Moral Judgments of Suicide

Tainting the Soul: Purity concerns predict moral judgements of suicide
Rottman, Kelemen and Young

Blog by Dr Elise Holland

Last week in our moral psychology discussion group, we read the above publication:

The authors begin by bringing up the notion of suicide and how it represents an interesting case for dyadic accounts of morality. According to dyadic accounts, moral acts require an agent who enacts the deed, and a patient who is the recipient of said deed. However, when it comes to suicide, the agent and patient are essentially the same. The harm is self-directed, not other-directed. So why then, is suicide judged to be immoral?

It has been suggested that suicide is immoral in that even though it’s self-directed, it causes unseen harm to others – so for instance it harms the deceased person’s family and friends, their community, or even God. So people still perceive other victims involved in suicide.
Do perceptions of harm actually play a role in moral wrongness judgements of suicide, or do they just reflect people’s post hoc justifications of why suicide is wrong? The authors suggest that while people might explicitly argue that suicide is wrong because it causes harm to others, in reality what is truly driving their moral judgements is not harm but instead purity.
Providing some preliminary evidence for the idea that moral judgements about suicide might be closely related to purity, research has shown that people who believe that their life is sacred or that their life belongs to God are typically more disapproving of suicide. So perhaps it’s not so much about harm, but more about the impurity of tainting the soul.

So the paper consists of 2 studies designed to explore how harm and purity concerns feed into moral judgements of suicide. And what the authors hypothesise is the tendency to value purity but not harm, will predict the extent to which individuals morally condemn suicide.

Designed to test this hypothesis. Specifically, they looked at how individuals rated a suicide or homicide obituary in terms of harm and purity, and how these ratings predicted individual differences of moral wrongness. And again, what they expected was that judgement of moral wrongness would be predicted by the extent that participants perceived the suicide as impure, but not by the extent that they perceived the suicide to have caused harm.
As for the homicide obituaries, which obviously involved an other-directed act, they predicted the opposite effect. So they thought that harm judgements would drive moral wrongness, but not purity judgements.
MTurk study of 174 US participants
Participants read 8 fabricated obituaries of men and women who had either taken their own lives or who had been killed by someone else. The description of the deceased was exactly the same across conditions – except in one condition it was stated that the deaths were due to suicide, and in the other condition, they were all due to homicide.
After reading each obituary, each participant answered 5 questions about each obituary – how morally wrong the death was, how angry it made them feel, how disgusted it made them feel, and importantly, if it caused harm, as well as whether the purity of the victim’s soul was tainted as a result of the death.
After this, participants completed an ‘explicit justification task’ – which asked them the extent to which:
• Suicide/homicide is wrong because it directly hurts other people
• And suicide/homicide is wrong because it disrespects the sacredness and purity of the self.

They expected that for both suicide and homicide, people would claim that harm concerns were much more influential in their wrongness judgements.
Participants also completed a couple more measures – MFQ measures of harm and purity.
Finally, participants completed a trait anger and trait disgust measure, as well as political conservatism and religiosity. They included these latter 2 ideological variables because they expected these would predict moral condemnations of suicide but not homicide.


Figure 1 – mean ratings of the obituary ratings. Homicide was more wrong than suicide
*Homicide judged to be more harmful than suicide; and suicide judged to be more impure than homicide.
*Both emotions of anger and disgust were more strongly experienced for homicide than suicide.
The next step was to examine what predicted moral wrongness judgements of the homicide and suicide obituaries. To do this, they used linear regression for the suicide judgements, but logistic regression for the homicide moral judgements because of the highly skewed data. (see tables 4 and 5 for output)

Firstly regressed moral judgements of wrongness onto ratings of harmfulness and impurity. For suicide, it was purity that significantly predicted wrongness ratings, whereas harm did not. And for homicide, the opposite effect occurred – harm predicted wrongness judgements. So this is exactly what they were expecting – purity is driving moral wrongness judgements for suicide.

They also ran regression analyses using the emotions of anger and disgust as IVs instead of the purity and harm ratings. As predicted they also found that disgust predicted wrongness ratings of suicide, but not anger. But they couldn’t do this same analysis for homicide wrongness ratings, due to issues of multicollnearity.

In regression number 3, they regressed judgements of moral wrongness onto MFQ purity and harm ratings. Mirroring the harm and purity obituary ratings, they found that greater general concerns about purity predicted wrongness judgements for the suicide obituaries, but not greater concerns about harm. And vice versa for the homicide obituaries.

And for number 4, for the trait emotions – trait disgust predicted moral judgements of the suicide, but not trait anger. For the homicide obituary, neither trait anger or disgust predicted moral wrongness.

They then created a composite of harm/anger and purity/disgust and found that the purity/disgust composite predicted moral judgements for the suicide, but not harm/anger. For the homicide, harm/anger predicted moral wrongness, and oddly enough purity/disgust negatively predicted moral wrongness, although this was only marginal.

Finally, they entered all 8 predictor variables into the regression at once - Purity ratings and general concerns about purity predicted moral judgements of suicide, while harm ratings predicted moral judgements of homicide.

They then looked at the impact of the ideological variables (religiosity and political conservatism) and how this affected moral wrongness judgements. They found that both religiosity and political conservatism predicted wrongness ratings for suicide, bot not for homicide. So participants who are more religious and more politically conservative are more likely to believe that suicide is wrong.

But what’s interesting is that political conservatism and religiosity didn’t account for the link between purity concerns and moral judgements of suicide. Even when controlling for political conservatism and reliogisity, obituary purity ratings and general purity concerns still predicted moral wrongness judgements of the suicide. And even restricting the sample to just non-religious liberals, they found that purity and MFQ purity were still significant predictors of moral judgements.

To look at this further they conducted a number of mediation analyses (in the supplementary materials). What they find is that the link between ideological variables (political conservatism and religiosity) and judgements of suicide wrongness are mediated by both obituary purity ratings and MFQ purity concerns.

In terms of the explicit justifications – they found what they expected. At an explicit level, participants rated the harm principle overall to be a more relevant justification than the purity principle for BOTH suicide and homicide. So participants are explicitly saying suicide is wrong because of the harm it causes, even though this doesn’t reflect the underlying source of their judgements – that it’s actually purity concerns that is impacting moral judgements.

Preliminary evidence that suicide is considered morally wrong because it is perceived to taint the soul, and not because of the harm it causes.

Study 2

Wanted to extend upon study 1 by addressing one of its potential limitations – which was that the harm measure used (“Did Bob’s homicide/suicide cause harm?”) may have been a bit too vague or broad to capture how participants actually construed the harm.

Again using obituaries, this time around they asked participants about 3 types of harm – harm being done to others, to the self, and to God.

89 mturk participants. This time there was no homicide obituaries – all participants read 8 suicide obituaries.
Same wrongness and purity measures in study 1, but 3 measures of harm – to self, to others, to God.
Explicit justification task, and ideological variables just as study 1.

Results …
Once again, mirroring study 1, purity ratings independently predicted judgements of moral wrongness. However harm ratings did not – and this was across all 3 types of harm. Harm to others did not predict moral wrongness ratings, nor did harm to self, or harm to God. They take this to suggest that suicide is considered more morally wrong when concerns about purity rather than harm are elevated.

And the explicit justification and ideological results match on to the findings from study 1 as well.

General discussion
The authors also run 4 replication studies which are briefly featured in the supplementary material.

The authors conclude that suicide is considered wrong to the extent that it taints the soul – not the degree to which it is perceived as harmful. They consider it a purity-based violation. So they suggest that these results call into question harm-based or dyadic theories of morality a la Kurt Gray.

The authors offer some suggestions for future research
• Different measures of disgust (e.g., physiological), considering the problematic nature of linguistic measures of emotions
• Extend their research to other forms of suicide – assisted suicide, suicide terrorism, and honour suicide.
• They also think it’s important to look at factors that attenuate moral condemnation of suicide. Because factors that reduce moral judgements of harm-based transgressions differ from those that mitigate moral judgements of purity-based transgressions.

Add new comment

Filtered HTML

  • Web page addresses and e-mail addresses turn into links automatically.
  • Allowed HTML tags: <a> <em> <strong> <cite> <blockquote> <code> <ul> <ol> <li> <dl> <dt> <dd>
  • Lines and paragraphs break automatically.

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.
This question is to help prevent automated spam submissions.
Enter the characters shown in the image.