1In her monograph on the celebration of marriage, Y. Delsaut (1976) describes the situation of the young newlyweds. She focuses on the ambiguous attitude of the groom, torn between his desire both to maintain close relations with his own family and to make a good impression on his wife and family-in-law. The young man exhibits a complex range of feelings: he wants to be integrated in both groups while clearly indicating that he belongs exclusively to neither. How should this dual family commitment be managed? Herein lies the challenge facing young couples. While the egalitarian concept of the couple in our society calls for equality of treatment for both sets of kin, achieving such equality is not always simple. By describing the support provided by the couple to the two kin groups, we will seek to understand the structure and the limits of this precarious equilibrium.
2This study is based on a statistical analysis of the kinship networks and support section (Réseaux de parenté et entraide, RPE) of the permanent survey on living conditions of households (Enquête permanente sur les conditions de vie des ménages, PCV) conducted by INSEE, the French statistical office, in October 1997. It covers almost 6,000 households (of which 3,380 comprise a couple with or without other household members) whose reference person, aged at least 15, lives in metropolitan France. This is the first survey in France to cover support provided to affines. [1] Hence, while traditional analyses of family support are limited to services exchanged between blood relatives, this survey enables us to compare the support provided to both kin groups.
3Before presenting our results, we will briefly describe the challenges of using a statistical approach to analyse support offered by couples to their respective families, and then outline our hypotheses regarding preferential support for the woman’s family over that of the man.
I – Family support: strengths and weaknesses of a statistical approach
4Following early work on the role of exchange in society (Malinowski, 1922 ; Mauss, 1925), the first landmark studies on the question of family support in modern western societies were published primarily by British and American anthropologists, notably Young and Willmott (1957), Bott (1957) and Firth (1956). In France, [2] the sociological study of support provided and received by families began later, with research that moved beyond Parsons’ predictions of an increasingly isolated nuclear family (Parsons and Bales, 1955). The surveys by A. Pitrou (1976), L. Roussel and O. Bourguignon (1976) and C. Gokalp (1978) show, on the contrary, that family support reflects the persistence of ties between the conjugal unit and other relatives, despite the individualization of family relations. These groundbreaking studies initiated a new dynamic in the sociology of the family (Segalen, 1991).
5But the debate on the functions of family support did not really take off – in terms of research projects at least – until the early 1990s, with the rediscovery of poverty in Europe and the crisis of the welfare state. This gave a more political colouring to scientific discussion (Martin, 1996), and the temptation was strong to see kinship networks as a new form of social solidarity (Déchaux, 1996). In the quest to resolve the crisis of the welfare state, the use of the term family solidarity reveals a shifting of the issue from the public to the private sphere (Beck, 1992). Yet research results offer little to confirm the idyllic vision of family support. The lower socioeconomic categories are those where exchange is least developed (Déchaux, 1994; Déchaux and Herpin, 2004a), and a combination of public and family support is needed to protect the most disadvantaged (Paugam, 2007).
6Alongside the inequalities by social group in support exchanged, research on family support highlights the structuring nature of this type of exchange. Using data from the "kinship networks and support" section of the PCV survey [3] of October 1997, we will discuss the strengths and weaknesses of a quantitative approach to family support with respect to the three objectives of research on this theme, namely, to analyse the plasticity of family support, to determine the functions and purpose of support and, through the prism of family support, to identify practices and norms of behaviour within the family.
1 – Analyse the plasticity of family support
7The services rendered are many and varied. From car journeys provided by a son to his mother who no longer drives, to home-cooked dishes taken away by a student after a week-end with the family, the range of services is immense. [4] In a statistical survey, listing and recording these services in all their forms is an especially difficult task. Unless open questions are used, collecting information on a limited number of services rendered and/or received does not cover the full scope of opportunities provided by the family network. The RPE survey identifies ten types of support that can be provided by individuals to their family or family-in-law. [5] We grouped them into four categories by type of resource mobilized. The first group includes support of a "manual" nature: DIY, housework, childcare and shopping. The second group is "material": car loans, cash loans and cash gifts. The third involves "cultural skills" and covers help with schoolwork and with general paperwork. The fourth and last group covers "moral support", though the survey description does not detail the actual nature of this support.
8Services are wide-ranging because they are adapted to individual needs. So how, for example, can we compare childcare during school holidays with childcare after school hours? Which is most important? The same problem arises when two different types of support are compared. Who helps more, the person who gives a large lump sum to pay for a car, or the gardener who regularly gives away homegrown produce? In other words, measuring the intensity of family support is a difficult task, so no hierarchy of services rendered is given in our data analysis.
9For researchers, the third difficulty arising from the plasticity of family support is that of measuring frequency. First, we must rely on the respondent’s memory. Yet, the greater the accuracy required, the more difficult it may be to give a precise answer. Moreover, certain services are not necessarily provided on a regular basis. [6] One way around this problem is to distinguish between "mobilized" support (Paugam and Zoyem, 1997), i.e. actually provided, and "mobilizable" support, i.e. that a respondent knows he/she can count upon. However, this solution is only partial and, by omitting the quantification of services rendered, we are denied an important dimension in the understanding of family support, namely the volume of services exchanged. The time scale chosen in the RPE survey is that of the most recent twelve months. The questions are formulated as follows: "have you provided a service of this type (e.g. DIY) to this type of relative (e.g. father) over the last twelve months?". The answer can be: "very often", "often", "sometimes", "never", "don’t know" or "not applicable". With these approximate answers, it is impossible to measure the volume of each type of support provided, so we chose to use only two modalities: did or did not provide support. Our figures express the fact that an individual gave a particular type of support over the preceding year, with no indication of frequency.
2 – Determine the functions and purpose of family support
10Identifying the functions and purpose of family support involves considering the preferences (Herpin and Verger, 1996) and needs of those who give it and those who receive it. For the beneficiary, the purpose of support is two-fold. On the one hand, services rendered in the form of "subsistence" support (Pitrou, 1976) or "compensatory" support, (Paugam and Zoyem, 1997) aim primarily to protect the beneficiary from economic insecurity. The purpose of "promotional" (Pitrou, 1976) or "emancipatory" support (Paugam and Zoyem, 1997), on the other hand, is to give the recipient more time to focus on other activities and hence to integrate society more effectively. [7] For the provider, the motives for offering support may be relational. Giving help maintains ties, but also creates a debt which must be repaid, either by support of a similar nature or by implicit gratitude (an "emotional debt"). In this sense, it is appropriate to speak of "mutual" support, since the emphasis is on reciprocity. It is a great pity that the RPE focuses exclusively on support provided without any mention of support received. But even if we had information on both, could we truly measure reciprocity? Prouteau and Wolff (2003) believe not, notably because reciprocity is often deferred. In the absence of longitudinal data and with no knowledge of how individuals view their actions, any attempt to measure reciprocity would indeed be risky.
3 – Through the prism of family support, identify practices and norms of behaviour within the family
11Family support adds a new dimension to the analysis of intergenerational relationships within the kin group. By naming a chapter of her book "the dual transmission circuit", C. Attias-Donfut (1995) suggests, for example, that family support is by no means a one-way process, even if the "pivot" [8] generation is the most strongly solicited. In a quantitative perspective, longitudinal survey data are needed to distinguish age effects from generation effects. As the RPE survey is cross-sectional, our analysis will not focus on relationships between generations. While family support sheds light on vertical relationships, it also operates horizontally, and support between siblings has not been forgotten (Coenen-Huther, Kellerhals and von Allmen, 1994). Among immediate blood relatives, it is between brothers and sisters that the ties of family support are weakest (Déchaux and Herpin, 2004b). Last, studies of solidarity have also investigated new family structures, and notably the effect of recomposition on family support (Cadolle, 2003). An implicit hierarchy of support relationships – and hence more globally of family bonds – thus emerges here.
II – Measuring the balance between kin groups: methods and hypotheses
12Although analysis of family support sheds new light on contemporary kin relations (Debordeaux and Strobel, 2002), the relationships between individuals and their spouse’s family remain largely unexplored. This "omission" is hardly suprising, since surveys tend to reflect definitions of the family prevailing at the time. [9] To our knowledge, the RPE survey is the only national statistical survey that deals with kinship by alliance. However, it does not distinguish, for example, between married children on the one hand and sons- and daughters-in-law on the other. We know nothing about brothers’ and sisters’ partners, so certain types of kin relations cannot be examined. The specific identity of kinship network members is nonetheless a central issue. For this reason, we set strict boundaries for the definition of blood relatives and affines. Blood relatives include the mother, father, maternal and paternal grandparents, brothers and sisters, mother’s brothers and sisters, father’s brothers and sisters, nephews, nieces and cousins of the respondent. Affines include the mother, father, maternal and paternal grandparents, brothers and sisters, mother’s brothers and sisters, father’s brothers and sisters, nephews, nieces and cousins of the respondent’s spouse. In addition, the relatives included in our study all live separately from the respondent.
13The purpose of this article is to identify any differences in the way individuals in a union treat blood relatives and affines. For this purpose, we first sought to determine whether French people offer the same services to their own family and to their spouse’s family. For each type of help, we compare the percentages of persons having helped at least one member of their own family and one member of their spouse’s family, respectively, over the last twelve months. We then looked at the conditions in which one kin group might take precedence over the other. As we cannot reason in quantitative terms, the degree of lateral orientation [10] of kin support is measured by the difference in the number of types of support an individual provides to his/her own family and to his/her spouse’s family. We thus obtain four groups: those who help neither their own or their spouse’s family, those who support both kin groups equally, those who favour their own family and those who favour their spouse’s family.
14Before comparing the support provided to each kin group, we controlled for the relative distance of parents and parents-in-law’s homes from the respondent’s home. Table 1 shows that the proportions of men and women living closer to their parents than to their parents-in-law are equivalent to the proportions of men and women in the opposite situation. There is no clear evidence of matrilocality. [11]
Relative geographical distance of parents and parents-in-law (%)
Relative geographical distance of parents and parents-in-law (%)
15If both families are equidistant from the respondent, then neither should receive preference, an observation explained notably by the absence of differentiation in our filiation system. [12] This, in any case, is the hypothesis we wish to verify.
III – Results
1 – Types of support and preference for own family by sex
16Our first observation is a strong preference for own family (Table 2), with 62.1% of respondents having helped at least one blood relative over the last twelve months, versus just 43.7% having helped a relative by alliance over the same period. This preference is strong for respondents of both sexes and of all occupational categories.
Support provided to own and spouse’s family over the previous twelve months by sex and occupational category (%)
Support provided to own and spouse’s family over the previous twelve months by sex and occupational category (%)
17Beyond this family preference, men and women provide support in the areas traditionally associated with each sex. Women generally deal with tasks such as shopping, housework and childcare: over the preceding twelve months, 40.5% of female respondents had been shopping for at least one member of their own family, and 23.9% for at least one member of their spouse’s family, versus 30.9% and 21.2% of men respectively. Be it for blood relatives (Fougeyrolles-Schwebel, 1994) or for affines, women more often perform domestic tasks than men, while men are over-represented in DIY, a traditionally masculine domain: 30.7% had done DIY jobs for their own family and 25.1% for their spouse’s family, versus 15.2% and 8.9% for women. We also note that moral support is a female speciality. There is little gender distinction, however, in help to either kin group with school work (outside the home) or administrative paperwork, or in financial gifts and loans or car lending. At aggregate level, the same gender differences are observed in support to blood relatives and to affines.
18The individual’s social position – captured here via the six INSEE socio-occupational categories [13] – is also a strong factor of differentiation in the types of support provided to own and spouse’s family. Persons in higher-level occupations make particular use of their cultural capital: over the twelve previous months, 34.2% of men in this category had helped a family member with administrative paperwork, and 19.6% had helped a member of his spouse’s family. Likewise, 7.5% of women in higher-level occupations had provided help with schoolwork to at least one blood relative outside the home, and 3.8% to an affine. Clerical and manual workers, for their part, participate more in practical tasks: 47.6% of male manual workers had done DIY jobs for a family member (36.8% for a member of the spouse’s family), and 20.4% of female manual workers had cared for the children of at least one of their blood relatives (12.2% for an affine). The help offered by individuals to members of their own and their spouse’s family thus corresponds to their capacities and their personal capital.
19Over and above persistent social and gender divides, we observe a clear difference in levels of investment between blood relatives and affines. While both men and women help their own family more than their spouse’s family, women provide much more support than men to their own family, while men offer more help than women to their spouse’s family. [14] This is notably the cases for support requiring material and manual resources (Table 2). For example, 44.5% of men provided manual assistance to their own family, versus 50.8% of women (–6 points for men), while 37.4% offered offered the same type of assistance to the spouse’s family, versus only 31.6% of women (+6 points). Moreover, the male-female gap in support to the spouse’s family is larger among lower socioeconomic groups (13 points among manual workers versus 3 points among higher-level occupations), a fact which suggests, at aggregate level, a stronger propensity among couples to favour the female line, notably in these social categories. This result corroborates the findings of recent qualititive surveys (Le Pape, 2006).
2 – Sex and inequality of treatment between kin groups
20To verify inequality of treatment between blood relatives and affines, individuals can be divided into three groups: individuals who provide at least one type of additional support to their own family, those who provide at least one type of additional support to their spouse’s family, and those who help both kin groups equally. [15] First, we see that men and women behave very differently (Table 3). Fewer men than women offer more types of help to their own family (the model coefficient is –0.40), though men are more present in the group of persons favouring the spouse’s family (+0.30) and in the group which favours neither kin group (+0.37).
Influence of gender on orientation of support to relatives by kin group (model coefficients)
Influence of gender on orientation of support to relatives by kin group (model coefficients)
21Age, for its part, only has an effect among men. Older men are more likely to favour their spouse’s family (Table 4). [16] Men aged 50-59, for example, are more likely than men aged 40-49 to belong to the second group (the model coefficient is +0.54), though we cannot know whether this is an age effect (men are progressively "drawn" or integrated into their family-in-law) or a generation effect. For women, no coefficient relating to age is significant in our regression.
Determinants of the orientation of support to relatives by kin group (model coefficients)
Determinants of the orientation of support to relatives by kin group (model coefficients)
22All other things being equal, inactive men and female manual workers have a higher propensity to offer at least one type of additional service to their spouse’s family. Compared with female clerical workers, female manual workers are also less inclined to give priority to blood relatives. This finding may simply reflect their more limited capacity to offer diversified forms of support, with help provided mainly in the form of domestic tasks.
23Family structures, on the other hand, have a much clearer effect among women than among men. When a women is the eldest sibling in her family, she is much less likely (coefficient of –0.33) to favour her spouse’s family in terms of diversity of support, and gives preference to her blood relatives (+0.18). Conversely, position in the sibship has no significant influence among men. One hypothesis for explaining this difference is that men have a weaker sense of obligation towards their family. Moreover, when their mother lives in closer proximity, women treat their affines more unfavourably than men do.
24Last, since we are reasoning in terms of the diversity of forms of support provided and not their quantity, it is not surprising to observe that the size [17] of the two kin groups has practically no effect on the lateralization of support. On the other hand, seeing a members of one’s family at least once a week, for example, substantially increases the likelihood of providing different forms of support to both blood relatives and affines, even more so for women (+1.18) than for men (+1.09). Socialibility and exchange are very intimately linked in the system of family solidarity (Jonas, 2006). All in all, these figures show that in relation to the diversity of support provided, the woman’s blood relatives are more likely to be favoured than those of her partner. On the basis of our selected indicator, the hypothesis of equality of treatment is thus invalidated.
Conclusion
25As early as the 1950s, British anthropologists (Firth, 1956 ; Bott, 1957) began to question whether the western filiation system was truly undifferentiated. They focused on the strength of the mother-daughter relationship to explain why couples have more frequent contacts with the woman’s family than with the man’s. This preference for the maternal line is observed in the analysis of help exchanged, such as care of grandchildren, often a source of unvoiced competition between the grandparents on both sides, with the maternal side most frequently taking precedence (Le Pape, 2006). These qualitative studies generally agree upon a tendency towards matrilaterality [18]. Our aim here was to obtain statistical confirmation of this lateral orientation of family support using the kinship networks and support survey, the first survey to provide data on the practices of individuals with respect to their own family and that of their spouse.
26This study shows clearly that men and women have very different attitudes to their affines. Men more frequently provide more types of support to their spouse’s relatives than do women, and they are more likely than women to favour their spouse’s family over their own family, especially after age 50. These behavioural differences suggest that women have a more exclusive, if not stronger, attachment to their blood relatives than men. They more often maintain a distance from their spouse’s kinship exchange system, preferring to reserve their availability (and perhaps their loyalty) for their own kin group. In the couple, men tend to follow their spouse’s lead in kin relations, focusing attention on relatives by alliance rather than on their own family. It is this dual phenomenon which underlies the tendency towards lateral orientation of kin support.
Notes
-
[*]
Observatoire sociologique du changement, CNRS-FNSP, Paris – Université Paris X Nanterre.
-
[**]
Observatoire sociologique du changement, CNRS-FNSP, Paris – Université Lumière Lyon II.
Translated by Catriona Dutreuilh. -
[1]
Affinity by marriage signifies kinship ties created through marriage. Affines are members of the family-in-law.
-
[2]
Our study is limited to the French context. Readers interested in a more international perspective should refer to the report by Jean-Hugues Déchaux (2001), and a study by Catherine Bonvalet and Jim Ogg on the various aspects of European surveys on family solidarity (2006).
-
[3]
For simplicity, we will call it the kinship networks and support survey or RPE survey.
-
[4]
The most widely studied services are those in the domestic sphere, childcare for example (Daniel, 2004). Family support in finding a job or a place to live are also widely studied (Segalen, 1980; Bonvalet, 1991). Cash loans or gifts have likewise been examined in depth. Though money is never the term of the exchange (in this sense, family support is characterized by the absence of any cash payment), it may be lent or given. The main beneficiaries of these cash transfers are young people, especially those in high-income social groups and students (de Barry, Eneau and Hourriez, 1996).
-
[5]
The forms of support included in the RPE survey are probably those most easily identifiable by the individuals concerned.
-
[6]
Because she is busy at work for several months, a mother may rely very heavily on the family for childcare over this period, then not at all after her workload has decreased.
-
[7]
However, as pointed out by Jean-Hugues Déchaux (1996), the boundary between these two purposes is never clearcut, since integration is a form of protection and protection facilitates integration.
-
[8]
In the "3 generations" survey (1995), the pivot generation comprises individuals aged 49-53 who must support both ageing parents (Weber, Gojard and Gramain, 2003; Le Borgne-Uguen and Pennec, 2006), and children who are starting out in adult life (Chambaz and Herpin, 2005; Strobel, 1997).
-
[9]
The relative lack of interest shown by sociologists for the spouse’s family can be explained by the instability of unions, and the increasingly ephemeral nature of ties with affines.
-
[10]
Lateral orientation of kin support indicates whether one or other family line receives preferential treatment. It may indicate preference for blood relatives, for affines or no preference for either.
-
[11]
Matrilocality signifies living near to the woman’s family than to the man’s.
-
[12]
i. e. neither family line has any legitimate justification for taking precedence.
-
[13]
The categories are: farmers, higher level occupations, intermediate occupations, clerical and sales workers, manual workers and economically inactive.
-
[14]
By a composition effect, we can assume that women’s stronger investment in their family (compared with men) and men’s stronger investment in their spouse’s family (compared with women) leads to a tendency among couples to favour the female line. However, like many statistical surveys on exchanges between kin, the RPE survey does not consider both members of the couple, but only one partner selected at random. Consquently, this tendency among couples to favour the woman’s family cannot be fully verified.
-
[15]
As this study focuses exclusively on comparing equality of treatment between blood relatives and affines, we excluded the group of individuals who provide no support to either kin group.
-
[16]
The logistic regression models were estimated using the standard method under SPSS (all variables are entered simultaneously in the model).
-
[17]
Kin group size is determined by counting the number of parents, maternal and paternal uncles and aunts, brothers and sisters, nephews and nieces, and maternal and paternal grandparents of the individual and his/her partner who are still alive and who live in a different home from that of the couple to which the individual belongs.
-
[18]
Matrilaterality should not be confused with matrilinearity. Matrilinearity is used in kinship terminology to designate a particular form of exclusive filiation. Matrilaterality is a sociological term signifying the tendency of couples to favour the woman’s line in a system of undifferentiated filiation.
