- Email: [email protected]
Time of day influences food acceptability
Appetite, 1984,5, 109-116
Time of Day Influences Food Acceptability LEANN LIPPS BIRCH, 'JANE BILLMAN and SUSAN SALISBURY RICHARDS University of Illinois To determine whether a food's acceptability changes with time of day and degree to which a food is appropriate for a given mealtime, preferences were obtained both in the morning and the afternoon for a set of foods which included items appropriate "for breakfast", "for dinner", or for either mealtime. A second purpose of the study was to determine whether such cultural rules regarding food appropriateness had been acquired by preschool children and whether the pattern of shifting acceptability with time of day predicted for adult subjects would also appear in the preference data of the children. Participants included 25 adults and 30 three- and four-year-old children. All subjects were seen twice, once in the morning and once in the late afternoon for two identical sessions. Half of the subjects in each age group were seen first in the morning, half were seen first in the afternoon. In these sessions food preferences for the set of foods varying in mealtime appropriateness were assessed and the foods were categorized as "for breakfast" of "for dinner". The preference data of both the adults and the children showed significant preference shifts with time of day, with breakfast items more preferred in the morning than the afternoon and dinner items more preferred in the morning than the afternoon. These shifts were larger for adults than for children but not significantly so.
Food acceptability is dynamic, not static and unchanging. Momentary changes in acceptability occur during the course of eating a food, with indicators of acceptability generally declining as amount consumed increases (Booth, 1982, Booth, 1977; Cabanac, 1971; Rolls, 1982). Research has also demonstrated that over days, weeks, or months, longer term changes in acceptability can result from repeated exposure (Birch & Marlin, 1982; Pliner, 1982) and associative conditioning to the context (Birch, Zimmerman & Hind, 1980; Birch, Birch, Marlin & Kramer, 1982; Birch, Marlin & Rotter, 1984) and consequences (Booth, 1977; Booth, Mather & Fuller, 1982) of consumption. The experiment reported below was designed to investigate yet another possible factor contributing to systematic changes in food acceptability: time of day. Particularly in affluent western cultures, time of day has a major impact on food availability and selection, as a result of culturally based rules that designate foods as appropriate or inappropriate for particular mealtimes (see Rodin, 1980 for a discussion of this issue). For example, in the United States people typically do not eat pizza and green salad at breakfast time or cold cereal and orange juice at dinner time. It was hypothesized that for adults any systematic changes in a food's acceptability from early to late in the day
This research was supported by the Science and Education Administration of the U.S.D.A. under grant 5901-04108-0077-0 from the Competitive Research Grants Office, by a grant from the National Dairy Council, and by a Hatch grant administered through the Illinois Agricultural Experiment Station. Sincere thanks go to the children and teachers of the Child Development Laboratory. Requests for reprints should be sent to Leann Lipps Birch, Child Development Laboratory, 1105 West Nevada, Urbana, IL 61801, U.S.A. 0195-6663/84/020109 + 08 $03'00/0
© 1984 Academic Press Inc. (London) Limited
110
L. L. BIRCH ET AL.
would be related to commonly held, accepted cultural beliefs about the food's appropriateness for particular mealtimes, with acceptability being highest at the time of day when it is culturally most appropriate to consume the food. During the first few years of life, children learn a great deal about the cultural and social meanings of food and the purposes of eating; evidence is accumulating that this learning is related to the acquisition of food preferences during this period (cf. Birch, 1980; Birch et al., 1980; Birch et al., 1984). Cultural beliefs about food appropriateness must be learned by the members of a cultural group. The present experiment was also intended to provide initial information on (1) whether such cultural rules had been learned by young children (three- to five-year-olds), and (2) whether these young children's affective reactions to foods would show systematic changes with time of day related to generally held cultural rules of food appropriateness. To obtain information on these questions, adults and young children were first asked to provide preference rankings on a set of foods, which varied in the time of day for which they were appropriate. These rankings were done twice, once near breakfast time and once near dinner time. The same subjects were then also asked to place each of the foods in the set into one of the two categories: "for breakfast" or "for dinner".
METHOD
Subjects
Participants included 30 preschool children, 15 males and 15 females, ranging in age from three years 0 months to five years two months at the initiation of the study (mean age four years 0 months). All children were enrolled in laboratory preschool programs and were from middle-class families. Preschoolers were not included in the sample if they (1) refused to participate, (2) were unavailable for the second session, (3) failed to perform adequately on the preference assessment task or to indicate to our satisfaction that they knew when during the course of the day breakfast and dinner were served. Participants also included 25 adults, 18 of whom were parents of preschool children enrolled in the programs; the remainder were preschool staff members. The adults ranged in age from 23 to 44 years, with a mean age of 33 years. All but three were females. Materials
The eight foods used in the preference assessments and the sorting tasks were orange juice (Kraft), sliced banana (Dole), scrambled eggs (Hillfarm), Cheerios (General Mills), frozen green beans (University), cheese pizza (Eisner), macaroni and cheese (Stouffer), and whole wheat bread (Brownberry). Foods that required it were prepared according to the manufacturer's directions and were served at the appropriate temperature. Scrambled eggs were prepared with one teaspoon of water to one egg and cooked in a microwave oven just prior to each assessment session. These particular foods were selected from among a larger set used in pilot work in which "middle American" adults categorized the foods according to the mealtime for which they were most appropriate. Pilot work indicated that the foods varied in the extent to which they were seen as appropriate for inclusion in morning or afternoon meals. Three of the foods were appropriate for breakfast but not for dinner (Cheerios, scrambled eggs, orange juice), three of the foods were appropriate for dinner but not for
TIME OF DAY
111
breakfast (pizza, green beans, and macaroni and cheese), and two foods were seen as appropriate for either morning or afternoon meals (whole wheat bread, bananas).
Procedure All subjects were seen twice; once in the morning between 0800 and 1000 hrs and once in the afternoon between 0330 and 0530 hrs. Half ofthe subjects in each age group were seen first in the morning, half were seen first in the afternoon. Children had been served either a meal or a snack 1-2 h prior to the morning and afternoon sessions. Information on the adults' meal and snack patterns was not available. The interval between the two sessions was not less than two days and not more than two weeks. The subjects were not told the purpose of the study and the data collection was presented as a part of a larger project designed to obtain information on the food preferences and eating behaviors of preschool children, their families, and teachers. Each of the two sessions took no more than about 15 minutes and included a preference assessment followed by a sorting task in which subjects were to categorize the foods in the set as "for breakfast" or "for dinner". The sorting task was always the second of the two tasks in order to avoid drawing attention to rules regarding meal appropriateness prior to assessing the subjects' preferences.
Preference Assessment, Children The children were seen individually by a female adult experimenter who was familiar to them. The child and the adult were seated at a table in a room adjacent to the classroom. On the table were three 8" x 10" drawings of cartoon faces. The expressions on the faces were designed to illustrate someone who had "just eaten something liked", "just eaten something not liked (hated)", and someone who "had just eaten something that was just okay, not good and not bad", respectively. All children had attended a preliminary training session on the meaning of the expressions. The experimenter discussed the faces again with each child until it was apparent that the child understood the meaning of each picture. The child was then asked to "point to the face that you make when you taste something you like" (child responds). "Point to the face that you make if you taste something you did not like" (child responds). "Now, point to the face you make when you taste something that is just okay, not good and not bad" (child responds). The experimenter then placed a tray with the eight food items on the table between the pictures and the child. Each food was presented in a clear 3 oz plastic cup. The child was asked to select a food, taste it and name it. If the child was unable to identify the food, the experimenter provided the correct name. After the child named the food, s/he was asked in random order "Did you like it? Not like it? Or was it just okay?" (child responds). "Put the cup in front of the face that tells me you (according to the child's response) tasted something you liked, did not like, or was just okay." This procedure was followed for all eight food items. After the child had placed all the cups in front of the appropriate faces, the experimenter focused the child's attention on the set of items in front of the "like" face. The child was asked "Tell me the food you liked the very best." As the child indicated a food, it was removed from the set: the child indicated his/her preference for all foods within a set in this manner before the experimenter continued to the next grouping of foods. The process was then repeated for neutral and disliked groups until a complete
112
L. L. BIRCH ET AL.
rank order was obtained. This procedure has been shown to yield data that are reliable and valid predictors of consumption (Birch, 1979 a, 1979 b). Preference Assessment, Adults
The adults were seen individually and were presented with a tray with samples of the eight foods in clear plastic cups. The experimenter identified the foods to be tasted. Each adult read a set of instructions, indicating that they were to taste the foods in a self-selected order, record their tasting order on the sheet, and write the name of the food in one of three columns, each with a cartoon face printed at the top, corresponding to "like", "neutral", and "dislike", and to then rank order the foods within each category. The procedure was designed to be as similar as possible to that used with the children, and differed only in that the adults recorded their own preferences. Food Categorization, Children
Upon completion of the assessment procedure, the three 8" x 10" pictures were removed from the table. The child was then told "Now we are going to play the second part of the game." Two sheets of 8t" x II" plain white paper were placed in front of the child. One paper had breakfast printed on it along one of the 11" sides; the other had "dinner" in the corresponding position. The sheets of paper were then identified for the child "this is the breakfast sheet. All the foods you just tasted that can be eaten for breakfast you place on this sheet," and "This is the dinner sheet. All the foods that you just tasted that can be eaten for dinner you place on this sheet." The experimenter then pointed to each sheet of paper and stated "Breakfast foods here, dinner foods here." The child was then asked "Show me the paper where you will put all the foods you can eat for breakfast" (child responds). "Now, show me the paper where you will put all the foods you can eat for dinner" (child responds). Once the child could correctly identify the papers, the tray with the eight foods was presented to the child, who was then asked "Where would you put these foods? Put the foods you can eat for breakfast in the breakfast paper and put the foods you can eat for dinner on the dinner paper." The child then placed each of the eight foods into one of the two categories. Food Categorization, Adults
Following the preference assessment, the adults completed the appropriateness questionnaire. Again this was designed to be as similar as possible to the procedure used with the children. The adults were asked to write the name of each item in one of two columns, one corresponding to "breakfast" and one corresponding to "dinner". Subjects were instructed to place each food into the meal category in which it was most appropriate to eat it, and to place each food in only one meal category.
RESULTS
The categorization data indicated that, across individuals, the adults were completely consistent in their categorization of six of the eight foods in the set: 100% of the adults categorized three ofthe foods (scrambled eggs, orange juice, Cheerios) as "for breakfast" and another three foods (pizza, green beans, macaroni and cheese) as "for
113
TIME OF DAY
dinner". The corresponding data for the children were somewhat less consistent, but did indicate that young children have already acquired some information about the mealtime appropriateness of the foods in the set. The proportions of children categorizing scrambled eggs, orange juice, and Cheerios as "for breakfast" were 70, 73, and 93%, respectively, and the proportions of children categorizing pizza, green beans, and macaroni and cheese as "for dinner" were 87, 95, and 75%. As expected from the pilot work, categorization of whole wheat bread and bananas was less consistent across adults, although 81 and 89% of the adults, respectively, categorized these foods as "for breakfast". The corresponding proportions for the children were 65 and 75%. Because the purpose of the experiment was to investigate whether commonly held cultural beliefs about the appropriateness/inappropriateness of a food for a particular mealtime would systematically influence the affective reactions to that food at those mealtimes, the analyses presented below will focus on those six foods categorized by 100% of the adults in the sample as "for breakfast" or "for dinner". To obtain information on this issue, for each subject the preference ranks assigned by the subject in the morning to the three breakfast foods were summed, as were the corresponding ranks assigned to these same foods in the afternoon. The same two sums of ranks (morning and afternoon) for the three dinner foods were also obtained. These data are presented in Figure 1. As shown in Figure 1, both the data from adults and preschool children showed the predicted preference shifts with time of day: breakfast foods were more preferred in the morning than in the afternoon, while the reverse was true for the dinner foods. The statistical analyses were performed on an "index of appropriateness" calculated for each subject. This index was calculated by summing the differences between the morning and afternoon judgments for both the breakfast foods and the dinner foods: [(LBr am - LBr pm ) + (LDi pm - LDi am)]. Preliminary analyses for order effects indicated no differences between the subgroups within each age group that were seen first either morning or afternoon, and these data were combined for the subsequent analyses. The obtained indices were analyzed using a Mann-Whitney test to determine whether the type of food (breakfast or dinner) by time of day (morning or afternoon) interaction
21-
t
2
31-
I
3
"0
~
2 ~
a.
aE
I
u
41-
n
51-
I-
c:
~
a:
6
-
7 B
Breakfast foods
Dinner foods
I 4f51-
:r B
Breakfast foods
Dinner foods
FIGURE 1. Summed preference ranks for three breakfast and three dinner foods, obtained in the morning (0) and afternoon (.), children (n=30) and adults (n=25).
114
L. L. BIRCH ET AL.
differed for adults and young children. The preschool children's categorization of the six foods for which the adults had clear mealtime appropriateness beliefs indicated that these children had already acquired information about the mealtime appropriateness of those foods. Such information is necessary but might not be sufficient to produce the hypothesized preference shifts with time of day. Somewhat surprisingly, the results of this analysis indicated no age x type offood x time of day interaction. That is, there was no age difference in the pattern of preference shifts; this finding is reflected in Figure 1, which shows the same pattern of preference shifts from morning to afternoon for adults and children (i.e., a decrease in preference for the breakfast foods and an increase in preference for the dinner foods from morning to afternoon), despite age differences in overall relative preferences for breakfast and dinner foods (see below). To test the time of day x food type interaction separately for the two age groups, Wilcoxan tests were performed on the appropriateness indices of the adults and of the children. In both cases, the time of day x food type interaction was significant: for the adults z = - 2-49, p < 0,02, while for the children it just reached significance z = - 1'96, p=0·05. There was no significant difference in the size of the preference shifts for the breakfast versus the dinner foods for either adults or children. However, the children showed a clear preference for the breakfast foods over the dinner foods at both times of day, while the adults tended to show an overall preference for the dinner foods. When the children and the adults were dichotomized based on whether they (1) preferred the dinner foods to the breakfast foods, or (2) preferred the breakfast foods to the dinner foods (using the summed morning and afternoon ranks), 21 of 30 children showed an overall preference for the breakfast foods, while only 9 of 25 adults did so (one subject included in this cell had tied summed ranks for the breakfast and dinner foods). A X2 analysis test for independence performed on these data was significant, X2 (1) = 6'35, p<0·025.
DISCUSSION
Culturally based rules of cuisine regarding the appropriateness of foods for particular mealtimes influence food acceptability. The data of the present experiment have demonstrated that the mealtime context in which a food is presented can systematically influence the dynamic acceptability relationship between food and person. While systematic changes in acceptability within a meal have been demonstrated in animals other than man (Booth, 1982; Rolls & Rolls, 1982), the changes in acceptability noted in the present study have their basis in cultural rules and are therefore unique to humans. Although these changes in acceptability with time of day are uniquely human, these data do not exclude the possibility that such cultural rules may be influenced by underlying circadian variations in physiological biases that predispose the selection of foods varying, for example, in the balance of carbohydrate, fat and protein they contain. Humans exhibit prominent circadian rhythms in plasma insulin and glucagon, among other variables known to influence eating behavior (Armstrong, 1980), which could contribute to such systematic shifts with time of day in food acceptability and, in turn, to the development of cultural rules consistent with such biases. These changing affective reactions to foods, shown to shift in a direction consistent with the culture's rules of cuisine regarding food appropriateness, can function to
TIME OF DAY
115
reinforce these existing cultural rules, hence increasing the chances that rules of cuisine will be maintained, both within and across generations. When one cultural group is integrated into another, the cuisine of the group is often maintained even after other aspects of the culture have all but disappeared. The strong affective component may be primarily responsible for the persistence of cuisine; the individual eats a pattern offoods consistent with the cultural rules of cuisine not just because she has knowledge that it is appropriate, but also because what is appropriate has become the most acceptable in that context. By contributing to the likelihood that different foods are eaten at different mealtimes, these systematic shifts in acceptability shown to occur with shifts in mealtime context can function in combination with within meal changes in acceptability (e.g., sensory specific satiety and conditioned state dependent preferences) to increase variety in the diet. Surprisingly, the affective shifts noted in the present study and found to be related to rules of cuisine regarding food appropriateness are already operative in the food preferences of children as young as three years. Casual observation suggests that children are not explicitly taught these rules offood appropriateness; in most cases they must be inferred from the repeated association of food and context that the children experience in meal patterns generated by their elders. The children in the present study possessed knowledge of these rules despite the fact that both parents and children reported that on some occasions the children's meal patterns violated the rules. In informal interviews following the experiment, parents reported that their children sometimes did eat breakfast foods at dinner, although the adults themselves did not do so. Reasons given by parents for these "rule violations" included: foods on the dinner menu believed age inappropriate (usually due to cost or spiciness) or not accepted by the child, children eating dinner separately from parents, breakfast foods well accepted and easy to prepare. Some of the children who categorized breakfast and dinner foods appropriately then volunteered that they could (and sometimes did) eat breakfast foods for dinner (the reverse pattern was essentially never seen). The fact that the children did eat breakfast foods at dinner is probably related to the children's smaller time of day shifts and overall preference for breakfast foods regardless of time of day. Although there are no data at present on the etiology of these systematic shifts in acceptability that occur with time of day, it is possible to speculate on how such patterns might be acquired. Affective reactions to foods can be associatively conditioned to cue complexes that include aspects of the individual's internal state, organoleptic characteristics of foods (Booth et aI., 1982) as well as the social context in which foods are presented (Birch et al., 1984). It is likely that mealtime context could become a part of the cue complex that is conditioned through the consistent repeated experience of eating a food in one particular mealtime context. REFERENCES
Armstrong, S. A. A chronometric approach to the study of feeding behavior. Neuroscience and Biobehavioral Reviews, 1980, 4, 27-53. Birch, L. L. Dimensions of preschool children's food preferences. Journal ofNutrition Education, 1979, 11, 77-80. (a) Birch, L. L. Preschool children's food preferences and consumption patterns. Journal of Nutrition Education, 1979,11,189-192. (b). Birch, L.L. Effect of peer model's food choices and eating behaviors on preschoolers' food preferences. Child Development, 1980, 51, 489-496.
116
L. L. BIRCH ET AL.
Birch, L.L., Birch, D., Marlin, D. W., & Kramer, K. Effects of instrumental consumption on children's food preference. Appetite, 1982,3, 125-134. Birch, L.L., & Marlin, D. W. I don't like it; I never tried it: Effects of exposure to food on twoyear-old children's food preferences. Appetite, 1982, 3, 353-360. Birch, L. L., Marlin, D. W., & Rotter, 1. Eating as the "means" activity in a contingency: Effects on young children's food preference. Child Development, 1984, 55, 431-439. Birch, L.L., Zimmerman, S., & Hind, H. The influence of social-affective context on preschool children's food preferences. Child Development, 1980, 5/, 856-861. Booth, D. A. How nutritional effects of foods can influence people's dietary choices. In L. M. Barker (Ed.). The psychobiology ofhuman food selection, Pp. 67-84, Westport, CT. Avi, 1982. Booth, D. A. Appetite and satiety as metabolic expectancies, In Y. Katsuke, M. Sato, S. F. Takage, & Y. Oomura (Eds.). Food intake and chemical senses, Pp. 317-330. Tokyo: University of Tokyo Press, 1977. Booth, D. A., Mather, P., & Fuller, 1. Starch content of ordinary foods associatively conditions human appetite and satiation indexed by intake and eating pleasantness of starch-paired flavours. Appetite, 1982, 3, 163-184. Cabanac, M. Physiological role of pleasure. Science, 1971, 173, 1103-1107. Pliner, P. The effects of mere exposure on liking for edible substances. Appetite, 1982,3,283-290. Rodin, J. Social and immediate environmental influences on food selection. International J oumal of Obesity, 1980,4,364-370. Rolls, B. 1., Rolls, E. T., & Rowe, E. A. The influence of variety on human food selection and intake. In L. M. Barker (Ed.). The psychobiology of human food selection, Pp. 101-122. Westport. CT: Avi, 1982. Rolls, E. T. & Rolls, B. 1. Brain mechanisms involved in feeding. In L. M. Barker (Ed.), The psychobiology of human food selection, Pp. 33--62. Westport, CT: Avi, 1982.
Received 27 September, 1983; revision 22 December, 1983
Time of Day Influences Food Acceptability LEANN LIPPS BIRCH, 'JANE BILLMAN and SUSAN SALISBURY RICHARDS University of Illinois To determine whether a food's acceptability changes with time of day and degree to which a food is appropriate for a given mealtime, preferences were obtained both in the morning and the afternoon for a set of foods which included items appropriate "for breakfast", "for dinner", or for either mealtime. A second purpose of the study was to determine whether such cultural rules regarding food appropriateness had been acquired by preschool children and whether the pattern of shifting acceptability with time of day predicted for adult subjects would also appear in the preference data of the children. Participants included 25 adults and 30 three- and four-year-old children. All subjects were seen twice, once in the morning and once in the late afternoon for two identical sessions. Half of the subjects in each age group were seen first in the morning, half were seen first in the afternoon. In these sessions food preferences for the set of foods varying in mealtime appropriateness were assessed and the foods were categorized as "for breakfast" of "for dinner". The preference data of both the adults and the children showed significant preference shifts with time of day, with breakfast items more preferred in the morning than the afternoon and dinner items more preferred in the morning than the afternoon. These shifts were larger for adults than for children but not significantly so.
Food acceptability is dynamic, not static and unchanging. Momentary changes in acceptability occur during the course of eating a food, with indicators of acceptability generally declining as amount consumed increases (Booth, 1982, Booth, 1977; Cabanac, 1971; Rolls, 1982). Research has also demonstrated that over days, weeks, or months, longer term changes in acceptability can result from repeated exposure (Birch & Marlin, 1982; Pliner, 1982) and associative conditioning to the context (Birch, Zimmerman & Hind, 1980; Birch, Birch, Marlin & Kramer, 1982; Birch, Marlin & Rotter, 1984) and consequences (Booth, 1977; Booth, Mather & Fuller, 1982) of consumption. The experiment reported below was designed to investigate yet another possible factor contributing to systematic changes in food acceptability: time of day. Particularly in affluent western cultures, time of day has a major impact on food availability and selection, as a result of culturally based rules that designate foods as appropriate or inappropriate for particular mealtimes (see Rodin, 1980 for a discussion of this issue). For example, in the United States people typically do not eat pizza and green salad at breakfast time or cold cereal and orange juice at dinner time. It was hypothesized that for adults any systematic changes in a food's acceptability from early to late in the day
This research was supported by the Science and Education Administration of the U.S.D.A. under grant 5901-04108-0077-0 from the Competitive Research Grants Office, by a grant from the National Dairy Council, and by a Hatch grant administered through the Illinois Agricultural Experiment Station. Sincere thanks go to the children and teachers of the Child Development Laboratory. Requests for reprints should be sent to Leann Lipps Birch, Child Development Laboratory, 1105 West Nevada, Urbana, IL 61801, U.S.A. 0195-6663/84/020109 + 08 $03'00/0
© 1984 Academic Press Inc. (London) Limited
110
L. L. BIRCH ET AL.
would be related to commonly held, accepted cultural beliefs about the food's appropriateness for particular mealtimes, with acceptability being highest at the time of day when it is culturally most appropriate to consume the food. During the first few years of life, children learn a great deal about the cultural and social meanings of food and the purposes of eating; evidence is accumulating that this learning is related to the acquisition of food preferences during this period (cf. Birch, 1980; Birch et al., 1980; Birch et al., 1984). Cultural beliefs about food appropriateness must be learned by the members of a cultural group. The present experiment was also intended to provide initial information on (1) whether such cultural rules had been learned by young children (three- to five-year-olds), and (2) whether these young children's affective reactions to foods would show systematic changes with time of day related to generally held cultural rules of food appropriateness. To obtain information on these questions, adults and young children were first asked to provide preference rankings on a set of foods, which varied in the time of day for which they were appropriate. These rankings were done twice, once near breakfast time and once near dinner time. The same subjects were then also asked to place each of the foods in the set into one of the two categories: "for breakfast" or "for dinner".
METHOD
Subjects
Participants included 30 preschool children, 15 males and 15 females, ranging in age from three years 0 months to five years two months at the initiation of the study (mean age four years 0 months). All children were enrolled in laboratory preschool programs and were from middle-class families. Preschoolers were not included in the sample if they (1) refused to participate, (2) were unavailable for the second session, (3) failed to perform adequately on the preference assessment task or to indicate to our satisfaction that they knew when during the course of the day breakfast and dinner were served. Participants also included 25 adults, 18 of whom were parents of preschool children enrolled in the programs; the remainder were preschool staff members. The adults ranged in age from 23 to 44 years, with a mean age of 33 years. All but three were females. Materials
The eight foods used in the preference assessments and the sorting tasks were orange juice (Kraft), sliced banana (Dole), scrambled eggs (Hillfarm), Cheerios (General Mills), frozen green beans (University), cheese pizza (Eisner), macaroni and cheese (Stouffer), and whole wheat bread (Brownberry). Foods that required it were prepared according to the manufacturer's directions and were served at the appropriate temperature. Scrambled eggs were prepared with one teaspoon of water to one egg and cooked in a microwave oven just prior to each assessment session. These particular foods were selected from among a larger set used in pilot work in which "middle American" adults categorized the foods according to the mealtime for which they were most appropriate. Pilot work indicated that the foods varied in the extent to which they were seen as appropriate for inclusion in morning or afternoon meals. Three of the foods were appropriate for breakfast but not for dinner (Cheerios, scrambled eggs, orange juice), three of the foods were appropriate for dinner but not for
TIME OF DAY
111
breakfast (pizza, green beans, and macaroni and cheese), and two foods were seen as appropriate for either morning or afternoon meals (whole wheat bread, bananas).
Procedure All subjects were seen twice; once in the morning between 0800 and 1000 hrs and once in the afternoon between 0330 and 0530 hrs. Half ofthe subjects in each age group were seen first in the morning, half were seen first in the afternoon. Children had been served either a meal or a snack 1-2 h prior to the morning and afternoon sessions. Information on the adults' meal and snack patterns was not available. The interval between the two sessions was not less than two days and not more than two weeks. The subjects were not told the purpose of the study and the data collection was presented as a part of a larger project designed to obtain information on the food preferences and eating behaviors of preschool children, their families, and teachers. Each of the two sessions took no more than about 15 minutes and included a preference assessment followed by a sorting task in which subjects were to categorize the foods in the set as "for breakfast" or "for dinner". The sorting task was always the second of the two tasks in order to avoid drawing attention to rules regarding meal appropriateness prior to assessing the subjects' preferences.
Preference Assessment, Children The children were seen individually by a female adult experimenter who was familiar to them. The child and the adult were seated at a table in a room adjacent to the classroom. On the table were three 8" x 10" drawings of cartoon faces. The expressions on the faces were designed to illustrate someone who had "just eaten something liked", "just eaten something not liked (hated)", and someone who "had just eaten something that was just okay, not good and not bad", respectively. All children had attended a preliminary training session on the meaning of the expressions. The experimenter discussed the faces again with each child until it was apparent that the child understood the meaning of each picture. The child was then asked to "point to the face that you make when you taste something you like" (child responds). "Point to the face that you make if you taste something you did not like" (child responds). "Now, point to the face you make when you taste something that is just okay, not good and not bad" (child responds). The experimenter then placed a tray with the eight food items on the table between the pictures and the child. Each food was presented in a clear 3 oz plastic cup. The child was asked to select a food, taste it and name it. If the child was unable to identify the food, the experimenter provided the correct name. After the child named the food, s/he was asked in random order "Did you like it? Not like it? Or was it just okay?" (child responds). "Put the cup in front of the face that tells me you (according to the child's response) tasted something you liked, did not like, or was just okay." This procedure was followed for all eight food items. After the child had placed all the cups in front of the appropriate faces, the experimenter focused the child's attention on the set of items in front of the "like" face. The child was asked "Tell me the food you liked the very best." As the child indicated a food, it was removed from the set: the child indicated his/her preference for all foods within a set in this manner before the experimenter continued to the next grouping of foods. The process was then repeated for neutral and disliked groups until a complete
112
L. L. BIRCH ET AL.
rank order was obtained. This procedure has been shown to yield data that are reliable and valid predictors of consumption (Birch, 1979 a, 1979 b). Preference Assessment, Adults
The adults were seen individually and were presented with a tray with samples of the eight foods in clear plastic cups. The experimenter identified the foods to be tasted. Each adult read a set of instructions, indicating that they were to taste the foods in a self-selected order, record their tasting order on the sheet, and write the name of the food in one of three columns, each with a cartoon face printed at the top, corresponding to "like", "neutral", and "dislike", and to then rank order the foods within each category. The procedure was designed to be as similar as possible to that used with the children, and differed only in that the adults recorded their own preferences. Food Categorization, Children
Upon completion of the assessment procedure, the three 8" x 10" pictures were removed from the table. The child was then told "Now we are going to play the second part of the game." Two sheets of 8t" x II" plain white paper were placed in front of the child. One paper had breakfast printed on it along one of the 11" sides; the other had "dinner" in the corresponding position. The sheets of paper were then identified for the child "this is the breakfast sheet. All the foods you just tasted that can be eaten for breakfast you place on this sheet," and "This is the dinner sheet. All the foods that you just tasted that can be eaten for dinner you place on this sheet." The experimenter then pointed to each sheet of paper and stated "Breakfast foods here, dinner foods here." The child was then asked "Show me the paper where you will put all the foods you can eat for breakfast" (child responds). "Now, show me the paper where you will put all the foods you can eat for dinner" (child responds). Once the child could correctly identify the papers, the tray with the eight foods was presented to the child, who was then asked "Where would you put these foods? Put the foods you can eat for breakfast in the breakfast paper and put the foods you can eat for dinner on the dinner paper." The child then placed each of the eight foods into one of the two categories. Food Categorization, Adults
Following the preference assessment, the adults completed the appropriateness questionnaire. Again this was designed to be as similar as possible to the procedure used with the children. The adults were asked to write the name of each item in one of two columns, one corresponding to "breakfast" and one corresponding to "dinner". Subjects were instructed to place each food into the meal category in which it was most appropriate to eat it, and to place each food in only one meal category.
RESULTS
The categorization data indicated that, across individuals, the adults were completely consistent in their categorization of six of the eight foods in the set: 100% of the adults categorized three ofthe foods (scrambled eggs, orange juice, Cheerios) as "for breakfast" and another three foods (pizza, green beans, macaroni and cheese) as "for
113
TIME OF DAY
dinner". The corresponding data for the children were somewhat less consistent, but did indicate that young children have already acquired some information about the mealtime appropriateness of the foods in the set. The proportions of children categorizing scrambled eggs, orange juice, and Cheerios as "for breakfast" were 70, 73, and 93%, respectively, and the proportions of children categorizing pizza, green beans, and macaroni and cheese as "for dinner" were 87, 95, and 75%. As expected from the pilot work, categorization of whole wheat bread and bananas was less consistent across adults, although 81 and 89% of the adults, respectively, categorized these foods as "for breakfast". The corresponding proportions for the children were 65 and 75%. Because the purpose of the experiment was to investigate whether commonly held cultural beliefs about the appropriateness/inappropriateness of a food for a particular mealtime would systematically influence the affective reactions to that food at those mealtimes, the analyses presented below will focus on those six foods categorized by 100% of the adults in the sample as "for breakfast" or "for dinner". To obtain information on this issue, for each subject the preference ranks assigned by the subject in the morning to the three breakfast foods were summed, as were the corresponding ranks assigned to these same foods in the afternoon. The same two sums of ranks (morning and afternoon) for the three dinner foods were also obtained. These data are presented in Figure 1. As shown in Figure 1, both the data from adults and preschool children showed the predicted preference shifts with time of day: breakfast foods were more preferred in the morning than in the afternoon, while the reverse was true for the dinner foods. The statistical analyses were performed on an "index of appropriateness" calculated for each subject. This index was calculated by summing the differences between the morning and afternoon judgments for both the breakfast foods and the dinner foods: [(LBr am - LBr pm ) + (LDi pm - LDi am)]. Preliminary analyses for order effects indicated no differences between the subgroups within each age group that were seen first either morning or afternoon, and these data were combined for the subsequent analyses. The obtained indices were analyzed using a Mann-Whitney test to determine whether the type of food (breakfast or dinner) by time of day (morning or afternoon) interaction
21-
t
2
31-
I
3
"0
~
2 ~
a.
aE
I
u
41-
n
51-
I-
c:
~
a:
6
-
7 B
Breakfast foods
Dinner foods
I 4f51-
:r B
Breakfast foods
Dinner foods
FIGURE 1. Summed preference ranks for three breakfast and three dinner foods, obtained in the morning (0) and afternoon (.), children (n=30) and adults (n=25).
114
L. L. BIRCH ET AL.
differed for adults and young children. The preschool children's categorization of the six foods for which the adults had clear mealtime appropriateness beliefs indicated that these children had already acquired information about the mealtime appropriateness of those foods. Such information is necessary but might not be sufficient to produce the hypothesized preference shifts with time of day. Somewhat surprisingly, the results of this analysis indicated no age x type offood x time of day interaction. That is, there was no age difference in the pattern of preference shifts; this finding is reflected in Figure 1, which shows the same pattern of preference shifts from morning to afternoon for adults and children (i.e., a decrease in preference for the breakfast foods and an increase in preference for the dinner foods from morning to afternoon), despite age differences in overall relative preferences for breakfast and dinner foods (see below). To test the time of day x food type interaction separately for the two age groups, Wilcoxan tests were performed on the appropriateness indices of the adults and of the children. In both cases, the time of day x food type interaction was significant: for the adults z = - 2-49, p < 0,02, while for the children it just reached significance z = - 1'96, p=0·05. There was no significant difference in the size of the preference shifts for the breakfast versus the dinner foods for either adults or children. However, the children showed a clear preference for the breakfast foods over the dinner foods at both times of day, while the adults tended to show an overall preference for the dinner foods. When the children and the adults were dichotomized based on whether they (1) preferred the dinner foods to the breakfast foods, or (2) preferred the breakfast foods to the dinner foods (using the summed morning and afternoon ranks), 21 of 30 children showed an overall preference for the breakfast foods, while only 9 of 25 adults did so (one subject included in this cell had tied summed ranks for the breakfast and dinner foods). A X2 analysis test for independence performed on these data was significant, X2 (1) = 6'35, p<0·025.
DISCUSSION
Culturally based rules of cuisine regarding the appropriateness of foods for particular mealtimes influence food acceptability. The data of the present experiment have demonstrated that the mealtime context in which a food is presented can systematically influence the dynamic acceptability relationship between food and person. While systematic changes in acceptability within a meal have been demonstrated in animals other than man (Booth, 1982; Rolls & Rolls, 1982), the changes in acceptability noted in the present study have their basis in cultural rules and are therefore unique to humans. Although these changes in acceptability with time of day are uniquely human, these data do not exclude the possibility that such cultural rules may be influenced by underlying circadian variations in physiological biases that predispose the selection of foods varying, for example, in the balance of carbohydrate, fat and protein they contain. Humans exhibit prominent circadian rhythms in plasma insulin and glucagon, among other variables known to influence eating behavior (Armstrong, 1980), which could contribute to such systematic shifts with time of day in food acceptability and, in turn, to the development of cultural rules consistent with such biases. These changing affective reactions to foods, shown to shift in a direction consistent with the culture's rules of cuisine regarding food appropriateness, can function to
TIME OF DAY
115
reinforce these existing cultural rules, hence increasing the chances that rules of cuisine will be maintained, both within and across generations. When one cultural group is integrated into another, the cuisine of the group is often maintained even after other aspects of the culture have all but disappeared. The strong affective component may be primarily responsible for the persistence of cuisine; the individual eats a pattern offoods consistent with the cultural rules of cuisine not just because she has knowledge that it is appropriate, but also because what is appropriate has become the most acceptable in that context. By contributing to the likelihood that different foods are eaten at different mealtimes, these systematic shifts in acceptability shown to occur with shifts in mealtime context can function in combination with within meal changes in acceptability (e.g., sensory specific satiety and conditioned state dependent preferences) to increase variety in the diet. Surprisingly, the affective shifts noted in the present study and found to be related to rules of cuisine regarding food appropriateness are already operative in the food preferences of children as young as three years. Casual observation suggests that children are not explicitly taught these rules offood appropriateness; in most cases they must be inferred from the repeated association of food and context that the children experience in meal patterns generated by their elders. The children in the present study possessed knowledge of these rules despite the fact that both parents and children reported that on some occasions the children's meal patterns violated the rules. In informal interviews following the experiment, parents reported that their children sometimes did eat breakfast foods at dinner, although the adults themselves did not do so. Reasons given by parents for these "rule violations" included: foods on the dinner menu believed age inappropriate (usually due to cost or spiciness) or not accepted by the child, children eating dinner separately from parents, breakfast foods well accepted and easy to prepare. Some of the children who categorized breakfast and dinner foods appropriately then volunteered that they could (and sometimes did) eat breakfast foods for dinner (the reverse pattern was essentially never seen). The fact that the children did eat breakfast foods at dinner is probably related to the children's smaller time of day shifts and overall preference for breakfast foods regardless of time of day. Although there are no data at present on the etiology of these systematic shifts in acceptability that occur with time of day, it is possible to speculate on how such patterns might be acquired. Affective reactions to foods can be associatively conditioned to cue complexes that include aspects of the individual's internal state, organoleptic characteristics of foods (Booth et aI., 1982) as well as the social context in which foods are presented (Birch et al., 1984). It is likely that mealtime context could become a part of the cue complex that is conditioned through the consistent repeated experience of eating a food in one particular mealtime context. REFERENCES
Armstrong, S. A. A chronometric approach to the study of feeding behavior. Neuroscience and Biobehavioral Reviews, 1980, 4, 27-53. Birch, L. L. Dimensions of preschool children's food preferences. Journal ofNutrition Education, 1979, 11, 77-80. (a) Birch, L. L. Preschool children's food preferences and consumption patterns. Journal of Nutrition Education, 1979,11,189-192. (b). Birch, L.L. Effect of peer model's food choices and eating behaviors on preschoolers' food preferences. Child Development, 1980, 51, 489-496.
116
L. L. BIRCH ET AL.
Birch, L.L., Birch, D., Marlin, D. W., & Kramer, K. Effects of instrumental consumption on children's food preference. Appetite, 1982,3, 125-134. Birch, L.L., & Marlin, D. W. I don't like it; I never tried it: Effects of exposure to food on twoyear-old children's food preferences. Appetite, 1982, 3, 353-360. Birch, L. L., Marlin, D. W., & Rotter, 1. Eating as the "means" activity in a contingency: Effects on young children's food preference. Child Development, 1984, 55, 431-439. Birch, L.L., Zimmerman, S., & Hind, H. The influence of social-affective context on preschool children's food preferences. Child Development, 1980, 5/, 856-861. Booth, D. A. How nutritional effects of foods can influence people's dietary choices. In L. M. Barker (Ed.). The psychobiology ofhuman food selection, Pp. 67-84, Westport, CT. Avi, 1982. Booth, D. A. Appetite and satiety as metabolic expectancies, In Y. Katsuke, M. Sato, S. F. Takage, & Y. Oomura (Eds.). Food intake and chemical senses, Pp. 317-330. Tokyo: University of Tokyo Press, 1977. Booth, D. A., Mather, P., & Fuller, 1. Starch content of ordinary foods associatively conditions human appetite and satiation indexed by intake and eating pleasantness of starch-paired flavours. Appetite, 1982, 3, 163-184. Cabanac, M. Physiological role of pleasure. Science, 1971, 173, 1103-1107. Pliner, P. The effects of mere exposure on liking for edible substances. Appetite, 1982,3,283-290. Rodin, J. Social and immediate environmental influences on food selection. International J oumal of Obesity, 1980,4,364-370. Rolls, B. 1., Rolls, E. T., & Rowe, E. A. The influence of variety on human food selection and intake. In L. M. Barker (Ed.). The psychobiology of human food selection, Pp. 101-122. Westport. CT: Avi, 1982. Rolls, E. T. & Rolls, B. 1. Brain mechanisms involved in feeding. In L. M. Barker (Ed.), The psychobiology of human food selection, Pp. 33--62. Westport, CT: Avi, 1982.
Received 27 September, 1983; revision 22 December, 1983