Child Indicators Research

, Volume 3, Issue 1, pp 1–10

Child Well-Being and Intergenerational Inequality

Editorial

During recent decades the research into children’s well-being, in a broad sense, has developed into a mature research field with conferences, wide networks, and also a new Journal, Child Indicators Research. We have witnessed a timely and necessary focus on young people’s living conditions at large, a growth of specialized studies in various aspects of young people’s lives, and recent attempts at summarizing their well-being in a multidimensional and international perspective (Bradshaw et al. 2006; UNICEF 2007; Bradshaw and Richardson 2009).1 This development has, as outlined by Ben-Arieh (2008), several notable features: The young have increasingly been seen as active purposeful beings; their conditions here-and-now have been emphasized—‘well-being’ has more and more complemented the traditional perspective of ‘well-becoming’; focus has been moved from young people’s problems—or even young people as a problem—to positive outcomes, and enabling conditions; and young people have to an increasing extent been used as informants.

This development has set new standards for studies of young people, and the research reported in this volume is anchored in the view of young people as rapporteurs of their own conditions. Several of the contributions also stress their role as conscious actors, capable of benefitting from resources of various kinds. However, resources may or may not be available. The young—just like adults, but probably to a larger extent—are dependent on opportunities for realizing their goals and ambitions. These opportunities hinge on resources and actions of others: ranging from the ‘inner circle’ of young people’s environment, particularly the family, to ‘outer circles’ like the neighbourhood and historical time (Bronfenbrenner 1994). Much research has established the association between characteristics of the family of origin and child outcomes—what could be called inequality of opportunity—and the study of intergenerational processes leading to such inequality is crucial for understanding why some children fall behind.

Research focusing on intergenerational processes is prevalent within many disciplines and is characterized by different theoretical stances. This volume of CIR is concentrating on what could be called “structural inequalities”, i.e., differences that depend on durable and fundamental characteristics of parents, such as their educational qualifications, their occupations and social class, their ethnic origin, their incomes, and their marital (or family) status; characteristics that form, via different mechanisms, an essential part of the proximate opportunity structure of children.

1 The Need for Plausible Theories

Particularly within social stratification research, studies of income or social mobility between generations have attracted a lot of attention over the recent decades (for reviews, see Björklund and Jäntti 2009; Breen and Jonsson 2005). There is a strand within this research tradition that focuses almost exclusively on intergenerational associations (e.g., between occupational or income origins on the one hand, and destinations on the other) as indicators of societal conditions. In particular, the interest focuses on inequality of opportunity as indicated by the association between parents’ and children’s socioeconomic positions. Change over time or differences between countries in inequality of opportunity are often in turn taken as implications of institutional change or variation, or as the results of policy change or reforms (e.g., Erikson and Goldthorpe 1992).

The challenge for social stratification research into intergenerational mobility is to pay more attention to understanding why children tend to end up in their parents’ footsteps, and more generally to form theories about inequality among the young. And, indeed, here is the obvious connection point with research into the well-being of young people; for it seems also to be of utmost importance to account for variations in their well-being by scrutinizing the role of parents and household conditions during childhood. For example, we would need to know what processes lie behind the lingering importance of parental socioeconomic position and the well-being of their children, as well as their later attainment in school and in the labour market. The same obviously goes for ethnic origin and family type (such as parental divorce and remarriage).

Much research has been devoted to studying school processes and theories in connection to the family–school nexus, emphasizing mechanisms such as social capital and the transmission of skills, resources, and aspirations (e.g., Entwistle et al. 1997; Erikson and Jonsson 1996; Portes and Rumbaut 2001). Besides a literature within developmental psychology, which normally does not take the intergenerational perspective, there are however still many gaps in our knowledge about younger children—although several studies do show that inequalities in “ability” show up already during the pre-school years (e.g., Feinstein 2003; Becker 2009).

While previous research in intergenerational processes has been much engaged in cognitive abilities in children and their educational attainments (e.g., Sewell et al. 1980), there has been a recent upsurge in other types of explanations. For example, adoption data and sibling data have been used to estimate the proportion of intergenerational income correlations that are due to ‘nature’ and ‘nurture’ (e.g., Björklund et al. 2005), showing that genetic factors may well be behind a non-trivial part of what turns up as a parent–child correlation (some of which are of course transmitted via cognitive ability). Also, non-cognitive factors such as personality traits have been put forward as a transmitter of socioeconomic status across generations, though it is (just like for IQ) questionable whether the intergenerational correlations are strong enough to mediate much of the traditional mobility measures (such as occupational or income mobility) (e.g., Osborne Groves 2005, Table 7.2).

A very general theory explaining why children from advantaged origins show higher well-being sees parents’ interaction with their children as essential. This is of course the pillar of socialization theory (either in the form of role modeling or internalization of values, aspirations, and norms), but also the basis of economic family investment models (Becker and Tomes 1986). Parents with more resources (e.g., higher education and income) may either spend more time with their children—‘invest’ more, as economists would say, perhaps by trading child quantity for ‘quality’—or they may be more effective in transmitting skills and aspirations because of their superior resources. Others would prefer to think of parent–child interactions as ‘involvement’ rather than investments, and to care about the quality of the relationship (e.g., the emotional content, see Pungello et al. 2009) and the form of interaction (such as the often noted effect of reading to children, cf. Brooks-Gunn and Markman 2005). Nonetheless, a common frame of reference is the parent–child interaction, and how that differs according to parental characteristics and resources. This, in turn, is something that would demand more theoretical thought and ‘process’ data—such as time use data (e.g., Stafford and Yeung 2005; Bonke and Esping-Andersen 2009), or on-site observations (e.g., Lareau 2002).

2 The Need for Good Data

Given the great and renewed interest in intergenerational processes, one would perhaps think that there is an abundance of studies, done in many countries, that would aid in an understanding how parental characteristics and resources influence their children’s well-being. True, there are many commendable cross-country studies that aim at such comparisons (e.g., Jäntti et al. 2006; Currie et al. 2007; Shavit et al. 2007; Breen et al. 2009; Jonsson et al. 2009; Elo 2009). However, while the results of such comparisons are vital for an understanding of intergenerational processes, most of them use single indicator outcome variables (often attainment variables), while studies that cover many facets of young people’s lives in an intergenerational perspective are rare indeed. Comparative studies that take a multidimensional approach in analyzing young people’s well-being, on the other hand, are not able to cover the intergenerational dimension (e.g., Bradshaw and Richardson 2009).

Most large comparative studies also suffer from what we could call “the worst common denominator problem”—in order to achieve comparability they normally use relatively crude measures and few outcomes and they tend to align to the country or data-set with the crudest definitions and most limited data. The same data problem often occurs on a national basis when we want to trace changes over time, for example, in trying to monitor how children’s well-being develops over time. Such monitoring is not only required by the United Nation’s Convention on the Rights of the Child (CRC), but also a litmus test of the level and progress of the welfare of any modern nation. It takes an extra-ordinary effort to come up with (again, often crude) indicators that are comparable over time (see, for the US, for example, race differences analyzed by Land et al. 2007).

Comparability of data is of course not the only issue in studying young people’s well-being, not even the only data problem. As far as data are concerned, there exist today several excellent international data sets that provide indicators of young people’s health, school achievement, and other aspects of their lives, such as household poverty (e.g., PISA, TIMMS, HBSC, SILC). The main weakness of these data, for the purposes of studying intergenerational processes, is that they either do not provide measures of parental characteristics of sufficient reliability (or, sometimes, validity), or they lack measures of child outcomes as reported by children themselves. Although we traditionally faced problems with the latter (e.g., Ben-Arieh 2005), we should not underestimate the former. Measuring structural background variables, except perhaps ethnicity and family structure, pose formidable problems, particularly with younger children. They are unlikely to give sufficiently precise information on parents’ income, occupation, or education (and errors can be expected to vary between countries and over time, e.g., with changes in the educational system). Crude indicators such as home or car ownership could be used, but are far from ideal.

Most of the available data sets also do not contain a longitudinal element. This makes it difficult to study how well-being develops over childhood and adolescence, and whether the family background impact on changes over children’s age. Moreover, longitudinal (panel) studies give us the opportunity of approaching interpretations in causal terms, as we can study whether changes in parental or household characteristics are associated with changes in child outcomes—there are other ways of identifying causal models but a large panel data set is an excellent start (e.g., Wooldridge 2002).

So, could we muster any data-set that both have reliable measures of parental characteristics (preferably reported by parents themselves, or taken from high-quality administrative registers) and relevant indicators of their children’s well-being (reported by children themselves, or taken from school records etc.)? There are at least four types of data-sets that could do it. First, comparative large-scale studies could include parental questionnaires to supplement the child reports (a strategy actually followed in some countries already now). Second, longitudinal cohort studies, the queen of all data sets, would seem especially relevant for intergenerational analysis. A cohort study that starts at (or preferably some time before) birth and then follows children up to adult age, with repeated measures of outcomes, is the best tool for describing, monitoring, and analyzing young people’s well-being. It is an ideal design for studying intergenerational processes, especially if recurrent data collections are done of parents’ and households’ characteristics, as well as with day-care personnel and teachers. To monitor change over time, one needs ideally repeated cohort studies, which are available for England, with the 1946, 1958 and 1970 cohorts as early studies, and the Millennium cohort as the most recent (Dex and Joshi 2005). The future of cohort studies is difficult to predict, however, as they are expensive and require collaboration between researchers in different disciplines (which should be a strength, but may well turn out to be an obstacle).

Third, there are longitudinal data-sets that are not as large as cohort studies, but still large enough so one can generalize results to the population. Some data-sets that contain direct information from both parents and children are focused on a limited set of outcome variables, but extremely valuable for those (e.g., NELS, NLSY, and Add-Health in the USA; UGU in Sweden). One of the most used general data-sets is PSID in the USA, where it is possible to follow consecutive cohorts (and which also has a child supplement). Two European studies with a general content that come closest to the ideal are BHPS in England and Child-LNU/ULF in Sweden. These surveys both interview adults and children (11–15, and 10–18, respectively), thus maximizing the accuracy of information and extending the scope of possible child outcomes (to areas such as social relations, psychological well-being, and political resources, for which indirect indicators are difficult to gather). These studies (like the PSID) profit from the fact that they take an existing sample of adults as a point of departure, and then in addition interview the children in the household, reducing the marginal costs for child interviews dramatically. The great promise they hold for the future is to point to a simple, inexpensive, and effective method of collecting data on parents’ characteristics and their children’s well-being at the same time: Integrating the approach of BHPS or LNU/ULF into EU/SILC would generate rich data for young people of the ages 10–18, approximately, from almost 30 European countries.

Fourth, there are more specialized studies, typically with smaller sample sizes. These might not always be possible to generalize, and do rarely cover many dimensions of young people’s well-being, but are likely to be of great theoretical value as they can be designed for a special question.

3 Contributions to this Volume

The contributions to this special volume all make headway in the areas discussed above. They all use data collected from children, with parental information (in all cases but one), to study intergenerational processes and inequality among the young. Summary statistics of the content and the data used in the articles are found in Table 1.
Table 1

Summary of studies in the special issue of Child Indicators Research, on Child Well-being and Intergenerational Inequality

Author

Type of inequality

Outcome

Child age

Data set

N of cases

Country

Biedinger

Ethnic

Cognitive differences

3–4

PEMC

1,057

Germany

Becker

Ethnic

Language development

3–4

PEMC

1,064

Germany

Jonsson and Östberg

Ethnicity, origin class, family type

Multidimensional, level-of-living

10–18

Child-LNU

1,304

Sweden

Robson

Family type

Happiness, self-esteem

11–15

BHPS

15,000

England

Låftman

Family type

Social support, health, material conditions

11/13/15

HBSC

95,335

24 countries

Smyth et al.

Origin class & education

Reading and Math test scores

9

Growing up in Ireland

8,578

Ireland

Jones and Hansen

Origin education, social housing, income, ethnicity

Cognitive differences

Up to 5

MCS

8,131

England

Abbreviations and home-pages for the data-sets

PEMC Preschool Education among Migrant Children www.mzes.uni.mannheim.de

LNU The Swedish Level-of-Living Survey www.snd.gu.se, www.sofi.su.se

BHPS The British Household Panel Survey www.iser.essex.ac.uk/survey/bhps

HBSC Health Behaviour in School-aged Children (WHO) www.hbsc.org

Growing up in Ireland www.growingup.ie

MCS Millennium Cohort Study www.cls.ioe.ac.uk

It is easy to conclude that the studies reported here are based on excellent data—of the four different kinds just discussed. Biedinger and Becker use a data set on ethnic minority children of German and Turkish parents to scrutinize whether ethnic minority pre-school children lag behind, and, if so, why. This is a specialized, longitudinal, study which they are responsible for themselves, and although it is not representative for Germany as a whole, it is indeed large for a study of that kind, and one of the most promising in the area of early ethnic inequality. The longitudinal element also facilitates causal modelling, which Becker profits from in her fixed effects estimations of early language development.

Jonsson and Östberg (Sweden), and Robson (England) use nationally representative data (in Robson’s case merging a large number of surveys) that take a more general approach towards studying young people’s well-being. The Swedish contribution shows an example of a study of young people that collects data both from parents and children, with a multidimensional design. Robson uses the panel element in the BHPS to model the causal effect of family change on child outcomes, which is a rare possibility and thus a very strong analysis in the field of family studies where most researchers are limited by cross-sectional data.

Smyth, Whelan, McCoy, Quail, and Doyle present a first analysis based on the new Irish cohort study “Growing up in Ireland”. Though still naturally confined to cross-sectional data, their article shows the great potential of this kind of design. Especially, their data provide the opportunity of merging information gathered from parents and teachers with that from young people themselves—which, as the authors point out, does make a difference as the structural inequality experienced by pupils is mediated by their own attitudes and actions. This data-set will be one of Europe’s most important for studying young people’s well-being in the years to come, provided continuous funding. Also Jones and Hansen use data from a cohort study, the Millennium Cohort Study (MCS). Together with the 1958 and 1970 cohorts the MCS constitutes one of Europe’s richest data sets for studying young people’s well-being and intergenerational processes. Though the MCS is still new, there are already longitudinal data, and Jones and Hansen profit from the panel design in applying ‘value added’ models in their study of small children’s cognitive development.

Finally, Låftman uses the largest data-set in this volume, by merging the HBSC data for 11-, 13-, and 15-year-olds from 24 countries in her study of children to single mothers. While HBSC suffers from not having direct parental information, single parenthood is a structural feature of children’s lives that can be reliably measured via questions to children. Låftman also benefits from an advantage with the HBSC data-set, namely that it contains information on other aspects of young people’s lives than health.

Many of the results in the articles in this volume are of great importance and both of academic and policy interest. Although the reader is advised to consult each article, some of the more general conclusions can be useful as an introduction to this. One such conclusion is that parental influences start early—in three of the articles inequalities are visible already in preschool years. Biedinger (on cognitive differences) and Becker (on language development) show that children of Turkish parents are doing worse already at ages 3–4 in Germany; and Jones and Hansen demonstrate that children to some ethnic minorities as well as to low-income earners are disadvantaged at age 3 and 5. These studies all suggest that parental involvement, such as reading, or doing activities with children, are partly behind these early inequalities—but neither study shows an especially large effect, suggesting that much of the early transmission of advantage may be of a subtle kind. Not surprisingly, given these results, inequalities among older children are prevalent, as documented by Smyth et al. for Ireland for test scores at age 9. For slightly older children this is also shown by Låftman for many countries, and for Sweden by Jonsson and Östberg along several dimensions of the level-of-living.

Though all studies document intergenerational influences of several structural characteristics on crucial indicators of well-being, the importance of family structure and ethnic minority origin may be of special relevance. Ethnic differences among young people are evident in Becker’s and Biedinger’s study of early child development in Germany as well as in Jones’ and Hansen’s study of British children; and Jonsson and Östberg show that ethnic minority 10–18-year-olds in Sweden experience material deprivation. Children to single parents have worse material conditions too in the Swedish study (and in addition suffer from weaker social relations). More disturbingly perhaps, children to single mothers face a multifaceted disadvantage (in material conditions, health, and social support) in a large number of countries according to Låftman’s analysis. Another strong result, shown by Robson, is that changes in the family structure do seem to lead to decreases in happiness as well as in economic conditions—this goes both for leaving a two-parent family and entering a step-family. On the other hand, none of these changes had any impact on self-esteem, so it is possible that long-term effects of family dissolution and reconstitution are relatively modest.

One worrying societal change is that the disadvantaged groups just discussed, children who experienced their parents’ divorce and children of ethnic minority parents, are ones that have recently grown or are currently growing in many Western countries. Particularly for ethnic minority children, the fear is of course that the unfavourable socioeconomic position that many (though certainly not all) of them experience (Heath et al. 2008) is topped up by discrimination if not in school, later in the labour market (Riach and Rich 2002). The same may in fact also be the case for children of working class origin (Jackson 2009), and it is also clear from the results in this volume that Western societies have far from done away with traditional socioeconomic inequality. In fact, a good case could be made for claiming that many (but not all) of the disadvantages faced by children to immigrants and by children who experienced family disruption are fundamentally socioeconomic.

4 Future Directions

Inequality among the young emanates to a large extent from their proximate surrounding, primarily their family of origin, meaning that resources (and disadvantages) are transmitted during childhood. The results shown in this volume underline that growing up in single parent households, in ethnic minority families, and with parents whose educational qualifications, occupations, and social class positions are unfavourable, on average means a lower level of living. The theories behind, supported by empirical results, emphasize that parent–child interaction already at young ages is important (both quantity and quality, and also form of interaction). Also the transmission of material resources—or the security afforded by such—is likely to be vital for children’s chances in life.

Much of the research into young people’s well-being is guided, explicitly or implicitly, by a desire for improving the living conditions of the young. By monitoring their well-being it is possible to safe-guard against a general deterioration. But what is amply demonstrated in this volume is that it is likewise of great relevance to study the dispersion of young people’s well-being, and how it is related to their family of origin. Equalizing opportunities do not only lead to a more just society: the average level of well-being will also be improved by raising the standard of those worst off (assuming that well-being is not a zero-sum game). Monitoring inequality among the young is therefore crucial. However, when researchers want to move from monitoring to suggesting policy measures, the demands on data and analysis grow. Even though, for example, reducing poverty appears to be the royal road to equalizing well-being among the young, income redistribution (would it be politically feasible) may still not be a particularly efficient way of improving the conditions of children in poverty—the causal effect of income on children’s outcome appears to be modest, and non-monetary effects may be more important (Mayer 1997; Blau 1999).

For children whose parents are unable to provide the non-material support that is conducive of well-being, an intriguing issue is to what extent, and via what measures, this can be counterbalanced. It is obvious that almost everyone would support the idea of a “legitimate parental partiality” (Swift 2005): it is in fact, for society at large, a good thing that parents make great efforts in providing for their children, such as reading to them, even if this also means that more skilful parents transmit advantages from one generation to the next. In order to compensate children with less involved, or less skilful, parents, the policy solutions are often public daycare or preschool programmes. There is also some evidence that such measures could lead to positive outcomes, particularly the provision of high-quality daycare for deprived children (cf. Heckman et al. 2006; Barnett 2008).

The research into early inequality is important for verifying which child-care arrangements might be conducive to child well-being. However, to take the view of intergenerational inequality as a whole, we also need to study how inequality grows—or, perhaps, contracts—over the early life-course. It is probable that parental influence continues over the childhood, because, for example, well-educated parents can provide help with their children’s school-work, and rich parents can buy their children a prestigious college education. Longitudinal studies starting at an early age are needed to elucidate how school politics, or perhaps various forms of family policy, can ensure that children who are off to a bad start in life do not fall behind even more.

To conclude, a promising future direction of studies of young people’s well-being would be to expand research into monitoring inequality among the young; and to follow children up to their early adulthood in order to understand how and when the parental influence plays out. Thus, combining the perspectives of well-being and well-becoming in an intergenerational perspective would seem especially fruitful. While we will probably not find a panacea for improving the situation of those worst off, such a perspective might nonetheless inform social policy to advance young people’s living conditions.

Footnotes
1

I will henceforth refer to the target group as ’young people’ rather than ’children’ (which is common for 0–18-year-olds, the age-group in focus). ‘Children’ then refers to a generation, which is practical when treating intergenerational processes.

 

Acknowledgements

The papers in this volume emanate from a conference on “The Transfer of Resources across Generations: Family, Income, Human Capital and Children’s Wellbeing”, organized by Elizabeth Thomson and Jan O. Jonsson, both of Stockholm University. The conference was held in Vadstena, Sweden, June 9–13 2008, funded by European Science Foundation (ESF) in collaboration with Linköping University, and co-funded by Riksbankens Jubileumsfond (RJ), the Swedish Research Council (VR), and the European Consortium for Sociological Research (ECSR). I am grateful to organisers and contributors to that conference, including authors, commentators, and representatives from the ESF, as well as to Asher Ben-Arieh, the Editor of CIR. My own research was generously funded by the Swedish Council for Working Life and Social Research (FAS 2004–1975).

Copyright information

© Springer Science+Business Media B.V. 2010

Authors and Affiliations

  1. 1.Swedish Institute for Social Research (SOFI)Stockholm UniversityStockholmSweden

Personalised recommendations