WHERE HAS FAMILY TIME GONE? IN SEARCH OF JOINT FAMILY ACTIVITIES AND THE ROLE OF THE FAMILY MEAL IN 1966 AND 1999 Inge Mestdag Jessie Vandeweyer The idea that family meal time is disappearing is gaining growing attention in Western societies. This article investigates to what extent family time has decreased and what place the family meal has within family time. Belgian time-budget data gathered in 1966 and 1999 were used to answer these research questions. Analyses show parents were spending less time together as a family and also on family meals, especially on working days. Nevertheless, the growing number of dual-earner families was not responsible for the decline in family meal time between 1966 and 1999. Keywords: family time; family meal; time use; change 1966-1999; Belgium Since the 1850s, life within a nuclear family has been considered an important ideal in Western society. As the impact of industrialization on society grew, the middle class began to attach growing importance to the idea of the family and the home. Industrialization promoted an ideology that identified work and the wider public area with the male and the private, domestic area with the female. The home was the place a man could come home to, where the private, caring family could protect itself against the heartless, competitive world.1 And it was women's work to make the house into a home. In contrast to housekeeping, homemaking was considered a labor of love that Inge Mestdag is a research assistant for the Fund for Scientific Research-Flanders (Belgium; F.W.O.-Vlaanderen). She graduated as a sociologist in 2001. In 2002, she was awarded the yearly prize of the Flemish Association of Sociology (WS—Vlaamse Vereniging voor Sociologie) for her article based on her licentiate 's thesis. In the first volume of Food and History (2003), she published on her thesis with the article "Introducing Italian Cuisine into Flemish Home-meal Cooking in the Twentieth Century: An Analysis of the Flemish Cook Bible Ons Kookboek (1927-1999)." Currently, she is preparing a Ph.D. dissertation on the disappearance of the traditional meal in Belgium. Jessie Vandeweyer works as a research assistant at the Vrije Universiteit Brüssel. She graduated as a sociologist in 2000. She specializes in time-budget research, labor, and social relations. Currently she is working on a Ph.D. dissertation on voluntary career interruption, specifically social integration, time use, and time structuring of time-credit system users in Flanders. (Major publications available only in Dutch; conference papers available in English.) Journal of Family History, Vol. 30 No. 3, July 2005 304-323 DOI: 10.1177/0363199005275794 © 2005 Sage Publications 304 Mestdag, Vandeweyer / ROLE OF THE FAMILY MEAL 305 involved the creation of myths, symbols, and rituals that turned a house into a home.2 Family life had not only its special location in the family home, but also its special time. Starting in the second half of the nineteenth century, Victorian Protestant culture emphasized the importance of hard work. Together with the greater symbolic importance attached to family life, the smaller amount of real time families could spend together resulted in a growing importance attached to special, symbolic times, such as family meals and childrens' bedtime. For the first time in history, there was a need for quality time with the family.3 It was also around that time that Sunday came to be considered the archetypal day for the family.4 Although the ideology of the family was only responding to a middle-class reality, this falsely monolithic concept of the family5 has had an important influence throughout society. Spending adequate time with family members is deemed necessary for both personal and collective well-being. As such, the ideal of spending quality time with other family members is an important guiding principle for modern Western families. Recent research has shown that even though family time is subject to serious competition because of changing structures of work and the family, ideas on family time seem resistant to change. Although they hardly ever become real, ideas on family time are still vividly cherished and wished for. Despite innumerable frustrations, ideals on family time live on as powerful cultural standards.6 The family meal was also attached great value ever since family life became central. Before that time, it was unthinkable that the family would reserve a special time to share a meal.7 Qualitative research by DeVault shows that in the early 1980s, parents were still concerned about establishing a regular meal pattern in their families. Such a pattern was considered an important organizing entity within family life, with family meals setting the rhythm of family life. Family meals remain a central ritual within family life, guaranteeing the continuation of a family.8 At the moment, there is considerable concern over the decline of the family meal. However, figures from various studies on the percentage of families sharing meals vary considerably.9 The different survey methods of the various studies may account for their varying results in considerable measure.10 The way people remember things often differs from the way things were actually done. For that reason, we often think things were better in the past. Although they are rather recent developments, spending quality time with the family and sharing family meals are considered to be part of tradition and are awarded some kind of eternity that seems to be threatened now. However, the idea that family time is in danger is not entirely new. The same idea cropped up in the 1920s and the 1940s. In fact, it has been a hot item ever since Western society started to attach great importance to family life in general.11 In this article, we wish to focus on how much time people actually spend together as a family and to what extent family time has actually declined in Belgium. Time-budget data are used to answer these research questions. First, we focus on how time spent with the family has changed over time. Second, we move on to the variety of activities families do together and how much time these take in the total family time budget. More particularly, we focus on the family meal, which takes the largest part of the total family budget. Next, we study to what extent family time and family meal timing have changed. Finally, we look for the impact of structural changes that might account for the decline in family time and family meals between 1966 and 1999. 306 JOURNAL OF FAMILY HISTORY / July 2005 DATA To get a better view on how family time has evolved during the past decades, we used data from two Belgian time-budget studies. Time-budget studies give us a reliable view on how people really spend their time.12 The first time-budget study was conducted between January and April 1966 and contains time-budget data from 2,077 Belgians aged between nineteen and sixty-five.13 All respondents kept a diary during twenty-four hours. This study was part of the Multinational Time Budget Study of 1965.14 The second study was conducted between December 1998 and January 2000 and contains time-budget data from 8,382 Belgians aged between twelve and ninety-five originating from 4,275 families. All respondents kept a diary during one working day (Monday through Friday) and one weekend day (Saturday or Sunday). Both original data sets were weighed by age, sex, and education level of the respondent. As we were interested only in family time for this study, we decided to remove all respondents that were not members of a traditional family. As such, our database contains only time-budget data from parents who live as a couple with at least one child. In the two time-budget studies different age limits were applied to define children. In 1966 and 1999, children were defined as children living in the household, younger than nineteen and younger than sixteen, respectively. As there was no further indication of children's age in 1966, we were unable to overcome this problem. In the 1999 database we selected only one respondent per family. We used this smaller random sample instead of the total 1999 database to overcome interdependence between respondents (partners from the same family). The fact that we were interested in time spent with one's family provided all the more reason to do this.15 As such, we relied on individual data to define family time. We defined family time as the time spent on each activity with other family members present, as indicated in the individual diaries. Thus, family activities were not defined as activities which were done by all family members at the same time, as appeared from all the family members' time budget. RESULTS Is family time on the decline? In several countries doing time-use research we discern a general trend of increased time devoted to child-care activities by parents.16 Under debate is whether this trend is a result of an actual increase in child-caring activities or rather due to a growing awareness of the importance of child-rearing activities reported by those filling in the time-use diary.17,18 How much time is spent in the presence of children is not always clear due to different methodologies used to measure contact time with children.19 Recent social developments such as the rise of the labor force participation of women, however, could suggest that the overall time spent in the presence of children has reduced, thus shifting from quantity time to quality time with children. In Belgium, the time for joint family activities has declined (Table 1). Belgians spend significantly less time with their partner and children in 1999 than in 1966. On an average working day, the time spent with both partner and children has declined by twenty-five minutes. There seems to be an overtaking maneuver on Saturdays, though, with Belgians spending more than one hour more with both partner and children in Table 1 Mean Duration of Time Spent Alone, with Partner and/or Children, and Others (Working Day and Weekend, 1966-1999) Duration Working Day Saturday Sunday 1966 (n = 786) 1999 (n = 661) 1966 (n = 122) 1999 (n = 309) 1966 (n = 108) Alone 3:30 3:34 (ns) 3:23 3:13 (ns) 2:42 With children, without partner 2:02 2:06 (ns) 1:58 2:15 (ns) 1:49 With partner, without children 2:05 1:57 (ns) 2:40 2:38 (ns) 3:11 With children and partner 2:12 1:47*** 3:19 4:31* 4:47 With others, without children and without partner 0:57 1:01 (ns) 1:26 1:03 (ns) 1:25 Note: Statistical significance of the difference between 1966 and 1999 for time spent alone and in company. *p < .01. **p < .001. ***p < .001. 1999 (v\ = 353) 3:00 (ns) 1:56 (ns) 2:46 (ns) 4:28 (ns) 0:47 »»» 308 JOURNAL OF FAMILY HISTORY / July 2005 1999 than in 1966. This is probably due to the fact that paid labor occurred less on Saturdays in 1999 than in 1966. On Sundays the difference is not significant. The time spent with children solely or with the partner solely has not changed significantly between 1966 and 1999. In the following paragraphs we will concentrate on the time spent as a complete family, that is with both partner and children.20 This time is clearly under a lot of pressure. By consequence it is a hot issue in the media today. What Do Families Do Together? The scarce time spent by family members in each other's company is predominantly used for sharing meals (Table 2). Eating is still the social moment par excellence for keeping up with the latest news and events of other family members. This holds true for both 1966 and 1999 and on every day of the week, although the share of eating in the total family-time budget has declined significantly during the past thirty years. Watching television has had a strong hold on this second position for more than thirty years now, on all weekdays. Competition is coming up, though, from traveling. On Saturdays in 1999 families spend more time traveling together than watching TV together. Although watching TV is the second-most important family activity after the meal on weekdays and Sundays in 1966 as well as in 1999, the time spent on this activity as a family as well as its share in the total family-time budget declined. This is not due to diminishing hours of viewing. On the contrary, time spent on watching television (for the Belgian population between 19 and 65 years old) increased by forty-one minutes on a weekday, by fifty-three minutes on a Saturday, and by thirty-three minutes on a Sunday. The same trends can be seen for our subsample of parents. By consequence, watching TV has lost importance as a joint family activity. Instead, we see a rise in solitary television watching and one parent watching TV with the children. The same trends go for working days and weekend days. This finding goes hand in hand with the rising number of television sets in a household.21 Meanwhile, traveling with the whole family went from almost nonexistent to 12 percent to 14 percent of the total amount of family time on all weekdays. Performing household tasks came third in rank in 1966. In 1999, it came third on Saturdays and fourth on working days and Sundays, when more leisurely activities are done instead in presence of the family members: watching TV on Saturdays and traveling on working days and Sundays. However, Saturdays are marked by significantly more time spent in the presence of the family to purchase goods. The time spent with the complete family has reduced on an average workday from two hours and twelve minutes in 1966 to one hour and forty-seven minutes in 1999. Nevertheless, Table 2 shows that significantly more time is spent on playing with children as a family activity. On working days this activity took up four minutes, or 3 percent, of the total family-time budget in 1999, while it was nonexistent in 1966. Previously this task was probably performed by only one of the parents, most likely the mother. Saturdays seem to be suitable to catch up on lost time as a family. In 1999, one hour more is spent in the presence of all family members than in 1966 (Table 1). As a result time spent on child care and playing with children quadrupled on Saturdays, while its share in the total family-time budget tripled. Although the time spent with the complete family has not changed significantly on Sundays, child care and playing with children doubled. The share of child care in the total family-time budget in 1999 has consequently risen on every weekday, and especially on the week- Table 2 Time Spent with Partner and Child at Most Important Activities Working Day Saturday Sunday 1966 1999 1966 1999 1966 1999 (N = 786) (N = 661) (N = 122) (N=. 309) (N = 108) (N = 353) h, mm % h, mm % Significance h, mir, % h, mir, % Significance h, mir, % h, mm % Significance Eating 0,51 38 0,27 25 ,„ 1, 05 32 1, 03 23 ns 1,23 29 1,05 24 „ Watching TV or video 0,28 22 0, 16 15 *»* 0, 35 17 0, 24 9 ns 1, 11 25 0,36 13 *»* Household work 0, 17 13 0, 11 11 ** 0, 30 15 0, 27 10 ns 0,23 8 0,23 9 ns Traveling 0,01 1 0, 13 12 *»* 0, 05 3 0, 37 14 *»* 0, 12 4 0,33 12 *»* Childcare, playing, and reading combined 0,08 6 0, 11 9 ns 0, 06 3 0, 25 9% *«* 0, 10 4 0,21 8 * Care only 0,06 5 0,07 6 ns 0, 06 3 0, 14 5 * 0,06 2 0, 11 4 ns Playing and reading only 0,01 1 0,04 3 *»* 0, 00 0 0, 11 4 *»* 0,03 1 0, 10 4 * Purchase of goods 0,00 0 0,04 4 *«* 0, 02 1 0, 17 6 *«* 0,00 0 0,04 1 ns Reading 0,08 6 0,01 1 *»* 0, 16 8 0, 03 1 *»* 0, 13 5 0,05 2 ** Other activities 0, 19 14 0,20 19 — 0, 35 18 — 1,07 23 1,07 24 — Total 2, 12 100 1,47 100 *** 3, 20 100 4, 32 100 ** 4,48 100 4,35 100 ns Note: Statistical significance of the difference between 1966 and 1999 for time spent with partner and child for different activities, h : *p < .05. **p < .01. ***p < .001. hours; mm = minutes. 310 JOURNAL OF FAMILY HISTORY / July 2005 end, implying that more of the child-care activities are taken up by both parents together. Consequently, childcare is to be found in the top five of activities done as a complete family in 1999 on every weekday, which was not the case in 1966 (only on working days). While eating and watching television are the most important activities of the family as a whole, watching TV is also the most popular activity done by partners when children are not present. Both in 1966 and 1999 approximately one-third of the time spent with the partner is taken by watching television (table not shown). Eating (9 percent to 14 percent of the total time budget of partners) and household tasks (from 10 percent to 18 percent of the total time budget of partners) are often performed in the presence of the partner as well. On the contrary, household tasks with only the children present have declined. In 1966 spending time with children during a household activity such as cooking, doing the dishes, ironing, and cleaning took the larger share of the time budget with children (40 percent on working days and Saturdays, 20 percent on Sundays). In 1999, this activity has been overtaken by child care on all weekdays (accounting for between 23 percent and 29 percent of the time spent with children). Consequently, time with only children present has changed from children in "passive" presence of the parent to "active" interaction between the parent and the children. The Place of the Family Meal in the Family-Time Budget It is striking how the family meal, however still being the most important family activity, has lost its importance, especially on working days and to a lesser degree on Sundays. On an average working day in 1966, fifty-one minutes were spent on eating together, 38 percent of the total family-time budget (Table 2). A family meal in 1999 took only twenty-seven minutes, a quarter of the total family-time budget on an average working day. Even on Sundays, the archetypal day for family time, the family meal lost much in influence. The Sunday meal in the presence of the family was almost twenty minutes shorter than in 1966, lasting one hour and five minutes. On Saturdays, there was a status quo with about one hour spent on the family meal. However, as family time increased significantly on Saturdays, the share of the family meal decreased in the total time spent as a family, accounting for a quarter of total family time in 1999, but one-third of the total time spent as a family in 1966. The decline of the family meal resulted not only from the reduced time spent on eating with partner and children and its decreasing share in total family time, but also from the declining number of commensal occasions with partner and children. This held true for both weekends and working days. On working days the average number of eating occasions shared with partner and children decreased significantly from more than one and a half (1.56) daily family meals to less than one family meal a day (.88). On Saturdays and Sundays the number of family meals was higher than on weekdays. In spite of Sundays' importance with regard to the family and Sunday lunch,22 the family meal was unequivocally robbed of its strength. On average, only 1.68 meals were taken as a family on that day, while in 1966 2.29 meals were family meals on Sundays (Table 3). As these average numbers of family meals are rather abstract, we tried to clarify them by adding some frequencies reporting the daily number of family meals. Although from a historical viewpoint three daily meals is a fairly recent development,23 it is considered the standard pattern in most Western (European) countries and Mestdag, Vandeweyer / ROLE OF THE FAMILY MEAL 311 Table 3 Mean Number of Family Meals and Percentage of Belgian Parents Taking Zero to Six Daily Family Meals on Working Days, Saturdays, and Sundays (1966 vs. 1999) Working Day Saturday Sunday 1966 1999 1966 1999 1966 1999 (n = 786) (n =661) (n = 122) (n = 309) (n = 108) (n = 353) Mean number of family meals 1.56 0.88*** 1.80 1.62 (ns) 2.29 1.68*** % taking zero family meals 13.7 39.8*** 12.5 25.2** 8.7 21.3** % taking one family meal 37.9 39.2 (ns) 26.8 19.6 (ns) 14.1 21.2* % taking two family meals 30.1 14.3*** 35.4 28.2 (ns) 28.9 29.6 (ns) % taking three family meals 16.1 6.1*** 19.7 22.8 (ns) 37.4 22.9** % taking four family meals 2.0 0.4** 4.9 3.0 (ns) 10.2 4.5* % taking five family meals 0.2 0.2 (ns) 0.7 1.1 (ns) 0 0(ns) % taking six family meals 0 0(ns) 0 0(ns) 0.8 0(ns) Note: Statistical significance of the difference between 1966 and 1999 for mean number of family meals and percentage of Belgian parents taking zero to six daily family meals. *p < .05. **p < .01. ***p < .001. also in our country.24 However, not all meals are family occasions. History shows that breakfast has not always been a commensal occasion. Recently, however, concern is rising that not only breakfast, but also lunch and dinner are becoming solitary occasions.25 This concern is especially strong in the United States, though it has found a considerable echo in Western Europe too. In the case of the present-day Western family, the intensity of domestic commensality may be considered an indicator of the integration of the family group, of the degree to which family life resists pressures from parents' occupations and childrens' schooling.26 Everywhere, efforts are rising to promote regular family meals. As appears from Table 3, on working days and Saturdays one or two daily family meals were most common in 1966. No daily family meal was exceptional, accounting for only 13.7 percent on working days and 12.5 percent and 8.7 percent on Saturdays and Sundays, respectively. In 1999, no daily family meal was much more widespread. Four parents in ten did not manage to share a meal with their partner and their children on a working day. On Saturdays and Sundays respectively, a quarter and one in five failed to have a family meal. In 1999, none or only one daily family meal became the general practice on working days, both accounting for about 40 percent of Belgian parents. On Saturdays and Sundays two daily family meals were most common. In 1966, three daily family meals was still most common, accounting for 37.4 percent of Belgian parents. In 1999, the share of Belgian parents having three family meals on Sundays was almost cut in half, leaving the majority sharing a family meal only once (21.2 percent) or twice (29.6 percent). Timing of Joint Family Activities Figures 1A through 3B show us the course of family time and family meals during the day in 1966 and 1999, for working days, Saturdays, and Sundays. To take into account the significance level of the differences between 1966 and 1999, we performed analyses of variance. These significance tests compared hourly mean percentages of Belgian parents engaged in family time or family meals from 1966 and 1999. The results of these analyses will be discussed in the following paragraphs. The daily 312 JOURNAL OF FAMILY HISTORY / July 2005 Sg ? E Time of day (Working day) Figure 1A. Family Time by Time of Day on Working Day C IĎ Q) OJ W E c >, Q) — "D E o S 60 40 ■ 20 ■ o 4 "66 -99 Time of day (Working day) Figure 1B. Family Meal by Time of Day on Working Day Time of day (Saturday) Figure 2A. Family Time by Time of Day on Saturday Time of day (Saturday) Figure 2B. Family Meal by Time of Day on Saturday Mestdag, Vandeweyer / ROLE OF THE FAMILY MEAL 313 -ô E 0 Time of day (Sunday) Figure 3A. Family Time by Time of Day on Sunday ? E Time of day (Sunday) Figure 3B. Family Meal by Time of Day on Sunday rhythm of family time and family meals on working days, Saturdays, and Sundays will be discussed subsequently. Working days. Joint family activities typically took place around eating times. This was not surprising because eating was the most important family activity. This held true especially for working days. Peak times occurred around 7:30 a.m., 12:30 p.m., and 7:20 p.m. However, there appeared to be significant differences between 1966 and 1999 in the percentage of Belgian parents engaged in family time and family meals. On working days in 1999 significantly fewer parents spent time with their children in the early morning. In 1966,14.3 percent of Belgian parents were spending time with partner and children between 7:00 a.m. and 8:00 a.m. (Figure 1 A). In 1999 this held for barely 6 percent. The decline in family time in the early morning also appeared from the decline in family breakfast (Figure IB). With only 2.6 percent (one-third of 1966), there was a significant decrease in parents eating with children and partner between 7:00 a.m. and 8:00 a.m. Later in the morning, though, significantly more parents engaged in family time in 1999 than in 1966. At noon, the same held true as in the early morning: only 10 percent of Belgian parents shared time with partner and children between noon and 1:00 p.m.; that is less than half the number recorded in 1966. Again, the decline of family time was reflected in the share of Belgian parents sharing meals with their families: in 1966,18 percent shared meals at noon; in 1999 this held for less than 6 percent. After the noon peak, the percentage of Belgian parents engaged in family time plunged in 1966, while there was a gradual decline in 1999. As such, in the afternoon (2:00 to 5:00 p.m.), significantly more parents spent time with their families 314 JOURNAL OF FAMILY HISTORY / July 2005 in 1999 than in 1966. However, in 1999 significantly fewer parents engaged in a 4:00 family snack between 4:00 and 5:00 p.m. The same held true between 5:00 p.m. and 6:00 p.m., announcing the later start of family dinner in 1999. In the evening (as from 6:00 p.m.), fewer parents engaged in family time in 1999. Between 6:00 p.m. and 7:00 p.m. more than one out of three Belgian parents were spending time with partner and children in 1966. In 1999, this held for barely a quarter of Belgian parents. Moreover, the percentage of parents sharing family meals was halved in 1999, 9 percent instead of 17 percent in 1966. Again, we conclude that meals are being timed later than before. Between 8:00 p.m. and 9:00 p.m., almost 5 percent of Belgian parents shared a family meal in 1999. In 1966, that was only just more than 2 percent. After the evening peak of 7:20 p.m., the share of parents engaged in family time remained rather high until 10:00 p.m. After 10:00 p.m., there was no difference in the share of parents engaged in family time. Saturdays. On Saturdays, family time started about an hour later in 1999 than in 1966 and lasted one hour longer in the evening (Figures 2A and 2B). In 1999, significantly fewer parents spent time with their partner and children between 7:00 a.m. and 8:00 a.m., while significantly more parents spent time with partner and children between 9:00 a.m. and noon. In contrast to what we diagnosed in 1966, there was no early-morning peak in family time in 1999. Instead, there was a gradual increase in the share of parents engaged in family time as the morning moved on. On Saturdays, 5 percent to 7 percent of Belgian parents were engaged in family time between 9 a.m. and noon. In 1999, in that same time span, 16.5 percent to 18.5 percent were engaged in family time. In contrast to working days and Sundays, there was no significant difference in parents engaged in family time at noon, between noon and 2:00 p.m. On the other hand, there was a significant decline in the percentage of parents engaged in family meals between noon and 1:00 p.m. The percentage engaged in family meals declined from almost a quarter to more than 16 percent between 1966 and 1999. In the afternoon (2:00 p.m. to 6:00 p.m.), more parents spent time with their families. In contrast to more family time, significantly fewer parents engaged in family meals between 4:00 p.m. and 6:00 p.m., testifying to the general disappearing of the 4:00 snack and the later start of the family dinner in 1999. There was no significant difference between 1966 and 1999 in the percentage of parents engaged in family time between 6:00 p.m. and 8:00 p.m. Mealwise, our data showed that dinner on Saturday was timed later. As from 7:00 p.m., the share of parents engaged in family meals was significantly higher in 1999, although their share is significantly smaller between 6:00 p.m. and 7:00 p.m. Between 8:00 p.m. and 10:00 p.m., the share of parents engaged in family time was significantly higher in 1999. After 11:00 p.m. the percentage of parents engaged in family time plunged in 1966, while it remained rather high in 1999. Moreover, it appeared that more parents were engaged in family time in 1999 than in 1966. This was especially true on Friday nights or rather early Saturday mornings. While in 1966 almost no parents were spending time with partner and children, in 1999 more than 4 percent of parents spent time with children between midnight and 1:00 a.m. The following hour this held for 2.4 percent of parents in 1999. Sundays. As on Saturdays, Belgian parents started their Sundays later as family days. In 1999, this trend was more pronounced than in 1966 (Figure 3A). Between 7:00 a.m. and 10:00 a.m., significantly fewer parents engaged in family time. The per- Mestdag, Vandeweyer / ROLE OF THE FAMILY MEAL 315 centage did not even cross the threshold of 20 percent before 10:00 a.m., while in 1966 family time reached its first peak at 8:50 a.m. with more than 34 percent of parents spending time with partner and children. Moreover, significantly fewer parents engaged in family breakfast in 1999. With about one in five Belgian parents engaged in family time between 10:00 a.m. and noon, there was no difference in family time. At noon, on the contrary, there was a significant decline in the parents engaged in family time between 1966 and 1999. Between noon and 2:00 p.m., the share of parents engaged in family time declined from 43 to 46 percent to around 39 percent. Again the decline of family time was reflected in the share of Belgian parents sharing meals with their families. On Sundays the percentage engaged in family meals declined from a quarter between 14 and 17 percent. However, between 2:00 p.m. and 3:00 p.m., almost one out of twelve parents is engaged in a family meal, indicating that family dinner is taken later in the afternoon on Sundays (Figure 3B). Again, as on Saturdays, significantly more parents spent time with their families on Sundays between 2:00 p.m. to 6:00 p.m. In 1999, Belgians spread their joint family activities more equally all over the afternoon. More than 30 percent of families were together between noon and 9:00 p.m. with a peak at 6:30 p.m. (48 percent). Strangely enough, in 1966, family time reached its summit more than an hour later, at 7:50 p.m., with 53 percent. From then on, family time dropped slowly but steadily. However, on Sunday evenings between 9:00 p.m. and 10:00 p.m. significantly more parents spent time with their partners and children in 1966 than in 1999. THE DECLINE IN FAMILY TIME AND FAMILY MEALS: STRUCTURAL OR BEHAVIORAL CHANGES? The preceding paragraph makes it clear that the average time spent with family and on family meals was on the decline, especially on working days. Of course, these are average numbers. As such, they highly depend on how the population was composed in 1966 and 1999. There was serious reason to believe that the decline in average family time was due to a differential composition of the population in both years, rather than to a decline of family time within the population in general. In particular, the decline of average duration and frequency of family time and family meals might be due to the fact that the share of housewives within the Belgian population has seriously diminished. While the share of housewives in the 1966 questionnaire data amounted to 35 percent, the Belgian data from 1999 contained only 7 percent housewives. As housewives are more available throughout the day, the odds are that they spend more time and share more meals with both partner and children than working women. Other structural changes such as the number of children, the occupational class, and education level might have had an influence on the shift in family time between 1966 and 1999. On the other hand, the changes in family time and family meals might (also) be due to behavioral changes between 1966 and 1999, that is, within specific structural categories of people.27 To describe the evolution of family time and family meals for specific structural categories of the population, we performed analyses of variance. These analyses were complemented with some regression analyses, controlling for the structural factors that changed between 1966 and 1999. The fact that the 1999 questionnaire was not designed to be in line with the 1966 questionnaire severely limited our options to allow for a wide range of structural factors. In our analyses we took into account age, educa- 316 JOURNAL OF FAMILY HISTORY / July 2005 tion level, and occupational class. Furthermore, the number of working hours per week, the flexibility of the work schedule, and the number of children in the household were used in the analyses. Unfortunately, we were ignorant on a man's partner's activity status from the 1966 questionnaire. This questionnaire provided information only on the activity status of the respondent and that of the head of the household. In the case of a working male respondent, the respondent's and household head's occupational status were identical. The 1966 questionnaire provided us with both partners' occupational status only as long as the respondent was not the household head, in point of fact the male working partner. Moreover, the 1966 questionnaire had almost no data from unemployed male respondents. As such, we could not involve men's activity status as an independent variable in the regression analysis for men. Due to the shortcomings of the 1966 questionnaire, we performed separate analyses for women (always with working partner) and men (always working). As the number of respondents was too small for weekend days in 1966, regression analyses were restricted to working days for regression analyses. Analyses of variance were performed for both weekend days and working days, but due to the small numbers in 1966, differences were often found to be insignificant. In the following paragraphs we focus on the results of the analyses of variance, as these provide us with a clearer description of the changes between 1966 and 1999 with regard to family time and family meals. Wherever relevant, the results from the regression analysis were included in the description of the results. Table 4 shows us that on working days women and working men spent less time with partner and children in 1999 than in 1966. This decline was not significant for women in dual-earner families, however. For women in breadwinner families, on the contrary, there was a significant and much more alarming decline in family time. Unemployed housewives with a working partner spent about half an hour less with their partner and children in 1999 than in 1966. Our regression analyses corroborated this finding. Women's activity status could not explain women's decline in family time between 1966 and 1999. Although all the analyses confirmed that employed women spent less time with their families than unemployed women, the decline in average family time between 1966 and 1999 was not due to the differential composition of the population; to wit, increasing female employment. Nevertheless, the year of measurement did not add much to the variation in time spent with family. Neither did women's activity status, the number of children, the number of hours worked per week, the education level, or the flexibility of the work schedule. The changes in time spent with family was thus not due to the aforementioned structural changes. With respect to the evolution of family time on weekends, again the hypothesis of the overtaking maneuver was corroborated. Women in dual-earner families spent significantly more time in the presence of their partner and children on Saturdays in 1999 than in 1966. On Sundays, we noticed a decline for employed men and women in dual-earner families and a slight increase for women in breadwinner families, although none of these evolutions in family time was significant. With respect to the duration of the family meal (Table 5), there was a significant decline for both employed and unemployed women on working days. Women in dual-earner families spent about a quarter of an hour less on family meals in 1999 than in 1966. Women in breadwinner families showed an even greater, significant decline in family meal duration. The evolution in family meal duration was so drastic for housewives in breadwinner families that they were no longer in a more advantageous situa- Table 4 Duration of Family Time on Working Days, Saturdays, and Sundays by Individual/Family Occupational Category and Sex Duration Family Time Working Day Saturday Sunday 1966 1999 1966 1999 1966 1999 Woman in dual-earner family Woman in breadwinner family Employed man 2:02 (n = 96) 2:23 (n = 261) 2:02 (n = 384) 1:45 (n = 506) {ns) 1:50 (n = 152)** 1:39 (n = 749)** 2:12 (n= 19) 3:53 (n = 40) 3:23 (n = 54) 4:19(n = 240)* 4:21 (n = 76) {ns) 4:17(n = 364) {ns) 5:29 (n= 10) 4:10 (n = 26) 4:58 (n = 68) 4:39 (n= 265) (ns) 4:17 (n = 76) {ns) 4:29 (n = 386) {ns) *p < 05. **p < 01. ***p < .001. Tabfe 5 Duration of Family Meals on Working Days, Saturdays, and Sundays by Individual/Family Occupational Category and Sex Duration Family Meals Working Day Saturday Sunday 1966 1999 1966 1999 1966 1999 Woman in dual-earner family Woman in breadwinner family Employed man 0:46 (n = 96) 0:53 (n = 261) 0:49 (n = 384) 0:29 {n= 506)** 0:26 (n= 152)*** 0:26 (n= 749)*** 0:40 {n= 19) 1:16 (n= 40) 1:07 (n = 54) 0:59 (n = 240) {ns) 1:08 (n = 76) (ns) 1:04(n = 364) {ns) 1:22 (n= 10) 1:06(n = 26) 1:27(n = 68) 1:05 (n = 265) {ns) 1:01 (n=76) {ns) 1:05 (n = 386)* *p < 05. **p < 01. ***p < .001. 318 JOURNAL OF FAMILY HISTORY / July 2005 tion than their working counterparts. The same conclusions hold for the number of family meals. Table 6 shows an even more pronounced decline in the number of family meals for housewives in a breadwinner family than for women in dual-earner families. As a result, in 1999 the average number of family meals was almost equal for unemployed and employed women. In 1966, on the contrary, unemployed women still had a higher average number of daily family meals on working days. Again, these findings were corroborated in the regression analyses. Although all the analyses confirm that employed women spent less time on family meals than unemployed women, the decline in average time spent on family meals between 1966 and 1999 was not due to the rise in female employment. Working men also reported a significant decline in time spent on family meals and in the average number of family meals, both on working days and Sundays. Employed men have almost halved their number of family meals on working days. As such, the evolution in the number of family meals is most alarming for this category. The regression analysis for time spent on family meals allowed us to explain much better for working men (R2 = 19.0 percent) than for women (R2 = 8.0 percent). The same held true for the number of family meals. Moreover, in contrast to what we found for women, men's time on and the number of family meals was highly sensitive to social class. Indeed, craftsmen, managers, professionals, and white-collar workers spent significantly more time on family meals and had a higher number of family meals than unskilled workers. Moreover, working men with a degree of higher or middle education also spent significantly more time on family meals and had significantly more family meals. Obviously, the standard image of the bourgeois family gathered around the table still finds wide application,28 at least for working men. For the higher social classes, dinnertime was the outstanding moment for gathering as a family.29 Moreover, the number of children also had a significant effect on working men's family meals. Having more than one child significantly contributed to more time spent on family meals as well as to a higher number of family meals. For working women, on the contrary, the decline in family meals and family meal time could be attributed neither to the decline in the number of children, nor to the rise in female employment. Nevertheless, we may conclude that the average decline in family meals between 1966 and 1999 was not due to the differential composition of the population in terms of occupational class or education level. On the contrary, we found that for men more time was spent on family meals within the social classes that grew during the last third of a century. Obviously, behavioral changes and structural changes, other than the ones we could allow for here, account for the changes in family meals between 1966 and 1999. On Saturdays and Sundays, we notice a slight though insignificant increase in the average number of family meals for women in dual-earner families. Unemployed women in breadwinner families, on the contrary, show a significant decline in the number of family meals. As for the duration of family meals, the decline in the average number of family meals is not significant on Sundays, except for unemployed men. SUMMARY AND DISCUSSION In this article we have tried to assess how family time evolved in Belgium between 1966 and 1999. We used Belgian time-budget data from respondents who lived as a couple with at least one child. We retained the time-budget data from only one respon- Table 6 Number of Family Meals on Working Days, Saturdays, and Sundays by Individual/Family Occupational Category and Sex Working Day 1966 1999 Woman in dual-earner family 1.44 (n = 96) 0.92 (n = 506)*** Woman in breadwinner family 1.50 (n = 261) 0.87 (n= 152)*** Employed man 1.61 (n = 384) 0.84 (n = 749)*** *p < 05. **p < 01. ***p < .001. Number of family meals Saturday Sunday 1966 1999 1966 1999 1.26 (n = 19) 1.56 (n = 240) (ns) 1.64 (n= 10) 1.72 (n = 265) (ns) 2.23 (n= 40) 1.52 (n = 76)* 1.86 (n = 26) 1.68 (n = 76) (ns) 1.74 (n= 54) 1.55 (n = 364) (ns) 2.53 (n = 68) 1.67 (n = 386)** 320 JOURNAL OF FAMILY HISTORY / July 2005 dent per family. As such, our definition of family time was based on the time spent on each activity with other family members present, as indicated in the individual diaries. We found that time spent with both partner and children declined seriously and significantly on an average working day. However, the situation was not as alarming as appeared at first sight. There was an overtaking maneuver on Saturdays, with a significant increase in family time. Rhythm diagrams also corroborated this finding. In 1966 Sundays were typical for family time, while in 1999 Saturdays had to catch up for lost family time on working days. Next to the overtaking maneuver on the weekend, it also appeared that parents spent more time on activities that involved active interaction as a family. The share of child care in the total family-time budget rose significantly on every weekday, especially on the weekend, implying that more of the child-care activities are taken up by both parents together. In contrast to 1966, in 1999 child care was in the top five of activities done as a complete family on every weekday. Moreover, the nature of time with only children present changed substantially. Passive presence during parents' household activities turned into "active" interaction between the parent and the children, during playing and child-care activities. Housewives, becoming more rare today, could previously manage household and child-care activities, being a "time buffer" against unforeseen circumstances.30 Therefore, children were more likely to be in the presence of their mothers on working days. Our data confirm that the total time in presence of children (without the partner) has not changed, despite the rise of dual eamership. This suggests that after a hard day of work, time is still freed up for children and especially for "quality time" with children. Moreover, the decline in family time between 1966 and 1999 for women cannot be attributed to the rise in female employment. Women in breadwinner families still spend a little more time in the presence of partner and children than working women in dual-earner families, but the decline in family time was much more alarming for housewives than for working women. Obviously, the declining number of housewives in the Belgian population is not the reason for an average decline in family time. In addition, we have tried to assess the role of the family meal within family time. Eating is still the most important activity done as a family. Nevertheless, its share in the total family-time budget has declined significantly during the past thirty years. We also diagnosed that in 1999 family time is less characterized by peak moments around meal times than in 1966. Traditional meal times have lost importance. In general, meals are taken later, but peaks are never attaining the same levels they did in 1966. Moreover, the number of commensal occasions with partner and children have also declined since 1966, resulting in less than one family meal a day on average on working days. No daily family meals was a much more widespread practice in 1999, both on working days and weekend days. In 1999, one daily family meal had become the general practice on working days, while two daily family meals were most common on Saturdays and Sundays. In 1966, the ideal of three daily meals was still the most common practice on weekends. Most surprisingly, the decline in the average number of family meals and the average duration of the family meal was again more pronounced for unemployed women than for their working counterparts. As a result, with regard to family meals housewives lost the advantageous position they held in 1966. The evolution of family meals for employed men is even more alarming. As with family time in general, the decline in family meals for women could not be attributed to the rise in female employment. For men, we found that social class still had a significant effect on the number of family meals and the duration of family meals. Working men with a Mestdag, Vandeweyer / ROLE OF THE FAMILY MEAL 321 higher education level or higher occupational class spent more time on family meals and have more family meals than their counterparts with a low education or unskilled laborers. Nevertheless, the mutual differences between the datasets from both years prevented us from going deeper into the factors that could explain for the decline in family meals and family time between 1966 and 1999. For example, we could not assess the impact of children's age or the occupational status of both parents, factors that made a serious difference in family time, as shown in previous studies. Moreover, these findings need to be situated within the larger frame of societal changes. In 2003, almost one in eight households were single-parent families. Traditional households, that is, a couple living with unmarried children, still outnumber single-parent families, but their share within the total number of households is rapidly declining. One-parent families probably have their own dynamics as for family time and family meals, which would certainly complete the general picture of time spent with family members. NOTES 1. Marjorie De Vault, Feeding the Family. The Social Organization of Caring as Gendered Work (Chicago: The University of Chicago Press, 1994), 13-15. 2. John Gillis, A World of Their Own Making: Myth, Ritual, and the Quest for Family Values (Cambridge: Harvard University Press, 1997), 122. 3. Gillis, A World of Their Own Making, 81, 87. 4. J. K. Daly, "Deconstructing Family Time: From Ideology to Lived Experience," Journal of Marriage and the Family, 63, no. 2 (2001): 283-94. 5. Thorne (1982) in De Vault, Feeding the Family, 15. 6. Daly, "Deconstructing Family Time," 283-85. 7. Gillis, A World of Their Own Making, 88. 8. DeVault, Feeding the Family, 38^-0. 9. John Gillis, "Making Time for Family: The Invention of Family Time(s) and the Reinvention of Family History," Journal of Family History 21, no. 1 (1996): 4. 10. Jean-Pierre Poulain, "L'espace social alimentaire," Cahiers de nutrition et de diététique 34, no. 5 (1999): 273-74. 11. Gillis, "Making Time for Family," 5. 12. Nelly Kalfs, Hour by Hour: Effects of the Lata Collection Mode in Time Use Research (Amsterdam: Nederlands Instituut voor Maatschappij- en Marktonderzoek, 1993). 13. Claude Javeau, Les vingt-quatre heures du Beige Bruxelles, Editions de l'Institut de Sociologie, 1970, 17. 14. The data from 1966 were collected by Philip J. Stone and are available through the archive of the Henry A. Murray Research Center of the Radcliffe Institute for Advanced Study, Harvard University, Cambridge, Massachusetts (Producer and Distributor). This data set was made accessible in 1983 [computer data], 15. In an earlier version of this article, presented at the 26th International Association for Time Use Research (IATUR) conference in Rome, Italy, October 27-29 2004, we performed our analyses using the full 1999 data set, thus including more than one respondent per household. We compared the results based on the full data set with the ones presented here, originating from a random sample containing only one respondent per household. It appeared that the results were only slightly different in some cases, with average durations and numbers and significance levels being somewhat higher or lower. The conclusions with regard to the comparison to the 1966 data were identical to the ones we drew based on the first analyses. Nevertheless, we decided to present the results from the random sample here, as this was more correct, method- 322 JOURNAL OF FAMILY HISTORY / July 2005 ologically speaking. Moreover, the use of the random sample from the 1999 database also allowed us to overcome the problem of significance estimation, as the sizes of the 1966 and 1999 databases were more comparable. 16. For America: John Robinson and Geoffrey Godbey, Time for Life: the Surprising Ways Americans Use Their Time, (Philadelphia: Pennsylvania State University, 1997). For the Netherlands: Andries van den Broek, Koen Breedveld, et al., "Het tijdsbudget. Het tijdsbeslag van verplichtingen, herstel en verplaatsingen," in Trends in de tijd. Een schets van recente ontwikkelingen in tijdsbesteding en tijdsordening, ed. A. van den Broek, K. Breedveld, et al. (Den Haag, Nederland: Sociaal Cultureel Planbureau, 2001), 9-28. For the United Kingdom: Jonathan Gershuny, Changing Times: Work and Leisure in Post-industrial Society (New York: Oxford University Press, 2000). 17. Ignáce Glorieux and Joeri Minnen, "Kinderen maken het verschil. Over de invloed van kinderen op het tijdsbestedingspatroon van ouders," Gezinsbeleid in Vlaanderen 33, no. 1 (2004): 2-10. 18. Tanj a van der Lippe, Is er werkelijk een toename in de tijd besteed aan kinderen ? (Paper presented at the Elf de sociaal-wetenschappelijke studiedagen, Amsterdam, April 22 and 23, 2004). 19. Time spent with children can be measured from an activity-based approach or a context approach. The former measures the time spent on child care and other activities for which children's presence is assumed. The latter measures the time spent on any activity for which children's presence was reported. 20. However, there is no guarantee that an activity with partner and children is an activity in which all children of the household are involved. For families with only one child, we can be sure that the complete family is together; for families with more than one child, there is no such guarantee. 21. In 1966, 28 percent of the Belgian population between nineteen and sixty-five years old did not possess a television set. In 1999 only slightly more than 1 percent of the Flemish population between sixteen and seventy-five years old did not possess a television set. More than 66 percent possessed only one television set. A quarter of the Flemish respondents had two television sets in the house and more than 6 percent even had three or more sets at its disposal. 22. For France: Yvonne Serville and Claire Trémoliěres, "Recherches sur le symbolisme et la signification du repas familial," Cahiers de nutrition et de diététiquel, no.l (1967): 48-59. For the United Kingdom: Michael Nicod, "Gastronomically Speaking," in Nutrition and Lifestyles, ed. M. Turner (London: Applied Science Publishers), 56-57; Anne Murcott, "Cooking and the Cooked: A Note on the Domestic Preparation of Meals," in Sociology of Food and Eating, ed. A. Murcott (Aldershot, Gower, 1983), 178-179; Nickie Charles and Marion Kerr, Women, Food and Families (Manchester: Manchester University Press), 172-73. 23. For Vienna, Austria: Robert Rotenberg, "The Impact of Industrialisation on Meal Patterns in Vienna, Austria," Ecology of Food and Nutrition 11: 25-35; For France: Jean-Louis Flandrin, "Mealtimes in France Before the Nineteenth Century," FoodandFoodways 6, nos. 3^1 (1996): 261-82; Claude Grignon, "Rule, Fashion, Work: The Social Genesis of the Contemporary French Pattern of Meals," Food and Foodways 6, nos. 3^1 (1996): 205^-1. 24. Our data confirmed that the average number of daily meals for the Belgian population between nineteen and sixty-five years old has diminished significantly between 1966 and 1999, no longer attaining the general standard of three daily meals. Nevertheless, the three-meal pattern is still considered a standard. For Belgium: Nest Mertens, "Het ontbijt, onze "jongste" maaltijd," Mededelingsblad Academie voor Streekgebonden Gastronomie. Periodiek voor voedingsgeschiedenis, streekgastronomie en toerisme 15, no. 4/60: 37^-5; For the Netherlands: Jozien Jobse-van Putten, "Onderscheid moet er zijn. Culturele heterogeniteit in de Nederlandse voeding," Volkskundig Bulletin. Tijdschrift voor Nederlandse Cultuurwetenschap 21, no.l (1995): 69. Mestdag, Vandeweyer / ROLE OF THE FAMILY MEAL 323 25. Felipe Fernandez- Armesto, "Meals Make Us Human. Never Mind Obesity, It's the Loneliness of the Fast-food Eater That Matters," The Guardian, September 20, 2002. 26. Claude Grignon, "Commensality and Social Morphology: An Essay of Typology," in Food, Drink and Identity. Cooking, Eating and Drinking inEurope Since the Middle Ages, ed. P. Scholliers (Oxford, Berg, 2001), 25. 27. Oriel Sullivan and Jonathan Gershuny, "Cross-national Changes in Time-use: Some Sociological (Hi)stories Re-examined," British Journal of Sociology 52, no. 2 (2001): 335. 28. Annick Sjögren-de Beauchaine, The Bourgeoisie in the Dining-room: Meal Ritual and Cultural Process in Parisian Families of Today (Stockholm, Insitutet for foMivforskning, 1988). 29. Nicolas Herpin, "Le repas comme institution. Compte rendu ďune enquéte explora-toire," Revue francaise de sociologie, 14 (1988): 503-21. 30. Mark Elchardus, De gemobiliseerde samenleving. Tussen de oude en de nieuwe ordening van de tijd (Brüssel: Koning Boudewijnstichting, 1996).