Contextualizing sacrificial dilemmas within Covid-19 for the study of moral judgment.

"Sacrificial dilemmas" are the scenarios typically used to study moral judgment and human morality. However, these dilemmas have been criticized regarding their lack of ecological validity. The COVID-19 pandemic offers a relevant context to further examine individuals' moral judgment and choice of action with more realistic sacrificial dilemmas. Using this context, the purpose of the present study is to investigate how moral responses are influenced by the contextualization of the dilemma (i.e., contextualized or not within the Covid-19 pandemic). By comparing two versions of one dilemma, Experiment 1 revealed that the more realistic version (the one contextualized within the Covid-19 pandemic) did not elicit more utilitarian responses than the less realistic version (the one not contextualized within the Covid-19 pandemic). In Experiment 2, we examined more specifically whether both the perceived realism of the dilemma and the plausibility of a utilitarian action influence moral responses. Results confirmed that the contextualization of the dilemma does not make any difference in moral responses. However, the plausibility of an action appears to exert an influence on the choice of action. Indeed, participants were more inclined to choose the utilitarian action in the plausible action versions than in the implausible action versions of the dilemma. Overall, these results shed light on the importance for future research of using mundane and dramatic realistic dilemmas displaying full information regarding a sacrificial action and its consequences.

Introduction

Today, the lack of perceived realism of scenarios used to study moral judgment and human morality constitutes one of the main criticisms of research in this area [1]. In a well-known example, the classic trolley dilemma [2, 3], individuals are told that a runaway trolley speeding down the tracks is about to kill a group of five workers who are unable to move out of the way in time. The only way to prevent the deaths of these five is to pull a lever to redirect the trolley to another track where it will kill only one person. Is it morally acceptable to kill one person to save five others? Bauman et al. [1] argue that people’s responses to this type of sacrificial dilemma have low external validity because the scenarios used are unrealistic in three distinct ways: they are low in experimental, mundane and psychological realism. According to the definitions of Aronson et al. [4], a study is low in experimental realism when the situation is not engaging for the participants, is not impactful to them or is not taken seriously; and low in mundane realism when the situation is in no way similar to what people encounter in everyday life. In this respect, classical sacrificial dilemmas are low in experimental and mundane realism. Indeed, it seems unusual to imagine being in a situation like the one described in the trolley dilemma and to take these scenarios seriously. Moreover, in everyday life, it is rare to make choices that involve the sacrifice of human life. Finally, hypothetical dilemmas do not elicit the same psychological processes as other real-world moral situations (i.e., they are low in psychological realism).

In sum, the commonly used sacrificial dilemmas are different from more realistic moral situations and, therefore, the trade-offs between deontological principles (e.g., do not kill) and utilitarian consequences (e.g., kill one to save five) may depend on the perceived realism of the situation. In this respect, the coronavirus (Covid-19) pandemic could be viewed as an opportunity to examine moral decisions about life and death with more realistic dilemmas. In this viral pandemic context, several countries have been confronted with shortages of ventilators, intensive care beds and medical personnel. The question of the allocation of scarce medical resources was therefore raised [5]. For example, you are an emergency room physician with five dying patients who have been admitted and one patient in critical condition. Would you remove the patient from a ventilator to provide it to others? From a utilitarian point of view, approving the sacrifice of one life in order to save the lives of five others is acceptable with respect to the number of lives saved. By contrast, from a deontological point of view, it is not acceptable to sacrifice the life of one person in order to save others because the value of each human life is paramount and no one has the right to sacrifice a life, regardless of any benefits that may arise in doing so.

The two types of sacrificial dilemmas (i.e., traditional trolley dilemmas and dilemmas related to the Covid-19 pandemic) correspond to problematic situations that involve a conflict between utilitarian and deontological principles. Thus, the question is to know whether subjects’ hypothetical moral responses are predictive of the moral responses they would display in real-life dilemmas (e.g., those experienced during Covid-19). In line with this question, some researchers have tried to increase the realism of sacrificial dilemmas and, consequently, their ecological validity. For instance, one solution proposed by Greene et al. [6] was to ask participants to “suspend their disbelief” while reading the scenarios and to exclude from study those who reported lack of success (i.e., those who reported being unable/unwilling to do it). In line with this, Christensen and Gomila [7] conclude their review of moral dilemma tasks by suggesting that participants be instructed that the dilemmas they are about to read are similar to those likely to occur in real-life.

Solutions to the lack of perceived realism

Virtual reality moral paradigms

In an attempt to improve the realism of hypothetical dilemmas, some researchers introduced moral paradigms under Virtual Reality (VR) [8-13]. Virtual Reality allows users to be immersed in a variety of moral situations that are designed to be more realistic and believable than hypothetical versions. Because immersive VR affects human behavior as a real equivalent environmental experience [14-16], it provides an alternative for studying moral responses in a more ecological way. Note also that the realistic simulation of sacrificial dilemmas allows for the examination, in real-time, of rule-violating behavior (e.g., involving physical harm) that would be impossible to study in real world settings. Navarrete et al. [10] were the first to integrate, in immersive VR, the trolley problem consisting in switching the trolley’s direction to kill one person instead of five. They found that 90% of the participants chose the utilitarian solution, a behavior pattern that is congruent with judgments of participants from previous studies using written scenario descriptions. In another study, the participants were the drivers of the train and the same pattern of results was reported [17]. By contrast, further research found a greater endorsement of utilitarian responses in VR compared to the same scenario described textually [8, 12, 13]. Niforatos et al. [11] repeatedly exposed participants to VR and text format versions of the trolley dilemma and concluded that the VR enactment of a moral dilemma fosters utilitarian decision-making. In these studies, people become more utilitarian when they are placed in an immersive virtual environment compared to a situation in which they mentally visualize the dilemma through reading the scenario (i.e., text format). This difference could be explained by the fact that the VR format of the dilemmas could elicit artificial gaming behaviors where sacrificial action may be trivialized [8]. Another possibility that deserves to be considered is that in paper-based dilemmas, participants simply express a moral judgment, whereas in VR versions, they actually perform the action: two moral responses that are likely underlined by distinct psychological mechanisms [18]. There is a gap between what people think they would do when faced with a hypothetical sacrificial dilemma and what they would do when faced with a more realistic scenario. In line with this conclusion, Bostyn et al. [19] found that participants are more willing to be utilitarian when confronted with a real-life sacrificial dilemma. In their study, participants had to make the real-life decision to administer a very painful electroshock to a single mouse in order to save five. In a real decision, 84% of participants shocked the mouse, whereas in a hypothetical version of the same dilemma, only 66% of participants predicted that they would do that.

In the VR studies and the mouse scenario described above [19], moral dilemmas were enacted and participants experienced the feeling of “being there” (e.g., [14, 16, 20]). In these studies, they also carried out harming actions (i.e., killing or letting die) and probably experienced an illusion of reality compared to conditions in which they mentally visualized dilemmas (i.e., in which dilemmas were presented in text format). Because of their ability to provide a feeling of “being there”, VR scenarios increase experimental realism [21] and thus could contribute to increasing the perceived realism of sacrificial scenarios compared to the paper-based equivalent. However, Weber et al. [22] claimed that the immersed illusion (i.e., having a sensation of being in a real place) is not sufficient to create a virtual world that is perceived as realistic; the scenario must also be close to real life. More precisely, the plausibility of the environment, addressed as “the overall credibility of the scenario being depicted in comparison with expectations” ([20], p. 3549) is a contributing factor that modulates perceived realism of virtual environments. It’s probably the plausibility of scenarios that constitutes the main gap between reality and the traditional moral-dilemma paradigm. Indeed, most sacrificial dilemmas correspond to outlandish scenarios that differ considerably from moral situations that people face in real life. As Bauman et al. [1] argued, “trolley problems are unrealistic and unlike anything people encounter in the real world” (p. 545). Consistent with this criticism, Terbeck et al. [13] suggested that future VR studies might investigate moral situations that are more plausible.

Context-related studies

Gold et al. [23, 24] (see also [25]) have tried to overcome the lack of realism of sacrificial dilemmas by presenting trolley problems that elicit actions and moral judgments about decisions likely to occur in real life. They used a computer animation where participants were asked to make decisions that would influence the amount of money donated to children living in an orphanage in northern Uganda. In this Orphan scenario, participants decided whether varying amounts of money, denominated in meals, would be taken away from either five children or one. For this purpose, they had the option of clicking a switch to divert a meal from one child to five others. To increase the realism of the scenario, participants were shown photos and short biographies of some of the orphans and were told that the donation to the orphanage would depend on their decisions during the experiment. Gold et al. ([23], study 2) also used a standard trolley scenario. In this version, a lower percentage of participants said that they would take action (i.e., click the switch to save the five) than in the corresponding real-life scenario (with orphans). Note that this difference in utilitarian responses between the two scenarios is not necessarily due to perceived realism. In fact, there are two obvious differences between the two scenarios that may influence participants’ responses: the type of victims (i.e., children in an orphanage vs. men working on a train track) and the harm inflicted (i.e., depriving of meals vs. killing).

As many researchers have pointed out, contextualizing the action (e.g., killing or letting die in sacrificial dilemmas) can help to study moral responses as they are made in real life (e.g., [26-28]. In Watkins and Laham’s [29] research, participants responded to trolley problem scenarios in two different contexts (i.e., in a war context and a peace context). The findings demonstrated that the context changes judgment on sacrificial dilemmas: killing one person to save five was judged more acceptable in a war context than in a peace context. Following Watkins and Laham’s [29] research, Christen et al. [26] wanted to understand moral judgment in more real-life trolley problem scenarios involving military, firefighting, and surveillance missions. In their study, participants made decisions about remotely piloted aircraft that resulted in sacrificial outcomes.

This type of dilemma patterned after the trolley problem scenarios is expected to be more realistic than the decontextualized ones because it is transplanted into contexts that often involve life and death decisions (i.e., in which the harmful action is more plausible). The moral decision in the trolley problem dilemma situated in a war against terror (whether or not to redirect a missile) may seem more plausible compared to the same kind of decision in the classical scenario (whether or not to redirect the trolley). However, with the trolley problem, even if a context is provided, it is difficult to imagine oneself in such a situation because the scenario described does not correspond to any lived experience. It is this lack of mundane realism that Bauman et al. [1] point out in their criticisms of sacrificial dilemmas. They argue that “trolley problems also lack mundane realism because the catastrophes depicted in sacrificial dilemmas differ considerably from the type and scale of moral situations people typically face in real life” (p. 542).

Because the COVID-19 pandemic is a real-world crisis that has confronted us with decisions about life and death, it provides a unique opportunity to address previous criticisms (the lack of mundane realism in particular) and thus to investigate moral judgment in real-world dilemmas.

Contextualizing sacrificial dilemmas with the Covid-19 pandemic

Dilemmas posed by the Covid-19 pandemic involved trade-offs between human lives. Indeed, the massive influx of patients and shortages of healthcare resources, such as anti-virals, intensive care unit beds and mechanical ventilators, led doctors to sometimes make moral decisions about which patients to admit to the Intensive Care Unit (ICU) and for which ones to provide lifesaving resources [30, 31]. Many studies have investigated public attitudes toward ethical principles underlying these pandemic triage dilemmas (e.g., [32-34]). For some authors, sacrificial dilemmas related to the Covid-19 pandemic were perceived as an alternative to traditional sacrificial dilemmas and as a unique opportunity to overcome criticisms about the lack of realism of hypothetical moral scenarios. According to Kneer and Hannikainen [35]: « Sacrificial dilemmas related to the Covid-19 pandemic—henceforth triage/critical care dilemmas—have both experimental and mundane realism; they are real-life situations with which most participants are at least indirectly acquainted. Consequently, we suspected that their level of psychological realism would also be high” (p.7). In one of their studies ([35], study 3), they investigated the effect of realism with dilemmas patterned after the trolley problem scenarios contextualized, or not, within the medical triage context. They observed a greater utilitarian tendency in response to triage dilemmas compared to dilemmas unrelated to Covid-19: Participants were more likely to disconnect an oxygen tank used to treat a single coronavirus patient in order to save five others than to order firefighters to stop rescuing a single person trapped in a burning house in order to save five others in another burning house. Note that this tendency to favor utilitarian response in triage dilemmas, relative to dilemmas unrelated to the Covid-19 pandemic, is not necessarily due to a difference in realism between the two types of dilemmas. Indeed, the two types of dilemmas also differed regarding the nature of the moral response: medical in the triage context and non-medical in the others.

From our perspective, the solutions provided to the lack of realism of sacrificial dilemmas are still insufficient for two reasons. First, in the vast majority of studies conducted on moral judgment, researchers did not control the perceived realism of the scenarios. They started from the assumption that the situations and actions described were perceived as realistic by the participants. Kneer and Hannikainen [35], for example, argued that sacrificial dilemmas related to the Covid-19 pandemic are high in experimental, mundane and psychological realism compared to dilemmas unrelated to Covid-19, and hypothesized that the former are perceived as more realistic than the latter. But they did not measure the perceived realism of the two types of dilemmas. Second, perceived realism is defined as a one-dimension construct. In virtual reality studies, perceived realism refers to the subjective perception of being present in the depicted scenario (i.e., experimental realism, according to Bauman et al. [1]), whereas in context-related studies, perceived realism refers to the degree to which the depicted scenario could possibly occur in the real word. Finally, in studies conducted during Covid-19, perceived realism refers to the scenario’s resemblance to participants’ current real-world experiences (i.e., mundane realism, according to Bauman et al. [1]). Based on this analysis of the strengths and weaknesses of previous researches conducted on moral judgment, the aim of the present study was to investigate more thoroughly the issue of realism and its influence on participants’ moral responses to sacrificial dilemmas.

Perceived realism

In a media and more specifically in a narrative environment, perceived realism is the degree to which the narrative environment reflects the real world [36, 37] and conveys coherence and genuineness [38]. Busselle and Bilandzic [38] place this construct at the center of their model of narrative comprehension. They explain that a perceived lack of realism affects the understanding of the narrative and can alter attitudes and behaviors toward the narrative. In the current state of knowledge, scholars argue that perceived realism is a multidimensional construct (e.g., [38-40]). They agree on three dimensions necessary to create a realistic narrative: plausibility, factuality and typicality (e.g., [41]). Plausibility concerns the possible occurrence of the story in reality (i.e., how likely it is that the event in the scenario could possibly happen in real life); factuality concerns actual occurrence of the story in reality (i.e., how well the event in scenario depicts something that really happened) and typicality refers to the degree to which the story is perceived to be similar to events in one’s real life (i.e., how well the event in the scenario reflects people’s past and present experiences) [39, 40]. Some of the previous research (the context-related studies) uses scenarios that meet certain realism requirements, such as plausibility and factuality, but do not correspond to what most people actually experience (typicality). In this regard, the Covid-19 pandemic triage dilemmas are scenarios in which all three dimensions of perceived realism could be simultaneously included. Thus, they may provide a unique opportunity to study the role of perceived realism on moral judgment.

Because realism is a multi-faceted construct, Körner et al. [42] also raised the question of the plausibility of utilitarian actions in hypothetical sacrificial dilemmas. In the footbridge scenario, for example, the only way to save the lives of the five workers is to push a stranger off a footbridge and onto the tracks below where his body would stop the trolley. Readers of this variant of the trolley problem may find the situation plausible but the action implausible (it seems implausible that a man’s body would be sufficient to stop a trolley). Körner et al. [42] showed that plausibility of action, specifically plausibility of stated consequences (i.e., how probable is it that the action will achieve the desired outcomes?) and plausibility of alternatives (i.e., how probable is it that there are other, feasible and reasonable actions to achieve the desired outcomes?), influence moral judgments. Participants judged the utilitarian action (i.e., killing one person to save five) to be more appropriate in plausible action dilemmas compared to implausible action dilemmas. Thus, the higher proportion of utilitarian responses usually reported in the trolley scenario (an impersonal dilemma) compared to the footbridge scenario (a personal dilemma) could be the result of a more plausible action in the former dilemma than in the latter. Namely, it seems more plausible to save the lives of the five workers by pulling a lever to redirect the trolley to another track than by trying to stop it by pushing a man off a footbridge and onto the tracks. In line with this interpretation, Shou et al. [43] showed that participants generally had greater certainty that the five people would survive if they decided to kill the individual in impersonal dilemmas compared to personal dilemmas: an additional result that invites us to consider the plausibility of a sacrificial action in the study of moral judgment.

In the two present studies, we had two main objectives: 1) We investigated the role of perceived realism on moral responses by distinguishing and measuring different dimensions of perceived realism in sacrificial dilemmas; 2) We attempted to address criticisms about the lack of realism of sacrificial dilemmas by comparing different versions of the triage dilemma. In study 1, we specifically investigated the perceived realism of the scenario with a version contextualized within the Covid-19 pandemic (the supposedly realistic version) and another version not contextualized within the Covid-19 pandemic (the supposedly less realistic one). We hypothesize that a scenario perceived as realistic should lead to more utilitarian responses than one perceived as less realistic. In study 2, we examined whether moral responses depend on the perceived realism of the scenario as well as on the plausibility of utilitarian actions. To that aim, we used realistic and less realistic triage dilemmas that differed in action plausibility.

Experiment 1

Method

Participants

One hundred and fifty-five French participants took part in this study during the COVID-19 crisis, between January 9 and January 17, 2021. However, 15 of them (those who reported having a close relative or friend that had died from the virus) were excluded from analyses, leaving the present sample at n = 140 (126 female, M = 29.51, SD = 11.31). This exclusion criterion was determined before data collection. Indeed, to the extent that the implication of a close relative influences participants’ decisions in sacrificial dilemmas (see Tassy et al., [18]), this exclusion criterion ensured that, in the context of the pandemic in which the study was conducted, the potential victim(s) of the decision had no proximity (i.e., affective or genetic) to the participant who had to make a decision. Participants were not paid for their participation.

Materials

The TRIAGE dilemma was presented in two versions: a version contextualized within the Covid-19 pandemic (the supposedly realistic version) and another version not contextualized within the Covid-19 pandemic (the supposedly less realistic version) (Table 1). The TRIAGE dilemma presented the option of taking away an oxygen reserve or antibiotics from one person to save five others. The two versions were similar with regard to the following dimensions: the dilemma involved killing one person in order to save several others, the number of people saved was identical (N = 5) and the potential victims were unknown and not affectively or genetically related to the participants (according to our exclusion criterion). In addition, the utilitarian response was associated with the withdrawal of life-saving resources (i.e., an oxygen reserve or antibiotics) in the health domain which ensured that the two versions were indistinguishable in regards to ethical principles. Finally, the two versions of dilemma contained exactly the same number of words. The main difference between the two versions of the dilemma concerned the contextualization of the scenario (contextualized or not within the Covid-19 pandemic).

Table 1

Contextualized and non-contextualized versions of the TRIAGE dilemma used in experiment 1 (translated from French).

Versions contextualized within COVID-19

Versions not contextualized within COVID-19

You are the department head of a hospital in eastern France. A new coronavirus from China which causes respiratory irritation has appeared. Every day you receive more and more new patients with breathing problems. You don’t have enough oxygen for all of the patients. Five new patients are admitted to the hospital’s intensive care unit. Their health condition requires immediate hospitalization and the administration of oxygen for the next 15 days. There is no more oxygen available, and you have no way to get it. The only way to save the five patients is to take an oxygen tank from one of your patients who is in critical condition. If you do that, the patient will die but the other five will be saved.

You are the department head of a hospital in eastern France. A new bacterium has contaminated the water of the city. Every day you receive more and more new patients with intestinal disorders and blood poisoning. You do not have enough antibiotics for all of the patients. Five new patients are admitted to the hospital’s intensive care unit. Their health condition requires immediate hospitalization and a dose of antibiotics. There are no more antibiotics available, and you have no way to get some. The only way to save the five patients is to take antibiotics from one of your patients who is in critical condition. If you do that, the patient will die but the other five will be saved.

Procedure and measures

Participants were tested using an online questionnaire created on the platform Qualtrics (https://www.qualtrics.com). Note that French Law does not require approval of an ethics committee when data from a survey are collected and analyzed anonymously.

After giving their informed consent, participants were randomly assigned either to the contextualized version of the TRIAGE dilemma (N = 70) or to the non- contextualized one (N = 70). The dilemma was briefly presented by stating that it refers to a serious situation that could be seen as unpleasant but requires making a difficult choice.

After reading the dilemma, participants successively provided two moral responses: a moral judgment and a choice of action. Participants were asked to be as honest as possible in their responses, knowing that there was no right or wrong answer. Immediately afterwards, they were again instructed to carefully read the same version of the dilemma presented earlier and to evaluate the dilemma from a moral perspective by answering various questions (i.e., perceived realism and plausibility of stated consequences). Finally, participants provided demographic information and were asked the following yes/no question: “have you lost a close relative or friend to COVID-19?”. This question was asked at the end of the experiment and not at the beginning in order to avoid raising the saliency of the pandemic context, especially among participants assigned to the version of the dilemma which was not contextualized within the Covid-19 pandemic.

Moral responses measures

Because there is a huge difference between what one judges as morally acceptable and what one actually does (see Tassy et al., [18]), participants responded successively to two questions, one targeting moral judgment and the other targeting choice of action. They rated to what extent the utilitarian action was appropriate or not (i.e., moral judgment task). For each version, the question was “How appropriate is it for you to take [the oxygen / antibiotics] of one of your patients in order to save the five?”. They were also asked whether they would perform the utilitarian action (i.e., choice of action task). For each version, the question was «Would you take [the oxygen / antibiotics] of one of your patients in order to save the five?” These two questions were answered on a 6-point scale (1 = not at all; 6 = definitely) with higher scores being closer to utilitarian responses.

Perceived realism measures

To check the effectiveness of our realism manipulation, we also assessed the perceived realism of the two versions of the TRIAGE dilemma. In line with authors who argue that realism perceptions are multidimensional (e.g., [38-40], our assessment of perceived realism included three sub-dimensions: perceived plausibility, typicality, and factuality. The question related to plausibility was “How probable do you think it is that the event in the scenario could possibly happen in real life?” The question related to factuality was “How probable do you think is it that the event in the scenario depicts something that really happened?” The question related to typicality was “How probable do you think is it that the event in the scenario reflects people’s past and present experiences?” (see [41] for similar measures). Responses to these three perceived realism measures were all rated on a 6-point scale (1 = not at all, to 6 = definitely).

Plausibility of stated consequences measures

Because the plausibility of the sacrificial action, especially the plausibility that the stated consequences will occur, might influence moral responses (see [42]), participants answered the following question “how probable do you think is it that this action would save the five people?" (see [42] for a similar measure). Responses were rated on the same 6-point scale (1 = not at all, to 6 = definitely).

Results

Manipulation check

A one-way analysis of variance (ANOVA) was run on three measures of perceived realism (i.e., plausibility, factuality and typicality) to evaluate the differences between the two dilemma versions (Contextualized vs. non contextualized within the Covid-19 pandemic).

The manipulation check confirmed that the two versions differed in perceived realism. A significant main effect was observed for plausibility F(1, 138) = 15.94, p < .001, η = .10, for factuality F(1, 138) = 23.14, p < .001, η = .14, but also for typicality F(1, 138) = 19.61, p < .001, η = .12. Participants perceived the dilemma as more realistic when it was contextualized within the Covid-19 pandemic (M = 5.30, SD = 1.13; M = 5.44, SD = 1.05; M = 5.43, SD = 1.11) than when it was not contextualized (M = 4.40, SD = 1.51; M = 4.37, SD = 1.54; M = 4.43, SD = 1.53).

Pearson’s correlation analysis was performed to explore the correlation between the three dimensions of perceived realism. The results showed that plausibility was positively associated with factuality and typicity and that factuality was positively related to typicity, ps < .001 (see Table 2).

Table 2

Correlations between the three dimensions of perceived realism.

Dimensions	1	2	3
1. Plausibility	-	-	-
2. Factuality	.52	-	-
3. Typicity	.65	.67	-

Note. All ps < .001. A Pearson correlation coefficient ranging between 0.5 and 1 indicated a strong correlation.

Moral responses

To explore the effect of perceived realism on participants’ moral responses, a repeated-measure ANOVA was conducted with the Type of dilemma (Contextualized vs. non contextualized within the Covid-19 pandemic) as the between-subject factor, and the Type of moral response (Judgment vs. Choice of action) as the within-subject factor. This analysis revealed a significant main effect of Type of moral response, indicating that responses to choice of action (M = 3.93, SD = 1.50) were overall more utilitarian than responses to judgment (M = 3.41, SD = 1.54), F(1, 138) = 17.88, p < .001, ηp2 = 0.11. No other effects were significant.

To assess the relationship between perceived realism and moral responses, we also conducted Pearson’s correlation analyses between each of the three measures of realism (plausibility, factuality and typicality) and moral responses (moral judgment and choice of action). None of these measures were significantly correlated with moral judgment or with choice of action.

In order to further test the impact of perceived realism on moral responses, we conducted two multiple regression analyses predicting judgment and choice of action separately. For each analysis, we entered the three measures of perceived realism (i.e., plausibility, factuality and typicality) as predictors. Multiple regression analysis showed that none of the perceived realism measures was a significant predictor for moral responses (all ps > .10, see Table 3 for a summary of regression results).

Table 3

Multiple regression analysis on judgment and choice of action with perceived plausibility, factuality and typicality of dilemma as predictors.

Predictors	Beta	SE (B)	t	p	Semi-partial correlation
Predicting Judgment
Plausibility	0.039	0.125	0.310	0.757	0.027
Factuality	-0.061	0.126	-0.485	0.629	-0.041
Typicity	0.127	0.141	0.897	0.371	0.076
Predicting Choice of action
Plausibility	0.120	0.121	0.997	0.321	0.085
Factuality	-0.009	0.122	-0.077	0.939	-0.007
Typicity	0.044	0.137	0.321	0.749	0.028

Plausibility of stated consequences

A one-way ANOVA was run on this measure with the Type of dilemma (Contextualized vs. non contextualized within the Covid-19 pandemic) as the between-subject factor. No significant difference was found between the two versions of dilemma. Overall, the probability of the action saving the five people was rated as medium (M = 3.76, SD = 1.44). Moreover, Pearson’s correlation analysis showed that the plausibility of stated consequences and moral responses were positively correlated. That is, the more plausible participants rated the stated consequences, the more utilitarian their moral judgment and choice of action were (r = .18, p < .05 and r = .22, p < .01, respectively).

Discussion

Contrary to studies dealing with the influence of realism on moral responses in sacrificial dilemmas, in this first experiment we assessed the perceived plausibility, factuality, and typicality (three dimensions of perceived realism) of each employed dilemma. Although the two versions of the TRIAGE dilemma differed on these three dimensions, the more realistic version (the one contextualized within the Covid-19 pandemic) did not elicit more utilitarian responses than the less realistic version (the one not contextualized within the Covid-19 pandemic). This suggests that perceived realism did not influence moral responses. Subsequent regression analysis showed that none of the three dimensions of perceived realism was a significant predictor for moral judgment or choice of action.

One possible explanation is that participants’ moral responses are not based on what they perceive as realistic but on their beliefs about the plausibility of the stated consequences. According to Körner et al. [42] (see also [43, 44]), these beliefs may influence the expected usefulness of the sacrificial action and consequently their moral decision-making. Note that the two versions of our TRIAGE dilemma were only distinguished according to perceived realism and did not differ as to the plausibility of the stated consequences, which were considered moderately plausible in both versions. Like other sacrificial dilemmas employed in moral judgment studies, our scenarios displayed low information regarding sacrificial actions and their consequences. This lack of plausibility action information in both versions of the dilemma may have led all participants to infer that the sacrificial action (i.e., killing one patient) would not with certainty produce the stated positive outcome (i.e., saving five others). Consistent with this reasoning, Kusev et al. [45] demonstrate that, in a moral dilemma, displaying full information regarding moral actions and their consequences resulted in an increase of utilitarian responses.

In Experiment 2, we examined whether both perceived realism of the scenario and plausibility of utilitarian action influence moral responses in a sacrificial dilemma. We assume that plausibility, but not necessarily dimensions of perceived realism, influences moral responses. Based on the findings of Körner et al. [42] (see also [43, 44]), but also according to our data showing that the more plausible participants rated the stated consequences, the more utilitarian their moral judgment and choice of action were, we suppose that a scenario with a plausible action should lead to more utilitarian responses than the same scenario with an implausible action, regardless of the perceived realism of the dilemma.

Furthermore, by distinguishing moral judgment and choice of action, we found that the tendency to produce utilitarian responses was higher for the choice of action than for judgment (see [18] for similar results). In experiment 2, we investigated whether the perceived realism of the scenario and the plausibility of the utilitarian action differentially affect these two distinct processes underlying moral responses.

Experiment 2

334 undergraduate students from the University of Montpellier 3, France, took part in this second study during lockdown (April 03rd-May 03rd, 2021). In accordance with the exclusion criterion defined before data collection (see Experiment 1 for a similar criterion), 14 participants (those who reported having a close relative or friend that had died from the virus) were excluded from the analyses. The sample size was therefore n = 320 (271 female, M = 20.03, SD = 4.55).

Material

We used the same TRIAGE dilemma as in Study 1. This dilemma was modified to create four versions (Table 4). The type of dilemma (contextualized or not within the Covid-19 pandemic) and the plausibility of action (plausible or implausible) varied orthogonally. Specifically, we used both the contextualized and non-contextualized versions of the dilemma described in experiment 1. Each of these two scenarios were modified to create a plausible action version and an implausible action version. In order to manipulate the plausibility of action, we referred to the methodology used by Körner et al. [42] in which two aspects of plausibility were manipulated: the plausibility of the stated consequences and the plausibility of alternatives (two aspects influencing moral responses in an independent and additive manner). The two aspects were combined; that is, in the plausible action versions, the consequences of the utilitarian option were certain (i.e., this solution guarantees a 100% probability of saving the five) and the utilitarian option was the only way to save people (i.e., there were no other possible alternatives). In the implausible action versions of the dilemma, the consequences of the utilitarian option were not certain and better alternative actions could be imagined. In summary, we had four versions of the TRIAGE dilemma: two versions contextualized within the COVID-19 pandemic with either a plausible or implausible action; and two versions not contextualized within the COVID-19 pandemic with either a plausible or implausible action. Finally, both the plausible and implausible action versions of the dilemma contained exactly the same number of words.

Table 4

The versions of the TRIAGE dilemma added in experiment 2 (translated from French) in order to examine the plausibility of an action.

Versions contextualized within COVID-19

Versions not contextualized within COVID-19

You are the department head of a hospital in eastern France. A new coronavirus from China which causes respiratory irritation has appeared. Every day you receive more and more new patients with breathing problems. You don’t have enough oxygen for all of the patients. Five new patients are admitted to the hospital’s intensive care unit. Their health condition requires immediate hospitalization and the administration of oxygen for the next 15 days. There is no more oxygen available, and you have no way to get it. The only way to save the five patients is to take the oxygen tank from one of your patients who is in critical condition.This solution is possible because the sharing of oxygen would suffice to save them. Only this solution guarantees a 100% probability of saving them.If you do that, the patient in critical condition will die but the other five will be saved.There is no other possible alternative. Transferring patients to another hospital is not feasible because they would not survive the transportation time. The patient in critical condition cannot share the oxygen because it would cause his or her death, and requisitioning oxygen from another hospital would take too much time.Time is short and you know that this is the one and only solution to save the five patients.

You are the department head of a hospital in eastern France. A new bacterium has contaminated the water of the city. Every day you receive more and more new patients with intestinal disorders and blood poisoning. You do not have enough antibiotics for all of the patients. Five new patients are admitted to the hospital’s intensive care unit. Their health condition requires immediate hospitalization and a dose of antibiotics. There are no more antibiotics available, and you have no way to get some. The only way to save the five patients is to take antibiotics from one of your patients who is in critical condition.This solution is possible because the sharing of antibiotics would suffice to save them. Only this solution guarantees a 100% probability of saving them.If you do that, the patient in critical condition will die but the other five will be saved.There is no other possible alternative. Transferring patients to another hospital is not feasible because they would not survive the transportation time. The patient in critical condition cannot share the antibiotics because it would cause his or her death, and requisitioning antibiotics from another hospital would take too much time.Time is short and you know that this is the one and only solution to save the five patients.

The procedure and measures are similar to those used in experiment 1. Participants were tested using an online questionnaire created on the platform Qualtrics (https://www.qualtrics.com). They were informed that their responses would remain anonymous according to the Data Protection law.

After giving their informed consent, participants were randomly assigned to one of the four versions of the TRIAGE dilemma. First, they read the dilemma and successively provided two moral responses: a moral judgment and a choice of action. Second, we measured the perceived realism of the dilemma and the plausibility of the action for each version. In this second part, participants carefully reread the same version of the dilemma and evaluated it by answering the same questions as in experiment 1 (i.e., three questions related to perceived realism and one question related to the plausibility of the stated consequences). In addition, based on the findings of Körner et al. [42] and in order to check the effectiveness of our plausibility manipulation, we also assessed the plausibility of better alternative actions with the following question: “How plausible is it that there are no better alternative actions—no reasonable actions to [achieve outcome = save the five patients]?”. Then, participants had to rate it on the same 6-point scale as the other questions. Finally, participants provided demographic information and were asked “have you lost a close relative or friend to COVID-19?” (yes/no question). As in Experiment 1, we asked this question at the end of the experiment in order to avoid raising the saliency of the pandemic context, especially among participants assigned to the version of the dilemma which was not contextualized within the Covid-19 pandemic.

To check the effectiveness of our realism manipulation, we conducted separately 2 X 2 analyses of variance (ANOVAs) with the Type of dilemma (Contextualized vs. non contextualized within the Covid-19 pandemic) and the Plausibility of action (plausible or implausible) as between-subject factors, on the three measures of perceived realism (i.e., plausibility, factuality and typicality).

As in Experiment 1, we observed a significant main effect of the type of dilemma on each of the three dimensions of perceived realism indicating that the contextualized versions within Covid-19 were judged more realistic than the non-contextualized ones. Specifically, participants perceived the dilemma as more plausible, more factual and more typical (M = 5.15, SD = 1.38; M = 5.27, SD = 1.09; M = 5.65, SD = 0.70) when it was contextualized within the Covid-19 pandemic than when it was not contextualized (M = 4.50, SD = 1.45; M = 4.90, SD = 1.18; M = 5.04, SD = 1.11), ps < .01. It should be noted that the perceived realism of the scenario was not influenced by the plausibility of the action, regardless of the type of dilemma (Contextualized vs. non-contextualized within the Covid-19 pandemic). In short, there was no main effect of Plausibility of action, nor significant interaction with Type of dilemma on the three dimensions of perceived realism (i.e., all versions of the dilemma were judged equally plausible, factual, and typical). Once again, Pearson’s correlation analysis showed that these three dimensions of perceived realism were positively and significantly correlated, ps < .001.

Judgments of plausibility of action (i.e., plausibility of stated consequences and plausibility of better alternative actions) were also entered separately into 2 (Type of dilemma: Contextualized vs. non contextualized within the Covid-19 pandemic) X 2 (Plausibility of action: plausible vs. implausible) between-subjects ANOVAs. These analyses showed a significant main effect of the plausibility of an action, indicating that the stated consequences were rated as more certain in the plausible action versions (M = 4.87, SD = 1.51) than in the implausible action versions of the dilemma (M = 3.56, SD = 1.65), F(1,316) = 54.43, p < .001, η = 0.15. In addition, participants judged it more likely that there were no better alternative actions in the plausible action versions (M = 3.12, SD = 1.61) than in the implausible action versions of the dilemma (M = 2.20, SD = 1.43), F(1, 316) = 29.54, p < .001, η = 0.09. Regarding the manipulation check, no other effects were significant.

To explore the effect of both perceived realism and plausibility of action on individuals’ moral responses, a repeated-measure ANOVA was conducted with the Type of dilemma (Contextualized vs. non contextualized within the Covid-19 pandemic) and the Plausibility of action (plausible or implausible) as between-subject factors, and the Type of moral response (Judgment vs. Choice of action) as within-subject factor. The Type of moral response main effect was significant, F(1,316) = 32.33, p < .001, η = 0.09. As in Experiment 1, responses for choice of action (M = 3.97, SD = 1.46) were overall more utilitarian than responses to judgment (M = 3.47, SD = 1.45). The Plausibility of action main effect was also significant, F(1,316) = 7.69, p < .01, η = 0.02, indicating that moral responses were more utilitarian in the plausible action versions (M = 3.91, SD = 1.20) than in the implausible action versions of the dilemma (M = 3.53, SD = 1.22). A significant Plausibility of action x Type of moral response interaction showed that participants tended to be more utilitarian for the plausible action versions compared to the implausible action versions of the dilemma in their choice of action only (Fig 1), F(1,316) = 6.13, p = .01, η = 0.02. Post hoc analysis (Scheffé test) revealed that participants were more inclined to choose the utilitarian action in plausible action versions than in implausible action versions of the dilemma, p = .004. Their moral judgment did not differ according to the plausibility of action, p = .80. In addition, in plausible action versions of the dilemma, they appeared to be more utilitarian in their choice of action than in their judgment, p < .001. No other effects were significant.

Fig 1

Mean moral responses as a function of type of moral response (judgment vs. choice of action) and plausibility of sacrificial action (plausible vs. implausible).

Higher scores (max = 6) are closer to utilitarian responses.

To further test the impact of perceived realism and plausibility of action on moral responses, we conducted two multiple linear regression analyses. For each regression calculation, we entered the three measures of perceived realism (plausibility, factuality and typicity) and the two measures of plausibility of action (Plausibility of stated consequences and Plausibility of better alternative actions) as predictors. As in Experiment 1, results showed that none of the perceived realism measures was a significant predictor of moral responses: perceived plausibility, factuality and typicity of the dilemma did not predict judgment or choice of action (all ps > .10, see Table 5 for a summary of regression results). Interestingly, both perceived components of the plausibility of an action (i.e., plausibility of stated consequences and plausibility of better alternative actions) predicted individuals’ choice of action. Indeed, a significant regression equation was found, F(5,314) = 6.07, p < .001), with an R2 of .09: The more participants judged the consequences of the utilitarian action as certain and rejected the possibility of better alternatives, the more they endorsed the utilitarian action. Note that the plausibility of the stated consequences predicted choice of action better than the plausibility of alternatives. Neither of these two measures was a significant predictor of moral judgment.

Table 5

Results of linear regression analyses predicting moral choice ratings from ratings of perceived realism and plausibility of action.

Predictors	Beta	SE (B)	t	p	Semi-partial correlation
Predicting Judgment
Plausibility	0.124	0.069	1.782	0.076	0.100
Typicity	0.060	0.110	0.544	0.587	0.031
Factuality	-0.086	0.086	-0.994	0.321	-0.056
Plausibility of stated consequences	0.066	0.052	1.249	0.213	0.070
Plausibility of better alternative actions	0.051	0.056	0.907	0.365	0.051
Predicting Choice of action
Plausibility	-0.038	0.068	-0.555	0.579	-0.031
Typicity	0.125	0.108	1.154	0.249	0.065
Factuality	-0.034	0.084	-0.397	0.692	-0.022
Plausibility of stated consequences	0.190	0.051	3.697	< .001	0.204
Plausibility of better alternative actions	0.111	0.055	2.032	0.043	0.114

In this second experiment, we examined whether both the perceived realism of the scenario and the plausibility of the utilitarian action influence moral responses in a sacrificial dilemma. In investigating perceived realism, we again distinguished different dimensions of perceived dilemma realism (i.e., plausibility, factuality and typicity). As in Experiment 1, we found that none of these dimensions of perceived realism was a significant predictor for moral judgment or choice of action. As expected, plausibility of action influenced moral responses: participants were more inclined to choose the utilitarian action in plausible action versions than in implausible action versions of the TRIAGE dilemma. It is noteworthy that this effect was not observed for moral judgments but only for choice of action. This result confirms and extends previous findings showing that plausibility of action, especially plausibility of stated consequences and plausibility of better alternative actions, influences moral responses in sacrificial dilemmas. Initial evidence for this effect came from Körner et al. [42], (see also [43, 44]), who showed that participants endorsed the utilitarian killing response if the stated consequences seemed certain and if it seemed plausible that there were no better alternatives to save the lives of five people. Furthermore, the fact that plausibility of action did not have a significant influence on participants’ moral judgment sheds light on the necessity of distinguishing judgment and choice of action in moral responses.

According to Tassy et al. [18] (see also [46]), moral judgment and moral choice may be underlied by distinct psychological mechanisms. In our studies, participants were asked to make a decision related to the allocation of healthcare resources. In the context of a severe pandemic, such as Covid-19, we know that maximizing survival rates is the most important aim (e.g., [5, 30, 47]. Therefore, it can be assumed that, faced with a TRIAL dilemma, participants were able to distinguish between what is morally acceptable (i.e., judgment) and what decisions must be made to manage the crisis (i.e., choice of action). Thus, they might judge the sacrificial action mainly according to moral rules (“it’s forbidden to kill”, “do no harm”). On the other hand, they might choose to act or not to act mainly according to the outcomes they could reasonably expect (i.e., “saving a maximum number of lives”). In other words, unlike moral judgment, the choice of action could result mainly from controlled cognitive processes, as identified in the dual-process theory [48-50]. In line with this theory, primitive emotional responses that prime us not to endorse harmful actions in the name of moral rules can be overridden by an effortful deliberate processing of outcomes that drive rational choices, seeking to maximize the wellbeing of a larger number of people. If people consider plausible the suggestion that killing one person will guarantee positive consequences (i.e., saving five others), they should rationally choose the sacrificial action. In line with this explanation, our results reveal that if taking [the oxygen / antibiotics] of one patient saves with certainty the five and if there is no other way (via alternative actions) to save them, people tend to be more utilitarian in their choice of action than when the consequences are uncertain and alternative actions exist. When faced with a scenario with a plausible action, they also tend to be more utilitarian in their choice of action than in their judgment. However, the plausibility of an action does not predict their moral judgment.

To explain moral responses, the rational perspective might not be sufficient to account for quick decision-making in real-world situations, such as the triage dilemma. Indeed, it could be the case that, in these mundane as well as dramatic realistic scenarios in which urgent decisions must be made, the choice of an action is driven by simple and fast intuitive processes. According to one intuitionist perspective, the Agent Deed Consequence (ADC) model [51, 52], people use and evaluate three kinds of intuitions while making moral decisions: the person who is doing something, specifically with respect to his traits or intention (the Agent-component, A), the deed or what is being done (the Deed-component, D), and the consequences or outcomes that resulted from the deed (the Consequences-component, C). The positive or negative intuitive evaluations of each of these components can be used simultaneously in a comparative framework in order to produce a positive or negative judgment of moral acceptability. Therefore, a wrong deed (e.g., killing or harming a person) may be more acceptable if the agent has good intentions and the action has a good consequence (e.g., saving five people). According to Dubljević [53] (see also [54]), this idea that different components of moral intuitions (A, D, and C) simultaneously drive moral responses is helpful to explain the flexible but stable nature of moral judgment. In our Experiment 2, the negative evaluation of the sacrificial action (taking [the oxygen / antibiotics] of one patient) could be counterbalanced by the positive evaluation of the consequences (being certain to save the five). Consistent with the predictions of the ADC model, our results suggest that the consequences or outcomes that resulted from the deed (the D component) may be less relevant in moral judgment than in moral choice.

Such a promising explanation should be considered with caution, and further research is needed to investigate the processes and mechanisms underlying judgment and choice of action in sacrificial moral dilemmas. These processes and mechanisms might differ depending on whether individuals are faced with real-world dilemmas or hypothetical dilemmas.

Limitations

Several limitations, inherent to the nature of the sacrificial dilemma used in our studies, should be noted.

First, although the TRIAGE dilemma in health care is not a new topic (similar dilemmas have arisen in organ transplantations, for example), the version based on oxygen allocation was considered more realistic than the one based on antibiotic allocation because France and other counties were actually faced with a critical shortage of the lifesaving machines during the Covid-19 pandemic. Outside of this pandemic context caused by a respiratory virus, our dilemma would have no ecological validity.

Second, even if our two versions of the dilemma (i.e., the contextualized one and the non-contextualized one) were strictly comparable and differed only in perceived realism, the lack of difference in moral responses between the two versions should be interpreted with caution. It does not necessarily mean that perceived realism does not influence the moral response. Indeed, in both versions, the moral response was a medical decision that may have interfered with the effect of perceived realism in a highly salient pandemic context at the time of the experiment. In this context, the moral rule (“it’s forbidden to kill”) can be outweighed by moral considerations such as the harm that will be prevented if these rules are set aside. Kahane [55], referring directly to medical triage, argues that “overruling a moral rule in emergency context when lives are at stake is part of commonsense morality” (p.556). In this sense, the duty of rescue may have functioned as a prescriptive (or injunctive) norm guiding participants’ judgments and choice of action. Further research should investigate the role of perceived realism on moral responses in other dramatic situations involving real life and death decisions.

Third, because triage issues in a pandemic context reflect a given society’s moral standards, the moral responses collected in the present studies are not necessarily the same in every country around the world.

Conclusion

Despite their limitations, the present studies are the first to investigate the role of perceived realism on moral response by measuring the realism of proposed scenarios. Contrary to previous studies, we compared different versions of the same sacrificial scenario (a triage dilemma contextualized or not within the Covid-19 pandemic) by controlling for the perceived realism of each version. Although the contextualized version was perceived as more realistic than the non-contextualized version, the moral responses were the same. Thus, our results do not allow us to conclude that there is an effect of perceived realism on moral responses but highlight one key factor: the plausibility of the sacrificial action. In our studies, regardless of the degree of realism, participants chose the utilitarian action if the stated consequences seemed certain and if it seemed plausible that there were no better alternatives to save the lives of five people. This result supports the findings of several authors (e.g., [42-45]) who had already highlighted the influence of the plausibility of sacrificial action on moral responses.

It is also worthwhile to highlight that distinguishing between the perceived realism of the scenario and the plausibility of the action leads to a reconsideration of criticisms made concerning hypothetical dilemmas: instead of focusing on the lack of realism (a factor that does not seem to influence moral responses), the certainty of action (a determinant of moral choices) deserves to be more thoroughly considered in moral dilemma studies. This is closely linked to an idea mentioned by Dubljević and Racine [51]: “In the case of hypothetical dilemmas, uncertainty related to the response options is silenced” (p.12). In that case, participants may therefore infer probabilities for themselves concerning certainty of outcomes. To solve this problem (i.e., subjective ratings of certainty), Shou et al. [43] suggested controlling or measuring subjective judgments regarding dilemma outcomes.

In addition, Bauman et al. [1] suggested to study moral responses in a more ecological way by creating “new scenarios involving the same types of trade-offs as sacrificial dilemmas, but presenting them in a way that is more consistent with how people might face these trade-offs in the real world” (p.546). As an example, Bauman et al. [1] were already referring to medical professionals who must make decisions about the allocation of scarce medical resources. In line with this recommendation, we suggest that future research should use both mundane and dramatic realistic dilemmas displaying full information regarding the three components needed to make moral decisions: the agent, the deed in itself and the consequences of the deed (Dubljević et al., [54]).

(XLSX)

Click here for additional data file.

18 Mar 2022

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Sara Rubinelli

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. In your Data Availability statement, you have not specified where the minimal data set underlying the results described in your manuscript can be found. PLOS defines a study's minimal data set as the underlying data used to reach the conclusions drawn in the manuscript and any additional data required to replicate the reported study findings in their entirety. All PLOS journals require that the minimal data set be made fully available. For more information about our data policy, please see http://journals.plos.org/plosone/s/data-availability.

Upon re-submitting your revised manuscript, please upload your study’s minimal underlying data set as either Supporting Information files or to a stable, public repository and include the relevant URLs, DOIs, or accession numbers within your revised cover letter. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories. Any potentially identifying patient information must be fully anonymized.

Important: If there are ethical or legal restrictions to sharing your data publicly, please explain these restrictions in detail. Please see our guidelines for more information on what we consider unacceptable restrictions to publicly sharing data: http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions. Note that it is not acceptable for the authors to be the sole named individuals responsible for ensuring data access.

We will update your Data Availability statement to reflect the information you provide in your cover letter.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: I Don't Know

Reviewer #2: I Don't Know

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: No

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: The experiments conducted are very interesting but I have some doubts about their methodology and purported realism, which is very important given that that’s the topic under investigation.

Lengthy introduction gives useful introduction to the literature in this area. However, there are a few shortcomings, mainly from the ethical perspective:

The language used here presents utilitarianism as unethical/an opponent of ethics. “From an ethical point of view” should say “from a deontological” point of view, or some other moral system. “From a utilitarian point of view, approving the sacrifice of one life in order to save the lives of five others is morally acceptable with respect to the number of lives saved. By contrast, from an ethical point of view, it is not acceptable to sacrifice the life of one person in order to save others because the value of each human life is paramount and no one has the right to sacrifice a life.”

It's not the “footbridge dilemma”, it’s the most famous variation of the trolley problem. And “Someone is standing on the tracks of a bridge” doesn’t make sense.

Why use the fat man example rather than the classic trolley problem? Using the fat man example instead makes it sound like you intended people to die when you deny them an intensive care bed, whereas that’s not the case.

More argumentation is needed to back up the claim that “emotion makes utilitarian decisions harder to make”. For many people the utilitarian choice is justified by an appeal to emotion.

Also, people might intuitively believe that they should act deontologically, but later on reflection regret not making the utilitarian decision.

It’s true that the footbridge dilemma is unrealistic. However the main trolley problem is much more realistic and that’s another reason it should be used in this paper instead.

Again, the “switch dilemma” isn’t “a version of the trolley problem”, it IS the trolley problem.

One drawback of the VR approach is that it risks making the experiment seem more like a game than an abstrac scenario.

I’m not the greatest fan of trolley problems, but it’s overstating it to say ““trolley problems are unrealistic and unlike anything people encounter in the real world”. Anyone who has swerved the car to avoid a squirrel or fox, putting their family at risk, has experienced a real life trolley problem. And as you argue in the paper, doctors on the ICU face similar issues.

Doctors have always been making ‘moral decisions about which patients to admit to the Intensive Care Unit (ICU) and for which ones to provide lifesaving resources” regardless of the pandemic.

ICU sacrifical dilemmas might seem more realistic than trolley problems, but they’re often greatly oversimplified in a way that makes them clinically unrealistic, even if they seem real to members of the public.

The footbridge dilemmas is also implausibly because it seems unlikely the decision-maker could life/shove such a heavy person, and also because there would be no way of knowing whether he would stop the train or not. In contrast, the main trolley problem is much more plausible as it lacks these “stupid” features.

Why were those who had lost a relative to Covid excluded after participation rather than before, if the criterion was decided before collection? That means that their time was wasted. An ethics committee would have picked up on this if the study had been reviewed.

The triage dilemma used in study 1 is unrealistic, at least compared to the ICU bed example. Hardly any hospitals ran out of oxygen and none that I’m aware of ran out of antibiotics. Given the emphasis on plausibility this is unfortunate. The oxygen example also differs from antibiotics (and the ICU bed) because it’s normaly certain that a patient will die without oxygen but that’s not the case for the other two. In addition, the antibiotics example seems implausible because five patients would need five times as much antibiotics as one patient, and antibiotics courses need to be taken for days. To me the oxygen scenario is less plausible than the ICU bed one, but the antibiotic one is deeply implausible which is a major issue given that it was used only in the example that’s supposed to be less plausible for the reason that it’s decontextualised. This highlights that it seems to be a methodological flaw to have oxygen in the contextualised example, but antibiotics in the noncontextualised one? That introduces an unnecessary variable that could act as a confounder. Also, looking at the examples it’s clear that they’re both contextualised; it’s just that one is a fictional and one a real-world context.

“Participants were asked to be as honest as possible in their 335 responses, knowing that there was no right or wrong answer.” Were participants told that there is no right or wrong answer in ethics? That’s a very contentious morally relativistic statement that many ethicists would disagree with.

The results section is very short and could use more textual description of the results.

Also, it would probably make sense to describe the Methods of both studies in one section, the Results in another, and then have a combined discussion.

It seems rather problematic to characterise a scenario where the outcome is uncertain as less plausible; in real life the outcome is often uncertain, ie, the more uncertain the more realistic it is. Yet the authors say “In the implausible versions of the dilemma, the consequences of the utilitarian option were not certain”; to me, that sounds very plausible.

Why does the “plausible” version of the Covid-19 scenario in study 2 say “a virus

causing respiratory irritation” rather than stating that it’s Covid-19?

The only difference between the plausible and implausible versions is the added explanation at the end. To many people that part would be obvious, so it’s not clear how it can be used to differentiate based on plausibility.

Also, it’s extremely implausible to say that “this solution guarantees a 100% probability of saving them” – there’s never a 100% probability of saving any patient, much less one who needs oxygen on the intensive care unit, where around half of patients die.

The same problem applies to the plausible version of the antibiotics example – it remains extremely implausible and the explanation that’s designed to increase plausibilty actually descreases it by making implausible statements. However, the antibiotic example because, while denying oxygen will almost certainly kill a patient, providing antibiotics only increases chances of survival.

Telling participants that one option guarantees 100% that five lives will be saved clearly reduces real-world plausibility, but increases certainty and makes the choice easier.

Given all these issues, I’m not sure how valid any of the results are.

“On the other hand, they might choose to act or not to act mainly according to the outcomes they could reasonably expect (i.e., “saving a maximum 618 number of lives”), rather than following their moral considerations”

Once again, this implies that utilitarian principles don’t count as moral considerations. A utilitarian would regarded maximising lives saved as the moral and ethical thing to do. The fact that a deontologist would disagree doesn’t mean that the utilitarian is wrong or unethical.

“Our results reveal that if taking [the oxygen / antibiotics] of one patient saves with certainty the five and if there is no other way (via alternative actions) to save them, people tend to be more utilitarian in their choice of action than when the consequences are uncertain and alternative actions exist.”

Of course they do – but certainty is deeply implausible, which is problematic for this study,.

Reviewer #2: These authors have conducted two experiments to explore factors that may affect hypothetical moral responses and actions. With regard to their finding that the respondents view the COVID-19 scenario as more realistic than the antibiotic scenario, it seems quite clear that the difference is attributable to the current pandemic. From my perspective as a clinician ethicist, those two scenarios are ethically indistinguishable. It is interesting that people tended to say that they would be more likely to act than to say that action was morally appropriate. The impact of certainty of the ability of action to save more people makes sense and adds support to existing literature. I found the authors' decision to divide the methods, results, discussion into two parts to be quite odd. They should unite them into one.

I am confused about one thing -- you said that consent is not required but you obtained it from the students. that seems desirable but inconsistent.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

4 May 2022

Please see the file entitled "response reviewers" we upload with the revised version of the manuscript. Regarding the data set linked to the results presented in the manuscript, of course we agree to share these data (Please see the uploaded file).

Submitted filename: RESPONSE.Reviewers.docx

Click here for additional data file.

31 May 2022

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Sara Rubinelli

Academic Editor

PLOS ONE

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

Reviewer #2: (No Response)

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: (No Response)

Reviewer #2: Partly

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: (No Response)

Reviewer #2: I Don't Know

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: (No Response)

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: (No Response)

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: (No Response)

Reviewer #2: Ultimately, I am not sure whether this adds significantly to the literature. I understand the difference between judgment and choice of action, but the problem is that the choice of action in this setting is only hypothetical. What one should care about is what people actually do when confronted with the dilemma. It is helpful that the authors assessed the perceived realism of their scenarios, but distressing that the scenarios made so little difference. Rereading this, I have further methodologic concerns. They need to say, first, what their recruitment and exclusion criteria were (and justify the latter, which was not done; it may not have been justifiable to exclude them only after they had already been asked to revisit their loss) and then they need to say what their analytical strategy was -- as it is, it reads almost like data mining (I hate to say it). And what I wanted you to do was present Experiment1 and then Experiment 2 and then discuss them together -- as it was, there was a good bit of redundancy. Moreover, unlike Reviewer 1, I thought that the introductory literature review was way too long.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

6 Jul 2022

Reviewer’s comments

I have re-checked this paper and I would advice to shorten the introduction by clearly introducing the topic, why it is relevant, what is the state of the art in the field and why this manuscript addresses that topic and with which purpose/objective.

Currently, the introduction is 13 pages and the reader gets lost about the focus, its relevance and ground in the literature.

Thank you for your comments that helped us to revise the manuscript. We agree that the introduction section was too long. As requested, we shortened and reorganized the introduction so that the way the manuscript now addresses the topic is made more salient. Three pages were eliminated from the introduction section.

Following your recommendations, we began by directly introducing the topic with the major criticism of dilemmas classically used to study moral judgment: their lack of realism (Bauman et al., 2014). This criticism calls into question the ecological validity of measured moral responses. In line with previous research that has attempted to address this criticism, the goal of the two studies presented in this article is to 1) propose more realistic sacrificial dilemmas (the triage dilemma by using the crisis context of Covid-19), but also to measure the realism of the scenarios used. Indeed, even if the literature proposes several solutions to improve this realism (solutions that we present and criticize in our state of the art), to our knowledge no study measures the perceived realism of the dilemmas used. It is thus impossible to determine the role of this factor on moral responses.

The contribution of our work is therefore to study moral responses in sacrificial dilemmas which are designed to be more realistic than those used so far, and also to control the perceived realism of the dilemmas studied in order to be able to draw scientific conclusions regarding the role of this factor on moral responses.

To this end, we propose a sequence of two experimental studies in which perceived realism is assessed in its different aspects: the realism of the scenario (i.e., its plausibility, factuality, and typicity) in Study 1, and the realism of the scenario associated with the plausibility of the sacrificial action in Study 2.

Finally, these studies contribute to shedding light on the question of the perceived realism of the dilemmas on two quite different aspects of moral response (Tassy et al., 2013): the judgment we make about a moral response (i.e., "is it morally acceptable?") and the intended action (i.e., the choice we would make if faced with a sacrificial dilemma). To the extent that, for obvious reasons, our participants could not actually act (i.e., could not sacrifice one person or let 5 persons die), measuring their intended action when faced with a sacrificial dilemma (contextualized within the Covid-19 pandemic or not) sheds light on the influence of perceived realism on choice of action.

Reviewer’s comments

For the other comments, I would better explain the methodology of the study and everything that can help understanding the results.

We agree that the elements provided in the manuscript relating to the exclusion criterion were insufficient. To remove any ambiguity on this subject, we added information in the method section:

Before conducting the study, it was clear to us that participants with a relative or close friend who had died of Covid-19 should be excluded. This exclusion criterion is based on Tassy et al. (2013)’s results: they showed that emotional or genetic proximity between participants and dilemma protagonists influences moral responses. Indeed, responses (and especially the choice of action) are less utilitarian when the potential victim (i.e., the one to be sacrificed) is a relative (e.g., a brother, a cousin, a friend) than when he/she is a stranger. Given the number of victims of the Covid-19 pandemic, the likelihood that participants had a relative or close friend who died of Covid-19 had to be considered. Of the 489 participants, 29 were excluded on the basis of this criterion. This allowed us to avoid any bias coming from closeness of relationship of the decision-maker with the potential victim, a factor determining moral responses.

We decided not to ask the question “have you lost a close relative or friend to COVID-19” at the beginning of the experiment for the following reason: Asking the question at the beginning would have raised the saliency of the Covid-19 pandemic context. It was indeed risky to ask this question at the beginning of the experiment for an obvious reason: in the condition with the non-contextualized version of the TRIAGE dilemma, participants were not expected to make the connection with the Covid-19 pandemic. In other words, raising the saliency of the covid-19 pandemic context could have interfered with the context variable we manipulated in this study (dilemmas contextualized and non-contextualized within Covid-19) and introduced a bias.

Of the 489 participants, 29 participants (15 in Study 1 and 14 in Study 2) were excluded on the basis of this criterion. All participants were volunteers and those who were finally excluded from the sample were not informed that their contribution to this study was not taken into account due to this exclusion criterion.

Reviewer’s comments

As for the English, once the paper has been revised I suggest rechecking the wording.

A native English speaker rechecked the wording.

Submitted filename: ResponseReviewersJuly2022.docx

Click here for additional data file.

10 Aug 2022

Contextualizing sacrificial dilemmas within COVID-19 for the study of moral judgment

PONE-D-22-03571R2

Dear Dr. Blanc,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Sara Rubinelli

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #2: (No Response)

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #2: I Don't Know

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #2: (No Response)

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #2: No

**********

12 Aug 2022

PONE-D-22-03571R2

Contextualizing sacrificial dilemmas within Covid-19 for the study of moral judgment

Dear Dr. Blanc:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Sara Rubinelli

Academic Editor

PLOS ONE