Attitudes, Care and Commitment: Pattern and Process

by Sarah Irwin
University of Leeds

Sociological Research Online, Volume 9, Issue 3,
<http://www.socresonline.org.uk/9/3/9/3/irwin.html>.

Received: 5 Dec 2003     Accepted: 26 Jul 2004    Published: 31 Aug 2004

Abstract

The paper develops a new analysis of attitudes to aspects of work and caring for children. The paper reports on analysis of data generated through a small scale survey designed and conducted with pilot study funding to follow up aspects of research by the ESRC Research Group for the study of Care, Values and the Future of Welfare. The survey research takes as a focus a specific point in the life course of parents: when they have children in the early years at primary school. The paper develops an analysis of the coherence of people's attitudes and their social location, with particular reference to social inequality. It also reports on new linked analysis of British Social Attitudes Survey data. In contrast to recent arguments of increasingly 'autonomous' subjectivities the research contributes to a broader theoretical understanding of the mutuality of subjectivities and extant social relations.

Keywords: Attitudes, Norms, Care, Parenting, Employment, Work-Life Balance, Schools

Introduction

1.1 Recent research in the areas of parenting, employment and childcare has stressed the importance of norms in shaping people's decisions and actions in these domains. There is a general perception that normative issues have been under-acknowledged and insufficiently researched, and there is a growing literature on diverse subjectivities and their link to new forms of diversity in family life, in childrearing and in gendered roles (eg. Hakim 1996, 2000, Duncan et al 2003, McRae 2003, Williams 2001, Himmelweit 2002, Marks and Houston 2003, Hattery 2001, Thomson 1995).

1.2 In some perspectives norms and values appear to be now more 'free floating' of social structure than in the past, such as within preference theory, (Hakim 2000) and individualisation theory (Beck 1992). In these perspectives we are experiencing a social transformation in which we are more freed from prior economic and cultural constraints, and reflexivity, agency and value are more central components of (late) modern social dynamics (cf. Bauman 1995, Beck 1992, Giddens 1991). These theories of a transformation in social dynamics may overstate the degree of change (eg. Irwin and Bottero 2000, Adkins 2003). What is particularly problematic in arguments of a shift to a new kind of reflexivity, a new salience of choice and value, is the presumption that the current salience of values entails their 'loosening' from structural processes. This is a feature of theses of individualisation (eg. Irwin and Williams 2002) and there are some parallels with preference theory where the separation has clear expression:

"Our thesis is that lifestyle preferences and values are becoming more important determinants of behaviour, relative to economic necessity and social structural factors" (Hakim 2000, pp80-81).

1.3 Rather than rethink our understanding of structural processes we are enjoined to see them as less relevant. This is arguably a feature of the sociological turn to agency more generally.

1.4 For her critics Hakim presents an over-individualised and inappropriately voluntaristic theory of human behaviour (eg. Duncan et al 2003, Blackburn et al 2002, Tomlinson 2003, Bruegel 1996, Ginn et al 1996). Hakim foregrounds the sociological significance of choices but social diversity means that the very contours of choice and constraint vary (eg. Duncan and Edwards 1999, Glucksmann 2000). Without analysis of experience and context, choice holds limited explanatory purchase.

1.5 Part of the impetus behind recent, principally qualitative, research in the areas of family, kin and commitment has been the concern that we need to more precisely contextualise individuals' values and the ways in which moral choices and negotiations are made in care decisions and employment decisions (eg. Duncan et al 2003, Williams 2002, Mason 2001, Silva and Smart 1999, Duncan and Edwards 1999). The emphasis within research by the ESRC Research Group for the study of Care, Values and the Future of Welfare (CAVA) has been on the context and texture of social interaction and on understanding the links between moral judgements and social practices.[1] The research helps to reveal the ways in which people are moral beings, embedded in webs of relationships and, when confronted with moral ambiguities in their conduct and relationships, seek to do 'the right thing'. The research then challenges models of uniform rationalities and usefully positions 'values' not as something simply held, possession like, by individuals, but as something often in process, evolved in concrete circumstances and contexts (Williams 2001, Mason 2000).

1.6 There remains a gap in our understanding of the general social landscape of diverse contexts. There is of course a wealth of quantitative attitudinal research on issues relating to family, work and welfare (eg Scott et al 1998, Jarvis et al 2000, Bonoli et al 2000). Such research offers important insights into diversity, and continuity and change in attitudes, yet these studies offer a very broad brush description of social diversity.

1.7 In some respects the current emphasis on norms and values within the literature is part of an ongoing critique of structuralist perspectives in which social location, allegedly, was seen to determine interests and values. However there is broad concern that we have inadequate tools for conceptualising and analysing structural processes (eg. Bradley 1996). The new interest in norms and values has been productive and has offered a range of insights but the emphasis on contexts has not been matched by research on their links with the broader social structure. Recently commentators have stressed the need to improve theories of how norms and the social order mesh together (Duncan et al 2003, Crompton 2002, Himmelweit 2002).

1.8 This paper offers a contribution to this end. It reports on a small scale research project designed to sociologically 'map' attitudes around work, care and parenting and their links to diverse contexts. Where appropriate, points of comparison are made with data from the International Social Survey Programme component of the 2002 British Social Attitudes Survey. I will explore links between attitudes, experience, and social circumstances with particular reference to advantage and disadvantage. The paper develops some general arguments regarding the articulation of attitudes and the social order. In contrast to some recent arguments of their disjuncture, it will be argued that diverse dispositions, and attitudes regarding 'the right thing to do' are closely linked to social location. The analysis feeds into a more general argument: that recent treatments of normative processes as autonomous from structural relations are inappropriate. Subjectivities are not more autonomous of the social interaction order than in the past, but remain an integral part of that order, even in a time of rapid change (see eg. Irwin 2003, Bottero and Irwin 2003). In short we should not 'conceptualise away structure', but construe and analyse it as a dynamic process in which the subjective and normative dimensions of social life are meshed with extant material social relations.

1.9 To clarify the point of entry into the research question some definitions are helpful. Norms provide a framing set of assumptions on which people draw, usually implicitly, in their choices or judgements about forms of moral conduct and in their perceptions about appropriate behaviours, and rights and responsibilities. The contemporary reworking of gendered norms forms part of the backcloth to this paper, and is not directly addressed here (see Irwin and Bottero 2000, and Irwin 2003 for more on this). Nevertheless such norms and their reshaping are important, manifest in particular in the repositioning of women and men in household resourcing and the more routine expectation, across an expanding section of the population, that women with young children will engage in paid employment. Values refer to beliefs which people hold, and may consciously reflect on, or which may inform (or be consciously worked out in) chosen modes of conduct. Values will have a varying salience in respect of the shaping of social action. Much social action, even in a context of aggregate social change, may be fairly routinised or pragmatic, rather than necessarily value driven. Values per se are not tackled directly in this paper. The empirical analysis of this paper focuses on attitudes and dispositions. Attitudes are more specific than values, and in many respects may be considered as 'generated data', offered in response to direct questions about how things are or should be. Responses to attitudinal statements, perceptions of how things are, respondents' priorities and so on, tap into dispositions, that is to orientations that may be more mutable than are values. However, responses are not mere 'artefacts' of imposed meaning, and reveal a clear pattern of covariation with material and situational factors, a covariation which requires us to re-interrogate recent arguments of disjuncture between subjective and structural processes.

Research Design and Method

2.1 The new empirical evidence reported on in this paper was funded as a small follow-on project (called here 'Life as a Parent') which would generate data to complement research done by the ESRC Research Group on Care, Values and the Future of Welfare (CAVA) based at the University of Leeds (Williams 2001). The Life as a Parent research into attitudes, care and commitments took as a focus the experiences, perceptions and behaviours of parents of young school children aged 4 to 7 years. This diverse 'group' was a key site for research, targeted for two reasons. Firstly, it is one of the successive cohorts entering the family building period which has authored new relations to employment and family. To illustrate, in the years 1981, 1991 and 2001 the full time employment rates of women with children aged 0 to 4 increased from 6% to 13% to 18%. Across the same years the part time employment rate of women with children aged 0 to 4 increased from 18% to 29% to 36% (OPCS figs cited in Walby 1997, and ONS 2003). This group has been cited as the most rapidly growing group in the labour force, and the trend here has been of much interest. What happens, though, when young children are at school? Amongst women with children aged 5 to 10 participation rates are higher. Across the years cited, full time employment rates for the group increased from 13% to 22% to 26%, and part time employment rates stayed constant at 43% to 44% across the period (Walby 1997 and ONS 2003). Because of higher participation rates amongst mothers with school aged children there is a tendency to assume that childcare problems 'fall away', when in fact they may become complex and problematic for many (eg. Skinner 2003; Windebank 1999). The transition to primary school is a transition not just for children but also for their parents. Precisely how parents manage this transition remains under-researched and generally oblique.

2.2 The research sought to develop a more nuanced approach to the subjective aspects of social experience than is usual in a survey approach, paying particular attention to context. The sample locales were chosen with reference to specific dimensions of social diversity, through targeted schools. A purpose here was to explore if school neighbourhoods and residential locales contained support networks and reference groups of relevance to understanding people's attitudes and decisions regarding appropriate care of children. Attention to context was also a feature of the questionnaire design.

2.3 The research comprised 102 interviews with parents of children aged 4 to 7 and attending 7 identified schools in specific locales across Leeds. Schools were used as a point of access where possible. Locales were chosen to provide diversity along a range of socioeconomic indicators, including rates of unemployment, incidence of lone parenthood, and employment rates amongst women. School intakes were profiled with reference to ethnic diversity, percentage of pupils eligible for free school meals, and percentages registered as having special educational needs. Three locales were chosen, two of which correspond to CAVA locales in the Leeds area (Duncan 2000). 'North East Leeds' and 'North Leeds' locales contain relatively high proportion of mothers in full time employment compared to the Leeds average, and a relatively high proportion of full time to part time female workers. 'North Leeds' is a predominantly white middle class area. North East Leeds is more heterogeneous with a significant proportion of disadvantaged households and with a high percentage of minority ethnic and of lone parent households, alongside its more advantaged middle class neighbourhoods. The third locale was chosen as a complement: This 'West Leeds' locale has a relatively low percentage of mothers in full time work, and a relatively low ratio of full time to part time female workers, relative to the Leeds average. It is a predominantly white, working class area. In terms of social inequality the overall sample was 'middle ranging' - it encompasses a range of unequal situations, but not those in the most disadvantaged areas where unemployment rates are extremely high since I wanted to select circumstances in which there was a realistic possibility of work for most residents, and some genuine scope for making choices and decisions in respect of employment participation[2].

2.4 Interviews were around 1 hour long, based on a structured questionnaire survey which included a range of questions covering household / family composition; life course event histories; employment and care arrangements; and a series of questions relating to attitudes and evaluations, including structured questions, vignettes and open ended questions. Interviews were with the parent with whom contact was made: this turned out to comprise 96 women and only 6 men.[3] It had been an initial expectation that a non-quota sample would generate a somewhat larger sub-sample of men. Because there are so few fathers in the sample we cannot meaningfully explore gendered differences and similarities from within this data set. Most of the paper will accordingly focus on the experiences, perceptions and commitments of mothers. However, this should not be taken to imply that mothering entails a wholly different set of values and commitments than does fathering. We can see through the data very clearly the depths of mothers' commitments to mothering and to caring for their children. Often this is where mothers' primary commitments lie, a theme which echoes that elsewhere in the literature. However, it may be very commonly the case that fathers' primary commitments lie also with their children. However, women and men are still typically differently positioned such that both the content of, and the ways they meet, their commitments may vary. Clearly there is much to be gained from more research allowing a fuller understanding of fathers' as well as mothers' perceptions and attitudes. The following analyses are based on mothers purely as a consequence of how the sample 'fell out' - in large part a reflection of mothers' and fathers' distinctive social positions in respect of the care of school children.

Dispositions, Experience and Context

The Importance of Time

3.1 Duncan and his colleagues have challenged what they see as a flaw in government intervention and policy around welfare to work, work-life and childcare issues. People do not react individualistically and uniformly to economic incentives, but rather through diverse moralities and social negotiations regarding for them the right thing to do (Duncan and Edwards 1999, Duncan et al 2003). Others have argued also that social policy is out of line with the preferences and aspirations of many parents (eg. Dean 2001, Williams 1999). In particular, they argue, parents and especially mothers' commitments to caring for their children means that work as a route to citizenship (as promoted by New Labour policies and rhetoric) marginalises care. There has been an argument that an increase in women's employment and continuities in ideologies of good mothering intensify the double burden many women carry, and generates new moral dilemmas as well as levels of guilt amongst many mothers (eg. Saugeres and Duncan 2002, Dean 2001, McKie et al 2001). Other research demonstrates high levels of stress experienced by women with career aspirations and significant domestic commitments (Crompton et al 2003).

3.2 Having time to care, properly, is a theme that emerges from the cited research as a priority for parents and this is echoed by the 'Life as a Parent' sample. Respondents were asked to look at a list of items and state which was most important and which was second most important to them. The initial fixed choice list generated a distribution of 72 stating as most important: 'to have the amount of time I want for spending with my child/children', and 23 stating as most important: 'to have enough money to bring up my child/children in the way I want'.[4] Respondents were then asked 'Is there anything in your life you feel is important but isn't on this list?'. They were asked what this was and where there would place it in their priorities, that is first, second or lower. 48 respondents identified something important to them that is not on the list, of which 21 ranked it their first priority. The final list of priorities was as follows:

3.3 Clearly time for children is the priority for the majority of respondents. The female profile of the sample means it is not possible to analyse male responses - there were only 6 fathers who were interviewed. It is worth noting that 3 out of these prioritised time for children, and 2 prioritised money. It seems plausible that fathers as much as mothers will prioritise time for children, although the content of this time may often look different.

3.4 Amongst the mothers we can see a diversity of priorities which relates in part to their own employment status. Amongst the 28 women who are full time homemakers, 16 prioritise time for children, compared to 2 who prioritise money for children. Amongst the 42 women who work part time, 28 prioritise time for children, compared to 5 who prioritise money for children. Amongst women who currently work full time the numbers are low but we can note that of 19 such women 7 prioritise time for children compared to 8 who prioritise money for children. Time for children also is the majority preference (at an equivalent level) amongst those with pre-school children and those with only school aged children. This preference and its generality is unsurprising yet the evident value of time to our respondents - as of others with children - deserves underlining.

Attitudes and Experience

3.5 In line with other research, the evidence from the 'Life as a Parent survey' reveals a clear pattern of co-variation between people's attitudes and their behaviours in respect of gender and parenting roles, work and care (cf. Alwin et al 1992, Thomson 1995, Marks and Houston 2002, Hattery 2001, McRae 2003). Alwin and colleagues were agnostic regarding the the nature of causality here although many writers have placed an emphasis on beliefs (attitudes or preferences) as having a significant role in shaping behaviours (eg. Marks and Houston 2002, Hattery 2001, Hakim 2000). They are all aware that stated attitudes may be a rationalisation of behaviour, or possibly created through experience. However, they all offer evidence that attitudes, or in Hakim's argument preferences, play a significant role in shaping people's work and care decisions. Further, attitudes are not deemed a small addition to standard socio-economic indicators of employment participation amongst mothers. Thompson concludes that whilst social and economic factors differentiate mothers who work and mothers who do not, and availability of childcare is very important:

"...what most distinguishes working from non-working mothers is their attitude towards women and work" (Thomson 1995, p. 83).

3.6 This echoes Hakim's argument that women have choices in a way they did not in the past (Hakim 2001). However, analysis of causal direction, from beliefs to behaviours, is not definitive within the literature but somewhat speculative. For example, Thomson identifies diversity of belief within her three groups of full time workers, part time workers and homemakers, and an association between within-group variation and women's stated preparedness to return to work if they had access to ideal childcare. This patterning suggests to Thomson that attitudes are not merely a reflection of own labour market position, and so she argues that women's beliefs condition, and partly determine, their behaviours. This is a plausible enough statement, but it does not necessarily follow the empirical evidence she presents, in which various factors remain uncontrolled. It may be that there is a co-incidence of attitude and experience that can be explained only with reference to prior variables. Indeed it is likely that the notion of a general causal model is inappropriate: causality may be context specific. For the moment I will explore co-variation between stated attitudes and respondents' experiences, a pattern of association which parallels that revealed in other studies and data sets. This coherence can help us to shed light on the shape of social diversity.

3.7 To preface the empirical analysis, firstly the evidence reveals a clear association between respondents' attitudes to mothers' responsibilities and their own circumstances and experiences. Additionally, those with a consistently pro maternal care ('homemaker') attitude are in similar social positions[5]. Furthermore, evidence on seemingly homogeneous evaluations across subsamples of middle and working class women shows, on closer inspection, a pattern of diversity in line with differential opportunities. There is no evidence here of attitudes being 'disembedded' from social location.

3.8 The first analyses below show evidence of a strong association, as we would expect, between attitudes and experience, and circumstances. The two attitudinal questions referred to here were part of a battery of questions drawn from the International Social Survey Programme component of the British Social Attitudes Survey (included in 1989, 1994 and 2002). A number of these questions were also asked in a self completion questionnaire handed out at the end of the Life as a Parent survey interview. In all the tables presented below the terms 'agree' and 'disagree' are aggregates for responses to both 'agree' and 'strongly agree', and 'disagree' and 'strongly disagree'. Table 1 shows the distribution of women's responses to the statement shown.


Table 1. Responses by mothers in the survey to the attitude statement: A pre-school child is likely to suffer if his or her mother works, as they vary by own experience: Life as a Parent data[6]

3.9 In the Life as a Parent data there is an association between respondents' attitudes towards the likelihood of a child suffering if their mother works and their own pattern of working when they had pre-school children. Again this echoes 2002 BSAS data, as shown in Table 2. There, amongst women with pre-school and/ or primary school age children, 40% of homemakers disagree that a child would suffer if his/her mother works in contrast to 63% of working mothers holding such a view. Interestingly amongst homemaker mothers only 1% strongly disagree with the statement (not shown) compared to 23% of working mothers.


Table 2. Responses by mothers in the survey to the attitude statement: A pre-school child is likely to suffer if his or her mother works, as they vary by own experience: BSAS (ISSP) 2002 data

3.10 There is also quite a strong association between attitudes and respondents' qualification levels.[7] In the Life as a parent data, amongst those with qualifications at 'A' level and higher only 5/38 agreed that a pre-school child will suffer if their mother works. In contrast, amongst those with lower level qualifications, or no qualifications, 23/58 agreed that the child would suffer in this situation. White women were more inclined to agree with this attitude statement than were Black women: the ratio of agree to disagree stood at 23 : 35 for White British women and 2 : 8 for Black women. [8]

3.11 The next question also parallels one in the BSAS. Respondents were asked 'Do you think that women should work outside the home full time, part time or not at all under these circumstances?', with a list given as shown in Table 3.[9]


Table 3. Stated preferences across female respondents in response to the question 'Do you think that women should work outside the home full time, part time or not at all under these circumstances?' (Life as a Parent data)

3.12 A clear majority see full time work as appropriate for women without dependent children, and part time work as the appropriate pattern for women whose (youngest) child is at school. A substantial minority (30%) recommend the homemaker option for mothers of pre-school children. This is the pattern amongst White British women for whom the ratio of responses favouring full time work, part time work, and staying at home were 2 : 39 : 25. In contrast Black women eschew the stay at home option, the equivalent figures standing at 4 : 5 : 1.

3.13 There is a clear, although not general, association between respondents' attitudes and their own behaviour when they themselves had pre-school children. Working mothers favour work (clearly part time work even where they worked full time) and homemaker mothers favour homemaking (see Table 4). So for example less than one fifth (12/65) of women who worked felt that a mother of pre-school children should stay at home. In contrast just over half of homemaker women (16/30) felt that a mother of pre-school children should stay at home.

3.14 In the 2002 BSAS survey data there is also a marked association between respondents' attitudes and experiences. There, amongst working women with children aged 11 and under, those favouring working to staying at home stood at 57% to 26%. Amongst homemaker women with children aged 11 and under, those favouring working to staying at home stood at 29% to 52%.

3.15 There is also a marked association between attitudes towards work and care and level of qualifications achieved. In the Life as a Parent data those who are highly qualified are, as we would predict, much more likely to favour women working when they have a pre-school child, with 24 / 38 favouring this pattern and only 7/38 favouring the stay at home option. Those with lower qualifications conversely are fairly evenly divided.[10]


Table 4. Stated preferences across female respondents to whether of not a mother of pre-school children should work, by their own labour force status at the time of having pre-school children. (Life as a parent data)

3.16 Let us turn now to some issues in the care of school aged children. As we saw in Table 3 most women state that in general women should work when their youngest child is school age, the vast majority suggesting that (in response to an abstract and generalising question) such women should work part time. Whilst an abstract statement about what 'women in general' should do suggests a 15% 'approval rating' for full time employment amongst this group in the 2002 BSAS data, in practice nationally 26% of women with children aged 5 - 10 work full time (or did so in 2001, reported by ONS 2003). This may reflect a greater liberalism than the compulsion implied by the attitude statement. Often people may wish they could answer such generalising questions with an 'it depends...' response. Vignettes offer a more detailed scenario to respondents and arguably give us a slightly more subtle insight into respondents' opinions.

3.17 One of the vignettes used within the Life as a Parent survey sought to illuminate how people see the best way of balancing employment and childcare commitments by a mother of a school age child, when confronted with more detail and contextual information about options, desires and constraints.

3.18 The vignette was as follows.

I will now read out another dilemma.
Sue and David have one child and they do not plan to have any more. The child is about to start reception year at primary school. David works full time and Sue works part time. An opportunity to apply for a new job has come up and Sue would very much like this job, but it is only available on a full time basis. However, their child minder could take the child to school and pick her up and look after her until 6 o'clock each day. Sue is trying to decide whether or not to apply for the job. What should she do? Should she:
A. Apply for the job?
B. Stay in her part time job so she can drop off and collect her daughter and be with her after school?
C. Do something else?
If answer is - do something else: what would that be? (open ended)

3.19 Of the 96 female respondents, 87 gave as their answer A or B, and of these a two thirds to one third majority answered B (a ratio of 58 to 29). Women who are currently working full time, and the most highly qualified women, are marginally more likely than not to favour the full time option for the vignette character, in contrast to the part timer and homemaker respondents who clearly favour the part time option. Table 5 shows this distribution.

3.20 There is quite a strong relationship also between responses and whether or not respondents' youngest child is school age. Amongst those with only school aged children, 22/58 respondents favour Sue taking the full time option, in contrast to 7/38 who do so amongst those who have both pre-school and school aged children. It could be that the patterning of responses is partly related then to experience in this domain, although the evidence here is not clearcut. There is no clear patterning by ethnicity. For example White British women favour the full time to the part time option by a ratio of 23 to 47, and Black (Black British and Black Caribbean) women by a ratio of 3 to 8. This echoes the pattern of preferences revealed by Black women to the generalised attitudinal statement regarding labour force status of mothers of school aged children. One might expect a more clear preference for full time options amongst Black women given both the higher full time participation rates of black women (Lindley, Dale and Dex 2004), and a more typically routinised experience of combining extensive hours of paid work and motherhood (Duncan et al 2003). It is notable that in the sample here only 2 out of 11 black women were themselves in full time employment.


Table 5. Women's responses to vignette by labour force status and by qualification level

3.21 So far the evidence suggests a broad correspondence between circumstances and attitudes which is as we might expect. However, some writers as we have seen argue a growing importance of choices and preferences in shaping behaviour and actions. Where this has been subject to empirical research it has been argued that it is where attitudes and preferences are most strongly and consistently held that they will have the greatest causal impact (eg. Hattery 2001, Hakim 2001). There is a risk though that such attitudes and preferences are treated a-sociologically: we cannot locate them or analyse their provenance. For example, Hakim identifies a homemaker group, that is women who have a primary, and orienting preference for homemaking over work. Such women, she argues, exist across the social spectrum. This is taken as evidence of the relative fixity, and historical continuity, of such a preference, and as suggestive that such a preference is primary. It is certainly plausible that some women will indeed seek only a full time homemaker role, seemingly against the odds (at least of current sociological predictive capacities). However, we need be cautious about what can be read into this. We cannot always 'locate' values but this stems from the complexity of their provenance, rather than their randomness. In general we can still uncover connections between social location and perceptions. As with Hakim's evidence, the data here reveals homemakers to be spread across a diverse social spectrum. As discussed next however, the Life as a Parent sample shows those with the most consistent homemaker attitudes were all in very similar social circumstances.

3.22 In the following I take responses to the three survey questions discussed above to identify respondents who were most 'pro maternal care' in their attitudes, that is those who identified the 'homemaker' solution to the two attitude statements and who identified as best the part time work option within the vignette. Out of the sample of 96 mothers, 14 fell into the category of pro maternal care. We might see these women as being the most committed homemakers, at least in their general attitudes (preferences were not directly addressed in the questionnaire). For commentators such as Hakim (2001) and Hattery (2001) it is amongst women with consistent homemaker commitments we might see as most likely to realise their preferences. In my own sample it is particularly notable that members of this 'pro maternal care' subsample were all concentrated in relatively constrained circumstances. None held qualifications above O level or GCSE level. For the remainder of the sample, 38 / 82 did so. Taking qualification level as an indicator of social position we can presume that in general employment opportunities will be relatively limited for this group, a likely constraint then on their actions. An analysis across the school catchments shows that 10 / 14 pro maternal care respondents were concentrated in two school neighbourhoods, both in the West Leeds white working class neighbourhoods. The average age of these women when they had their first child was 22, in marked constrast to the average age of 26 at first birth for the rest of the sample. Again this is consistent with a pattern of relatively limited opportunities within employment. Such women have typically worked at some point at least and social class membership (based on current, or last held job) shows these women to be concentrated in RG social class III non-manual, or lower. Only 1 / 14 belongs to class I or II, compared to 28 / 82 of the rest of the sample. These women then appear to be relatively constrained in respect of their employment opportunities, but they are not necessarily amongst the most disadvantaged. They typically live with a spouse or partner, and the latter are typically working. These women with 'pro maternal care' attitudes are less likely than average to be working and where they do work, it is part time. In only one case is it described as being 'for essentials', in contrast to over half of the rest of the sample of working women.

3.23 In general then we can say that those who are consistently 'pro maternal care' in their attitudes are positioned very similarly to one another. They are white, working class, typically with a partner who is working. Their social circumstances limited their choices with respect to employment opportunities but also to some extent facilitated consistent pro maternal care attitudes. It is of interest to note within the 2002 BSAS data that it is within the lower half of household income groups, but not amongst the lowest, where mothers are most likely to express the most 'pro maternal care' attitudes (with respect to the two BSAS attitude statements discussed earlier). Again this is consistent with seeing such attitudes as associated not simply with 'constraint', but with a combination of constraint and a perceived relative adequacy of household resources.

3.24 The argument of Hakim is that because ostensibly similar preferences cross the social spectrum, so circumstances clearly do not determine choices. She argues that choices now are more primary in shaping outcomes in mothers' work and care decisions. I would argue that we cannot adequately understand social diversity if we disconnect choices, and perceptions, from circumstances. Continuities between attitudes and social location is in clear evidence within the data sets examined, as elsewhere. Diverse vantage points are continuous with diverse social locations. The evidence should help to reorient us to reflect further on the coherence of subjective evaluations and how people are positioned in social space. Can we then use evidence on perceptions to help us improve the ways in which we can analyse social diversity? The next sections of the paper explore this question with a focus on neighbourhood contexts, and on class and aspects of the life course.

Attitudes, Context and Reference Groups

3.25 One strand of the research related to the extent to which specific locale and network based contextual factors relate to attitudes and evaluative judgements. Would it be the case for example that perceptions of 'the right thing to do' are partly formulated through patterns of association and reference groups as these may be salient to people yet not adequately captured through standard indicators of socio-economic difference? This was part of the rationale for a sample based around school catchment areas and also lay behind a question within the questionnaire inviting people to describe the circumstances of three friends or acquaintances known to them. An objective was to explore neighbourhood, and networks, as an under-researched dimension of how subjective judgements may link to the experience of peers, particularly neglected within survey research.

3.26 Across the school catchments we do see some patterning of attitudes which appears suggestive of an effect over and above that captured by standard socio-economic indices of difference. For example, in responses to the vignette described in section 3.2, respondents at 4 schools reveal a preference for the part time as opposed to the full time 'solution'. This is marked in one school neighbourhood with a high proportion of black respondents, in the Leeds North East ward (School A). Conversely, another school in the same ward (School B) revealed a majority preference for the full over the part time work option (11 to 8). This school has a relatively middle class intake, and relatively high percentage of qualified women. Whilst numbers are low the evidence suggests that some aspect of the locales is salient over and above standard indices of diversity, such as qualification level, partnership status, whether or not youngest child is school aged, and social class. Patterns of childcare used by respondents in both 'catchments' are not suggestive of markedly different resources (most of course draw on relatives and friends, although there is more experience with childminders (and with formal care) at School B. It may be simply that experience in particular domains shapes people's sense of the appropriateness of different courses of action. Speculatively, it may be (also) that their judgements are partly shaped by their sense of the experience of those around them.

3.27 Elsewhere, across two of the schools in North Leeds we can again see an example of contrasting attitudes which vary by school in a way unexplained by standard indicators such as class, current employment status or qualification level. Focusing again on responses to the vignette described in Section 3.2, of those women who are qualified to A level and above: in one school (Leeds North, School C) 5/7 respondents favour the full time option, whilst in another school (Leeds North, School D) 4 / 4 respondents favour the part time option. These latter respondents hail from the same neighbourhood within North Leeds as the women Saugeres and Duncan characterise as having clear 'primarily mother' moral rationalities (Saugeres and Duncan 2002). The numbers here are very small of course but it is notable that the data does square with evidence on the salience of neighbourhood contexts (Duncan and Edwards 1999, Duncan 2003).

3.28 Data from the Life as a Parent survey also offers some support to the expectation that patterns of social association link to people's own experiences: people tend to associate and interact with people in like situations. There was an association between friends circumstances and respondents' own current circumstances. For example women currently in employment identified 50 friends/acquaintances who were in the labour force, and only 7 who were full time homemakers. In contrast female homemakers identified 62 friends/acquaintances in the labour force and 22 homemakers. This pattern probably reflects different patterns of association across the sample.

3.29 The extent to which reference groups and local networks - of friends, contacts and informal kinds of support - are significant shapers of experience, norms and values particularly in the context of care and employment, is an area deserving of further research.

Perceptions and the Life Course

3.30 Duncan et al (2003) maintain that understandings of good mothering transcend class and income differences. It is interesting in light of this to pause a little longer with the vignette as described above, and a follow up question. I have described the tendency for respondents with degree level qualifications to favour the full time option more than the other groups. Still, though, nearly half the graduates favour the part time option. This pattern echoes the argument of Duncan and his colleagues - that a group of highly qualified middle class women are seemingly not distinguishable from working class women in their homemaker orientation. It also seemingly echoes Hakim's argument that the same preferences can cross the social spectrum. Does the data imply that we have uncovered a uniform construction of maternal responsibilities that cuts across class based social inequalities? The analysis below suggests that such the construction is not uniform, but hides a diversity within, which corresponds closely with different social locations. The diversity reflects differing perceptions of how care commitments (practical and moral) should vary over a child's early years.

3.31 After providing an answer to the vignette, respondents were asked: 'Why do you think this is the better solution'?[11] Those who had said that 'Sue should stay in her part time job so she can drop off and collect her daugher and be with her after school' were then asked:

'Would it have made any difference to your answer if rather than starting at primary school the child had been at school a few years?', and then: 'why do you think that?'

3.32 For these respondents the distribution of responses disaggregated by qualification level, is as follows:

3.33 Amongst those who favour the part time option - we can see again a clear qualification level gradation on notions of what might be appropriate behaviour when the child is older (but still at primary school). It is notable that those with degree level qualifications stand apart in almost uniformly altering their recommendation as the child grows older. We can explore some of the substance of this pattern through considering the responses to the subsequent open ended question.

3.34 All the following quotes are taken from respondents who say that the character in the vignette should stay in her part time post. The first set of responses is from respondents qualified to degree level (6 out of 7 of these were from 'North Leeds').

3.35 Having stated they preferred the part time 'solution' to the vignette dilemma they were asked why:

Q: Why do you think this is the better solution?

A: To get a bit of the best of both worlds. I think kids like being taken to school and parents also like doing it. You can become very disconnected if you never pick up the kids.

(White, married, degree, 3 children aged 6 and under, homemaker /student).

Q: Why do you think this is the better solution?

A: Because she can go for the full time job later when her child is settled and more confident. Reception year is very important and can set a precedent for the rest of the years at school. By 6pm both you and the child would be too tired and have no real time for each other

(White, married, degree, 2 children, works 16-20 hrs).

Q: Why do you think this is the better solution?

A: To be involved in the school. To know the child's friends, and to know she is settled. Its very tiring to carry on until 6.

(White, married, degree, 3 children, works 16-20 hrs).

3.36 Of these and the other high qualified respondents opting for the part time solution, 6/7 say that their part time recommendation would change if the child had been at school for a few years. The following responses were made when they were asked why:

'Children become more independent as they become older' (White, married, degree, 3 children aged 6 and under, homemaker /student)
'The children's lives become much more independent. They get themselves to school and back independently' (White, married, degree, 3 children, works part time 26-30 hours)
'It might make a difference if the child is settled and has after school facilities' (Married, white, degree, 2 children, works 16-20 hours).
'[The] child would already be settled at school, and secure. The child would be older physically and emotionally to cope with the long day' (White, married, 3 children, degree, works 16-20 hrs)

3.37 This kind of response is not completely 'contained' within the high qualification category, although as we have seen it relates strongly to it. The following quote shows a similar attitude held by a woman who had no qualifications, but also at a North Leeds school. This respondent's part time recommendation would change to the full time one if the child had been at school for a few years:

'Because when they are little they like the security more. When the are older they are more confident and know (their) mum's coming back at a certain time. The younger they are the more they want their mum there'. (White, married, no qualifications, 4 children, homemaker)

3.38 The following quotes are from mothers with children at schools in the West Leeds neighbourhoods. Respondents here generally held few qualifications, and as we saw earlier people here tend to be more 'pro maternal care' in their attitudes and relatively constrained in their employment opportunities. Like the respondents cited above, these women also recommend the vignette character should give priority to time with her child, but in contrast they see this as the right thing to do, it seems, throughout the child's primary school years. Their recommendation would not change if the child had been at school for a few years. For example:

' [She] needs to spend time with her children [it] doesn't matter what age they are'. (White, married, no qualifications, 5 children, homemaker)
' [You] don't have kids to give to somebody else. It's a long time for a child to be with a childminder'. (White, Cohabiting, NVQ level qualifications, 2 children, works part time (11-15 hours))
' You still need to spend time with your children whatever age they are' (White, cohabiting, no qualifications, 3 children, homemaker).
'Because kids need their mum at home whatever age they are. They need them to cook their tea and stuff' (White, lone parent, no qualifications, 2 children, never worked).
'Your child has to come first. I have been offered full time work from September when my son starts but while he is (at) primary I will not do this' (2033) (White, cohabiting, unspecified vocational/professional qualification, 4 children, works 25-32 hours on shifts).

3.39 These women appear to believe that mothers' exclusive care and commitment, at least in the circumstance described, should extend right across a child's primary school years. This contrasts with responses by the highly qualified, more advantaged women described above. It seems very likely that the high qualified women possess opportunities for strategic employment decisions, and hold aspirations for themselves as workers, and careerists, independent of their commitments to their children. In contrast a more limited scope for strategic employment decisions is consistent with holding moral commitments which lie for much longer with the exclusive care of children. Thus perceptions which at first appear uniform, and independent of social difference, turn out to be closely related to social location, here with commitments seen through a temporal, life course lens, their patterning consistent with very different class related positions and likely trajectories.

Conclusion

4.1 Recently writers have highlighted both the importance of norms in shaping gendered work and care commitments, but also the pressing need to better understand how these mesh with social structural processes (eg. Crompton 2002, Himmelweit 2002). This paper offers an analysis of empirical evidence along just one dimension of this multifaceted research problem. It reveals a coherence of attitudes towards caring for children, and social location, the latter with particular reference to a differentiated structure of opportunity and constraint. There are clear continuities between social location and attitudes regarding 'the right thing to do' in respect of work and care.

4.2 In the paper I have analysed empirical data from a new small scale survey, and from the 2002 British Social Attitudes Survey, focusing on women who are mothers of primary and pre-school children, and argued that diverse attitudes are strongly associated with social circumstances. Seeming discrepancies between the two identified by other writers appear much more limited in light of the analysis of social diversity. The data presented here shows a clear alignment between attitudes and social location. We need to re-interrogate arguments of autonomy between norms and social location. The empirical links between attitudes and location explored in this paper square with, and can contribute to, a broader based conceptual understanding of social diversity in which norms, subjectivities and extant social relations are mutually made (eg. Irwin 2003, Bottero and Irwin 2003, Himmelweit 2002, Blackburn et al 2002, Glucksmann 2000, Pedersen 1995). Any new openness of values and dispositions in the current era should not lead us to see a new loosening of the subjective and the objective, but rather requires us to more vigorously research their articulation. This would contribute to a necessarily processual account of social structure.

Notes

1 We are grateful to the ESRC for funding the work of The ESRC Research Group for the Study of Care, Values and the Future of Welfare, which is based at the University of Leeds (award M564281001) See http://www.leeds.ac.uk/cava for more information.

2 The achieved sample in each ward reflected the target characteristics and corresponds with the generalised description of the locales as given. So for example (and providing percentage figures for ease of comparison only):
In broadly white middle class North Leeds with 26 respondents: 73% of respondents were working; 27% were homemakers, 85% were white British, 12% were lone parents, 54% had a degree and 73% had A level qualification or higher. Based on current occupation, or last occupation held where the respondent was not working at time of interview: 67% of women were in RG social class I and II, and 14% in manual work.

In heterogeneous NorthEast Leeds locale with 37 respondents: 65% were working; 19% were homemakers, 60% were white British, 30% were Black British and Black Caribbean, 51% were lone parents; 22% had a degree and 49% had A level qualification or higher. Based on current occupation, or last occupation held where the respondent was not working at time of interview: 29% of women were in RG social class I and II, and 34% in manual work.

In broadly white working class 'West Leeds' locale with 39 respondents: 56% were working, 39% were homemakers, 100% were white British, 18% were lone parents, none had a degree and 13% had A level qualification or higher. Based on current occupation, or last occupation held where the respondent was not working at time of interview: 14% of women were in RG social class I and II, and 40% in manual work.

3 Letters describing the research and inviting people to participate were distributed to all parents of children in reception and years 1 and 2 at 6 schools in the study locales. Fieldworkers of HI Europe research consultancy then recruited and interviewed respondents. Interviewers additionally doorknocked as a way to recruit respondents in an additional school 'catchment', interviewing people with children at this seventh school.

4 The initial fixed choice options, listed on a showcard, were as follows:

  1. To have a paid job, or the prospect of a paid job, which I enjoy doing
  2. To be successful in my career
  3. To have the amount of time I want for spending with my child/children
  4. To sometimes have some time to myself to do just what I want
  5. To have enough money to bring up my child/children in the way I want

5 Pro maternal care is used as a shorthand to denote an attitude that care should be provided by mothers full time excepting during school hours for school aged children.

6 I have included column percentages in the tables for ease of comparison. I would emphasise that numbers are often quite low and I have therefore kept in the absolute numbers so this is clear.

7 The full breakdown is as follows:
A pre-school child is likely to suffer if his or her mother works. (Life as a Parent survey)

BSAS (2002) data shows that amongst women with degree level qualifications (and children aged 11 and under) 74% disagree with the statement, in contrast to 50% without qualifications.

8 Respondents were asked 'To which of these groups would you say you belong' and were recorded as: White-British (83), White - other (4), Black-Caribbean (3), Black-British (8), Black-other (0), Indian (1), Pakistani (2), Bangladeshi (0), Asian-other (1), Chinese (0), Other (0).

9 A referee queried how respondents were interpreting this question. It is to be presumed that they interpreted the question in terms of what women 'ought to do' rather than in terms of whether they should be 'allowed to' follow different courses of action. The quite high 'can't choose' response may be a consequence of many respondents resisting the compulsion implied by the question.

10 The full breakdown is as follows: S2: Do you think that women should work outside the home full time, part tme or not at all under these circumstances: when there is a child under school age (Life as a parent survey)

The smaller group with no qualifications are more likely than the 'lower level qualification' group to favour the work options, but this is due to the relatively high incidence in the former group of black women with a strong work orientation.

11 The analysis in the text presents responses to open ended questions amongst respondents favouring the part time 'solution' to the vignette described in section 4.2. It is pertinent to record some illustrative responses amongst those who favoured the full time 'solution'. Three such responses are as follows:

"If she really wants the job it is something important to her. Being a mother is not all that she is. She can still strike up a healthy balance of work and family even if she is working full time";
"Because its what she wants to do. She has a childminder to cover the hours and would be able to afford the childminder";
"Because she wants the job and if she gets it she can turn it down. It's a time in her life to start a career for the future".

References

ADKINS, L. (2003) Reflexivity. Freedom or habit of gender Theory, Culture and Society, 20, 6, 21-42.

BLACKBURN, R. M., Browne, J., Brooks, B. and Jarman, J. (2002) Explaining gender segregation Brit. Journ. Soc., 53, 4, 513-36.

BONOLI, G., George, V. and Taylor-Gooby, P. (2000) European welfare futures. Towards a theory of retrenchment, Cambridge: Polity Press.

BOTTERO, W. and Irwin, S. (2003) Locating difference: class, 'race' and gender and the shaping of social inequalities, Sociological Review 51, 4, 463-83.

BRADLEY, H. (1996) Fractured identities. Changing patterns of inequality, Cambridge: Polity Press

BRUEGEL, I. (1996) Whose myths are they anyway?: a comment Brit. Journ. Soc 47, 1, 175-77.

CROMPTON, R., Brockmann, M. and Wiggins, R.D. (2003) A woman's place... employment and family life for men and women, in Park, A. et al (eds) British Social Attitudes. The 20th Report, National Centre for Social Research, London: Sage.

CROMPTON, R. (2002) Employment, flexible working and the family Brit. Journ. Soc 53, 4, 537-58.

DEAN, H. (2001) Working parenthood and parental obligation Critical Social Policy 21, 3, 267-86.

DUNCAN, S. (2003) Mothers, care and employment: values and theories,

CAVA Working Paper, May (http://www.leeds.ac.uk/cava)

DUNCAN, S. (2000) Localities paper 2, choosing case study areas for strand 3 research - localities and neighbourhoods, CAVA WP 17A.

DUNCAN, S. (2000) Localities paper 3, variations in parenting and partnering at the local level CAVA WP 17B.

DUNCAN, S., Edwards, R., Reynolds, T. and Alldred, P. (2003) Motherhood, paid work and partnering: values and theories, Work, Employment and Society 17, 2, 309-30.

DUNCAN, S. and Edwards, R. (1999) Lone mothers, paid work and gendered moral rationalities, Houndmills: Macmillan.

EGERTON, M. and Savage, M. (2000) Age stratification and class formation: a longitudinal study of the social mobility of young men and women, 1971-1991 Work, Employment and Society, 14, 1, 23-49

GLUCKSMANN, M. (2000) Cottons and casuals. The gendered organisation of labour in time and space. Durham: sociologypress.

HAKIM, C. (1996) Key issues in women's work. female heterogeneity and the polarisation of women's employment, London: Athlone Press.

Hakim, C. (2000) Work-lifestyle choices in the 21st century. Preference theory. Oxford University Press.

HATTERY, A. (2001) Women, work and family. Balancing and weaving. Thousand Oaks and London: Sage.

HIMMELWEIT, S. (2002) Economic theory, norms and the care gap, or: why do economists become parents?, in Carling, A., Duncan, S. and Edwards, R. (eds) Analysing families: Morality and rationality in policy and practice. London: Routledge.

IRWIN, S. (2003) Interdependencies, values and the reshaping of difference: gender and generation at the birth of twentieth century modernity British Journal of Sociology 54, 4, 565-84.

IRWIN, S. (2000) 'Reproductive regimes: changing relations of interdependence and fertility change', Sociological Research Online 5, 1 http://www.socresonline.org.uk/5/1/irwin.html

IRWIN, S. and Bottero, W. (2000) Market returns? Gender and theories of change in employment relations British Journal of Sociology 51, 2, 261-80.

IRWIN, S. (1999) Resourcing the family: gendered claims and obligations and issues of explanation, in Silva, E. and Smart, C.C. (eds) The new family? London: Sage.

IRWIN, S. (1995) Rights of passage. Social change and the transition from youth to adulthood, London: UCL Press.

IRWIN, S. and Williams, F. (2002) 'Understanding Social Values and Social Change: the Case of Care, Family and Intimacy', Paper presented to the ESPRN Conference on Social Values, Social Policies, 29 August -31 August 2002, University of Tilburg, Netherlands.

JARVIS, L. et al (2000) Women's Social Attitudes: 1983 to 1998. A report prepared for the Women's Unit, Cabinet Office, National Centre for Social Research.

LINDLEY, J., Dale, A. and Dex, S. (2004) Ethnic differences in women's demographic, family characteristics and economic activity profiles, 1992 to 2002 Labour Market Trends

MCKIE, L., Bowlby, S. and Gregory, S. (2001) Gender, caring and employment in Britain Jnl. Soc Pol. 30, 2, 233-58.

MCRAE, S. (2003) Constraints and choices in mothers' employment careers: a consideration of Hakim's preference theory, Brit. Journ. Soc. 54, 3, 317-38.

MARKS, G. and Houston, D.M. (2002) Attitudes towards work and motherhood held by working and non-working mothers Work, employment and society 16, 3, 523-36.

MASON, J. (2000) Researching morality. CAVA Working Paper 14, http://www.leeds.ac.uk/cava/papers

RUBERY, J. (1997) Wages and the Labour market Brit. Journ. Ind.Relns 35, 3, 337-62.

PEDERSEN, S (1993) Family, dependence and the origins of the welfare state. Britain and France 1914-1945. Cambridge University Press.

SAUGERES and Duncan (2002) 'Choice and ideology in mothers' work-life decisions', paper presented to the ISA Conference, Brisbane, Australia, August.

SCOTT, J., Braun, M. and Alwin, D. (1998) 'Partner, parent, worker: family and gender roles', in Jowell, R. et al (eds.) British and European social attitudes. The 15th Report, Aldergate: SCPR/Ashgate.

SILVA, E. and Smart, C.C. (eds) 1999 The new family? London: Sage.

SKINNER, C. (2003) Running around in circles. Coordinating childcare, education and work. Bristol: Policy Press for the Joseph Rowntree Foundation.

THOMSON, K. (1995) Working mothers: choice or circumstance? In Jowell, R. et al British Social Attitudes: the 12th Report.

TOMLINSON, J. (2003) How are women's decisions about transitions in and out of part-time work affected by their work/life balance and changing policy and employment contexts? PhD thesis, University of Leeds.

WALBY, S. (1997) Gender transformations London: Routledge.

WILLIAMS, F. (1999) Good-enough principles for welfare Journ. Soc. Pol. 28, 4, 667-87.

WILLIAMS, F. (2001) Changing Families - Changing Values? Paper given to the National Family and Parenting Institute at the Launch of Parents' Week, October 2001, available online: http://www.leeds.ac.uk/cava/papers/changingfamilies.htm

WINDEBANK, J. (1999) Political motherhood and the everyday experience of mothering: a comparison of the child care strategies of French and British working mothers Journal of Social Policy 28, 1, 1-25.