Contents - Previous - Next

This is the old United Nations University website. Visit the new site at http://unu.edu

Actual Meal Data

The following are some of the instruments used to collect actual meal data: (1) precise dietary records, including weighed and measured intake; (2) structured diaries of daily or special meal preparation and meal-consumption behaviour; and (3) recall of special events such as life-cycle rituals.

We originally selected the seven-day period for dietary record-keeping because it is commonly used for this purpose and because we had been told through the survey that the week was an important cycle in the community. However, we later discovered that for obtaining information on patterns, rather than nutrient intake, what is needed is a relatively large sample of all the types of food events in the repertory. In other words, large numbers of feasts on Tuesdays or Sundays are needed in order to develop a picture of a pattern, rather than one example of each day. Moreover, in order to investigate the degree to which ideals are carried out in real behaviour or mitigated by other constraints, we need numbers of each kind of event. Since we used seven-day dietary records of weighed and measured intake, this provided only one Friday, one Sunday, and one sample of a weekly cycle of days. These were not enough to test the discovered ideal principles of alternating formats or the rules for marking Fridays and Sundays. Moreover, we had no pool of feasts to analyse. Feasts occur frequently in most societies and account for significant intake, yet we did not have adequate samples of any one type of feast.

Intake data were analysed from 35 households in order to evaluate the survey information and to probe some ideas more intensively. In some cases, individuals were conscious of a pattern or rule and others were not. For example, the menu data indicated that most households actually practice an intra-weekly alternation of "gravy" (pasta) meals and "platters" (meat and potatoes), but very few had been able explicitly to recognize the pattern and state it in the interview. Real menu data were also useful for identifying patterns not consciously recognized by respondents, such as types of "eating out" events and the range of feast formats. People were not aware of the structural parallels between feast formats and everyday menus. For example, the ordinary gravy meals, Sunday dinner, and calendrical holidays were all similar in structure but increasingly elaborate. This structure was obvious in the analysis of observed and recorded menu material but was not readily recognizable in a survey.

We trained our informants to record their own menus. In the first enclave studied, the logistical problems involved in weighing and measuring were intrusive. Later, we developed structured diary forms that included descriptions of all shopping and preparation activities (including social interaction), designations of those present at the meal and where they sat, as well as comprehensive meal information including menu, recipes, style, order, and amount or servings. All that was missing were actual weighed and measured intake data.

Instead of seven-day information we should have sampled several such temporal sequences over the year. We should have had more than one week's data to test the principles of intra-weekly cycles and weeks from different seasons, We should have elicited menu data from feasts throughout the year to accumulate an adequate number of each type. One can even use recall data for recent major feasts that occur rarely, such as weddings and funerals.

It is important to indicate that we could not have analysed the actual menu data without the clues from survey material. The menu data would have appeared random and unpatterned without the suggested pattern rules developed from the survey. We were even able to use the pattern rules to analyse menu data collected from Italian-Americans in New Haven in 1935. These rules had not been obvious to the dietician who collected the menu information, since she used her own Anglo menu categories in collecting the data, and did not see that all meals fell into several structural types.

Discrepancies between Ideal Patterns and Real Behaviour

In analysing the real meals according to the ideal pattern rules, we encountered many discrepancies. Were the surveys useless statements of rules that were never followed? In the participant observation we discovered that the pattern rules did function but were not rigid. There were often many choices for a particular occasion and the outcome was a function of context: time, social participation, and situational pressures. For example, on the survey people listed one rule for Sunday meals and one rule for Fridays. The actual meals revealed little observance of this pattern. Participant observation revealed that there were actually several alternative ways to mark Sundays or Fridays, all of which were acceptable. Thus, while there was no single way to mark these meals, almost every household marked them in some special way (table 5).

Participant observation also allowed us to discover a weekend meal cycle that no informant had explicitly recognized in the interview phase, and we also discovered a new meal format. We observed how unanticipated irregularities in activities and social contexts led to regular, predictable, and expedient meal formats that had not been explicitly recognized by informants.

The expense in time and money involved in the intensive phase limited its use. Only four households were studied - one for two months continuously and three for one month each. However, the rewards in terms of insight made this an essential research component.

Table 5. Information provided by three kinds of data (percentages)

a. Awareness of relevance of other rules was derived from ethnography.
b. 50 per cent had fish or seafood; 50 per cent had meatless dish.
c. 65 per cent celebratory restaurant; 33 per cent Anglo mast meat.
d. 67 per cent Anglo roast meat.
Source: Goode, Curtis. and Theophano, 1984, p 215.

We paid the households a fee to cover the costs of the food eaten by the participant observers and to compensate for the time of the cook. The participant observers lived in their own homes but travelled to the households before breakfast and observed all food-related activities, including shopping, preparation, and mealtimes, every day for a month. The selection of households for this intimate kind of research could not be based on random sampling but was limited to those families who had the facilities and were willing to be subject to the presence of the ethnographer. The households selected had to satisfy criteria related to their position in the social networks of the community, since the role of ethnography was to provide information about social processes involved in menu negotiation. This is the process in which the cultural repertory of meal types and social pressure to conform is shaped by circumstances (household composition, income, parallel events, and activity patterns) to produce the actual selection of meal types and the food content of meals for particular occasions.

The use of formats also enabled us to understand change in this community. Change occurred by deleting formats, by adding new formats for newly recognized occasions, by adding formats to the range of alternatives for an old occasion, and through changing the rules for an old format.

We eventually described five weekday formats and six feast formats (table 6), which played a major role in food-item selection. We could indicate the way in which format selection influenced item selection by looking at how content-specific a given format was. This was possible after having analysed the variability in content for each format within each household and across households for each event (table 7). Table 8 indicates briefly some changes in formats in recent decades.

Insights developed from limited participant observation must be further tested on a broader sample of the population. For example, a hypothesis about the effects of household differences (in terms of resources, activities, roles, or social-network embeddedness) can be tested by comparisons using a survey developed and refined by participant observation. Furthermore, large amounts of actual menu data collected via diaries from a large sample of households as well as a diverse set of time periods (daily, weekly, and seasonal cycles) could be used to test the observance of pattern rules.

Table 6.

Table 7. Degree of content specificity

a. Significant negotiation for content.

Table 8. Historical continuity and modification

Feasts were highly significant in total intake since they occurred frequently in the community. There were at least six to eight special-occasion formats occurring in any two-week period for the households studied. During holiday seasons the number might double or triple. During two months in one household, guests were present for special events over 100 times, each calling for a limited range of formats. If we had not had data for time periods of more than a week, we would have missed the frequency of special occasions.

The Jerome Study

The third study using formats is Jerome's (1979) study of change in food-patterning among migrant Black households in a northern city. This used a method of describing meals that differed from that used in the Douglas and Nicod study (1974) and in the Russell Sage Project Study (1984). The meal formats were described primarily in terms of items rather than structure (rules for sequencing, presenting, and serving dishes). The format was used as a unit, helping to explain change along with five other sources of change: food supply, preservation techniques, food processing, preparation styles, and beliefs.

To collect data Jerome used a combination of structured intensive interviews with 63 households and seven-day dietary records for 23 households. Participant observation was sporadic. Most of the information about specific types of meals came from dietary records rather than observation, so they consisted of types differentiated by the observer rather than by informants, and they were item-based rather than defined by temporal and spatial presentation and methods of service.

Jerome's analysis focused on the everyday cycle of meals and Sunday dinner rather than on special events. She looked at the organization of breakfasts, lunches, and dinners and the ways they have changed in scale and content. Interview data provided some information about how special occasions were structurally related to daily meals (more items and courses). Each format had a core of items that were uniform among households and within households. Secondary items were those that were added to the core and were variable from house to house and meal to meal. Peripheral items were restricted to particular occasions, often holidays and ritual celebrations (tables 9, 10, and 11).

In one important instance, a set of contrasting dinner formats was distinguished by foodpreparation techniques, which in turn led to differentiated content (items). Dinners were referred to as "frying nights" or "boiling nights." Different items were used to produce a fried meal of fried meat with starch and vegetables or a one-pot boiled meal of greens and meat. The use of these formats, in terms of their placement in the daily and weekly cycle in the new setting, changed, leading to changes in food intake. This is another instance of the relevance of the format as a useful unit for understanding change.

The new dinner pattern emerges as a combination of the old breakfast (fried meats) and dinner (boiled vegetables or legumes) patterns. Instead of five "boiling" days as described in the traditional dinner pattern, there are now two or three; this leaves room for two or three "frying" days to replace the old breakfast pattern. The general absence of vegetables in the current "frying" pattern is probably accounted for by this shift, since there were no vegetables in the old breakfast pattern. Thus, changes in the nomenclature of meals and in meal times are expressed in meal components and in nutrient content (Jerome. 1979. p.301).

Table 9. Meal components of the lunch pattern

Source: Jerome, 1979, p. 288.

Table 10. Meal components and preparation methods: determinants of the dinner pattern

Source: Jerome, 1979, pp. 289-290.

Table 11. Components of the meal pattern and factors involved in determining the Sunday dinner

Source: Jerome, 1979. p. 291.

Contents - Previous - Next