CAPÍTULO IV. CONTROL INTERNO APLICABLE A LOS CICLOS DE TRANSACCIONES EN UNA ENTIDAD
4.5. CICLO DE INVENTARIOS
Numerous studies indicate a simple association between low self-esteem and suicide ideation in a variety of age and cultural groups (e.g. De Man et al., 1992, 1993; Gonzalez et al., 1998; Marciano and Kazdin, 1994; Reynolds, 1991; Sendbuehler et al., 1979; van Gastel et al., 1997; Vella et al., 1996). One interesting exception is the investigation by Beer and Beer (1992). Scores on the Coopersmith index were unrelated to suicide ideation, suggesting, as with eating disorders, that the association is with feelings about the self rather than with a set of evaluations of the self.
Others also find an association between low self-esteem and parasuicide (e.g. Boudewyn and Liem, 1995; Overholser et al., 1995; Petrie et al., 1988;
also reviews by Rittner et al., 1995; Yang and Clum, 1996). However, Nasser and Overholser (1999) found no relation of self-esteem to the lethality of the attempts.
A rather smaller number of longitudinal studies also find that self-esteem predicts later suidical ideation. These include McGee and Williams’s (2000) analysis of their New Zealand data and an investigation by Goldney et al. (1991) of
adolescents’ levels of self-esteem on extent of suicidal thinking eight years later. Kaplan and Pokorny (1976) found that, among 4,694 American high school pupils, a self-derogation measure was related to number of accidents reported one year later. They interpreted accidents as potential suicide attempts or at least failures to take proper care of oneself.
A small number of longitudinal studies have considered self-esteem as a predictor of later suicide attempts, though, in all cases, in
combination with other risk factors. Considered on its own, self-esteem does predict subsequent suicide attempts (e.g. Lewinsohn et al., 1994).
Kjelsberg et al. (1994) followed up 1,800 adolescents who had been psychiatric patients and found that about 2 per cent had committed suicide over that period. Those who did so had lower self-esteem than the average for the group at the time they were first studied. An examination by Seiden and Gleiser (1990) of chemists who had committed suicide indicated that they had lower than average self-esteem, though it is not entirely clear how this was determined retrospectively.
How can we determine whether self-esteem is not only a risk factor for suicide-related outcomes but also an important one? The first step is to recognise the many other risk factors and assess these simultaneously. Major risk factors for suicide include living alone, lack of social support, drug use, physical illness, depression, knowing others who have made suicide attempts or completed them, a history of prior attempts, experience of victimisation and economic difficulties. These may
also be the risk factors for parasuicide and suicidal ideation. Consequently, research into the causes of suicide-related outcomes should ideally assess the full range.
Many studies have assessed the impact of a range of factors. However, in examining these, one is struck by two things. The first is that, in several cases, no attempt is then made to decide on the relative importance of each factor. This is vital because some factors may be correlated with the outcome only because they are also associated with one of the other factors. So, for example, self-esteem may relate to suicide attempts only because it is also related to social isolation or victimisation.
And it may be related to these either because it is a causal factor – for instance, it influences the likelihood of vicitimisation which in turn is the direct cause of suicide attempts. Or it may be a correlated outcome – social isolation may depress self-esteem and lead to suicide attempts.
The second observation is that, when the appropriate statistical analyses are performed, whether or not self-esteem emerges as a predictor, and how significant a predictor it turns out to be, depend very largely on what other factors were included in the analysis. There are, however, consistent features of those studies in which self-esteem drops out of contention once other risk factors are considered.
Studies in which self-esteem survived as a predictor after taking into account the effects of other predictors include Kjelsberg et al.’s (1994) 15-year follow-up of adolescents who had been psychiatric inpatients. Self-esteem was one among a set of eight factors discriminating the suicides from the still living. Garnefski et al. (1992) found, in a study of Dutch 15 to 16 year olds, that suicidal thoughts and behaviours were related to low self-esteem but also and independently to use of drugs, experience of sexual abuse and feelings of
loneliness. Another Dutch survey, but on a much larger scale (Kienhorst et al., 1990) – a sample of 9,393 14 to 20 year olds – revealed a range of risk
factors independently predicting suicide attempts, including self-esteem but also family breakdown and alcohol and drug use among them.
Hershberger et al. (1997) looked at predictors of suicide attempts among lesbian, gay and bisexual youth. They found that the attempters had lower self-esteem but had also suffered more adverse consequences – loss of friends, victimisation, and so on – as a result of revealing their sexual orientation.
A distinguishing feature of studies in which self-esteem does not survive as a predictor or in which its significance at best is greatly reduced is that they included other measures of affect or feelings about the self. Petrie and Brook (1992) found that extent of suicidal ideation among parasuicides was best predicted by a variable which they called ‘sense of coherence’ and which seemed to reflect the degree to which their participants felt their lives were meaningful.
Previous history of suicide attempts, unemployment and living alone enhanced prediction.
Beautrais et al. (1999) found that a sample of suicide attempters under 25 were distinguished by higher neuroticism (which is also sometimes identified as ‘negative affect’), hopelessness, introversion, an external locus of control and low self-esteem. But low self-esteem ceased to be a significant predictor when the other factors were taken into account. In Kingsbury et al.’s (1999) study of overdosers, self-esteem was a correlate but its effect disappeared once level of depression was taken into account. The outcome of Marcenko et al.’s (1999) analysis of predictors of suicidal ideation was very similar. Roberts et al. (1998) examined predictors of suicidal ideation, including self-esteem, in a large (5,000+) sample of ten to 17 year olds. The surviving predictors were history of suicide attempts, depression and recent life stresses. Simons and Murphy (1985) found that self-esteem was correlated with occurrence of
suicidal behaviour in a sample of high school students but not when other factors were
considered. Emotional problems and involvement in delinquency were predictors for females;
employment problems were most potent predictors for males. Finally, De Man (1999) looked directly at the consequences for the association between self-esteem and suicidal ideation of taking into account the effects of depression. The consequence was to reduce substantially the strength of this association.
In some other studies including measures of affective variables in addition to self-esteem, the latter does survive as a predictor though typically not as the most important one. This is true, for example, in the Lewinsohn et al. (1994) study. Of the 1,500 14 to 28 year olds surveyed, 26 made suicide attempts one year into the study. Low self-esteem at entry predicted these attempts, but so did history of past attempts, suicidal ideation and depression. Similarly, in Yoder’s (1999) study of homeless youth and runaways, five factors distinguished those reporting suicidal ideation from those not doing so. The factors were self-esteem, depression, experience of physical abuse, experience of sexual abuse and having a friend who attempted suicide.
What all this suggests is that measures going by different names are to some extent assessing the same thing. Measures of depression, negative affect (or neuroticism), self-esteem, hopelessness,
fatalism, and locus of control are all attempting to assess a person’s feelings about themselves.
Depending on how each measure is constructed, the overlap may be almost complete or it will be more partial. It is a fair bet that an aspect of these feelings is implicated in suicidal thoughts and actions. But, whichever measure does the best job of assessing this particular aspect will emerge from the analysis as the predictor and others which measure it less well will fall away.