The available research supports the following conclusions about the role of low self-esteem in the social problems listed earlier.
With respect to a range of problems, no impact of low self-esteem is apparent. That is to say, the patterns of behaviour constituting or contributing to the problem cannot be attributed to lack of self-esteem. These problems include:
• crime, including violent crime
• racial prejudice
• teenage smoking
• child maltreatment.
In the case of racial prejudice, high self-esteem rather than low appears to be related to the outcome. In the case of violence, there are some indications that high self-esteem in combination with other factors carries a risk.
There is a second category of problems for which the influence of low self-esteem is not proven or its influence is very slight (the ‘not proven’ cases may merit further attention):
• educational underachievement
• alcohol abuse
• drug abuse.
With respect to four problems, low self-esteem does appear to be a risk factor. These are:
• teenage pregnancy and possibly unprotected sexual contact carrying a risk of sexually transmitted infection
• eating disorders
• suicide attempts, whether or not successful, and suicidal thoughts
• low earnings and extended unemployment (males).
In each of these cases, it is also evident, however, that low self-esteem is one among a number of risk factors and the indications are that, apart from the risk for females of teenage
pregnancy and for males of extended unemployment and lower earnings in their twenties, its impact is relatively minor. It would be productive to focus further research in each case upon the ways in which self-esteem interacts with other risk factors. The following questions in particular remain to be resolved:
1 Does self-esteem mediate the impact of certain of the other risk factors? If it can be shown to do so, then we would be in a better position to decide whether interventions should be directed at these more remote causes or at breaking the impact they have on low self-esteem
2 Does self-esteem operate as a risk factor quite independently of any others and, if so, to what extent does it affect risk compared to the other factors? If we can determine this more precisely, we are in a better position to decide whether resources should be focused upon this risk factor rather than others. Such a decision, however, should also consider the relative costs of alternative interventions.
3 Does self-esteem amplify or moderate the impact of other risk factors? If this can be determined, we will be in a better position to decide which of the factors, self-esteem or those others with which it interacts, is the more appropriate focus for intervention.
At this point, one might ask why low self-esteem does not have a wider range of negative effects. Or, rather, why has it not been possible to identify negative effects in so many of the areas considered? Has the science for some reason failed us because it does not confirm what intuition and common experience tell us is true?
The most obvious way in which it might have failed is in its detection of the presence of low self-esteem. In other words, the methods of
measurement have been inappropriate or insufficiently sensitive. It is true that some
researchers, notably Greenwald and Banaji (1995), have strongly criticised the methods most widely used to assess self-esteem. The substance of their criticism is that methods in which people answer direct questions about their feelings towards
themselves are more likely to detect what people are prepared to claim about their feelings than about the true character of these feelings. Answers will reflect either beliefs about what is the socially desirable thing to say, or the degree to which a person is prepared to be modest or boastful about themselves.
It has yet to be shown that the alternative method of measurement advocated by Greenwald and Banaji can satisfy the basic requirements of precision and validity, or that any of the other alternatives – such as observers’ ratings – can do so.
There is also a question about the practicality of these alternatives, particularly in large-scale
research with young people. But these problems are beside the point. If Greenwald and Banaji are right, then the facts that a widely employed measure such as Rosenberg’s scale is practical to use and that it achieves a high level of precision are not adequate justifications for its use.
On the other side of the argument, one can ask:
does the pattern of findings make sense? I think that, very largely, they do make sense and in the next chapter, where I look at the determinants of self-esteem, the justification for this confidence will, I hope, become clearer. But, to anticipate, the pattern indicates the following: people who have, or
admit to, negative feelings about themselves also treat themselves badly (and may be badly treated by others).
They do not tend to treat others badly.
The apparent anomalies of drug abuse, alcohol abuse and cigarette smoking are less anomalous if one considers that (a) the ‘abusers’ are not
primarily intent on self-harm and (b) they may be aware of health risks in their habits but their risk taking stems from confidence, however unrealistic, that they can beat the odds; these activities do not stem from lack of confidence. Pelham (1993), for example, reports that young adults with high self-esteem are the ones more likely to take up risky pursuits, such as riding motorcycles, and to take greater risks in these, such as driving too fast and driving while under the influence of alcohol.
The pattern also makes sense once one recognises (anticipating evidence reviewed in Chapter 3) that there is little connection between having positive or negative feelings about oneself on the one hand and one’s objective
accomplishments or failures on the other.
Greenwald is undoubtedly right that the commonly used measures of self-esteem detect differences in what people are prepared publicly to admit or claim about themselves. But it does not follow that these admissions or claims bear no relation to what they privately feel. What we can learn from this observation, however, is something quite informative about self-esteem. It partly describes how people behave publicly; it describes what they are and are not willing to say about themselves to others. Consequently, we should ask:
1 Why are some people willing to admit that they think themselves worthless?
2 What are the consequences of this kind of public admission?
Answers to both will be considered in the next chapter. But, to anticipate those to the second question, if you tell others you have no value, you are inviting them to treat you as if this were true.
This admission therefore carries a risk of victimisation.
One other way in which the science may have failed is in the detection of relatively hidden influences of low self-esteem. These influences could be hidden if self-esteem primarily operated as a moderator of the impact of other factors. This would be the case, for example, if low self-esteem, compared to high, increased levels of alcohol abuse among young people also exposed to high levels of stress, but decreased these levels among those not exposed to stresses. The influence could also be hidden if, for example, low self-esteem in
combination with a deferential attitude to authority
produced educational underachievement and if this effect also resulted from high self-esteem combined with a hostile attitude to authority.
I have already noted that moderator effects can be harder to spot. Moreover, despite their potential importance, in few studies have they been
systematically investigated or ruled out. There may therefore be a case for considering such effects in future research. But, if the necessary effort is to be expended, there should be a reasonable and
reasoned presumption as to where these effects will be found. This requires an informed understanding of what differences in self-esteem are and how they arise. I turn next, therefore, to these questions.
The scientific search for the determinants of self-esteem has been guided by the various theories as to its nature. The view of William James (1890) that self-esteem is success divided by pretensions potentially directs the search to two places. Does self-esteem result primarily from the degree to which a person succeeds in their aspirations?
Alternatively, does it depend primarily upon the nature of those aspirations? Both possibilities have been explored. Cooley’s (1902) emphasis on the perceived or anticipated reactions of others might seem to suggest that experience of others’
disapproval, or experience of hostility, rejection or stigmatisation by others will lead those who experience these reactions to devalue themselves.
In fact, much attention has been given to the consequences of belonging to low status categories.
What has emerged about the roots of self-esteem is not entirely what was anticipated. And this is leading to a reappraisal of the nature of self-esteem. Many of the factors which might be expected to result in low self-esteem do not do so. I will consider in turn: factors that have weak effects or none; factors that have modest effects; factors that have a more significant impact.