1. OVERVIEW OF THE BOOK

The book has a clear set of ‘learning objectives’ articulated at the start, viz., to:

• • describe the identified person bias,

• • appreciate its cognitive underpinning,

• • understand arguments in favour or against its moral justifiability, and

• • reflect on its practical import.

These objectives shape the structure of the volume. It opens with a short section on ‘social sciences’, which presents the cognitive underpinning of the identified victim effect, then moves on to the perspective of ‘ethics and political philosophy’ to explore the justifiability of the effect, and concludes with a series of applications to illustrate the impact of the effect and the role that research insights could play in policies, in particular in environmental legislation and healthcare priority setting.

The editors invite perspectives from multiple disciplines including, in order of appearance: cognitive psychologists, philosophers, medical ethicists, economists, political philosophers, population ethicists, bio-ethicists and lawyers. A common, useful thread in each chapter is the attention to their implications for public policy.

2. THE IDENTIFIED VICTIM EFFECT IN A DUAL PROCESSING FRAMEWORK

The first section of the book is entitled ‘Social Science’. The reader expecting a broad brush on the perspectives of social sciences such as anthropology, political science or sociology may be disappointed, because this section consists of just two chapters that focus on cognitive psychology. Nonetheless, as I argue below, the narrow focus of this first section gives this whole book a unique and insightful spin.

Deborah Small presents the current understanding of the cognitive underpinnings of the identified victim effect. Pioneering work in the 1990s suggested that the bias depended on the reference group size, i.e. on the proportion of those at risk that could be saved. For instance, saving all 33 miners who were trapped in the mine near Copiaco in Chile meant saving 100% of the reference group. The prediction from this seminal work is that, if we could have saved 33 identified miners from a very large group of trapped miners, our motivation to save them would have been weaker.

Small focuses on the subsequent developments of cognitive research and she frames the problem in terms of ‘System 1’ and ‘System 2’, an umbrella framework proposed by Kahneman and Frederick (Reference Kahneman, Frederick, Gilovich, Griffin and Kahneman2002) to capture the essence of multi-process theories. System 1 drives intuitive and emotional judgements, which may lead us to give priority to identified lives. System 2 drives a more deliberate judgement. Many psychologists suggest that we rely mostly on System 1 to go about our day. However, if sufficiently motivated, we might invest effort in thinking through a judgment and slowly invoke System 2 to inform our actions. Small unpacks the underlying mechanism of the ‘identified victim effect’ showing how different ways to describe a life at risk are more likely to activate a System 1 or a System 2 judgement. In particular, her research suggests that the key mechanism of the effect lies in psychological distance. If the potential beneficiaries of a particular action are described (a) vividly, (b) specifically and (c) representing a high proportion of a reference group, then the psychological distance between us and these potential beneficiaries will be reduced and the affective response heightened.

The dual processing framework is completely reinterpreted in the essay by Peter Railton. He provides an interesting, evolutionary account of the tension between System 1 and System 2.

Railton cautions against simplified descriptions of the dual processing framework, which equate System 1 with passion and irrationality and System 2 with calculated judgement and rationality. According to these simplified accounts, System 1 is like a primordial, animal response. When the System 1 response is in conflict with that of System 2, it is of ‘inferior quality’. Based on these simplified accounts, when confronted with the identified victim effect we ought to resist the instinctual response and suspend any action until we identify, through formal deliberation (System 2), what the morally correct action is.

In contrast, Railton argues that System 1 evolves over time to embody the rational, optimal judgement, which humans, as animals, have acquired through experience. System 1 is ‘designed for experience-based construction of expectation-based evaluative models of situation, actions and outcomes’ (33). System 1 contributes apt responsiveness to reasons, for instance through ‘gut feelings’.

A possible explanation for the identified victim effect is that we are rarely confronted with the choice between identified and statistical lives typical of experimental studies in cognitive psychology or thought experiments of moral philosophers. System 1 is hence currently unable to answer those exact questions and it answers, instead, ecologically valid proxies. For instance, experimental participants may be asked to choose between action A, which saves an identified person, or action B, which will certainly save one person at random from a group of five individuals. In fact, however, experimental participants may be simply drawing from experience of the real world, in which the lack of identifying information is usually associated with less certainty about the benefit or harm.

The provocative insight from this evolutionary account, in my understanding, is that we need to find ways to expose ourselves more frequently to the moral conflict of choosing between identified and statistical lives in order to build sufficient experience to inform our System 1 judgements. Then, and only then, we can revisit the identified victim effect. We may find that System 1 is in line with a rationalistic and consequentialist approach. Or we may find that we are better able to articulate the moral underpinning of favouring identified lives and embed these moral justifications in our System 2.

To speed up the process of reconciliation, Railton suggests looking into the development of a process similar to what flight simulators do in developing pilots’ intuitive response to emergency situations. It is, however, not clear how this ‘simulator’ can be practically organized to inform policy-makers.

3. ON NORMATIVE FORCE

The core of the book explores whether the identified victim effect taps into something morally relevant. The editors aimed to achieve a balance of views and describe the central section of the book as comprising two camps. On the first camp, four contributions claim there is no or very little normative force in the distinction (Dan Brock, Matthew Adler, Michael Otsuka and Nir Eyal) and four present arguments to support the opposite view (Norman Daniels, Caspar Hare, Marcel Verweij and Michael Slote).

I read these chapters through my disciplinary lenses of Neoclassical Economics and Decision Analysis. Both disciplines take a rationalistic and consequentialist stance, which is also the perspective taken by most authors in this book, especially those focusing on the normative underpinning of the effect. Unsurprisingly, I found their arguments elegant, logically convincing, and, at first, reassuring.

Four essays powerfully argue that an identified and a statistical life have the same value, and hence there is no moral force behind the identified victim effect. Brock revisits seven arguments that have been used to justify the identified victim effect (rule of rescue, urgency, aggregation, priority of the worse off, uncertainty, temporal discounting and special relationships). He dismisses each one in turn, showing that either they have no moral force or, if they tap into a morally relevant principle, this principle does not map directly onto the distinction between identified and statistical lives.

Adler engages the reader in a fascinating reflection on ethics and welfare theories. He argues that the key distinction is between ex-ante and ex-post approaches in comparing alternative states of the world. This distinction implies different ‘currencies’: expected well-being for the ex ante approach and final utilities for the ex post approach. He favours the ex-post approach, which implies rejecting the bias in favour of identified victims. Two other essays (by Michael Otsuka, Nir Eyal) analyse the nature of the risk of a statistical life and firmly reject the argument that the risk of harm to which an identified person is exposed should command any priority through a series of thought experiments.

These essays persuaded me that there is no difference in the value of a statistical and an identified life, but they failed to persuade me that the identified victim effect is a bias. A ‘bias’ is a systematic judgement error. For instance, I suffer from ‘attribution bias’: in considering how much of my time I invest in the smooth running of our household, at times I feel I do too much and my partner is not pulling his weight. Upon reflection, I realize that I had a distorted view, because I could recollect instances of my efforts more readily compared with those of my partner. By making my partner's contribution more salient to me, I can revise my judgement and come to a more balanced view in which System 1 and System 2 are not in conflict. My immediate response was a bias and there exists a de-biasing technique. The elegant and logically convincing argument presented at the beginning of this book section, however, did not effect a similar judgement revision with respect to the identified victim effect. On one hand, I think and feel there is no difference in the value of identified and statistical lives. On the other hand, I feel there is additional value attached to prioritizing rescue, which the rationalistic, consequentialist perspectives may be unable to capture.

In their essays, Marcel Verweij and Michael Slote put forward very different ethical perspectives, which may be able to capture the moral intuition behind the identified victim effect. They move our attention to the value we derive (as society or individuals) from the act of saving an identified versus a statistical life. Verweij considers ‘the importance of the fact that people are standing together, sharing hope and fear, and supporting each other in the face of – and fight against – disaster’ (145). This solidaristic perspective may give normative force to favouring identified lives, at least within the community that stands together. Slote argues that it would be fruitful to explore the role of empathy and emotions in morality. Slote's view is that ‘moral goodness is identical with having full and fully empathic concerns for others and that the wrongness of an action consists in its reflecting or exhibiting a lack of such concerns for others’ (155). Slote argues that, if our empathic concerns for an identified victim are stronger than the empathic concern of a statistical one, we ought to prioritize the former, and this priority is morally justified.

Have Verweij and Slote solved the tension? Am I persuaded that solidarity or empathy is the primitive for moral judgement? Can moral obligation follow from the actual pattern of our empathic concerns? I am not yet persuaded, but it seems their views should be more carefully investigated, because their policy implications are profound.

Let us consider for instance Slote's argument and relate his perspective to our current understanding of cognitive underpinning of the identified victim effect using, as suggested by Small, a dual processing framework. Policy analysts ought to describe statistical victims vividly, specifically and, whenever possible, highlight that the victims represent a high proportion of a reference group. This description will contribute to activating empathy towards statistical victims and consequently lead to better decisions. Failing to activate System 1 may lead to apathy and to morally wrong actions.

There is evidence to support perspectives such as Slote's and Verweij's. For instance, the National Institute for Health and Care Excellence (NICE) in the UK regularly consults a representative group of about 30 citizens to ensure public values are reflected in the Institute's methodology to develop its guidance. In 2006 this Citizen Council considered the moral standing of the identified victim effect. After two-and-a-half days of deliberation, they reported that the ‘importance of putting compassion above pure economics emerged in several answers. One [member of the Citizen Council] gave the example of the costly and ultimately unsuccessful rescue attempt of a whale stranded in the Thames earlier [that] year. “We would be diminished as a society if we hadn't made that attempt.” A minority believed firmly that life-saving should have no special priority over disease prevention programmes’ (NICE Citizens Council 2006; 9). It is unclear, however, how the principle of empathy or compassion could be operationalized in making policy recommendations.

What should policy analysts do in the absence of a definitive answer on the moral standing of the identified victim effect? A pragmatic option may lie in procedural justice, as outlined in the essay by Norman Daniels. He highlights that there is ‘reasonable disagreement’ among ‘reasonable’ people about the moral standing of the effect. He recommends a deliberative framework to debate policy decisions involving a trade-off between identified and statistical lives. The framework of ‘accountability for reasonableness’ (A4R; Daniels and Sabin Reference Daniels and Sabin2002) is an example of such a process. Daniels outlines elsewhere how A4R could be operationalized (Daniels Reference Daniels2008), but we still lack sufficient empirical evidence to evaluate its effectiveness in securing (perceived) fairness in the allocation of resources.

In principle, processes like A4R may play the role of the ‘simulator’ invoked by Railton. To play this role the process should take place in the context of particular policies, rather than as a consultation about normative principles, given that the current state of moral debate has not resolved the disagreement on principle. We should not be surprised, for instance, that the Citizens Council of NICE was not immune from the identified victim effect and could not confidently recommend NICE to ignore it in its guidance.