This post originally appeared on the University of Surrey blog on August 28 2015 and is reposted here with the permission of the author.
Discussion of goths and mental health has just made a high profile return to the news, following publication of a study in Lancet Psychiatry (Bowes et al, 2015) that indicates teenagers who identify as goths are more likely to be depressed or to self-harm than those who do not. Elaborating on an interview I gave to The Independent, this short piece tries to clarify what the study does and does not show and reflects on what inferences might be drawn.
Based on longitudinal survey data, the study identifies a statistically significant correlation between self-identification as a goth at the age of 15 and self-reported depression and self-harm at 18. Specifically, the study reports that 18% of strongly goth-identified participants met the study’s criteria for depression and 37% those for self-harm. This compares with figures of 6% and 10% respectively for respondents who did not identify with the subculture. The study also claims a ‘dose effect’, indicating the likelihood of both mental health outcomes increased in proportion to the strength of subcultural membership. Those who identified ‘somewhat’ with goth are reported as 1.61 times more likely to meet the criteria for depression while those who identified ‘strongly’ with the subculture were more 3.67 times as likely. A similar effect is reported for self-harm. While the authors recognise a number of limitations with the study, including with respect to what inferences can be drawn on direction of causation, they also claim their results ‘suggest that some peer contagion operates within the goth community’.
While its indication of a prevalence significantly higher than in other groups should be taken seriously, it is worth emphasising that the study does not show that most goths are either depressed or prone to self-harm – quite the contrary. This may be an obvious point but it is sometimes awfully easy to jump from one inference to another. Furthermore, in the case of those classified in the self-harm category, the study does not differentiate between types or levels of seriousness, nor does it show the regularity of such behaviour or whether it was recent. This is because this classification was based on answers to a single question: ‘have you ever hurt yourself on purpose in any way (eg, by taking an overdose of pills or by cutting yourself)’. Without underestimating the potential significance of minor or one-off forms of self-harming behaviour, this is not, it might be argued, a particularly high threshold and, crucially, it does not tell us whether the behaviour occurred within or before the aged 15-18 period focused on by the study.
This leads us to a more fundamental point, which is that, although the correlations reported in the study are potentially significant, they do not comprise evidence that being involved in the goth scene is a cause of depression or self-harm. As the authors of the study note, the figures are equally open to the conclusion that those with a propensity for depression or self-harm are disproportionately attracted to the goth scene. Such an interpretation would certainly find support in some of the qualitative research that has been carried out on goths, including my own, which has indicated that many goths become involved in the scene following periods of low status, social isolation or even bullying within mainstream school peer groups (Hodkinson 2002). Such studies indicate that such individuals tend to experience the goth scene as a positive environment, including with respect to the development of friendships, self-worth and a sense of belonging.
Importantly, Bowes et al did investigate the significance of their results in greater depth by making adjustments for a range of confounding variables, including previous depression, earlier experiences of bullying and mother’s education level. An increase in risk of both mental health outcomes in proportion to level of goth identification was still visible even after such adjustments, they report, albeit a smaller one. Importantly, though, the figures cited above and across media reports are non-adjusted and should be interpreted as such. It is also possible, as the authors acknowledge, that other confounding factors may not have been adjusted for at all. It is possible, for example, that individuals could be experiencing fluctuating levels of unhappiness without suffering direct bullying or meeting criteria for clinical depression at the time, for example.
Whilst the possibility that goth identification could itself contribute to depression or self-harm should not be discounted altogether, the nature of statistical observational studies such as this alongside the use of a self-harm threshold that does not indicate when such behaviour took place makes it difficult to be sure whether the symptoms identified, or some sort of propensity for them, pre-existed goth identification or developed subsequently. Meanwhile, even if involvement in the subculture can contribute in some way to depression or self-harm, there could be a number of explanations for this. Not least of these is that goths often are stigmatised and harassed because of their subcultural appearance and identity, something that may contribute to feelings of anxiety and rejection (Garland and Hodkinson 2014). Such an effect could, perhaps, be particularly pronounced if it serves as a continuation of patterns of bullying from earlier in life. This possibility is acknowledged by the authors of the study, though this is not reconciled with their earlier claim that the study specifically suggests evidence of ‘peer contagion’. A further important possibility, not considered by the authors, is that the social environment of the goth scene helps participants to overcome mental health stigmas and thereby to feel more comfortable reporting depression or self-harm than young people in other groups.
In summary, the study makes a useful contribution and signals that further research on the subject of subcultures, depression and mental health could be valuable. The authors are right, however, to be cautious (comments about peer contagion notwithstanding) about what the study does and does not show. For this reason, the suggestion that clinicians (or for that matter, parents, teachers or others) should treat indicators of goth identification as some sort of proxy for potential depression or self-harming (as O’Connor’s comment piece on the study for Lancet seems to infer) seems excessive, not least because such an approach could exacerbate the sense of societal stigmatisation goth participants can experience. At risk of veering a long way from my areas of expertise, perhaps it would be better to focus on rather more definitive symptoms.