Getting the Balance Right
Leon Feinstein

Introduction Governments around the world are eager to plough investment into education, both because of the importance of education to individuals and because of its role in national economic welfare. But has the emphasis on education meant that other important factors in childhood development been ignored? Leon Feinstein, a member of the Centre for Economic Performance based at the London School of Economics and Political Science, argues that it is crucial to get the balance right. The early psychological development of children has as much bearing on later productivity as their academic ability.

The Blair Government is understandably concerned to increase the literacy and numeracy of British children. There is a growing worldwide appreciation that education is important in itself and also as a way of enhancing children's life chances and the economic growth prospects of nations as a whole. But is the drive to improve educational standards leading to the neglect of other factors which can also affect employability, earnings and economic success?

The answer seems to be yes. Take two examples. There is now clear evidence that children with higher self-esteem at age ten get as much of a kick to their adult earning power as those with equivalently higher maths or reading ability. And better performance in tests of other aspects of psychological development are also well correlated with reduced risk of unemployment. These are two striking results from new research which has made use of British data to examine how the scores of psychological and behavioural ability at the age of ten can predict what happens when a child grows up and enters the labour market.

Psychological development matters

Popular demands for improving education standards are widespread. Some people advocate the extension of the school-leaving age; others argue for a reduction in class sizes. In many cases, proponents of more spending on education believe that such investment is worthwhile because education is a good thing in itself. But while they might not gain such widespread popular support, similar claims could be made about psychological development. Self-esteem, for example, is presumably a good thing in itself, provided that it is tempered by a realistic or balanced view of one's own attributes. Of course, the proponents of higher-education spending can also point to research showing that it brings wider economic benefits. What has been less clear until now is whether differences in individuals' psychological development have similar implications for productivity.

Our findings reveal that they do. The early psychological development of children has as much bearing as their academic ability on their later productivity: so much so that there may be grounds for arguing that school performance should be assessed not only in terms of maths, reading or science scores but also in terms of the success or failure in helping children to develop in other ways. There is as much of an economic or investment argument for helping children develop psychologically as there is for helping them develop academically. Indeed, there are important interactions between the two, and the fact that children are failing psychologically might in many cases have negative effects on their academic development.

Measuring psychological development

We've been able to track the impact of psychological development using the 1970 British Cohort Study. This interviewed the parents of all children born in the UK in the first week of April 1970. Children, parents, doctors and schools were subsequently questioned when the children were aged 5, 10, 16, 26 and, most recently in 1990, around the time of their 30th birthday. We were interested in three sets of questions asked when the subjects of the study were ten years old. First were standard maths and reading tests. Next were questions known as the Lawseq and Caroloc series, developed by the Cohort Study to test two particular psychological attributes, self-esteem and "locus of control". Both are important, well established notions in the psychological literature. They have been shown to predict later adolescent outcomes such as school performance and, when low, criminality or psychiatric disorder. The final set of questions was asked of the children's teachers and included questions relating to anti-social behaviour, relations with peers, attentiveness and the extent to which children were extroverted.


Self-esteem is reasonably self-explanatory. The psychological score was derived from questions such as "Do you think that other children often say nasty things about you?" or "Are there lots of things about yourself you would like to change?". Sixteen questions were used to create a reliable measure of self-esteem at the age of ten. Other studies have already shown that managers perceive workers with high self-esteem to have higher productivity in work because they use time more effectively, requiring less guidance and considering a wider range of solutions to problems. Self-esteem should, therefore, increase wages directly. It might also improve the chances of getting a job since candidates with higher self-esteem will be more confident in interviews and better able to sell themselves, as well as being more productive. More generally, it might be expected that children with higher self-esteem are better equipped both to set appropriate goals for themselves and to achieve the goals they set.

Self-esteem may be derived from awareness of genuine ability. Children with higher maths scores, for example, also have higher self-esteem. Nonetheless, there are children with high maths or reading ability and low self-esteem. It is these children that we focused on. Self-esteem is strongly associated with social class. Children from wealthier, more educated families have higher self-esteem. We sought to compare children from families with the same level of wealth and education. We found that in the case of two children from families with the same low level of income and with parents who left school at the minimum leaving age, the child with the higher self-esteem at age ten will earn more at age 26, even if the two children also have the same scores in maths and reading.

The fault lies not in our stars

The locus of control is a less well known notion. It refers to an individual's sense of control of their own destiny. Individuals with a high locus of control are better able to process information from the outside world, are concerned to improve both their circumstances and themselves and, finally, are more stable in response to external influences. It might be expected that such individuals will make better decisions about educational and career choices and have a higher degree of patience. Twenty questions were asked of the 10-year old children--including "When you get into an argument is it usually the other person's fault?" and "When someone is very angry with you, is it impossible to make him your friend again?". Strikingly, we found that the locus of control score is an important predictor of female wages but is less important for males.

Behavioural development

Assessing the significance of behavioural development--the questions asked of teachers--is more difficult. Some children classed as anti-social at the age of ten, for instance, might have resolved their difficulties and score very differently by the age of 16. It's also important to remember that some anti-social people earn high incomes! Interestingly, we found that boys considered to be anti-social at the age of ten are in general at greater risk of unemployment in early adulthood, whereas girls who scored similarly tend to earn more at the age 26 than their peers. Part of this discrepancy might be explained by an element of subjectivity in the test answers--what teachers consider to be natural aggression in boys could be seen as anti-social behaviour in girls, for example. Or it may be that more ambitious girls are less well accepted by their peers and so appear anti-social.

It is perhaps worth remembering at this point that these scores are proxies for features of personality which might be important. Our findings indicate a clear link between the test scores at age ten and subsequent performance in the labour market. They do not address problems associated with the tests themselves.

Why the family matters

Family was an important factor in performance in these tests. The higher the level of education of the parents, the better their children performed on all the age-ten tests. The same is true for the association with social class or income. The effects are not quite as strong for psychological scores as they are for academic ones, but they are large nonetheless. This link is perhaps not unexpected--but its extent is striking. The perpetuation of social inequality through the generations isn't just because children from poorer families get less education, poorer nutrition and worse housing: they also tend to have lower internal psychological support than children from richer families. This may make it even harder to confront the economic circumstances of early adulthood for which they will tend, in any case, to have less financial or other support than children from richer homes.

Another striking finding is that while girls do better than boys in the test of anti-social behaviour, they have lower self-esteem. These psychological/behavioural gender differences are stronger than the differences in academic abilities. The biggest gender difference is for attentiveness at the age of ten where boys do particularly badly.

Children with good academic scores tend to score highly on self-esteem, but it is not unusual to have high academic ability and low self-esteem. Self-esteem and locus of control, however, are strongly correlated, although, again, a significant proportion of children have high self-esteem and a low locus-of-control score, suggesting, perhaps, that their self-esteem is not well founded. Similarly there are children with low self-esteem and a high locus-of-control score, suggesting that they have psychological attributes that they don't value as much as they might.

The link with the labour market

Having established that the psychological-development tests did indeed provide clear indications about how children at the age of ten would develop in later life, we wanted to know what the link might be with performance at school and, more significantly, with performance in the labour market once the children reached adulthood. It seems clear from our findings that family influences--in particular the wealth and educational achievements of parents--continue to have a major influence on children's performance at school. But what about in the labour market? The findings for men and women are very different. In the case of men, those who only differ from each other in the degree of self-esteem recorded at age ten will, at the age of 26, nevertheless show significant differences in earnings. Someone with high self-esteem at the age of ten will be earning 5.6 percent more than his counterpart with low self-esteem in childhood. The same does not hold true for women, for whom locus of control and behavioural attributes are more important. Women who have better peer relations at the age of ten go on to earn far more than those with poor relations.

Attentiveness and anti-social behaviour

The risks of unemployment are also affected. Men who were introverted at the age of ten are more likely to experience a significant period of unemployment, as are anti-social boys. Those boys with self-esteem who become unemployed are less likely to experience prolonged unemployment. For women, the results are similar to those for wages. Attentiveness, good peer relations, locus of control and academic ability are all important in reducing the risk of short or long-term unemployment.

Family-friendly policies?

It will come as no surprise to many human-resource professionals to learn that people with higher self-esteem are more productive and earn more. Neither will it shock teachers to know that more attentive children tend to learn more. Our findings show conclusively that such intuitive beliefs can be supported by empirical evidence. But are these important psychological attributes at the age of ten randomly allocated? Or is there scope for policymakers to influence them?

Perhaps the most striking of our findings was the extent to which families rather than institutions seem to make the difference in influencing age-ten psychological and behavioural scores. This has clear implications for those who would wish to help people develop the childhood attributes which will help them in later life. Schools are generally large and imposing buildings where pupils are taught in one-year cohorts with classes of around 30. Schools are necessarily geared to helping pupils achieve good key stage and exam scores, and although this requires the development of ethical or moral attributes, they are not institutions created to help individual children achieve psychological growth. This has traditionally been the role of the family.

We found that children whose fathers are in manual, unskilled occupations have on average, low self-esteem, low locus of control, exhibit a high degree of anti-social behaviour and are inattentive even by age ten. Children in this group also have worse labour-market outcomes than other children; but their difficulties are made worse by their development up to age ten. It is not only social class of itself that causes these outcomes, but rather what comes with such categorisation in terms of parental education, income, housing, schooling and family size.

More important even than these general aspects of family upbringing, however, is the nature of the relationships between parents and children. In the questionnaire given to them as part of this study, teachers were asked to make two judgements about parental attitudes to the children: their interest in the education of the child and whether they exhibited any hostility to the child. Both variables are hugely important in influencing the age-ten scores. The effect of parental hostility on self-esteem and anti-social behaviour is overwhelming, much more important than even the absence of a parent, the nature of the school or the parents' social class.

It is also interesting that fathers' hostility is as important in influencing the age-ten maths scores as fathers' interest in education. We found both to be substantially more important than fathers' own education or than the type of school that the child attended. This shows how important the psychological background of the child is in helping them learn academic skills as well as psychological or behavioural ones.

Of course, our findings don't enable us to say anything about the nature of this parental hostility or low level of interest. It is possible that they are proxies for the stress of poor material circumstances in the home, though this is unlikely because the results already take income into account. It is also possible that teachers don't really observe these parental attitudes but assume them from the children's behaviour. Again, this isn't altogether a compelling argument. The most likely explanation is that it is the quality of the relationships formed by and with children that help them in developing the psychological attributes which help them achieve successful and productive economic lives.

The question for policy-makers is how to help children develop these attributes. Clearly, policies which help families are crucial: for example, support for parental leave, better post-natal or subsequent health and educational support for parents and finding new ways of involving parents in the education of their children. Our findings suggest that that there are economic as well as ethical rewards for such endeavours.

Relevant Links
CentrePiece Magazine

This article is taken from CentrePiece magazine published by the Centre for Economic Performance at the LSE. Copyright The London School of Economics and Political Science.