Background: A number of studies have suggested that social inequalities in health disappear or are attenuated in early adolescence possibly due to changing risk exposures. The present study examines social equalization in youth in a representative sample of British households with children aged 0–18 years. Methods: Secondary analysis of a cross-sectional survey of a representative sample of British households with children and youth with parent-reported less than good health, long-standing illness, and chest problems as outcomes. Results: Data were available on 15 756 children aged 0–18 years in 8541 households in the third sweep (2001) of the British government's Families and Children Study. Parent-rated health status, long-standing illness, and chest problems all showed social patterning among children who were 0–11 years of age. Among 12- to 14-year-olds, the social gradients in these outcomes noted in childhood associated with income, social class, and education were lost but inequalities in parent-rated health status and long-standing illness but not chest problems persisted associated with measures of household work status and wealth. Among 15- to 18-year-olds, income inequalities appeared to reassert themselves, particularly among girls, but gradients by maternal education noted among 0- to 11-year-olds were absent in both sexes. Inequalities persisted with measures of household worklessness and wealth. Conclusions: In this cross-sectional study, the social equalization in youth was noted for some health outcomes and by some measures of socioeconomic status but not for others. Inequalities in parent-rated health status and long-standing illness persist among young people in workless households and those experiencing severe material hardship.
equalization in youth
youth and health
The observation of attenuation of health inequalities in youth made in a number of studies1–8 has led to the social equalization in youth hypothesis. According to this hypothesis, the transition from childhood to youth is associated with a change from health inequality to relative equality. West2 postulates that equalization may be related to changing exposure to peer group, youth culture, and other influences in youth attenuating or reversing the influences of family socioeconomic status on health.
Not all studies support the equalization hypothesis. A study9 of self-rated health among 15-year-olds in 11 countries found no relationship with father's social class in 5 countries, a weak relationship in 4, and a strong relationship in 2. Based on Danish data from the same study, Due et al.10 reported a bivariate relationship between parents' socioeconomic position and physical and psychological symptoms in boys and girls aged 11, 13, and 15 years. Parents' participation in school and perceptions of bullying and safety at school accounted for this relationship. A social class gradient in limiting long-standing illness among children aged 0–9 years but not among young people aged 10–16 years was reported using data from the 1991–94 British General Household Survey.11 However, significant social differences among young people as well as children were present when measures, other than social class, reflecting material deprivation were used. Social gradients in mental health problems among young people have been reported.12–14
Equalization in early adolescence presents a challenge to the explanation for health inequalities and social gradients in health based on the accumulation of social risk exposures over the life course combined with the clustering of cross-sectional social risk exposures.15,16 If social equalization in youth does occur, it suggests either that this explanation of health inequalities and social gradients is incomplete or that peer group and youth culture pressures are strong enough to negate the effects of cumulative social exposures earlier in childhood. It is also possible that the relationship between health, age, and socioeconomic status varies with different health outcomes as suggested by Chen et al.17 or with severity of health outcome as acknowledged by West.2 Equalization may be an apparent rather than a real phenomenon resulting from the use of measures of socioeconomic status that poorly reflect social differentiation in early adolescence.4
In this paper, data from the 2001 sweep of the British Families and Children Study18 are used to explore the equalization hypothesis with parent-rated health, parent-reported long-standing illness, and chest problems as outcomes and a range of socioeconomic status measures as exposure variables. The dataset includes health outcome data among children and young people aged 0–18 years and a range of measures of family socioeconomic status in addition to social class allowing the impact of different components of socioeconomic status to be studied at different ages.
Secondary analysis of data on families that participated in the 2001 sweep of the Families and Children Study18 was undertaken. The Families and Children Study is a refreshed panel survey funded by the Department of Work and Pensions of the UK government. The first sweep (1999) included only low-income households with children aged 0–18 years. The second sweep (2000) was extended to include moderate-income households with children. The survey was extended in 2001 to include high-income families so that a representative sample of British families with dependent children could be included in the study. The sample was drawn from households in receipt of Child Benefit (a universal benefit available to families with children in Britain) in selected areas of the country. For the 2001 sweep, households were approached for participation in the study and recruitment was continued until the target number of households with an income distribution that is representative of British families with dependent children was obtained. Non-participation rates for the three sweeps ranged from 30 to 40%. The recruitment strategy ensured that the final sample was representative of households with dependent children.
Health-related data were available on all children in the families surveyed. Three health outcomes were included in the analysis:
Parent-rated health status over the last 12 months: Based on ‘(Since your baby was born/Over the last 12 months) would you say (child's name)’s health has been good, fairly good or not good?’ Responses were dichotomized into good versus fairly good/not good.
Long-standing illness or disability: Based on ‘Does (child's name) have any long-standing illness or disability?’
Long-standing chest illness or disability (chest, breathing problems, asthma, and bronchitis): A subset of the question on long-standing illness and disability.
Socioeconomic status variables
Current earnings and work status
Household income: Weekly income in £s was grouped into quintiles: lowest <£214; second £215–306; third £307–417; fourth £418–568; highest £569+.
Social class: Highest social class of the respondent and partner was categorized into seven groups: I, II, IIINM, IIIM, IV, V, and unclassified. These were collapsed to three categories for the analysis: I and II; IIINM and M; IV and V—unclassified participants were excluded as they were likely to be a heterogenous group.
Household worklessness: Households without a working adult.
Wealth and material resources
Housing tenure: Six categories (owned outright; mortgage; shared ownership; social tenant; private tenant; other arrangement) dichotomized to rented + other versus owner-occupied.
Access to car: Current driving licence + car access versus no access to car (with or without licence)
2002 FACS Hardship Index: Index developed for use in the analysis of the third wave of the Families and Children Study.18 Nine variables are included (two or more problems with accommodation and is unable to afford repairs if a home owner; lives in overcrowded accommodation; cannot afford to keep home warm; worries about money almost all the time and runs out of money most weeks; no bank account and has two or more debts; relative material deprivation score on food items in highest 7.5%; relative material deprivation score on clothing items in highest 7.5%; relative material deprivation score on consumer durables in highest 7.5%; relative material deprivation score on leisure activities in highest 7.5%). Family scores on hardship (zero to nine) were categorized into three levels of hardship: no hardship—zero on the nine-point scale; moderate hardship—one or two on the scale; severe hardship—three or more on the scale.
Maternal educational resources
Highest academic qualifications obtained: Four categories—none; General Certificate of Secondary Education (GCSE) or equivalent; A-level or equivalent; degree or equivalent.
Odds ratios with 95% confidence intervals were calculated for each outcome against socioeconomic status variables in the three age groups for boys and girls separately: 0–11 years; 12–14 years; 15–18 years. All odds ratios were adjusted for age within the age group and the number of dependent children in the household. All analyses were carried out in SPSS version 10.19
The 2001 Families and Children Study included data on 8540 households with 15 756 children and youth (51.5% boys and 48.5% girls). There were 10 325 (65.5%) children aged 0–11 years, 2594 (16.5%) youth aged 12–14 years, and 2837 (18.0%) youth aged 15–18 years. The sociodemographic characteristics of these children and their families are summarized in table 1. There is a trend for youth aged 15–18 years to live in more affluent households than youth aged 12–14 years who in turn tend to be living in more affluent households than children aged 0–11 years in the sample. In contrast, more youth aged 15–18 years live in households in which the mother has no qualifications—30.9% compared with 24.6% for youth aged 12–14 years and 18.7% for children aged 0–11 years. This suggests that mothers of young children, although tending to have higher educational qualifications, are living in less affluent households than mothers of young people.
Family sociodemographic status characteristics by age group
Socioeconomic status measures
Children 0–11 (n = 10 325)
Youth 12–14 (n = 2594)
Youth 15–18 (n = 2837)
Current earnings and work status:
V and IV
III M and IIINM
II and I
Working adult in household:
Access to car:
The prevalence of the outcomes of interest by age group is shown in table 2. There was a slight but statistically non-significant reduction in the prevalence of health status rated less than good by parents of boys aged 12–14 years compared with parents of 0- to 11-year-olds. No reduction was seen for girls. Among boys aged 15–18 years the prevalence increased back towards the 0- to 11-year level. Girls aged 15–18 years had a statistically significant increase in the prevalence of parent-reported fair/poor health. Girls aged 15–18 years were more likely to be reported as having less than good health than boys of the same age. Among younger children the reverse was found although the difference was of borderline significance.
Prevalence with 95% confidence intervals of health outcomes by age
Health outcome measure
Children 0–11 % (95% CI) [number]
Youth 12–14 % (95% CI) [number]
Youth 15–18 % (95% CI) [number]
Parent-rated health status (fairly good/poor)
17.6 (16.6–18.7) 
15.6 (14.6–16.6) 
14.6 (12.8–16.6) 
16.1 (14.2–18.2) 
16.9 (15.0–18.9) 
22.0 (19.9–24.3) 
Parent-reported long-standing illness/disability
17.5 (16.5–18.5) 
14.0 (13.0–15.0) 
16.7 (14.8–18.9) 
15.9 (14.0–18.0) 
17.3 (15.4–19.3) 
17.7 (15.8–19.8) 
Parent-reported long-standing chest problems
6.7 (6.1–7.5) 
6.4 (5.8–7.2) 
7.6 (6.3–9.2) 
6.7 (5.4–8.2) 
8.0 (6.7–9.5) 
7.3 (6.1–8.8) 
For long-standing illness/disability, prevalence remained stable among boys across the age groups but increased significantly among girls, with 17.7% of 15- to 18-year-old girls having long-standing illness/disability compared with 14.0% in the 0–11 age group. Boys aged 0–11 years were more likely to have long-standing illness/disability than girls of the same age but there were no differences in rates in the older two age groups.
The prevalence rates for long-standing chest problems showed non-significant trends of increase across the age groups for both girls and boys. There were no significant differences in prevalence between boys and girls in the three age groups.
The relationship of the outcomes of interest in childhood, early youth, and late youth with socioeconomic measures are shown in tables 3–5. Inequalities in outcomes among boys and girls aged 0–11 years were noted across the range of socioeconomic status measures except that long-standing illness/disability did not show an income gradient among boys and the association of low maternal education with chest problems in girls suggested a trend, but all the confidence intervals crossed unity.
Odds ratios (95% CIs) of parent-reported chest problems in childhood and youth by sex, age group, and socioeconomic status measures
SES measures a
Current earnings and work status:
V and IV
IIIM and NM
No working adult in household:
Housing tenure (rented):
No Access to car:
a: The most privileged group is the reference for each SES measure
In early youth, the association of less than good health with household worklessness, rented housing, no access to a car, and material hardship noted in the younger age group remained robust for both boys and girls. In both boys and girls living in a household in the lowest social class group remained significantly associated with parent-reported less than good health. A significant income gradient with less than good health was noted only among boys. Maternal education was no longer associated with less than good health for boys or girls.
Among boys and girls, the relationship of household worklessness, rented housing, lack of car access, and material hardship with long-standing illness/disability remains significant in early youth. The association with maternal education and income group was lost in both sexes and with social class in boys. Girls and boys aged 12–14 years from lowest social class households were at greater risk of long-standing illness/disability than girls from higher social class groups. None of the associations of chest problems with any of the socioeconomic status measures noted in 0- to 11-year-olds persisted into early adolescence except for an association with severe material hardship among boys.
In late youth, there was no re-emergence of an association of maternal education with any health outcomes. Less than good health and long-standing illness/disability showed a persistent relationship with the lowest social class group in boys and girls but chest problems only showed re-emergence of a significant association with household social class among girls. As noted for less than good health and long-standing illness/disability in both childhood and early youth household worklessness, rented housing, lack of car access, and material hardship retained an association with these outcomes among boys and girls in late youth.
This study confirms the social patterning of parent-rated health status, long-standing illness, and chest problems in childhood.20 The social patterning of these outcomes in early youth is less clear lending some support to the equalization hypothesis, particularly in relation to long-standing chest problems. However, for less than good health and long-standing illness/disability, although the gradients noted in childhood associated with income, social class, and education were either absent or inconsistent in early youth, persistent inequalities in both sexes through early into late youth were noted associated with measures of household worklessness and wealth.
The study suggests that the influence exerted by household income, parental social class, and maternal educational attainment on the health outcomes studied weakens as the individual passes from childhood to youth. Loss of gradients with household income, social class, and maternal education but persistence of differences associated with worklessness, car ownership, housing tenure, and severe hardship suggests a threshold effect such that only those young people in the most disadvantaged households show persistent inequalities beyond childhood. In addition, household worklessness and wealth are likely to reflect more long-term social influences than income and occupational status. In the UK, households with no working adult are amongst the poorest and are more likely than other groups to have been in persistent poverty.21 Insufficient wealth for home and/or car ownership identifies households in long-term financial hardship as does the hardship index. It may be that the influences of social factors that are working in the short term are more susceptible to attenuation by peer group and youth culture than those reflecting long-term social disadvantage.
Another interpretation could be that the small numbers in the older age groups mean that significant associations noted with dichotomized variables are lost when variables with more than two categories are considered.
In this study, different health outcomes behave differently with age and socioeconomic status. This is consistent with the hypothesis that health inequalities arise as a result of risk exposures, either cumulative or time-limited, that can occur at different points in the life course.22 It also supports previous studies showing that different components of socioeconomic status may exert different effects on the same outcome.4,23
Comparison with published literature
Goodman13 reported social gradients in self-rated health, depression, and obesity among 11- to 17-year-olds. Parental education, occupation, and household income were used as indicators of socioeconomic status. Although these findings are consistent with the present study showing a gradient in self-rated health beyond early childhood, they cannot be directly compared because of the different age groups studied and the use of parent-rated compared with self-rated health.
The findings of this study that parent-reported long-standing illness shows persistent social inequalities across all three age groups using measures of wealth and worklessness are consistent with those of Cooper et al.11 but not those of West2 summarizing studies of proxy- and self-reports of long-standing illness and Macintyre and West23 based on their reanalysis of the Twenty-07 Study data using non-occupational measures of social status.
Chest problems in this study appear to show a pattern of health, age, and socioeconomic status relationship consistent with the childhood-limited model proposed by Chen et al.17 although small numbers in the older age groups may distort some of the findings. West2 suggests that the evidence from the literature is conflicting; however, there is some evidence that serious respiratory conditions are socially patterned in adolescence but less severe problems such as wheezy bronchitis14 and a combined category of wheezing and asthma24 show no social variation at 16 years.
Strengths and weaknesses of this study
Examination of equalization is best done in a dataset that encompasses young children, early adolescents, and older youth and uses the same health and social status measures for all age groups. The dataset used in this study enabled such an examination so that it was possible to identify the health outcomes that did or did not show changing social risk with age against the same set of social status indicators.
The detailed measures of household socioeconomic status, particularly the measures of family hardship and work status, available in the Family and Children Study dataset made it possible to differentiate the effects of different dimensions of social status such as current income and social class from markers of more long-term wealth and parental educational resources.
The Family and Children Study third wave is a representative sample of British households with children aged 0–18 years. The study households cover the whole range of social circumstances including 20% of households with no working adult. This has allowed the study of equalization in a sample including households at the least and most disadvantaged ends of the British social spectrum. Bias associated with non-participation was minimized by the recruitment of participants until the sample contained a fully representative sample of British households with dependent children.
Cross-sectional studies have a number of weaknesses limiting interpretation of their findings. In relation to the study of equalization, a longitudinal cohort study of children as they pass from early childhood through adolescence would give the most powerful evidence with which to test the equalization hypothesis. The assumption made in this study that the sample of young people aged 12–14 years at the time of the study would have shown the same social gradient for the health outcomes studied between the ages of 0–11 years as the children studied in that age group is impossible to test. Similarly, it is not possible to predict that the sample will show similar socially patterning when they reach the age group of 15–18 years as those who were in that age group at the time of the study. Longitudinal cohorts, however, have the disadvantage that there is a long lag period before findings can be analysed.
The conclusions that can be drawn from this study are limited by the health outcome measures used in the Families and Children Study. Measures of general health status and non-specific long-standing illness are difficult to interpret. They do not lend themselves to in-depth study of the mechanisms by which health inequalities arise or are attenuated. A further problem in comparing these results with those of studies such as the West of Scotland Twenty-07 study1 using self-reported health outcomes in late childhood and youth is that parent-reported health measures were used. Whereas it may be suggested that self-reported health measures better reflect the ‘true' experience of young people, and self-reported measures cannot be used to study young children and maintenance of consistency of measure across age groups including very young children requires the use of parent-reported measures.
Further problems arise in interpreting the results due to small numbers of participants in some of the groups, particularly in the 12–14 age group. This makes interpretation of some of the findings difficult especially when predictor variables, such as income, have five subcategories.
The study suggests that the relationship of health with age and socioeconomic status varies with the health outcome studied and that the relationship varies with the measure of socioeconomic status used. It suggests that measures of household worklessness and wealth may be more sensitive markers of socially differentiated health outcomes in early adolescence than occupation, current income, and educational resources and that equalization may take place among young people at the top and middle levels of the social spectrum but not those in the most disadvantaged groups.
Is there evidence of social equalization in youth among children aged 0–18 years in a representative sample of British households with children?
Equalization occurs for some outcomes and using some socioeconomic measures but not for others.
Inequalities in parent-reported health status and long-standing illness persist in workless families and those in severe material hardship.
Policies aimed at reducing health inequalities should include young people particularly those living in workless and materially deprived households.
The authors thank Professor Alan Marsh and Dr S. McKay, the principal investigators in the Family and Children Study, and the UK Data Archive at the University of Essex for permission to access and use the data.
Siahpush M, Singh GK. A multivariate analysis of the association between social class of origin and current social class with self-rated health and psychological health among 16-year old Australians. Aust N Z J Med2000;30:653–9.
McMunn A, Bost L, Nazroo J, Primatesta P. Psychological well-being. In: Prescott-Clarke P, Primatesta P, editors. Health survey for England: the health of young people. Vol. 1: Findings. London: The Stationery Office, 1998.
Currie C, Hurrelmann K, Setterbulte W, Smith R, Todd J. Health and health behaviour among young people: health behaviour in school-age children: a WHO cross-national study (HSBC) International Report. Copenhagen: WHO, 2000.
Due P, Lynch J, Holstein B, Modvig J. Socioeconomic health inequalities among a nationally representative sample of Danish adolescents: the role of different types of social relations. J Epidemiol Community Health2003;57:92–8.