rates) and more than a 20% reduction in saturated fat intake within the general population. Approximately 40% (or 2 hours) of this annual education was devoted to nutrition instruction designed principally to achieve attitude and behavior change. That experience suggests that if only 10% of the total exposure is devoted to face-to-face communication, the changes in diet observed (i.e., a greater than 20% reduction in saturated fat intake) can still occur (Farquhar et al., 1990~. It is tempting to limit the face-to-face component of education to the least extent possible because of its cost. Its inclusion, however, is of great benefit for achieving and maintaining long-term dietary change. Additional experience in the Stanford Five-City Project derived from contests held at work sites and in the entire community indicates that incentives may help to maintain changes even in programs that rely largely on use of the media as the education method (King et al., 1987~. A person is more likely to adopt a new behavior if he or she is encouraged to do so through multiple channels, such as radio, televi- sion, print, point of purchase, small groups, schools, and work-site contests. For example, by taking advantage of the fact that parents often help their children with homework, a student's homework may be used to increase the parent's knowledge on nutrition and health. Coordinated programs with components that reinforce each other can be effective in reaching the most people (Farquhar et al., 1985a; King et al., 1987; Puska et al., 1981~. Successful maintenance of new dietary habits is dependent not only on individual factors such as knowledge, motivation, and skills but also on system factors, which comprise the physical and social environments of the individual. With individual factors, the nature of the health habit undergoing change has some relationship to main- tenance. For example, weight changes, especially those resulting from periodic calorie restriction, are associated with rather high recidi- vism rates, whereas weight loss achieved through exercise is better maintained (King et al., 1989~. Maintenance depends to a degree on
72 IMPROVING AMERICA'S DIET AND HEALTH the quality of the initial skills training program, e.g., its training in recidivism prevention skills, such as forewarning the participants of the social pressures that can precipitate a relapse. In addition, individuals who have changed their dietary habits through a phased approach that emphasizes the pleasure of new foods will often develop a new set of food preferences and then retain the new behaviors through internal motivation (Crouch et al., 1986; Fortmann et al., 1981; King et al., 1988; Meyer and Henderson, 1974~. The success of nutrition education in a group of people (e.g., students in a school, members of a community, or employees at a work site) is determined to a large extent by the initial proportion of that group who assimilated knowledge and changed behavior accordingly. The larger the proportion, the more likely that the newly adopted beneficial behavior will diffuse through the natural networks of the group (Rogers, 1983~. Maintenance is thus partly dependent on the degree of initial success, but apparently, it also requires the adoption of health promotion activities by community organizations, including schools, hospitals, public health agencies, and citizens' groups, to supply ongoing rein- forcement and reminders and to impart new knowledge and skills (Farquhar et al., 1985a, 1990~. The adoption and continued application of new technologies (such as use of principles of attitude change and social learning theory and the use of multiple channels of education) requires support from an organized structure of agencies and groups. For example, in order to ensure interorganizational collaboration, continuation could be sup- ported by a nutrition council or nutrition consortium of all the orga- nizations and individuals that impart nutrition knowledge and skills. An ongoing relationship between these community groups and exter- nal research and development organizations would ensure the dis- persion of new technologies to promote behavior change as they be- come available. In seeking comprehensive approaches that integrate various com- ponents (such as programs for schools or work sites or for point-of- purchase activities), it is necessary to identify all the subgroups within a population that require tailoring of nutrition education to their spe- cial needs. A successful program must take into consideration the characteristics of those who will receive the education. These include economic status, age, educational level, ethnic background, cultural values, health needs, health interest, and use of different media sources. The three basic levels of communication should be explored: the network, the organization, and the community (see Table 3-3~. Nu- trition education, in its broadest sense, should encompass all sub- groups reached through all three levels in order to achieve cost-effec
DETERMINANTS OF FOOD CHOICE TABLE 3-3 Sources of Health Communications 73 Network Level Extended social networks Peer groups Families Organization Level Work sites Restaurants Grocery stores Food producers and processors Institutional food providers Schools Mass media organizations Health-care organizations Public health organizations Government organizations Community Level Integrates network and organization levels, leading to change in public opinion, social norms, legislation, food production, and the social environment live and lasting change. This conclusion is based on the studies showing transitory effects in small groups, and on school or work- site programs conducted in isolation from broad, reinforcing community influences. The committee's review of these studies supports the common sense notion that change is more readily achieved if nutri- tion change programs include all three levels of communication. It is clear from the studies discussed above that well-designed nu- trition education programs are successful in a variety of settings and for a variety of people. Study results also demonstrate that unimagina- tive, information-only programs are not successful. The imagination of the learners must be captured; they must feel that the messages are personally relevant and that a stepwise course of action that avoids too much personal discomfort will yield tangible benefits (Bandura, 1986; Crouch et al., 1986; Killen et al., 1988; Meyer and Henderson, 1974~. Success requires both an adequate quantity and mix of effective instructional components. Appreciable changes in eating patterns have been maintained for 6 to 12 months in schools, work sites, and
74 IMPROVING AMERICA'S DIET AND HEALTH adult groups after approximately 7 to 15 hours of instruction given over a few months (Crouch et al., 1986; Killen et al., 1988; King et al., 1988; Meyer and Henderson, 1974). In addition, many people have improved their eating habits appreciably after being exposed to mul- tifactor, comprehensive, community-based programs that included prolonged and intermittent exposure to approximately 2 hours of nutrition education per year for approximately 2 to 4 years (Farquhar et al., 1977, 1990). Ingredients for success in both selected and general populations seem to require approximately equal proportions of (1) alerting, inform- ing, and changing attitudes; (2) step-by-step active learning of self- directed behavior change methods; and (3) prevention of recidivism. Key elements of the first category include transmission of knowledge concerning diet-disease links and the provision of evidence of their relevance to the individual so that personal attitudes are changed. The learner must also gain knowledge of high-risk dietary patterns and of his or her own eating patterns. To accomplish this, the individual must learn monitoring methods, gain confidence from early successes, identify internal and external barriers to change, learn how to resist social pressures to change, and practice new skills in restaurant menu selection, label reading, food shopping, and food preparation and tasting. This effort is assisted by continued social support and main- tenance incentives provided by others. SUMMARY The factors affecting food choices are numerous and complex. Some, such as inherent taste preferences and demographic trends, can be controlled little or not at all. Others more subject to modification include social norms, attitudes, skills, and availability of health-pro- moting foods. Over the past several decades, there have been important changes in food consumption patterns. Some of these changes are consistent with dietary recommendations (e.g., an increase in fish and vegetable consumption), but others are not (e.g., an increase in the consumption of high-fat ice cream). Similarly, recent changes in consumer attitudes and beliefs provide cause for both optimism and concern. Although there is a general trend toward recognition of the role of diet in disease prevention, surveys indicate that people are sometimes confused about which foods and food components are health-promoting and which are not. Nevertheless, a review of current theory and practice with respect to attitude and behavior changes suggests that modification of food preferences and eating patterns is possible, but will require more
DETERMINANTS OF FOOD CHOICE 75 than simply providing information to the population. People will need to be motivated to accept the information, see its personal relevance to them, integrate it into existing belief structures, acquire new skills and self-perceptior~s, and learn how to apply newly acquired atti- tudes to appropriate actions and to prevent recidivism. Various studies conducted within schools, at work sites, and in communities have indicated that intervention programs based on the communication/ persuasion model and the social learning model can be effective in producing substantial reductions in risk factors for diet-related diseases, particularly when they involve several components that reinforce each other arid include the mass media. It seems very reasonable to infer from these studies that new national programs that implement favorable regulatory and food supply changes will enhance the impact of com- prehensive education on the public's dietary patterns. REFERENCES Advertising Age. 1989. National ad spending by category. Advertising Age 60:8. Ajzen, I. 1989. Attitudes, Personality, and Behavior. Wadsworth, Florence, Ky. 150 PP Ajzen, I., and M. Fishbein. 1980. Understanding Attitudes and Predicting Social Behavior. Prentice-Hall, Englewood Cliffs, N.J. 278 pp. Arntzenius, A.C., D. Kromhout, J.D. Barth, J.H.C. Reiber, A.V.G. Bruschke, B. Buts, C.M. van Gent, N. Kempen-Voogd, S. Strikwerda, and E.A. van der Velde. 1985. Diet, lipoproteins, and the progression of coronary atherosclerosis: the Leiden Intervention Trial. N. Engl. J. Med. 312:805-811. Atkin, C.K. 1979. Research evidence on mass mediated health communication campaigns. Pp. 655-668 in D. Nimmo, ed. Communication Yearbook 3. Transaction Books, New Brunswick, N.J. Axelson, M.L., T.L. Federline, and D. Brinberg. 1985. A meta-analysis of food- and nutrition-related research. J. Nutr. Educ. 17:51-54. Bandura, A. 1977. Social Learning Theory. Prentice-Hall, Englewood Cliffs, N.J. 247 PP Bandura, A. 1982. Self-efficacy mechanism in human agency. Am. Psychol. 37:122- 147. Bandura, A. 1986. Social Foundations of Thought and Action: a Social Cognitive Theory. Prentice-Hall, Englewood Cliffs, N.J. 617 pp. Beauchamp, G.K., and M. Moran. 1982. Dietary experience and sweet taste preference in human infants. Appetite 3:139-152. Belasco, W.J. 1989. Appetite for Change: How the Counterculture Took on the Food Industry, 1966-1988. Pantheon Books, New York. 311 pp. Best, J.A. 1989. Intervention perspectives on school health promotion research. Health Educ. Q. 16:299-306. Best, J.A., S.J. Thomson, S.M. Santi, E.A. Smith, and K.S. Brown. 1988. Preventing cigarette smoking among school children. Annul Rev. Public Health 9:161-201. Birch, L.L. 1987. The acquisition of food acceptance patterns in children. Pp. 107-130 in R.A. Boakes, D.A. Popplewell, and M.J. Burton, eds. Eating Habits: Food, Physi- ology, and Learned Behavior. John Wiley & Sons, Chichester, Great Britain.
76 IMPROVING AMERICA'S DIET AND HEALTH Blackburn, H., R. Luepker, F.G. Kline, N. Bracht, R. Carlaw, D. Jacobs, M. Mittelmark, L. Stauffer, and H.L. Taylor. 1984. The Minnesota Heart Health Program: a re- search and demonstration project in cardiovascular disease prevention. Pp. 1171- 1178 in J.D. Matarazzo, S.M. Weiss, J.A. Herd, N.E. Miller, and S.M. Weiss, eds. Behavioral Health: A Handbook of Health Enhancement and Disease Prevention. John Wiley & Sons, New York. Blair, S.N., P.V. Piserchia, C.S. Wilbur, and J.H. Crowder.1986. A public health intervention model for work-site health promotion: impact on exercise and physical fitness in a health promotion plan after 24 months. J. Am. Med. Assoc. 255:921-926. Bogart, L. 1967. Strategy in Advertising. Harcourt, Brace, & World, New York. 336 PP Bruno, R., C. Arnold, L. Jacobson, M. Winick, and E. Wynder. 1983. Randomized controlled trial of a nonpharmacologic cholesterol reduction program at the worksite. Prev. Med. 12:523-532. Burros, M. January 6, 1988. What Americans really eat: nutrition can wait. New York Times. C1, C6. Calorie Control Council. 1989. 1989 National Survey. Conducted by Booth Research Services for the Calorie Control Council, Atlanta, Ga. Various pagings. Capps, O., Jr. 1986. Changes in domestic demand for food: impacts on Southern agriculture. South. J. Agric. Econ. 18:25-36. Carmody, T.P., S.G. Fey, D.K. Pierce, W.E. Connor, and J.D. Matarazzo. 1982. Behav- ioral treatment of hyperlipidemia: techniques, results, and future directions. J. Behav. Med. 5:91-96. Carmody, T.P., J. Istvan, J.D. Matarazzo, S.L. Connor, and W.E. Connor. 1986. Applica- tions of social learning theory in the promotion of heart-healthy diets: the Family Heart Study intervention model. Health Educ. Res. 1:13-27. Chaiken, S. 1987. The heuristic model of persuasion. Pp. 3-39 in M.P. Zanna, J.M. Olson, and C.P. Herman, eds. Social Influence: the Ontario Symposium. Vol. 5. Lawrence Erlbaum Associates, Hillsdale, N.J. Chaiken, S., and C. Stangor. 1987. Attitudes and attitude change. Annul Rev. Psychol. 38:575-630. Cialdini, R.B., R.E. Petty, and J.T. Cacioppo. 1981. Attitude and attitude change. Annul Rev. Psychol. 32:357-404. Contento, I.R., J.L. Michela, and C.J. Goldberg. 1988. Food choice among adolescents: population segmentation by motivations. J. Nutr. Educ. 20:289-298. Cooper, J., and R.T. Croyle. 1984. Attitudes and attitude change. Annul Rev. Psychol. v 35:395-426. Crouch, M., J.F. Sallis, J.W. Farquhar, W.L. Haskell, N.M. Ellsworth, A.B. King, and T. Rogers. 1986. Personal and mediated health counseling for sustained dietary reduction of hypercholesterolemia. Prev. Med. 15:282-291. Davis, C.M. 1928. Self selection of diet by newly weaned infants: an experimental study. Am. J. Dis. Child. 36:651-679. Davis, C.M. 1934. Studies in the self-selection of diet by young children. J. Am. Dent. Assoc. 21:636-640. Desor, J.A., O. Maller, and L.S. Greene. 1977. Preference for sweet in humans: infants, children, and adults. Pp. 161-172 in J.M. Weiffenbach, ed. Taste and Development: the Genesis of Sweet Preference. DHEW Publ. No. (NIH) 77-1068. National Insti- tutes of Health, Public Health Service, U.S. Department of Health, Education, and Welfare, Bethesda, Md. DHHS (U.S. Department of Health and Human Services). 1988. The Surgeon General's Report on Nutrition and Health. DHHS (PHS) Publ. No. 88-50210. Public Health
DETERMINANTS OF FOOD CHOICE 77 Service, U.S. Department of Health and Human Services. U.$. Government Print- ing Office, Washington, D.C. 727 pp. Drewnowski, A., J.D. Brunzell, K. Sande, P.H. Iverius, and M.R.C. Greenwood. 1985. Sweet tooth reconsidered: taste responsiveness in human obesity. Physiol. Behav. 35:617-622. Duewer, L.A. 1989. Changes in the beef and pork industries. Natl. Food Rev. 12(1):5-8. Dugas, C. 1985. Countermarketing: "bad" foods fight back. Ad Forum 6:18-22. Egger, G., W. Fitzgerald, G. Frape, A. Monaem, P. Rubinstein, C. Tyler, and B. McKay. 1983. Results of a large scale media antismoking campaign in Australia: North Coast "Quit For Life" programme. Br. Med. J. 287:1125-1128. Ehnholm, C., J.K. Huttunen, P. Pietinen, U. Leino, M. Mutanen, E. Kostiainen, J. Pikkarainen, R. Dougherty, J. Iacono, and P. Puska. 1982. Effect of diet on serum lipoproteins in a population with a high risk of coronary heart disease. N. Engl. J. Med. 307:850- 855. Ellison, R.C., A.L. Capper, R.J. Goldberg, J.C. Witschi, and F.J. Stare. 1989. Changing school food services to promote cardiovascular health. Health Educ. Q. 16:171-180. Elmore, J., C.B. Taylor, and J.A. Flora. 1982. Self-help nutrition kit with a nutrition booklet. Internal formative research report prepared for the Stanford Heart Dis- ease Prevention Program. Stanford, Calif. Erickson, J.L. 1989. Simplot bites back in micro snack war. Advertising Age 60:4. Fabsitz, R., M. Feinleib, and Z. Hrubec. 1980. Weight changes in adult twins. Acta Genet. Med. Gemellol. 29:273-279. Falciglia, G.A., and J.D. Gussow. 1980. Television commercials and eating behavior of obese and normal-weight women. J. Nutr. Educ. 12:196-199. Farb, P., and G. Armelagos. 1980. Consuming Passions: the Anthropology of Eating. Houghton Mifflin, Boston. 279 pp. Farquhar, J.W., N. Maccoby, P.D. Wood, J.K. Alexander, H. Breitrose, B.W. Brown, Jr., W.L. Haskell, A.L. McAlister, A.J. Meyer, J.D. Nash, and M.P. Stern. 1977. Com- mun~ty education for cardiovascular health. Lancet 1:1192-1195. Farquhar, J.W., N. Maccoby, and P.D. Wood. 1985a. Education and communication studies. Pp. 207-221 in W.W. Holland, R. Detels, and G. Knox, eds. Oxford Text- book of Public Health. Vol. 3. Investigative Methods in Public Health. Oxford University Press, London. Farquhar, J.W., S.P. Fortmann, N. Maccoby, W.L. Haskell, P.T. Williams, J.A. Flora, C.B. Taylor, B.W. Brown, Jr., D.S. Solomon, and S.B. Hulley. 1985b. The Stanford Five-City Project: design and methods. Am. J. Epidemiol. 122:323-334. Farquhar, J.W., S.P. Fortmann, J.A. Flora, C.B. Taylor, W.L. Haskell, P.T. Williams, N. Maccoby, and P.D. Wood. 1990. Effects of communitywide education on cardiovas- cular disease risk factors. J. Am. Med. Assoc. 264:359-365. Fazio, R.H. 1990. Multiple processes by which attitudes guide behavior: the MODE model as an integrative framework. Pp. 75-109 in M. Zanna, ed. Advances in Experimental Social Psychology. Academic Press, New York. Fazio, R.H., and M.P. Zanna. 1981. Direct experience and attitude-behavior consis- tency. Adv. Exp. Soc. Psychol. 14:161-202. Fletcher, H. 1899. Nature's Food Filter or What and When To Swallow. Herbert S. Stone & Company, Chicago. 29 pp. FMI (Food Marketing Institute). 1989. Trends: Consumer Attitudes & the Supermar- ket, 1989. Conducted for Food Marketing Institute by Opinion Research Corpora- tion. The Research Department, Food Marketing Institute, Washington, D.C. 65 pp. FMI (Food Marketing Institute). 1990. Trends: Consumer Attitudes & the Supermar- ket, 1990. Conducted for Food Marketing Institute by Opinion Research Corpora
78 IMPROVING AMERICA'S DIET AND HEALTH lion. The Research Department, Food Marketing Institute, Washington, D.C. 70 PP Fortmann, S.P., P.T. Williams, S.B. Hulley, W.L Haskell, and J.W. Farquhar. 1981. Ef- fect of health education on dietary behavior: the Stanford Three Community Study. Am. J. Clin. Nutr. 34:2030-2038. Freimuth, V.S. 1990. The chronically uninformed: closing the knowledge gap in health. Pp. 171-186 in E.B. Ray and L. Donohew, eds. Communication and Health: Systems and Applications. Lawrence Erlbaum Associates, Hillsdale, N.J. Freimuth, V.S., and W. Mettger. 1990. Is there a hard-to-reach audience? Public Health Rep. 105:232-238. Freimuth, V.S., S.L. Hammond, and J.A. Stein. 1988. Health advertising: prevention for profit. Am. J. Public Health 78:557-561. Garb, J.L., and A.J. Stunkard. 1974. Taste aversions in man. Am. J. Psychiatr. 131:1204- 1207. Glanz, K. 1985. Nutrition education for risk factor reduction and patient education: a review. Prev. Med. 14:721-752. Glanz, K., and R.M. Mullis. 1988. Environmental interventions to promote healthy eating: a review of models, programs, and evidence. Health Educ. Q. 15:395-415. Glanz, K., and T. Seewald-Klein. 1986. Nutrition at the worksite: an overview. J. Nutr. Educ. 18:S1-S12. Goldberg, M.E., and J. Hartwick. 1990. The effects of advertiser reputation and extremity of advertising claim on product evaluation. J. Consumer Res. 17:185-192. Green, L.W. 1975. Diffusion and adoption of innovations related to cardiovascular risk behavior in the public. Pp. 84-108 in A.J. Enelow and J.B. Henderson, eds. Applying Behavioral Science to Cardiovascular Risk: Proceedings of a Conference. American Heart Association, Dallas, Tex. Greene, C. 1988. A new look for supermarket produce sections. Natl. Food Rev. 11(4):1-5. Gutzwiller, F., B. Nater, and J. Martin. 1985. Community-based primary prevention of cardiovascular disease in Switzerland: methods and results of the National Research Program (NRP 1A). Prev. Med. 14:482-491. Harris, M. 1985. Good To Eat: Riddles of Food and Culture. Simon and Schuster, New York. 289 pp. Hewitt, L.E., and H.T. Blaine. 1984. Prevention through mass media communication. Pp. 281-326 in P.M. Miller and T.D. Nirenberg, eds. Prevention of Alcohol Abuse. Plenum Press, New York. Hjermann, I., K.V. Byre, I. Holme, and P. Leren. 1981. Effect of diet and smoking intervention on the incidence of coronary heart disease. Lancet 2:1303-1310. Janz, N.K., and M.H. Becker. 1984. The Health Belief Model: a decade later. Health Educ. Q. 11:1-47. Ketchum Communications. 1989. The Potato Board: How To Give a Greatly Misunderstood Food a Fresh Perception. Ketchum Communications, San Francisco, Calif. 1 p. Killen, J.D. 1985. Prevention of adolescent tobacco smoking: the social pressure resis- tance training approach. J. Child. Psychol. Psychiatr. 26:7-15. Killen, J.D., M.J. Telch, T.N. Robinson, N. Maccoby, C.B. Taylor, and J.W. Farquhar. 1988. Cardiovascular disease risk reduction for tenth graders: a multiple-factor school-based approach. J. Am. Med. Assoc. 260:1728-1733. King, A.C., J.A. Flora, S.P. Fortmann, and C.B. Taylor. 1987. Smokers' challenge: immediate and long-term findings of a community smoking cessation contest. Am. J. Public Health 77:1340-1341. King, A.C., K.E. Saylor, S. Foster, J.D. Killen, M.J. Telch, J.W. Farquhar, and J.A. Flora.
DETERMINANTS OF FOOD CHOICE 79 1988. Promoting dietary change in adolescents: a school-based approach for modi- fying and maintaining healthful behavior. Am. J. Prev. Med. 4:68-74. King, A.C., B. Frey-Hewitt, D.M. Dreon, and P.D. Wood. 1989. Diet vs. exercise in weight maintenance. Arch. Intern. Med. 149:2741-2746. Kinsey, J. 1983. Working wives and the marginal propensity to consume food away from home. Am. J. Agric. Econ. 65:10-19. Klesges, R.C., M.M. Vasey, and R.E. Glasgow. 1986. A worksite smoking modification competition: potential for public health impact. Am. J. Public Health 76:198-200. Kotler, P., and G. Zaltmann. 1971. Social marketing: an approach to planned social change. J. Market. 35:3-12. Kreuter, M.W., G.M. Christenson, and R. Davis. 1984. School health education re- search: future issues and challenges. J. School Health 54:27-32. Lantis, M. 1962. The child consumer: cultural. J. Home Econ. 54:570-579. Lasater, T., D. Abrams, L. Artz, P. Beaudin, L. Cabrera, J. Elder, A. Ferreira, P. Knisley, G. Peterson, A. Rodrigues, P. Rosenberg, R. Snow, and R. Carleton. 1984. Lay volunteer delivery of a community-based cardiovascular risk factor change program: the Pawtucket Experiment. Pp. 1166-1170 in J.D. Matarazzo, S.M. Weiss, J.A. Herd, N.E. Miller, and S.M. Weiss, eds. Behavioral Health: A Handbook of Health En- hancement and Disease Prevention. John Wiley & Sons, New York. Lefebvre, R.C., and J.A. Flora. 1988. Social marketing and public health intervention. Health Educ. Q. 15:299-315. Levy, A.S., and R.C. Stokes. 1987. Effects of a health promotion advertising campaign on sales of ready-to-eat cereals. Public Health Rep. 102:398-403. Levy, A.S., N. Ostrove, T. Guthrie, and J.T. Heimbach. 1988. Recent Trends and Beliefs about Diet/Disease Relationships: Results of the 1979-1988 FDA Health and Diet Surveys. Presented at FDA/USDA Food Editors Conference: December 1-2, 1988. Division of Consumer Studies, Center for Food Safety and Applied Nutrition, Food and Drug Administration, U.S. Department of Health and Human Services, Washington, D.C. Lewin, K. 1943. Forces behind food habits and methods of change. Pp. 35-65 in The Problem of Changing Food Habits. Report of the Committee on Food Habits 1941- 1943. Bulletin of the National Research Council, No. 108. National Academy of Sciences, Washington, D.C. Light, L., J. Tenney, B. Portnoy, L. Kessler, A.B. Rodgers, B. Patterson, O. Mathews, E. Katz, J.E. Blair, S.K. Evans, and E. Tuckermanty. 1989. Eat for Health: a nutrition and cancer control supermarket intervention. Public Health Rep. 104:443-450. Maccoby, N., J.W. Farquhar, P.D. Wood, and J. Alexander. 1977. Reducing the risk of cardiovascular disease: effects of a community-based campaign on knowledge and behavior. J. Community Health 3:100-114. Manoff, R.K. 1985. Social Marketing: New Imperative for Public Health. Praeger, New York. 293 pp. Marshall, J.F. 1971. Topics and networks in intra-village communication. Pp. 160-166 in S. Polgar, ed. Culture and Population: a Collection of Current Studies. Schenkman Publishing, Cambridge, Mass. Mayer, J. 1968. Overweight: Causes, Cost, and Control. Prentice-Hall, Englewood Cliffs, N.J. 213 pp. McGuire, W.J. 1964. Inducing resistance to persuasion: some contemporary approaches. Adv. Exp. Soc. Psychol. 1:191-229. McGuire, W.J. 1985. Attitudes and attitude change. Pp. 233-346 in G. Lindzey and E. Aronson, eds. Handbook of Social Psychology, 3rd ed. Vol. II. Random FIouse, New York.
80 IMPROVING AMERICA'S DIET AND HEALTH McGuire, W.J. 1989. Theoretical foundations of campaigns. Pp. 39-42 in R.E. Rice and C.K. Atkin, eds. Public Communication Campaigns, 2nd ed. Sage, Newbury Park, Calif. McGuire, W.J., and D. Papageorgis. 1961. The relative efficacy of various types of prior belief-defense in producing immunity against persuasion. J. Abnorm. Soc. Psychol. 62:327-337. Meyer, A.J., and J.B. Henderson. 1974. Multiple risk factor reduction in the preven- tion of cardiovascular disease. Prev. Med. 3:225-236. Meyer, A.J., N. Maccoby, and J.W. Farquhar. 1977. The role of opinion leadership in a cardiovascular health education campaign. Pp. 579-591 in B.D. Ruben, ed. Commu- nication Yearbook 1. Transaction Books, New Brunswick, N.J. Michela, J.L., and I.R. Contento. 1984. Spontaneous classification of foods by elemen- tary school-aged children. Health Educ. Q. 11:57-76. Michela, J.L., and I.R. Contento. 1986. Cognitive, motivational, social, and environ- mental influences on children's food choices. Health Psychol. 5:209-230. Miller, N., E. Wagner, and P. Rogers. 1988. Worksite-based multifactorial risk inter- vention trial. J. Am. Coll. Cardiol. 11:207A. Montagu, M.F.A. 1962. Prenatal Influences. Charles C Thomas, Springfield, Ill. 614 pp. Moskowitz, H.W., H.L. Jacobs, and S.D. Sharma. 1975. Cross-cultural differences in simple taste preferences. Science 190:1217-1218. Mothner, I. 1987. Our national food fight. We're eating lots of fruits and veggies, but more fats, snacks and sweets, too. Am. Health 6:48-49. MRFIT Research Group. 1982. Multiple Risk Factor Intervention Trial: risk factor changes and mortality results. J. Am. Med. Assoc. 248:1465-1477. Mullis, R.M., and P. Pirie. 1988. Lean meats make the grade a collaborative nutri- tion intervention program. J. Am. Diet. Assoc. 88:191-195. Mullis, R.M., M.K. Hunt, M. Foster, L. Hachfeld, D. Lansing, P. Snyder, and P. Pirie. 1987. The Shop Smart for Your Heart grocery program. J. Nutr. Educ. 19:225-228. Nader, P.R., J.F. Sallis, J. Rupp, C. Atkins, T. Patterson, and I. Abramson. 1986. San Diego family health project: reaching families through the schools. J. School Health 56:227-231. Nickens, H.W. 1990. Commentary: health promotion and disease prevention among minorities. Health Affairs 9:133-143. NRC (National Research Council). 1989. Diet and Health: Implications for Reducing Chronic Disease Risk. Report of the Committee on Diet and Health, Food and Nutrition Board, Commission on Life Sciences. National Academy Press, Washing- ton, D.C. 749 pp. Perry, C.L., R.V. Luepker, D.M. Murray, M.D. Hearn, A. Halper, B. Dudovitz, M.C. Maile, and M. Smyth. 1989. Parent involvement with children's health promotion: a one year follow-up of the Minnesota home team. Health Educ. Q. 16:171-180. Petty, R.E., and J.T. Cacioppo. 1981. Attitudes and Persuasion: Classic and Contem- porary Approaches. W.C. Brown, Dubuque, Iowa. 314 pp. Petty, R.E., and J.T. Cacioppo. 1984. Motivational factors in consumer response to advertisements. Pp. 418-454 in R.G. Geen, W.W. Beatty, and R.M. Arkin, eds. Human Motivation: Physiological, Behavioral, and Social Approaches. Allyn and Bacon, New York. Petty, R.E., and J.T. Cacioppo. 1986. Communication and Persuasion: Central and Peripheral Routes to Attitude Change. Springer-Verlag, New York. 262 pp. Popkin, B.M., P.S. Haines, and K.C. Reidy. 1989. Food consumption trends of US women: patterns and determinants between 1977 and 1985. Am. J. Clin. Nutr. 49:1307-1319.
DETERMINANTS OF FOOD CHOICE 81 Progressive Grocer. 1986. The cream also rises. Record growth for the ice cream business is being fueled by two new scoops: superpremiums and frozen novelties. Progressive Grocer 65:115. Puska, P. 1985. Effectiveness of nutrition intervention strategies. Pp. 39-46 in E.M.E. van den Berg, W. Bosman, and B.C. Breedveld, eds. Nutrition in Europe: Proceed- ings of the Fourth European Nutrition Conference. Voorlichtingsbureau voor de Voeding, The Hague, The Netherlands. Puska, P., K. Koskela, A. McAlister, U. Pallonen, E. Vartiainen, and K. Homan. 1979. A comprehensive television smoking cessation program in Finland. Int. J. Health Educ. Suppl. 22:1-28. Puska, P., J. Tuomilehto, J. Salonen, A. Nissinen, J. Virtamo, S. Bjorkqvist, K. Koskela, L. Neittaanmaki, L. Takalo, T.E. Kottke, J. Maki, P. Sipila, and P. Varvikko. 1981. Community Control of Cardiovascular Diseases: the North Karelia Project. World Health Organization, Copenhagen, Denmark. 351 pp. Puska, P., K. Koskela, A. McAlister, H. Mayranen, A. Smolander, S. Moisio, L. Viri, V. Korpelainen, and E.M. Rogers. 1986. Use of lay opinion leaders to promote diffu- sion of health innovations in a commurrity programme: lessons learned from the North Karelia project. Bull. W.H.O. 64:437-446. Putnam, J.J. 1989. Food Consumption, Prices, and Expenditures, 1966-87. Statistical Bulletin No. 773. Economic Research Service, U.S. Department of Agriculture, Washington, D.C. 111 pp. Ravussin, E., S. Lillioja, W.C. Knowler, L. Christin, D. Freymond, W.G. Abbott, V. Boyce, B.V. Howard, and C. Bogardus. 1988. Reduced rate of energy expenditure as a risk factor for body-weight gain. N. Engl. J. Med. 318:467-472. Roberts, D.F., and N. Maccoby. 1985. Effects of mass communication. Pp. 539-598 in G. Lindzey and E. Aronson, eds. Handbook of Social Psychology, 3rd ed. Vol. II. Random House, New York. Roberts, S.B., J. Savage, W.A. Coward, B. Chew, and A. Lucas. 1988. Energy expendi- ture and intake in infants born to lean and overweight mothers. N. Engl. J. Med. 318:461-466. Rodin, J. 1980. Social and immediate environmental influences on food selection. Int. J. Obesity 4:364-370. Rogers, E.M. 1983. Diffusion of Innovations, 3rd ed. Free Press, New York. 453 pp. Rootman, I. 1985. Using health promotion to reduce alcohol problems. Pp. 57-81 in M. Grant, ed. Alcohol Policies. WHO Regional Publications, European Series No. 18. World Health Organization, Copenhagen, Denmark. Rosenstock, I.M. 1974. Historical origins of the Health Belief Model. Health Educ. Monogr. 2:328-335. Rosenstock, I.M., V.J. Strecher, and M.H. Becker. 1988. Social learning theory and the Health Belief Model. Health Educ. Q. 15:175-183. Rossouw, J.E., P.L. Jooste, J.P. Kotze, and P.C.J. Jordaan. 1981. The control of hyper- tension in two communities: an interim evaluation. S. Afr. Med. J. 60:208-212. Rozin, P., and D. Schiller. 1980. The nature and acquisition of a preference of chili pepper by humans. Motiv. Emotion 4:77-101. Sallis, J.F., R.D. Hill, S.P. Fortmann, and J.A. Flora. 1986. Health behavior change at the worksite: cardiovascular risk reduction. Prog. Behav. Modif. 20:161-197. Schoenborn, C.A., and B.H. Cohen. 1986. Trends in Smoking, Alcohol Consumption and Other Health Practices Among U.S. Adults, 1977 and 1983. NCHS Advance Data from Vital & Health Statistics, No. 118. National Center for Health Statistics, Public Health Service, U.S. Department of Health and Human Services, Hyattsville, Md. 13 pp.
82 IMPROVING AMERICA'S DIET AND HEALTH Sclafani, A., and D. Springer. 1976. Dietary obesity in adult rats: similarities to hypo- thalamic and human obesity syndromes. Physiol. Behav. 17:461-471. Senauer, B. 1986. Economics and nutrition. Pp. 46-57 in What Is America Eating? Proceedings of a Symposium. Food and Nutrition Board, Commission on Life Sciences. National Academy Press, Washington, D.C. Sheppard, B.H., J. Hartwick, and P.R. Warshaw. 1988. The theory of reasoned action: a meta-arralysis of past research with recommendations for modifications and future research. J. Consumer Res. 15:325-343. Sherman, S.J. 1987. Cognitive processes in the formation, change, and expression of attitudes. Pp. 75-106 in M.P. Zanna, J.M. Olson, and C.P. Herman, eds. Social Influence: the Ontario Symposium. Vol. 5. Lawrence Erlbaum Associates, Hillsdale, N.J. Smallwood, D., and J. Blaylock. 1981. Impact of Household Size and Income on Food Spending Patterns. Technical Bulletin No. 1650. National Economics Division, Economics and Statistics Service, U.S. Department of Agriculture, Washington, D.C. 22 pp. Smith, K. 1990. Alar: One Year Later. A Media Analysis of a Hypothetical Health Risk. American Council on Science and Health, New York. 10 pp. Staats, A.W., and C.K. Staats. 1958. Attitudes established by classical conditioning. J. Abnorm. Soc. Psychol. 57:37-40. Steiner, J.E. 1977. Facial expressions of the neonate infant indicating the hedonics of food-related chemical stimuli. Pp. 173-190 in J.M. Weiffenbach, ed. Taste and Development: the Genesis of the Sweet Preference. DHEW Publ. No. (NIH) 77- 1068. National Institutes of Health, Public Health Service, U.S. Department of Health, Education, and Welfare, Bethesda, Md. Stone, E. 1985. School-based health research funded by the National Heart, Lung, and Blood Institute. J. School Health 55:168-174. Stone, E.J., C.L. Perry, and R.V. Leupker. 1989. Synthesis of cardiovascular behavioral research for youth health promotion. Health Educ. Q. 16:155-169. Strong, E.K. 1925. The Psychology of Selling and Advertising. McGraw-Hill, New York. 468 pp. Stunkard, A.J., M.R.J. Felix, and R.Y. Cohen. 1985. Mobilizing a community to promote health: the Pennsylvania County Health Improvement Program (CHIP). Pp. 143-190 in J.C. Rosen and L.J. Solomon, eds. Prevention in Health Psychology. University Press of New England, Hanover. Supermarket Business. 1986. 39th Annual Consumer Expenditures Study. Americans as artifacts: examining today's shoppers through their spending habits. Supermarket Business. September, pp. 69, 88. Szalai, A. 1972. The Use of Time: Daily Activities of Urban and Suburban Populations in Twelve Countries. Mouton Press, The Hague, The Netherlands. 868 pp. Telch, M.J., J.D. Killen, A.L. McAlister, C.L. Perry, and N. Maccoby. 1982. Long-term follow-up of a pilot project on smoking prevention with adolescents. J. Behav. Med. 5:1-8. Tuomilehto, J., J. Geboers, J.T. Salonen, A. Nissinen, K. Kuulasmaa, and P. Puska. 1986. Decline in cardiovascular mortality in North Karelia and other parts of Fin- land. Br. Med. J. 293:1068-1071. U.S. Department of Commerce. 1983. Statistical Abstract of the United States, 1982-83, 103rd ed. Bureau of the Census, U.S. Department of Commerce. U.S. Government Printing Office, Washington, D.C. 1042 pp. USDA (U.S. Department of Agriculture). 1953. Consumption of Food in the United States, 1902-1952. Agriculture Handbook No. 62. Bureau of Agricultural Econom- ics, U.S. Department of Agriculture, Washington, D.C. 249 pp.
DETERMINANTS OF FOOD CHOICE 83 USDA (U.S. Department of Agriculture). 1985. Nationwide Food Consumption Sur- vey Continuing Survey of Food Intakes by Individuals: Women 19-50 Years and Their Children 1-5 Years, 1 Day, 1985. Report No. 85-1. Nutrition Monitoring Division, Human Nutrition Information Service, U.S. Department of Agriculture, Washington, D.C. 102 pp. USDA (U.S. Department of Agriculture). 1989. Table on production and per capita consumption of soft drinks and soft drinks by flavor. Economic Research Service, U.S. Department of Agriculture, Washington, D.C. 2 pp. Weiffenbach,J.M. 1977. Sensory mechanisms in the newborn's tongue. Pp. 205-213 in J.M. Weiffenbach, ed. Taste and Development: the Genesis of the Sweet Preference. DHEW Publ. No. (NIH) 77-1068. National Institutes of Health, Public Health Service, U.S. Department of Health, Education, and Welfare, Bethesda, Md. White, S.L., and S.K. Maloney. 1990. Promoting healthy diets and active lives to heard-to-reach groups: market research study. Public Health Rep. 105:224-231. WHO (World Health Organization). 1986. Community Prevention and Control of Cardiovascular Diseases. Technical Report Series No. 732. Report of a WHO Expert Committee. World Health Organization, Geneva, Switzerland. 62 pp. Wilbur, C.S. 1983. The Johnson & Johnson Program. Prev. Med. 12:672-681. Wilhelmsen, L., G. Berglund, D. Elmfeldt, G. Tibblin, H. Wedel, K. Pennert, A. Vedin, C. Wilhelmsson, and L. Werko. 1986. The multifactor primary prevention trial in Goteborg, Sweden. Eur. Heart J. 7:279-288. Woteki, C.E. 1986. Methods for surveying food habits: how do we know what Americans are eating? Clin. Nutr. 5:9-16. Yankelovich, Skelly and White, Inc. 1985. Consumer Climate for Meat Products. Prepared for the American Meat Institute, Washington, D.C., and the National Live Stock and Meat Board, Chicago, Ill.