Consensus , conflict , and social citizenship Consensus and conflict in lay conceptions of citizenship : Why people reject or support maternity policies in Switzerland

A survey with a representative Swiss urban sample (N = 769) was carried out to study determinants of attitudes towards policy options in the context of women's rights. Two basic lay conceptions of citizenship are hypothesised to guide and justify perceived legitimacy of social rights. A consensus-based conception of citizenship is grounded on membership in a group that expects endorsement of common values. Group norms distinguish acceptable from unacceptable behaviour, and principles of deservingness define the scope of rights. Social rights acquire their legitimacy as devices to enforce common values. In contrast, a conflict-based conception of citizenship is grounded on a structural and hierarchical perception of the group. Social rights are seen as devices to challenge existing power relationships, and their justification lies in social change. Results show how attitudes towards maternity policies are related to social order, common group values, structural gender inequality and social change. Positioning on consensus-based citizenship predicted attitudes towards a restrictive (means-tested) maternity policy, whereas positioning towards conflict-based citizenship predicted attitudes towards unconditional and extensive policies (e.g., day nurseries and maternity leave). It is concluded that favourable attitudes towards maintenance or expansion of social rights in general, and of women's rights in particular, largely depend on perceptions of groups not in terms of their essential differences, but in terms of social inequalities.

Extended citizenship is typically the outcome of right claims advanced by groups or social movements (Hopkins, Reicher & Kahani-Hopkins, 2003;Isin & Wood, 1999). These collective demands are often the first step in a long process that ultimately should lead to fairer treatment for disadvantaged groups and individuals. For citizenship broadening policies are never granted in an unproblematic way.
They necessarily involve political debate, negotiation and collective mobilisation (Anthias & Yuval-Davis, 1992). Concrete success of claims depends on whether and how they are institutionally implemented at the policy level, and is to a large extent contingent upon their reception and perceived legitimacy in the public sphere, especially in contexts where new policies are subjected to popular vote, as is the case in Switzerland.
In this paper, we will analyse common sense justifications of social rights. Two lay conceptions of citizenship are hypothesised to shape perceived legitimacy of right claims, the first one based on value consensus in a community, and the second on perceived conflict between social categories. We will then demonstrate how these two modes of thinking predict perceived legitimacy of claims put forward by women, namely the right to lives less constrained by family duties and to enhanced professional integration. In the Swiss context, two policies that represent institutional responses to these claims are currently discussed in the public sphere, namely a collectively organised nation-wide maternity insurance and the development of day nurseries 1 .
Gender claims and social rights 1 It would have been interesting to integrate policy proposals involving fathers in domestic work and childcare, but they are virtually inexistent in Switzerland, and hardly object of political debate.
One major aim of social rights is to protect individuals against the economic consequences of ordinary and unpredictable events occurring during the life course, such as illness, accident, unemployment or old age. Furthermore, through correction and compensation of structural inequalities, social rights guarantee equal access to basic institutions in a modern society, for example to health care or education. In both cases, social rights are grounded in the idea that underprivileged positions in society stem from processes over which individuals have only limited influence and control (Roche, 2002). Implemented with various social policies, these rights are based on the principle of public (collective) responsibility, since many ordinary events result in economic hardship that exceeds the individuals' capacity to cope with them (Twine, 1994).
However, at a theoretical as well as at a political level, the family (including domestic work, childcare and intimacy) is located within the competence of the private sphere, as opposed to a public sphere in which principles of collective responsibility and intervention are more easily justified. Feminist movement and research (see for example Pateman, 1983;Okin, 1989Okin, , 1991Phillips, 1991;Charles, 2000) has vastly criticised the private / public dichotomy, because in their view it personalises, renders invisible and pushes essential gender questions, e.g. the sexual division of labour, maternity, abortion, domestic violence, professional inequality, and economic dependency, into the background of political priorities.
The deconstruction of this dichotomy has an obvious consequence: it implies a redefinition of collective responsibilities and thereby justifies the development of new social rights that were not incorporated in earlier welfare state models (e.g. women's rights). Claims therefore aim at challenging common boundaries between the public and the private, for example through state-managed maternity policies that give the hitherto private activity of mothering a public character. In Switzerland, as in other Western countries, current forms of citizenship and the social rights that derive from it essentially depend on waged work, and much less on domestic work and childcare. Yet, even if classical citizenship theories admit that social rights should allow individuals to become relatively independent from the labour market (e.g. the decommodification concept in Esping-Andersen's writings, 1990), the idea that these rights should also grant women a certain independence from men -in terms of material and symbolic resources -is much less acknowledged, with the exception of feminist writings (Abbott & Wallace, 1996;Orloff, 1993;Walby, 1997). The disadvantaged position of women in the labour market indeed promotes and legitimises the nature of the "matrimonial contract" which binds together most of the heterosexual couples (Delphy, 1998): the domestic work, primarily carried out by women, is not paid, socially discredited, and requires a permanent availability to which their job is subordinated.
The first social policy discussed in this study that responds to women's claim of enhanced autonomy concerns the development of day nurseries. In the current Western context, where the sexual division of labour (Kergoat, 2000) still heavily structures modes of family organisation, day nurseries should relieve women of family duties, and thus promote their professional integration. Yet, one of the problems of the current day care system in Switzerland, other than its insufficient capacities and its restrictive schedules, is that it is too expensive. This may discourage certain families from using the services of day nurseries, for even if their share is calculated as a function of their salary, it can easily cover if not the whole then at least a substantial part of the woman's salary. In this situation, some women feel that it is more worthwhile to diminish or even suspend their professional activity in order to take care themselves of the children. As a consequence, publicly financed free-of-charge day nurseries could, at least partially, promote women's autonomy, within the family as well as in the professional sphere (Charles, 2000).
The second policy addressing women's concern for professional integration is a state-guaranteed maternity leave. The principle of this maternity insurance, integrated in the Swiss constitution since 1945, prescribes a period during which the mother receives a salary or another benefit. Moreover, she has the right to return to her job if she has had one before maternity. But this principle is still not implemented at a national level. The concretisation of the principle raises various issues on which citizens are divided, particularly the length of the maternity leave, the definition of mothers who are granted the benefits (e.g. all mothers or only those with a professional activity) and concrete financing methods. During the last decades, three projects aimed at making the principle of this insurance a reality have been submitted to popular vote in Switzerland; yet, after three rejections (the last one in 1999), it has become apparent that there must be a deep suspicion towards these policies. In this context, it is therefore particularly important to study the determinants of public attitudes towards maternity policies.

Attitudes towards social policies
Much like public opinion research in general (Feldman, 1988;Skitka & Mullen, 2002), research on attitudes towards social policies and institutionalised solidarity has focused on sociological determinants (e.g. age, sex, education level), on individual value orientations (e.g. egalitarianism, freedom, authoritarianism, post-materialism;Gelissen, 2000, Van Orschoot, 2002, or on affective responses towards outgroups (Sears & Funk, 1990), particularly prejudice towards undeserving beneficiary groups (Gilens, 1999). But sociological factors do not have a direct, unmediated effect on attitudes, value orientations are not self-evident and require further justification, and deep-seated automatic judgements of outgroups by themselves do not reveal how people reason about the legitimacy of policies. In order to address these concerns in the context of maternity policies, we suggest that accounts of policy opinions need to take into consideration the specificity of lay reasoning in policy judgments (Clémence, Egloff, Gardiol & Gobet, 1994) on the one hand, and dynamics derived from group membership on the other. It is our general conjecture that attitudes towards social rights chiefly depend on how people perceive the political community (usually the nation) they refer to in legitimacy judgements of policies, on the subjective meaning of this community for their self-understanding and identity, and on the expectations they have of other community members. Together, these elements give rise to lay conceptions of citizenship that provide the rationale of the legitimacy of social rights and institutionalised solidarity. Since the focus of this lay citizenship approach is put on perceptions of the community and its members, policy attitudes should not only depend on the perceived legitimacy of social rights as such, but also on their scope, that is, the groups and individuals who benefit from the right (Opotow, 1996). Our general hypothesis is that two prototypical lay conceptions of citizenship-consensus-based and conflict-based-capture opposing perceptions of the community and guide lay reasoning on maternity policies.

Consensus-based citizenship
Individuals necessarily live in a permanent and durable relationship with others, in a family, in a city, or in a country. Each of these communities is structured and organised, and commonly agreed upon rules and values regulate their social order. Such communities exert a normative pressure insofar as their members are expected to adhere, at least minimally, to common values in order to ensure the groups' relative cohesiveness (Hogg & Abrams, 1988). Thus, a consensus concerning community-defining values and principles is necessary to ensure group stability and legitimacy of its order (see Zelditch, 2001). The resulting group norms prescribe acceptable and desirable behaviours that contribute to the maintenance of its order. Group members need to endorse similar criteria for drawing this necessary line between desirable and undesirable behaviour (Joffe & Staerklé, 2003), that is, they need to share the same values that define the moral boundaries of the community (Opotow, 1990).
Yet, even though norms and values as such are aimed at ensuring harmony and stability of the group, the community's continuous need to demonstrate their legitimacy may easily lead to rejection of group members who deviate from norms. For some members follow norms more closely than others.
Members who strongly identify with the group, for example, are more likely to accept its norms than members who do not feel close to the group. High identifiers are therefore also more likely to engage in an ongoing definition of moral boundaries (Duckitt, 1989), by pointing out members whose behaviour threatens the stability of the group (Abrams, Marques, Bown, & Henson, 2000;Marques, Páez & Abrams, 1998). These boundaries draw a line between "good" (those respecting common values) and "bad" members of the group (those calling into question common values or by endorsing other ones).
Hence, a consensus-based conception of citizenship is grounded on the normative expectancy to endorse group values. This requirement entails an ongoing motivation to distinguish community members as a function of perceived conformity with group values (Hopkins et al., 2003). Importantly however, not all values are useful to differentiate between "good" and "bad" members of a group. Since the core of a consensus-based conception of citizenship is to differentiate individual group members within a group, individualistic values bearing on various aspects of self-control such as hard work, self-reliance, autonomy, and discipline (Joffe & Staerklé, 2003) should be most likely associated with it. Unlike non-individualistic values such as equality, solidarity or benevolence that may foster diversity supporting group identities (Sanchez-Mazas, Van Humskerken & Casini, 2003), individualistic values are associated with dominant ideas about the social worth of individuals (Beauvois, 1994), and are therefore socially useful devices to differentiate individuals.
In a consensus-based conception of citizenship, good treatment needs to be deserved (Crandall & Beasley, 2001;Lerner, 1980). Accordingly, social rights are not absolute, but conditionally granted privileges which depend upon the social worth of individuals. They are not based upon the recognition of socially determined and structural hardship that calls for collective solutions (Wright & Tropp, 2002), but on a hierarchy between deserving and undeserving persons (Wacquant, 1999). Deservingness, or merit, is the basis of legitimacy of social rights. Because a consensus-based conception of citizenship develops in communities that stress unity and conformity, spontaneous or "hot" solidarity, directed towards similar and deserving ingroup members, is likely to be the dominant form of aid (Clémence et al., 1994; see also Stuermer & Kampmeier, 2003, for a related discussion of community participation). In other terms, social policies based on deservingness are a form of social control, because they have the capacity to regulate individual behaviour as a function of group norms.
In the case of maternity policies, value conformity is contingent upon the gendered structure of family relationships. Motherhood as such does not lead to perceptions of value violation and subsequent undeservingness. However, because unconditional maternity policies bestow on all mothers a certain degree of autonomy towards men and call into question the principle according to which domestic work is excluded from the economic domain, these policies can be seen as threatening the patriarchal family order in which men are the main breadwinners and women care for other family members while ensuring an upright education of children. A consensus-based thinking should therefore go along with the feeling that unconditional maternity policies contribute to a weakening of the patriarchal social order.
Because ingroup social control is likely to occur when group norms are undermined (Marques, Abrams & Serodio, 2001), we expect that perceived threat to traditional gender roles is positively linked to support for selective and conditional policies, in which hardship needs to be individually demonstrated to authorities in order to qualify for institutional support ("means-tested policies"; Skitka & Tetlock, 1993).
Since absolute social rights do not exist in such a consensus-based reasoning, the very principle of a maternity insurance should be delegitimised by a patriarchal vision of society: women ought to be confined to the private sphere where they can bring up the children in order to avoid that they become future delinquents. The will to withhold maternity insurance can thus be seen as a sanction against women's unacceptable behaviour that threatens the social order and the stability of the community (Abbott & Wallace, 1996;Wacquant, 1999).

Conflict-based citizenship
Prior research provides evidence that a consensus-based view of citizenship matches predominant forms of common sense thinking (Crandall & Beasley, 2001). Because intragroup processes that are its core rely upon individual responsibility, social problems occurring within the group are easily seen as a consequence of individual failings and weaknesses (Gilbert & Malone, 1995), a process anchored in the belief in a just world where individual destinies are deserved (Lerner, 1980). Alternatively, however, the legitimacy of social rights can be justified on grounds of a structural perception of the community that places less weight on individual behaviours than the consensus perspective (Zelditch, 2001). In this view, there is always a below and an above (Ridgeway, 2001). Priority is given to lessening the effects of power differences that jeopardize the democratic right to equal access to social rights (Roche, 2002). Unacceptable levels of inequality between social categories are seen as a sufficient motive to grant legitimacy to social rights. Actual, concrete access to rights is perceived as being conditional upon individuals' relative position in society, and social rights are seen as devices to ensure that this conditionality does not reach intolerable levels. Endorsement of common values is unrelated to the legitimacy of social rights. Instead, institutions best guarantee equality of access to rights.
This conception therefore focuses on the socio-economic position of groups in relation to other groups, recognizes structural and collective disadvantage, and grants legitimacy to feelings of entitlement that derive from membership in discriminated groups (Major, 1994). It assesses the extent to which social positions are legitimate and stable, and sees collective involvement and protest as a form of rightful political participation (Abrams & Randsley de Moura, 2002;Wright & Tropp, 2002). We refer to this conception of citizenship as "conflict-based", as it implies that the social order can be contested, and that conflict of interest between social categories and the resulting inequality are the essential driving force of lay definitions of citizenship.
Significance and implications of perceived inequalities between social categories should therefore guide the construction of social policy opinions. In the context of maternity policies, gender mainstreaming and discourses on gender parity have become common which may lead some people to think that everything possible has been done to make the principle of equality a reality (Roux, Perrin, Modak & Voutat, 1999). Resistance to women's right claims can therefore be accounted for with a perception of residual rather than structural inequality, even by those who suffer from its consequences (Roux, 2001).
Since a publicly financed maternity leave and free-of-charge day nurseries are foremost claims for addressing a structural disadvantage, support for these policies should be linked to the recognition of structural gender inequalities.
Furthermore, a conflict-based conception should be linked to attitudes favourable to social change, and social rights should be seen as devices for social change (Roche, 2002). As a result, when endorsing this citizenship conception, highest legitimacy should be granted to social policies that are unrelated to deservingness levels and independent from individual behaviour (e.g. work). In sum, we expect a conflict-based vision of citizenship to provide the normative foundation for the extension of the scope of social rights.
Our general conjecture holds that consensus-and conflict-based lay conceptions of citizenship are two major organising principles (Doise & Staerklé, 2002) that steer lay reasoning on the legitimacy of social rights. The opposition between the two common-sense forms of citizenship should be reflected in tensions about the perceived finality of social rights. In a conflict view of citizenship, rights should be legitimate devices to rectify unequal and unfair treatment, whereas in consensus-based thinking they should be privileges that need to be deserved (Appelbaum, 2001;Gilens, 1999).

Overview of survey and predictions
Public debates in Switzerland are mostly concerned with the scope and the concrete range of application of maternity rights. In this research, opinions towards four policy options are analysed. A maximal scope of the right is achieved when all women, independently of their personal situation and their behaviour, are granted the right to a maternity leave. A state-financed insurance for all mothers, and freeof-charge day nurseries are used as models of unconditional social policies that dissociate contributions from the protection against risks, and thus provide equal protection without consideration of individual inputs. A more restrictive form of maternity insurance is tied to certain conditions that must be met before protection becomes effective. We used a means-tested policy as a model of this restrictive form of protection, since it requires women to prove that they are in need in order to qualify for benefits. The final option is no maternity insurance at all. Given that unconditional maternity policies (publicly financed maternity leave and free-of-charge day nurseries) are destined to correct structural inequality, opinions towards them should be chiefly predicted by the endorsement of a conflict-based conception of citizenship. Because egalitarian social rights are often the outcome of collective mobilisation and protest against social hierarchies, greater receptivity of these policies should also be linked to attitudes favourable to social change. Consensus-based citizenship should be unrelated to opinions towards these two policy options, as deservingness-based delimitation of the scope of rights is not at stake.
Opinions towards the restrictive policy option (where need is individually assessed and protection conditionally granted) should be predicted with endorsement of consensus-based citizenship, since it is based on individual responsibility and sets up boundaries between deserving and undeserving beneficiaries.
Given that this policy option does not address the issue of inequality per se, opinions towards it should be unrelated to endorsement of conflict-based citizenship.
Finally, outright rejection of the principle of maternity insurance should be negatively linked to a conflict conception (because it implies refusal of recognition of structural inequality between men and women), and positively to consensus-based citizenship (because rejection is fuelled with ideas of private responsibility and perception of threats to the social order).

Method
Sample and procedure. A survey was carried out in four Swiss cities with a randomly selected sample, totalling 769 respondents who were part of a representative sample of the urban and working-age population 2 . Standardised questionnaires were administered in the presence of trained interviewers during the autumn of 2001, at respondents' homes. Filling in the questionnaire took between 60 and 90 minutes 3 . Anchoring variables. Sex, age, nationality, level of education, and perceived material insecurity were the five anchoring variables (see Doise, Clémence & Lorenzi-Cioldi, 1993) used in this study. These variables allowed measuring the impact of social positions that intervene in legitimacy judgements of social policies. Persons with a higher level of income and education, for example, should be opposed to the extension of social rights because they have little personal experience with welfare beneficiaries (Gelissen, 2000;Gilens, 1999). In the same vein, sex and age can be interpreted as variables assessing self-interest, because maternity policies directly concern younger women.
The sample was composed of 395 women and 374 men. Mean age was 41 years. 37% were aged between 20 and 34 years, 34% between 35 and 49 years, and 29% between 50 and 65 years. Eighty-three percent were Swiss nationals, whereof 14% had double nationality. The majority of migrants without Swiss 2 In the reported analyses, Ns vary slightly due to missing values. nationality had their origins within the European Union, totalling 13% of our sample. Socio-economic status was assessed with educational level 4 . Eight percent of the sample went through obligatory schooling, 32% had a certificate of apprenticeship, 12% a college degree (granting access to university), 22% a diploma of professional education (specialised high schools) 5 , and 26% obtained a university degree. Men had a higher educational level than women, just as the younger cohorts compared to the generation older than 50 years. In the context of this article focused on gender claims, it is important to note that 21% of the interviewed women, compared to 8% of the men, did not have a professional activity at the time of the study (students were excluded from this calculation). Furthermore, the mean rate of female professional activity was 76%, whereas the corresponding rate for men was 95%. This shows that most professionally active men had full-time jobs. Finally, 35% of the respondents of our sample had one or more children (M = 1.79).
Because perceived material insecurity reflects respondents' concrete situation in terms of available resources, it is also conceptualised as an anchoring variable, and measured with an item on perceived economic risk (likelihood "to be forced to tighten one's belt in order to make ends meet"; 1 = very unlikely, 6 = very likely); 38.6% reported low likelihood of economic risk and 60.7% report high likelihood of economic risk.

Measures
Unless otherwise noted, all scales used in this article ranged from 1 ("Totally disagree") to 6 ("Totally agree").
Predictors. Eight measures were used as predictors of opinions towards maternity policies. Three of them assessed a conflict-based conception of citizenship. Recognition of structural gender inequality was measured with an item stating "In Switzerland, a lot needs to be done in order to make equality between men and women a reality". A second item was created to measure support for civil society movements and attitudes favourable to social change. Respondents rated the importance of activities of "Movements defending women's rights" on a scale ranging from 1 (not important at all) to 6 (very important). Collective responsibility was measured with an item proposing unconditional support for a collectively managed welfare state. It posited that "Under no circumstances should the state cut social budgets; rather, if financial resources are lacking, some other way has to be found to cover social expenses (e.g. increase taxes of the wealthiest, other budgets)".
Five items were used as indicators of a consensus-based conception of citizenship. Common value threat was assessed with the item "To what extent does Switzerland run the risk of loosing its moral values, because good education is no longer a priority for parents ?" (1 = very unlikely, 6 = very likely).
Fear of delinquency was also measured with a probability item, stating "To what extent does Switzerland run the risk to become a less secure country because street delinquency is rising?". Deservingness was measured with the item "It is unfair that certain persons obtain help from the state, whereas others, in the same difficult situation, try to pull through on their own". A private conception of responsibility was captured with an item stating "The state should help persons in need only if the family is unable to support them financially". Finally, endorsement of equality of opportunity was determined with the item "In our society, everyone can, if (s)he wants so, do without the help of the state, because everyone has the same opportunity to succeed in life". Dependent variables: Maternity policies. Three maternity policy options were introduced, as well as an item proposing rejection of the very principle of maternity insurance. The first policy item proposed an unconditional maternity leave for all women, organised on the model of compensation of loss of income due to military duties: "Employers compensate the loss of salary due to military service, they could do the same with women in order to grant them the right to a maternity leave" 6 . The second measure concerns free-of-charge day nurseries, a policy from which all women could benefit: "There should be a new law that guarantees free-of-charge day nurseries so that women can get more involved in their professional activity". These two policies correspond to claims made by associations defending women's rights. A third item introduces the restrictive policy of maternity insurance only for needy women: "Maternity leave should only be granted to women who financially need it ". The last item on maternity leave rejects its very principle by making maternity a purely private issue: "No maternity insurance is needed, those who want children must take on financial responsibility, because maternity is a private matter". 6 In Switzerland, all men are subjected to military service until the age of 42 years (at the time the study was carried out). The loss of salary due to mandatory service is compensated (including for those without a professional activity) with a national income replacement scheme ("Fund for loss of earned income"), mainly financed with taxes and the economy, represented by the employers. To simplify item wording, only employers were mentioned. The maternity leave project currently prepared by federal authorities plans to pay a maternity allowance financed with these funds.

Conflict-based and consensus-based predictors
We first present descriptive statistics as well as a principal component analysis of the eight predictor variables. Means and standard deviations are displayed in Table 1.

INSERT TABLE 1 ABOUT HERE
On the whole, conflict-based predictors yielded higher agreement than consensus-based predictors. Strong agreement was found for the idea that equality between men and women was not yet Respondents were more reserved about endorsement of the principle of private responsibility (M = 2.84), and were also quite sceptical about the idea that everyone had the same opportunities in life (M = 2.52).
The structure of these predictors as evidenced with a principal components' analysis provides evidence for the existence of the two conceptions of citizenship in lay thinking, one focused on social conflict, the other on consensus. The analysis extracted two dimensions accounting for 45.2% of the total variance (Table 1). The first factor grouped the five items measuring a consensus-based conception of citizenship. Fear of delinquency and of loss of educational values went along with deservingness, equality of opportunities and a private conception of responsibility that mobilises the family. Equality of opportunities was associated with consensus-based thinking, but was also the only principle to be negatively linked to a conflict-based conception. This confirms that both dimensions are fairly independent from each other 7 . The second factor was defined with the three items referring to a conflict-based conception of citizenship. Recognition of structural inequality was associated with the activities of social movements and with the principle of public responsibility destined to correct inequality. 7 Principal components analysis with oblique rotation yields a correlation of r = -.21 between the two factors.
Analyses of variance were performed on the two factor scores to account for differential effects of social positions. Sex, nationality (two categories), age (three categories), education level (four categories) and material insecurity (two levels) were simultaneously used as sources of variation (Table 2).

INSERT TABLE 2 ABOUT HERE
Results show that concerning the consensus dimension, older respondents were the most likely to think in terms of common values and private responsibility. This feeling also characterised persons without post-obligatory education, and, to a lesser extent, those having completed an apprenticeship. No differences were observed between the two categories of post-obligatory schooling. On the whole, positionings developed as a function of education level: The higher the level, the less people were sensitive to consensus-based citizenship. Persons with a high level of material insecurity also more easily activated this conception. Finally, the analysis revealed no significant differences for categories defined by sex and nationality.
Women were clearly more sensitive to the conflict dimension. Generalisation of this result, however, is not warranted, because items involve specific gender issues (gender inequality and utility of women's associations). Furthermore, conflict-based reasoning prevailed for the generation aged between 35 and 49 years, especially when compared with the younger cohort, as well as for respondents with a high level of material insecurity. No significant differences were observed for nationality and education level.

Conflict-based and consensus-based predictions of opinions towards maternity policies
Regression analyses were performed to demonstrate how the components of consensus-based and conflict-based citizenship were linked to perceived legitimacy of policy options. Four linear regression analyses were performed, one for each policy under scrutiny (Table 3). Opinions towards each policy were predicted with the eight items assessing conflict-based and consensus-based citizenship, as well as with the five anchoring variables. Table 3 first indicates descriptive statistics for the criterion variables. A majority of respondents supported citizenship-extending policies, measured with a publicly financed insurance system (M = 4.71) 8 , and with free-of-charge nurseries (M = 4.17). They were more reluctant to support a policy grounded on category boundaries defined with the degree of need of women (M = 3.30), and they clearly rejected the 8 The endorsement of our respondents to a large-scale maternity insurance may seem surprising since the last project submitted to popular vote in 1999 has been rejected by the Swiss population (turn-out 45%). Yet, this vote has been idea that maternity is a private matter, and should therefore not be covered with institutional measures (M = 1.74).

INSERT TABLE 3 ABOUT HERE
Consistent with our predictions, unconditional policies (publicly financed maternity leave and free-of-charge day nurseries) were mainly predicted with the endorsement of a conflict-based conception of citizenship. Both were positively linked to the recognition of gender inequality, as well as to conceptions of public responsibility and favourable attitudes towards women's movements and social change, although the respective importance of these predictors varied. Consensus-based predictors intervened to a lesser extent: Fear of delinquency and support for the principle of equality of opportunities were linked to rejection of a public insurance. Rejection of day nurseries was related to both threat items (to fear of loss of common values, and, to a lesser extent, to fear of delinquency). Furthermore, women, younger persons, foreigners, and respondents with a high level of education tended to be more in favour of a public maternity leave. In the same vein, younger individuals, respondents of foreign origin and those who feel materially insecure supported free-of-charge day nurseries. Together, predictors accounted for 21% and 23%, respectively, of the variance of opinions.
Opinions towards the restrictive need-based policy (maternity insurance only for needy women), on the other hand, were predicted with two components of consensus-based thinking, without the intervention of conflict-based citizenship. Here, first of all the idea that the family should take care of its own members predicted support for the policy, as did the idea that a "good" education was a value to preserve. Swiss nationals, men, and respondents with a low educational level typically held opinions more in favour of this restrictive access to the right to a maternity leave. On the whole, the explained variance is lower than for the other policies (13%).
The downright refusal of the policy principle was strongly associated with conflict-based predictors, and to a lesser extent with consensus-based predictors, overall accounting for 26% of explained variance. On the one hand, respondents rejecting all three components of a conflict-conception were most likely to refuse the principle of a maternity insurance. On the other hand, the belief in equality of opportunities, endorsement of deservingness principles, and fear of delinquency were associated with the characterised by a clear cleavage between cities and the countryside: big cities have accepted the insurance (57% of Yes), and our respondents come from four of these cities. rejection of maternity insurance. Especially older people, and respondents with a low educational level were opposed to the principle of a maternity leave, and hence considered that it was a private matter.

Discussion
In this paper, we provide tentative evidence for the validity of the distinction between two types of lay justifications of social policies in the context of women's rights. We refer to these two modes of common sense reasoning as lay conceptions of "citizenship", because they shape perceived legitimacy of right claims. Consensus-based and conflict-based conceptions of citizenship can therefore be considered as organising principles of attitudes towards social rights (Doise, Clémence & Lorenzi-Cioldi, 1993). They represent two sets of common references towards which people take on a position (Clémence, 2001), and which subsequently guide lay reasoning on social and other forms of citizenship (for a related discussion of organising principles of citizenship, see Sanchez-Mazas et al., 2003).
The results of the regression analyses on the legitimacy of maternity policies presented an overall pattern consistent with our hypotheses. Legitimacy of unconditional policies destined to reduce inequalities was mainly determined by endorsement of conflict-based thinking, and to a lesser extent by consensus predictors. Legitimacy of the restrictive, need-based policy, on the other hand, was backed by components of the consensus-based conception: Feelings of common value threat and conceptions of family responsibility were associated with support for the individualisation of access to maternity rights. As expected, opinions towards this policy were independent of the structural perception of society. Rejection of day nurseries was associated with the two threat items. This may suggest that free-of-charge day nurseries are perceived as presenting a somewhat stronger challenge to the existing social order and the stability of the gender system than a public insurance, maybe because it implies that women's professional activity does not need to be subordinated to childcare. Denial of the very principle of maternity insurance, finally, was backed with arguments that minimise gender inequality and collective responsibility, and refute the need for social change. Since perceived levels of deservingness of beneficiaries are irrelevant when maternity leave is completely discarded, the link with rejection of conflict principles was stronger than with consensus-based arguments.
Furthermore, age was consistently linked to the opposition to citizenship extending policies. Even though other variables in the model controlled for differences in values and societal perceptions, age as such still accounted for a substantial part of variance. It can, at least partially, be interpreted as an effect of unmediated self-interest to the extent that older people grew up without maternity insurance, and won't be in a situation to take direct advantage from it. In contrast, foreigners, women and highly educated people tended to favour these policies.
Our results also showed these two lay conceptions appeared to be fairly independent from each other. Moreover, our urban sample distanced itself quite clearly from consensus-based discourses that try to take advantage of diffuse feelings of insecurity, and that emphasise individual and family responsibility to solve social problems. On the contrary, it was quite sensitive to inequalities that put their mark on social relations, and was favourable to their reduction through state intervention or through actions of social movements.

Conclusion
In conflict-based citizenship, lay reasoning is based on the perception of structural relationships and conflicting interests between dominant and subordinate categories. Consequently, endorsement of this conception goes along with the idea that illegitimate hierarchies and inequality between categories can be defied (Abrams & Randsley de Moura, 2001). Such perceptions call for protection and defence of disadvantaged groups and individuals. Here, social rights are seen as devices able to challenge existing power relationships. Their justification ultimately lies in social change.
A consensus-based conception of citizenship, on the other hand, can be characterised as a lay version of Talcott Parsons' (1951) theory of the social system. According to this conception, the society is perceived as a community of self-reliant individuals (particularly men) endorsing common values that ensure the stability of the system. Discrepancies within the community are understood as the outcome of differential value endorsement. Common values and norms become visible and gain their prescriptive power only when they are violated (Foucault, 1975;Young, 1999). Social policies based on deservingness may fulfil such a social control function, as they reward desirable behaviour and sanction undesirable behaviour. Hence, in a consensus-based conception of citizenship, social rights are seen as fair when they are founded on individual endorsement of common values. Social rights acquire their legitimacy essentially as devices to enforce common values.
In the case of nation-wide social policies, the community whose values are at stake is typically the nation. In this context, individuals endorsing a consensus-based conception of citizenship are likely to think that various social groups put the nation's order, stability and cohesiveness at risk, particularly from within (Duckitt, 1989). This positioning is at odds with the principle of social diversity, as it depicts groups and individuals suspected not to endorse common values as potential deviants or protesters, including professionally active women alleged to be "bad mothers" because their work diverts them from their primary responsibility.
The results of our survey have shown that representations associated with childcare and maternity leave relate to the moral and social order, to the redefinition of the meaning of family in contemporary society, and even to delinquency. On a political level, the two unconditional maternity policies are intended to provide social conditions favouring women's autonomy, on a professional level as well as in relation to their partners. Our results take us a step further inasmuch as the support for these policies is also a means to confer to maternity and to childcare a public status; that is, they become issues that are of concern for the entire society. The shifting and renegotiation of the boundaries between the public and the private leads to integrate in the public sphere-through a broader perspective of the political-issues that are still considered as personal by some (see Benhabib, 1996). Indeed, support for extensive maternity policies is tied to the recognition that women and men are social groups that are the outcome of a political relation that can be shaped through social policies. The opposition to day nurseries and to maternity insurance, on the other hand, is based on the idea that child rearing and education comes within private responsibility. This latter conception nevertheless contributes to the maintenance of many current social settings, particularly through the perpetuation of gendered positions and associated social roles. In this perspective of complementarity, status and social roles of women and men are not conceived as hierarchical and arbitrary, but as natural and equivalent (Delphy, 2001;Ridgeway, 2001;Roux et al., 1999).
The idea that the personal sphere is a territory safe from outside intervention leads to consider right claims on such private grounds as illegitimate. It is also a strategy to deny that the private sphere can be at the origin of discriminatory treatment. Yet, in the end, the recognition of groups is not sufficient by itself to guarantee their rights. Support for extended social citizenship is more likely to be promoted with perceptions of groups not in terms of their essential differences, but in terms of social inequalities. Note: All scales range from 1 to 6. Loadings smaller than .10 are not displayed. Means sharing different subscripts differ at p < .05 (Tukey HSD); * = p <. 05, ** = p <. 01, *** = p <. 001 Table 3 Descriptives and determinants of maternity policy opinions (linear regression) Publicly financed maternity leave