How does parental time relate to social class in a Nordic welfare state?

Time intensive parenting has spread in Western countries. This study contributes to the literature on parental time use, aiming to deepen our understanding of the relationship between parental childcare time and social class. Based on time-diary data (2010–2011) from Norway, and a concept of social class that links parents’ amount and composition of economic and cultural capital, we examine the time spent by parents on childcare activities. The analysis shows that class and gender intersect: intensive motherhood, as measured by time spent on active childcare, including developmental childcare activities thought to stimulate children's skills, is practised by all mothers. A small group of mothers in the economic upper-middle class fraction spend even more time on childcare than the other mothers. The time fathers spend on active childcare is less than mothers’, and intra-class divisions are notable. Not only lower-middle class fathers, but also cultural/balanced upper-middle class fathers spend the most time on intensive fathering. Economic upper-middle and working-class fathers spend the least time on childcare. This new insight into class patterns in parents’ childcare time challenges the widespread notion of different cultural childcare logics in the middle class, compared to the working class.


Introduction
Cultural shifts in parenthood ideals have engendered a time-intensive, interactive approach to parenting in Western countries (e.g. Hays, 1996;McLanahan and Jacobsen, 2015). Parental care time, especially in children's early years, is often viewed as essential for children's current and future well-being and skills development (e.g. Gracia and Ghysels, 2017;Monna and Gauthier, 2008). New childrearing norms have increased the time all parents spend on childcare, but studies find that highly educated mothers and fathers consistently spend more time on childcare and pursue childrearing strategies aimed at developing children's social, cognitive, and linguistic skills (e.g. Altintas, 2016;Altintas and Sullivan, 2017;Bonke and Esping-Andersen, 2011;England and Srivastava, 2013;Gracia, 2014Gracia, , 2015Sani and Treas, 2016;Sullivan, 2010). Socioeconomic gaps in parents' time use have also widened over time in some countries (Altintas, 2016;Altintas and Sullivan, 2017;Sani and Treas, 2016). The advantages of intensive parenting accumulating among the well educated have raised concerns about children's 'diverging destinies' (McLanahan and Jacobsen, 2015). However, there are competing claims; some studies indicate considerable similarity between the more highly and less highly educated in norms of parenthood and a narrowing of educational disparities in parents' time use (cf. Cha and Park, 2021;Ishizuka, 2019). Moreover, the intensive parenting discourse has been criticised for representing a deterministic line of thinking, reproducing negative images of working-class families (e.g. Dermott and Pomati, 2016;Lee et al., 2014).
The present study aims to advance our understanding of parental involvement and social stratification, by studying social class divisions in a Nordic welfare state, Norway. The study moves beyond the way socioeconomic differences are often conceptualised in the existing literature on time use (e.g. Altintas, 2016;Gracia, 2015;Gracia and Ghysels, 2017). Education level is often used as a proxy for social classbut class as such, has received little attention in this literature (Hook, 2015;Sani and Treas, 2016). Our study is inspired by Bourdieu's (1984) concept of the class structure that positions classes and class fractions in the 'social space'. Our analysis contributes to theories conceptualising class as a complex social division shaping family life, and how access to various resources or capitals as Bourdieu referred to them, may influence the level and style of parents' involvement (cf. Gillies, 2009). Moreover, our nuanced approach to class, paying particular attention to intra-class differences within the middle class, questions the widespread notion of different cultural logics of childcare in the middle class compared to the working class.
Using time-diary data (2010-2011), we examine parental childcare time based on a concept of class that incorporates parents' amount and composition of cultural and economic capital. We ask: What is the relationship between parents' childcare time and their class positions based on differences in cultural and economic capitals? Are the relationships similar for mothers and fathersor do class and gender intersect? Childcare time is measured by the total time reported spent on active childcare, and the time spent on the sub-categories of routine and developmental care.

Parents' childcare time and socioeconomic differences
A recurring question in the literature on parental time concerns the processes generating socioeconomic differences, especially the relative role of structure versus culture (e.g. Monna and Gauthier, 2008). Socioeconomic status may influence parents' ideas about parenthood and their ability to meet expectations of increasing time use on childrearing (Lee et al., 2014). Here, we review how socioeconomic differences in parental time have been conceptualised and explained in the time use literature, in which the emphasis has been on educational attainment.
Differences between educational groups can be mediated through various mechanisms, both structural and cultural. Several studies find that the impact of education remains significant even after controlling for such mechanisms (e.g. Bonke and Esping-Andersen, 2011;England and Srivastava, 2013). Among both mothers and fathers, the highly educated spend more time on childcare than the less educated even though they are more likely to be employed and work full time and income constraints do not mediate education effects (England and Srivastava, 2013). The less highly educated invest less time in their children, their parenting appears to be more gender traditional and there is also evidence of this polarised pattern in more egalitarian countries, such as Denmark (Bonke and Esping-Andersen, 2011).
Women's education seems more decisive than men's, suggesting that childrearing values mostly lie with women, whose values influence both their own and their partners' time spent on childcare (Bonke and Esping-Andersen, 2011;England and Srivastava, 2013). However, the growing emphasis on more intimate fathering practices and participation in child-related activities may have influenced subgroups of men differently (Sullivan, 2010).
Although Sullivan (2010) notes that there are 'conceptual difficulties' with interpreting the level of education as a variable, the association between education and time spent on children is commonly interpreted as the differing attitudes, values, and ideologies of differently educated parents. Indeed, the dominant interpretation of educational differences in parents' childcare time is cultural, rather than economic. Polarised parental time according to education is viewed as a bi-product of diverging norms and parenting preferences (e.g. Bonke and Esping-Andersen, 2011). Highly educated mothers and fathers are assumed to be more aware of the association between investing time and producing 'quality children' and are more motivated to conform to norms of intensive parenting (e.g. Sayer et al., 2004). Hence, educational differences are associated with the varying quality of parenting cultures (e.g. Sani and Treas, 2016).
Evidently, cultural interpretations of educational differences in this time use tradition are based on assumptions about preferences, not direct measures. Lareau's (2003) model of a classed cultural logic of childrearingdistinguishing between an intensive 'concerted cultivation' model in the middle class, involving investments in children's educational performance, enriching leisure activities and verbal exchanges, and a less involved model of 'natural growth' among parents in the working classis frequently cited to support culturalist interpretations (e.g. Altintas, 2016;Altintas and Sullivan, 2017;England and Srivastava, 2013;Gracia 2014Gracia , 2015Gracia and Ghysels, 2017;Sani and Treas, 2016).
Lareau's construct emerged from an ethnographic study of families in the US in the 1990s. The study's theoretical basis is a 'straightforward, if partial, empirical application of Bourdieu's broader theoretical model' of class position, where individuals in different social locations are socialised differently, using strategies to maintain or improve their situation and their children's (Lareau, 2003: 275-278). Lareau defines the middle class as those with a college degree or a job requiring one, with the working class defined as those who do not. She notes that a small qualitative sample does not allow for many different class categories and admits that the two class categories concealed important variations within them, including income (p. 261). However, income did not seem linked to differences in childrearing methods among middle-class parents.
Lareau's distinction between concerted cultivation and natural growth also features prominently in studies of intensive parenthood with a broader emphasis than time use studies, but some question the idea of binary class divisions (e.g. Cheadle and Amato, 2011;Chin and Phillips, 2004;Irwin and Elley, 2011;Sjödin and Roman, 2018;Vincent and Maxwell, 2016). Class differences do not necessarily stem from parents' different wishes in cultivating their children's talents and skills, but from differences in access to economic and cultural capital (Chin and Phillips, 2004). Some studies indicate great similarity in parenthood norms and time use (cf. Cha and Park, 2021;Ishizuka, 2019). Similarities in time use across classes can result from school policies expecting all parents to participate in their children's homework (Reay, 2005). Moreover, the distinction between two types of parentingconcerted cultivation versus natural growthis too simplistic (Cheadle and Amato, 2011). A binary class/educational construct may neglect diversity within class fractions and homogeneity may have been overstated in the middle and the working class alike (Irwin and Elley, 2011;Vincent and Maxwell, 2016).

Parental time, economic, and cultural capitals
To move beyond this binary cultural logic, our approach is based on a model of class where the amount and composition of economic and cultural capital not only differentiates between classes vertically, but also horizontally among middle-class fractions (cf. Bourdieu, 1984). Those who primarily base their position on economic capital (e.g. proprietors, business people, employers, and industrialists in the economic sector) can be differentiated from those possessing more cultural than economic capital, such as professors, academics, or incumbents of senior positions in the cultural sector. They have cultural capital not only because of their formal credentials or titles, but also because of their cultural resources in a wider sense. Cultural capital may consist of skills such as self-assurance, manners, lifestyle, and taste that is highly regarded and rewarded in institutional settings (e.g. Lareau, 2003). Professionals, such as physicians and lawyers, tend to have a balanced composition of capital, high incomes, and higher education. Similar divisions may be found on the medium middle class levels. Differences between economic and cultural capital are expected to be smallest on the lowest class levels, among skilled and unskilled workers, as they have little of either cultural or economic capital.
A key idea in Bourdieu's account of class is that members of privileged groups or classes attempt to retain their privileges over time and develop reproduction strategies involving transmitting resources over generations (Bourdieu, 1996). Transmitting cultural capital within the family is particularly important, meaning that cultural capital is more easily acquired the more cultural capital the parents possess. Transmission of academic and social skills occurs in parent-child interactions and through parents' facilitation of children's activities. Economic capital is essential, especially because wealthy families can purchase domestic help, freeing up time to transmit cultural capital (Bourdieu, 1986). More recent accounts have also emphasised the importance of economic capital to buy private tuition or developmental and high-status activities (Reay, 2005).
According to Bourdieu, class divisions are manifest in patterns of consumption and the division of labour between the genders (Bourdieu, 1984: 382-386). He argues that gender differences lessen as one moves up the social hierarchy and are likely to be smallest in classes with most cultural capital. Women in this class usually have little spare time and devote their time to childcare and transmitting cultural capital rather than to traditional domestic labour, which takes up more time for working-class women (Bourdieu, 1984: 185-186). Despite the comparatively small gender differences in the classes possessing most cultural capital, Bourdieu writes that it is primarily the mother's time that is freed up by economic capital. More recent studies have also emphasized gender differences in childrearing practices. According to Lareau, both parents in the middle classes participate in the concerted cultivation of their children, but mothers take on the greatest burden (Lareau, 2003: 51-52). Reay (2005) argues in her study that while there were hardly any differences by social class in the importance attached to schooling, children's schooling was primarily mothers' responsibility and there was little evidence of fathers' involvement.

The Nordic welfare state context
This description of gender differences does not fit well with the most recent accounts of the more symmetrical earner-carer family model which has developed in the Nordic countries.
Class differences and their impact on families and individuals can be modified by state policies (Crompton, 2006). The welfare state has a direct impact on parenthood and parental time by giving parents access to certain resources and services and an indirect one by contributing to changes in parenthood norms (Hook, 2010). These norms may be gendered and classed.
The Nordic earner-carer family policy model, expecting mothers and fathers to be both providers and carers, comprises generous parental leave and universal access to childcare and after-school care. In Norway, parental leave is 49 weeks at full pay, with 15 weeks (10 weeks in 2010) earmarked for fathers. Many children spend large parts of their day outside the home, nine out of ten children aged 2-5 are enrolled in childcare services and 60-80 percent of children in the lower grades are enrolled in after-school care, which will reduce variation in parents' time with children. Family policies in the Nordic countries have been found to increase economic opportunities, particularly for women without university-level education, thereby diminishing both gender and class differences (Korpi et al., 2013).
High levels of employment among mothers, coupled with compressed wage structures, probably reduce economic constraints on parents' time use (Bonke and Esping-Andersen, 2011). Still, more women than men work part-time, though part-time work among mothers has declined significantly (Ellingsaeter and Jensen, 2019). The Nordic welfare states also stand out as distinct regulatory environments, with regulated working hours and short work hour cultures (standard full-time work in Norway is 37.5 h/week) (Sayer and Gornick, 2012).
Gender differences in time use have diminished in the Nordic countries, and parenthood now affects men's and women's time use in childcare in a more similar way than previously (Neilson and Stanfors, 2013). A comparative study classified Norway as a 'high' childcare welfare state (based on 2000 data): Parents put in relatively high amounts of time compared to other countries, while the gender gap was relatively small (Sayer and Gornick, 2012). Although some qualitative studies have observed classed childrearing practices in line with 'concerted cultivation' versus 'natural growth' in Norway (e.g. Stefansen and Aarseth, 2011;Stefansen and Farstad, 2010), the earner-carer model has led to more standardised work-family adaptations, involving shared breadwinning and responsibility for childcare, thereby diminishing class differences (Bjørnholt and Stefansen, 2019).
Access to universal childcare facilitates mothers' employment across socioeconomic groups and it also equalises children's access to pedagogical resources in their early years. Norwegian mothers, across educational levels, consider childcare services as the best form of care for children aged two to five (Ellingsaeter et al., 2017), suggesting cultural homogenisation in childrearing ideals. Moreover, this childcare model also seems to impact on parents' cultural resources by transmitting a parenting style of concerted cultivation, influencing the overall parenting culture: for example, a Danish study has demonstrated how highly educated preschool teachers guided parents in cultivating their children's academic and social skills at home (Dannesboe et al., 2018). A Swedish study of parents' views of children's extracurricular activities found striking similarities between working-and middle-class parents, suggesting that preschools play a significant role in shaping a common parenting style (Sjödin and Roman, 2018). Hence, the simple distinction between middle class and working class cultural logics in understanding parental time input does not seem adequate for the Nordic context. However, policies enabling intensive parenthood may interact with gender and class. Studies suggest that parenthood and parental rights enjoy a strong position in Norwegian working life but expectations are gendered, for example, taking long parental leave is expected for mothers but not for fathers, who typically take only the reserved father quota (Nordberg, 2019). Studies have established that Nordic fathers spend more time on childcare than fathers in other welfare states (e.g. Craig and Mullan, 2011;Hook and Wolfe, 2012;Neilson and Stanfors, 2014). Nonetheless, there may be limits to the reach of a new fatherhood model. While Norwegian employers are generally positive towards father-friendly policies, such as the father quota (Hagen, 2017), there are indications of labour-market niches of a long-hours culture, e.g. within the financial sector. Moreover, elite corporate culture seems to preserve traditional business masculinity ideals, where the economic upper class emphasises tradition, strength, and competition (Halvorsen and Ljunggren, 2020). Moreover, a qualitative study of parents where the father belonged to the economic upper-class fraction (e.g. managers, executives in finance) has demonstrated a more complementary family model, where the father was devoted to his work, while the mother occupied herself with the family and everyday life (Aarseth, 2021).
In summary, parents' access to economic and cultural capitals in the Nordic welfare states aids the homogenisation of intensive parenthood, but actual adaptations may vary between class fractions.

Data
The analyses are based on the latest Norwegian time-use survey conducted in 2010-2011. We rely on a subsample of mothers and fathers in heterosexual unions who had at least one child under 16 years in the household. The respondents kept a diary for two consecutive days, and the sample was spread evenly throughout the year. The unit of analyses is a single day. Our first subsample comprises 891 diary days by employed mothers, and a small group of non-employed, approximating stay-at-home mothers. The second sub-sample comprises 867 diary days by employed fathers. We omitted the very few stay-at-home fathers from the sample (6 diary days), and for both genders, we omitted the few students, disability pensioners, and primary-sector workers (28/22/4 days for mothers and 10/12/24 days for fathers). These are too few to be analysed as separate class categories and cannot meaningfully be included in any of the categories we look at.
The diaries had fixed ten-minute intervals for which participants wrote down their most important activities, which were then coded. The dependent variables in the analyses and the time spent on various childcare activities are taken from the time diary. Demographic and socioeconomic background information for the respondent and partner was collected by an interview and from public registers. We have no diary information about the partner's temporal input. The respondents' social class, our principal independent variable, as well as the other independent variables, are based on the interview and register data.
The data have been weighted to adjust for bias in the response rate. The total net sample comprised 3975 persons (aged 9-79) and the response rate was 48 percent (for further details, cf. Holmøy et al., 2012).

Dependent variables: Time spent on childcare activities
We examine parental time spent on childcare based on measures widely used in the field (e.g. Altintas and Sullivan, 2017;Gracia and Ghysels, 2017;Hook and Wolfe, 2012) and present parents' time use in average number of minutes per day.
Active childcare is the key measure of parental childcare time, referring to the time slots when childcare is the respondent's most important activity, e.g. nursing and assistance, playing, talking, reading aloud, and escorting children to and from various arrangements. To capture differences in parental styles, an important distinction in the literature is made between two subcategories: routine childcare, which comprises nursing/physical care and assistance, such as feeding and dressing children, putting them to bed, and developmental childcare which includes activities regarded as important for developing children's social, cognitive, or linguistic skills, such as playing, talking, reading aloud, helping with homework, and escorting children to and from activities. These latter kinds of activities are typically considered an indicator of concerted cultivation (for instance Altintas, 2016;Cha and Park, 2021). However, there is not necessarily a sharp distinction between the two types of childcare activities in developing children's skills. For instance, physical care and assistance are typically accompanied by chatting and other forms of communication. One example is parents putting their children to bed (routine childcare), a situation when talking and chatting is likely to be going on.

Independent variables: Classification of class and control variables
Our classification of class seeks to identify occupations on different levels and in different sectors, i.e. those with more cultural capital than economic capital, those with more economic capital than cultural capital, and those with a more balanced composition, e.g. professions. The classification is derived from the Oslo Register Data Class scheme (ORDC) (cf. Hansen et al., 2009). Because of the small sample size, we have to use a simplified version distinguishing between four class fractions: 1) working class (skilled/unskilled workers), which is the reference category in the multivariate analyses, 2) lower-middle class, 3) economic upper/upper-middle class (hereafter, economic upper-middle class), 4) cultural and balanced upper/upper-middle class (hereafter, cultural/balanced upper-middle class). In this simplified version, professions are grouped with the cultural fraction as they will presumably be more similar to them than the economic fraction with respect to cultural capital.
The classification is based on four-digit ISCO codes shown in the online Appendix 1. Occupational titles often suffice to assign people to classes, e.g. primary schoolteacher, machine operator, shop assistant, doctor, lawyer, or professor. The main exceptions to this are for specific titles in the business sector, which may be ambiguous with respect to class level, such as the title 'manager', which is found on various hierarchical levels. Large working-class occupations include shop assistants, cleaners, nursing assistants, warehouse workers, and drivers, while pre-school and primary school teachers, office workers, and police represent large occupational groups among lower-middle class. In the economic upper-middle class, large occupations include directors and chief executives, board members, managing directors, and technical and commercial sales representatives, while examples of large occupational groups in the cultural/balanced upper-middle class are teachers of secondary and university-level education, physicians, and journalists.
For mothers, we also include a fifth category of non-/home workers. The distribution for the various categories is shown in table 1. The diverging distributions for mothers and fathers reflect the gendersegregated labour market in Norway, with few women in the economic upper-middle class.
To isolate the relationship between the parents' class and childcare time, we control a number of demographic factors (age of youngest child, number of children, respondent's age) and employment factors (the respondent's weekly working hours, the partner's class position and weekly working hours) that are often correlated with parental time (see table 1 for the distribution of variables). We do not include working time schedule or purchase of cleaning services, as these variables in the analyses did not affect our main results (results are available upon request). We do not control the level/type of education and income because of the very high levels of multicollinearity between these variables and social class, which might create bias in the analyses and also conceal the effect of the class variable. Specific educations are required to enter some of the occupations in the cultural and balanced upper/ upper-middle class.

Analytical procedure
We analysed mothers and fathers separately, starting with bivariate analyses of the relationship between class position and time spent on the three measures: total, routine, and developmental childcare time (table 2). We then ran multivariate regression models for each variable, controlling for demographic and employment factors. We started with a model where we controlled for the age of the youngest child, the number of children, respondent's age, and day of the week. In the next model, we added the respondent's weekly working hours in order to explore whether this affected the association between the parents' class position and childcare time. Finally, in the most comprehensive model, we added the partner's class position and weekly working hours to see whether this impacted the effect of the respondent's class on her/his childcare time. Results from the first two models are provided in the online Appendix 2 (tables 1 (mothers) and 2 (fathers)). For the most comprehensive model, results are reported below in tables 3 (mothers) and 4 (fathers). We use standard ordinary least squares (OLS) regressions in the multivariate analyses. Although time-use data often contain a high number of zero observations, many researchers recommend OLS regressions because zero observations are not usually a result of censoring or truncation, but stem from the respondent not doing a certain activity as per the diary (for instance Altintas, 2016;Cha and Park, 2021;Hook and Wolfe, 2012;Stewart, 2013). To account for repeated observations (two days per respondent), we used robust standard errors. As the focus is on class differences in parents' childcare time, we comment primarily on the effects of the respondent's class in the multivariate analyses. We use the weighted data in the analyses and report the number of observations unweighted.

Results
Compared to fathers, mothers spend more time on active childcare, both routine and developmental, according to the descriptive statistics (table 2). On average, mothers spend 102 min per day on active childcaremost on routine childcare, 57 min; and 45 min on developmental activities. Fathers spend on average 61 min per day on active childcare; about three-fifths of mothers' time use, split rather evenly between routine and developmental care.
Turning to class differences among mothers, when all middle-class mothers are compared to workingclass mothers, the descriptive statistics reveal striking similarity in the time spent on childcare (table 2). Middle-class mothers spend 94 min per day on active childcare compared to 98 min by working-class mothers. Middle-and working-class mothers also spend about the same amount of time on routine childcare, 51 and 50 min per day, respectively, and on developmental childcare, 48 and 43 min per day, respectively.
However, some notable differences appear according to the descriptive statistics for mothers in the three middle-class fractions (table 2). While mothers in two middle-class fractionscultural/balanced upper-middle and lower-middle classare quite similar to working-class mothers, the rather small group in the economic upper-middle class fraction spends more time on active childcare, both routine and developmental. Mothers in the economic upper middle-class fraction spend 121 min on active childcare per day, compared to 92 min among mothers in the two other middle-class fractions. As might be expected, home-working mothers spend more time on active childcare than both working-and middleclass mothers, but this applies only to the routine type of care. Moreover, gender and class intersect. When comparing mothers and fathers in the various class fractions, the smallest gender difference is found in the lower-middle class fraction, while the largest is in the economic upper-middle class fraction. The multivariate analyses for mothers confirm that while mothers in the cultural/balanced uppermiddle and lower-middle class fractions do not differ significantly from working-class mothers in their time use, those in the economic upper-middle class fraction spend more time on total childcare than working-class mothers. In the most comprehensive model (table 3), the difference amounts to 37.9 min per day (p < 0.05). As for employed mothers, all the multivariate models provide approximately similar results for the association between mothers' class position and the time spent on childcare. Although there is a negative relationship between mothers' working hours and time spent on childcare, and a positive relationship between her partner's working hours and her childcare time (see table 3 and online Appendix 2, table 1), this does not alter the connection between mothers' class and time on childcare. However, the difference in childcare time between the small group of home-working mothers and mothers in the other class fractions is greatly attenuated in the multivariate analyses compared to the bivariate analyses (table 2), mainly because home-working mothers typically have younger children.
Regarding the two subcategories of active childcare, mothers in all the three middle-class fractions and home-working mothers spend more time on routine childcare than working-class mothers. In the most comprehensive model (table 3), the coefficients remain statistically significant, although the coefficient for the cultural/balanced upper middle class is only marginally significant (p < 0.1). As for time spent on developmental childcare, none of the middle-class fractions differs significantly from the workingclass mothers, and the same is true for home-working mothers.
Turning to a closer examination of intra-middle class differences in mothers' childcare time, all the multivariate models show that mothers in the economic upper-middle class fraction spend markedly more time on total active childcare than do mothers in the two other middle-class fractions (table 3 and  online Appendix 2, table 1). Statistically significant results for intra-middle class differences are denoted by superscripts in the tables and are explained below them. According to the most comprehensive model (table 3), the difference in childcare time between mothers in the economic upper-middle class and those in the cultural/balanced upper-middle class amounts to about 35 min per day (p < 0.05). The difference between mothers in the economic upper class and those in the lower-middle class amounts to about 28 min per day and is only marginally significant (p < 0.1). As for the two subcategories of childcare, there are no statistically significant differences between the three class fractions. We note, however, that the small number of observations among mothers in the economic upper-middle class in the analysis sample (N = 47 days) makes reaching statistical significance at conventional levels difficult. As for class differences among fathers, the descriptive statistics (table 2) show that all middle-class fathers taken together spend more time on average on active childcare (total) than working-class fathers (66 vs 53 min per day). The pattern applies to both routine (34 vs 26 min per day) and developmental childcare (32 vs 27 min per day). However, when we distinguish between different middle-class fractions, the descriptive statics show considerable variation. Fathers in the lower-middle class spend by far the most time on active childcare (total) (80 min per day), followed by fathers in the cultural/balanced upper-middle class (66 min per day). However, fathers in the economic uppermiddle class spend about the same amount of time on active childcare as working-class fathers (47 min per day). Similar patterns apply to both routine and developmental childcare. Somewhat moderated, most of the descriptive results hold in the multivariate analyses, although the results differ slightly across the various models (table 4, and online Appendix 2, table 2). Cultural/ balanced upper-middle class and lower-middle class fathers spend more time on active childcare than working-class fathers. The coefficient for the cultural/balanced upper-middle class fraction is not statistically significant in the most comprehensive model (table 4), but when the controls for the partner's class and working hours are excluded, it is significant at the 0.05-level (online Appendix 2, table 2). As for mothers, the results are strikingly similar across the multivariate models (table 4, and online Appendix 2,  table 2), meaning that differences in childcare time between middle-and working-class fathers are not linked to differences in fathers' working hours or the partner's class position or working hours.
Looking more closely at intra-middle class differences among fathers, the multivariate analyses confirm notable differences among middle-class fathers (see the notes below the tables for information on statistical significance for intra-middle class differences). According to all models, fathers in the cultural/balanced upper-middle class and those in the lower-middle class spend more time on active childcare (total) than economic upper-middle class fathers. In the most comprehensive model (table 4), the difference amounts to about 20 and 26 min per day (p < 0.01), or about 140 and 182 min per week, which, arguably is quite large, after adjusting for both fathers' working hours and the partner's class and working hours. The intra-middle class differences apply to routine childcare and in part also to developmental childcare. As for routine childcare, according to table 4, fathers in the cultural/balanced uppermiddle class and those in the lower-middle class spend on average about 13 and 15 min more on childcare per day, respectively, than economic upper-middle class fathers (p < 0.01/p < 0.05). Looking at developmental childcare, we find that the corresponding differences amount to about 8 and 11 min per day, with the latter being marginally statistically significant (p < 0.1), while the former one is not significant.

Discussion and conclusion
This article contributes to the literature on intensive parenting and socioeconomic differences in parental time use, by addressing shortcomings in the dominant approach in the time use literature, based mainly on educational attainment. We provide new insight into the role of social class in generating differences in parents' childcare time, by applying an approach to social class that links mothers' and fathers' cultural and economic capitals. Our study suggests that the prevalent notion of different cultural childcare logics in the middle class compared to the working class is inadequate in a Nordic welfare state context. Among mothers, the broad picture is that different amounts and combinations of economic and cultural capitals do not matter much for the time they spend on childcare. Class differences in the total time spent on childcare activities, and the time spent on the sub-category of developmental childcare activities, thought to reflect the concerted cultivation logic of the middle class, are negligible. Time use indicating concerted cultivation is thus not a class marker. However, working-class mothers spend less time on routine childcare than middle-class mothers. But, as argued above, the distinction between the two subcategories of childcare is not necessarily very sharp. Nevertheless, the time use literature's preoccupation with developmental childcare has largely overlooked potential differences in routine care, which is an important dimension of family relations and childrearing. Routine care thus warrants further exploration. See for example Kalil et al. (2012), who examine how mothers' composition of routine/basic care and other forms of care change according to children's age, varying with mothers' educational levels. Moreover, one class fraction, the small group of mothers in the economic upper-middle class, stands out in intensive motherhood practices, spending even more time on childcare activities than other mothers. We can only speculate why, though possibly mothers in this class fraction work in the competitive business/finance sector, which may lead to extra efforts to provide the optimal investments perceived as necessary for their children to succeed.
In general, these findings suggest that intensive mothering ideals have been incorporated by all mothers. Mothers' rather equal time input on childcare indicates that the Nordic welfare state context, involving access to both economic and cultural capitals through high maternal employment rates, universal preschool attendance, and compressed wage structures, has contributed to the homogenisation of motherhood practices. Structural change in mothers' labour market participation and preschool enrolment have diminished differences both in mothers' time available for childcare and the time children are available for mothers' care (see Bianchi, 2000).
By contrast, differences in economic and cultural capitals between fathers do matter. Although intensive parenthood seems to have spread considerably to fathers too, who contribute substantial care time to children, their time use is still less intensive than mothers', regarding both total childcare time and routine and developmental childcare. Lower-middle class fathers particularly, but also cultural/balanced uppermiddle class fathers, represent the vanguards of intensive fatherhood, especially by spending more time on routine childcare. This distinguishes fathers in these middle-class fractions from fathers in the economic upper-middle class fraction, who put in significantly less care time, as do working-class fathers. The spread of intensive parenthood is thus more uneven among fathers. The 'new father' seems to be found among fathers with middle or high levels of cultural capital. The position of fathers in the lowermiddle class may be enhanced by more standardised work routines and less competitive jobs facilitating intensive fatherhood. For example, one study finds that in a male-dominated occupation such as the police, taking parental leave is regarded as part of a police officer's social responsibility (Nordberg, 2019). By contrast, barriers may exist in other jobs where institutional and cultural logics may challenge caring fatherhood (e.g. Aarseth, 2021;Nordberg, 2019). The intra-middle class differences shown may indicate the perseverance of a 'business masculinity' in the business and financial sector (Halvorsen and Ljunggren, 2020), while a male breadwinner culture may still exist in the organisation and expectations of traditional male working-class occupations. Moreover, we note that fathers' childcare time in various class fractions is only modestly affected when the mothers' class is accounted for, suggesting that fathers' childcare practices in Norway are less influenced by their partners' socioeconomic characteristics compared to some other countries (e.g. England and Srivastava, 2013). Among other things, this may suggest that relatively time-intensive fatherhood is becoming the new normal in the cultural/balanced middle class fractions, i.e. more independent of external factors, for example, 'pushing' mothers.
The patterns revealed imply that intra-class differences and their intersection with gender will be vital to the further study of the role of social class in parental time use and its societal foundations. However, our study has some limitations. Although diary data from time-use surveys are recognised as the best source of parents' time allocation, small sample sizes limit the study's subgroups. Our findings cannot be generalised to child-rearing practices in families other than heterosexual; two-parent families and immigrant families are under-represented. Also, the class model had to be simplified, so we were unable to capture potential class differences at a more detailed level. Furthermore, long intervals between data collections mean that our time-use data are a decade old. The past decade has witnessed changes both in family patterns and policies, and whether this has influenced gender and class divisions in parents' time use could be pursued further when new data become available.
Finally, it is necessary to emphasise that similar time use does not imply that the family no longer is a site of social reproduction. In time use surveys, there may be unobserved class disparities in parents' time with children. While intensive parenthood norms may lead to similar time input, differences in children's outcomes may still be considerable because parents have different prerequisites for supporting their children, for example, helping children with their homework (see Reay, 2005). This aspect of parental time use in the Nordic context could be an important topic of future studies.

Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article. This study was supported by the Department of Sociology and Human Geography, University of Oslo, and CORE Centre for Research on Gender Equality, Institute for Social Research, Oslo.