PreprintPDF Available

Trolleys, Triage and Covid-19: The Role of Psychological Realism in Sacrificial Dilemmas

Authors:
Preprints and early-stage research may not have been peer reviewed yet.

Abstract and Figures

At the height of the Covid-19 pandemic, frontline medical professionals at intensive care units around the world faced gruesome decisions about how to ration life-saving medical resources. These events provided a unique lens through which to understand how the public reasons about real-world dilemmas involving trade-offs between human lives. In three studies (total N = 2298), we examined people’s moral attitudes toward triage of acute coronavirus patients, and found elevated support for utilitarian triage policies. These utilitarian tendencies did not stem from period change in moral attitudes relative to pre-pandemic levels--but rather, from the heightened realism of triage dilemmas. Participants favored utilitarian resolutions of critical care dilemmas when compared to structurally analogous, non-medical dilemmas—and such support was rooted in prosocial dispositions, including empathy and impartial beneficence. Finally, despite abundant evidence of political polarization surrounding Covid-19, moral views about critical care triage differed modestly, if at all, between liberals and conservatives. Taken together, our findings highlight people’s robust support for utilitarian measures in the face of a global public health threat, and illustrate how hypothetical scenarios in moral psychology (e.g. trolley cases) should strive for more experiential and psychological realism, otherwise their results might not generalize to real-world moral dilemmas.
Content may be subject to copyright.
This article has been accepted for publication in Cognition & Emotion, published
by Taylor & Francis (https://doi.org/10.1080/02699931.2021.1964940).
Trolleys, Triage and Covid-19: The Role of Psychological Realism in Sacrificial Dilemmas
Markus Kneer1 & Ivar R. Hannikainen2*
1 Department of Philosophy, University of Zurich, Switzerland
2 Department of Philosophy I, University of Granada, Spain
Author Note
The research was supported by a Swiss National Science Foundation grant
(PZ00P1_179912, PI Markus Kneer). Data and materials are available on the Open Science
Framework at: osf.io/dpsq9/?view_only=54a7c150e03d4d78819b2954cee3a240.
* Correspondence concerning this article should be addressed to Ivar Rodríguez Hannikainen,
Department of Philosophy I, Faculty of Psychology, Cartuja Campus, Universidad de Granada
(Spain), 18011. Contact: ivar@ugr.es
Word Count: 8737
TRIAGE DILEMMAS 2
2
Abstract
At the height of the Covid-19 pandemic, frontline medical professionals at intensive care units
around the world faced gruesome decisions about how to ration life-saving medical resources.
These events provided a unique lens through which to understand how the public reasons about
real-world dilemmas involving trade-offs between human lives. In three studies (total N = 2298),
we examined people’s moral attitudes toward triage of acute coronavirus patients, and found
elevated support for utilitarian triage policies. These utilitarian tendencies did not stem from
period change in moral attitudes relative to pre-pandemic levels--but rather, from the heightened
realism of triage dilemmas. Participants favored utilitarian resolutions of critical care dilemmas
when compared to structurally analogous, non-medical dilemmasand such support was rooted
in prosocial dispositions, including empathy and impartial beneficence. Finally, despite abundant
evidence of political polarization surrounding Covid-19, moral views about critical care triage
differed modestly, if at all, between liberals and conservatives. Taken together, our findings
highlight people’s robust support for utilitarian measures in the face of a global public health
threat, and illustrate how hypothetical scenarios in moral psychology (e.g. trolley cases) should
strive for more experiential and psychological realism, otherwise their results might not
generalize to real-world moral dilemmas.
Keywords: moral judgment; values; Covid-19; utilitarianism; politics.
TRIAGE DILEMMAS 3
3
Trolleys, Triage and Covid-19: The Role of Psychological Realism in Sacrificial Dilemmas
Early on in the Covid-19 pandemic, case rates and deaths saw exponential growth. Their
trajectory surpassed numerous governments’ predictions, and quickly depleted medical resources
at many intensive care units (ICUs) around the world. As a result, frontline medical professionals
faced what some imagined would remain an outlandish philosophical thought experiment: the
choice whether to sacrifice some patients in order to save a larger number of lives (Emanuel et
al., 2020). The Covid-19 pandemic brought the infamous trolley problem (Foot, 1967) to life,
and provided a unique opportunity for moral psychologists to explore how the public approaches
such tragic dilemmas when contextualized in ecologically valid settings involving triage at the
ICU. What should frontline doctors do when faced with the decision whether to sacrifice one
patient to prioritize saving more lives? The purpose of our present work is to explore this exact
question (see also Arora, Savulescu, Maslen, Selgelid & Wilkinson. 2016; Wilkinson, Zohny,
Kappes, Sinnott-Armstrong, & Savulescu, 2020).
Below, we first distill past findings drawn from the literature on hypothetical sacrificial
dilemmas, and echo previous concerns about limits to their ecological validitywhich our
studies then partly substantiate. Additionally, we review distinct models according to which
various contextual elements (e.g., personal susceptibility, period change) could modulate the
public’s moral attitudes. Turning to our studies, we document pronounced utilitarian tendencies
in triage contexts (Study 1), and test two distinct explanations for this effect in Studies 2 and 3:
period change and psychological realism, respectively. In closing, we elaborate on the practical
implications of these results for public health guidelines governing the allocation of critical care
resources, and highlight their theoretical implications for traditional approaches to the study of
moral cognition.
TRIAGE DILEMMAS 4
4
Sacrificial Dilemmas
Research in moral psychology relies heavily on “sacrificial dilemmas” (Bartels &
Pizarro, 2011) in which impending harm to a larger group of people can be avoided by harming a
smaller group instead. Among the most famous of these hypothetical scenarios are two versions
of the trolley dilemma, in which a runaway trolley will kill five workmen unless something is
done (Foot, 1967; Thomson, 1976, 1985). In the Bystander version, the trolley can be diverted
onto another track by operating a lever, where it would run over a single worker instead. In the
Footbridge version, the trolley can be stopped before it reaches the workmen by pushing a heavy
man off a footbridge and into the trolley’s path.
Contrary to utilitarian moral principles, people report conflicting intuitions about these
cases: choosing to save the larger number of lives in Bystander but not in Footbridge. Hundreds
of empirical studies have now employed sacrificial dilemmas (for a review, see Christensen &
Gomila, 2012) to characterize the psychological basis of this distinction (e.g., Bartels & Pizarro,
2011; Conway & Gawronski, 2013; Cushman et al., 2006; Greene et al., 2001, 2009; Patil, 2015)
and examine its generalizability across cultures (Awad et al., 2020) and birth cohorts
(Hannikainen et al., 2018). This large empirical literature has also inspired sophisticated
theoretical developments, from Greene and colleagues’ (2001, 2007) dual process theory, and
Mikhail’s (2007, 2009) universal moral grammar to the two-dimensional model of utilitarianism
advanced by Kahane, Everett and colleagues (Kahane et al., 2015 2018).
Among the most recurring findings in this literature is the association between affective
empathy and opposition to utilitarian sacrifice (Gleichgerrcht & Young, 2013). For instance,
deficits in empathy have been associated with stronger utilitarian tendencies (Patil & Silani,
2014), which may help to explain why psychopathic individuals report more favorable attitudes
TRIAGE DILEMMAS 5
5
toward utilitarian sacrifice than do psychotypical controls (Koenigs et al., 2012). This line of
research has inspired skepticism about the relationship between sacrificial dilemmas and
utilitarian thought in moral philosophy (but see Conway et al., 2018). Namely, utilitarian
reactions to hypothetical dilemmas may not reflect a concern for the greater good (Kahane et al.,
2015) but rather a reduced aversion to causing harm (Miller et al., 2014) that accompanies Dark
Triad traits (i.e., Machiavellianism, narcissism, and psychopathy; see Bartels & Pizarro, 2011;
Patil, 2015).
Experimental studies employing immersive methods to elevate affective engagement
have provided further nuance. If affective empathy primarily constrains utilitarian sacrifice, then
inducing greater engagement ought to elicit stronger affective reactions and further dampen
utilitarian tendencies. Yet, efforts to make sacrificial dilemmas more vivid and life-like--e.g.,
through the use of virtual reality (Patil et al., 2014; Francis et al., 2016), haptic feedback (Francis
et al., 2017), or the anticipation of actual harm to mice (Bostyn et al., 2018)--have all revealed
evidence to the contrary. Immersive research methods in fact strengthen people’s utilitarian
motivation, when compared to their responses to abstract narrations of those same dilemmas.
These results cannot readily be explained when associating affective empathy in the
opposition to harmful acts. According to the two-dimensional model of utilitarianism, utilitarian
tendencies manifest in the general population as two distinctand even unrelatedattitudes:
approval of instrumental harm and impartial beneficence. Reactions to sacrificial dilemmas are
linked primarily to beliefs about the permissibility of instrumental harm (Kahane et al., 2015
2018), while being largely unrelated to impartial beneficence. Could enhancing the ecological
validity of moral dilemmas elevate utilitarian tendencies by promoting specifically beneficent
motivations to act? Our studies pursue this hypothesis, which (if supported) could help reconcile
TRIAGE DILEMMAS 6
6
the well-known association between affective deficits and utilitarian sacrifice (in traditional, text-
based dilemmas) with evidence that more immersive experimental paradigms have the opposite
effect: i.e., of strengthening utilitarian motivation.
External Validity
In an influential paper, Bauman and colleagues (2014) raised concerns regarding the
external validity of research on sacrificial dilemmasthat is, the extent to which the findings
generalize to other, real-world situations (Campbell, 1957). Drawn from analytic philosophy, the
dilemmas tend to be “set in fanciful, sometimes absurd, contexts, and these artificial settings may
affect the way people approach the situation and decide what to do” (Bauman et al., 2014, p.
536). Among philosophers, such thought experiments are fruitfully employed to elicit intuitions
within the philosophical community. However, since people are not acquainted with the
discipline’s conventions and methods (Converse, 1964; Tenbrunsel & Messick, 1999), they may
approach these dilemmas with a very different motivation.
The external validity of sacrificial dilemma stimuli may be compromised by three
different factors (Aronson et al., 1998): First, they could fail to meaningfully engage the
participant, for instance, if they are met with amusement (experimental realism). Second, they
may be too dissimilar to the situations participants routinely encounter in their lives (mundane
realism). Consequently, participants might question central premises (e.g., could the heavy man
actually stop the trolley?), and be reluctant to accept the scenario (as seen, e.g., in Greene et al.,
2009). Third, the scenarios might also lack psychological realismpartly as a consequence of
low experimental and mundane realism: They may not elicit the same cognitive processes that
take place in real-world circumstances. Unrealistic scenarios might, for instance, fail to produce
characteristic reactions such as outrage toward transgressors or avoidance of those who manifest
TRIAGE DILEMMAS 7
7
different moral values (for evidence that trolley dilemmas are wanting in this regard, see
Baumann et al., 2012, Study 3).
Sacrificial dilemmas related to the Covid-19 pandemichenceforth triage or critical
care dilemmasmay overcome some of these limitations (Buckwalter & Peterson, 2020;
Wilkinson et al. 2020; Romero-Rivas et al., 2021; Reeskens et al. 2021; Francis & McNabb,
2021; Antoniou et al., 2021). First, due to a shortage of supplies and staff, medical personnel at
numerous intensive care units around the worldparticularly during the first wavefaced
actual sacrificial dilemmas. Second, the death toll due to coronavirus rose rapidly, and received
continuous press coverage. Third, in addition to the health burden of the virus itself, the
pandemic has also brought about changes in lifestyle, mental health, occupational status, and
family life that have dramatically affected everyone. It stands to reason, then, that triage
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. And indeed, transgressors of health and security
measures tend to witness the full spectrum of reactions (e.g., outrage, contempt, disgust, see
Rozin et al., 1999) known to accompany moral misbehavior. Disagreements over the correct
political response to the pandemic as well as individual distancing measures have engendered
heated debate, manifestations drawing large crowds, and destroyed many a relationship, as
research on prejudice toward morally dissimilar others would predict (Skitka et al., 2005; Cole
Wright et al., 2008).
TRIAGE DILEMMAS 8
8
Prevalence, Proximity and Period
Susceptibility to severe coronavirus infection varies across time and locations (i.e., with
fluctuation in contagion and case rates) and also across individuals (i.e., depending on health
factors). This variation provided two distinct tests of the psychological realism hypothesis.
When we ran the first wave of studies, in mid-April 2020, the bulk of infections and
deaths were limited to a handful of US states, including New York, Pennsylvania and Michigan.
Most other states had extremely low infection rates, many vacant hospital beds, and barely any
coronavirus-related deaths. Thus, in New York City, which was seeing hundreds of daily deaths
and where hospitals were at maximum occupancy, a triage dilemma would plausibly have been
much more realistic than in Wyoming, which at the time had a total of 296 confirmed cases,
plenty of hospital bed availability, and only two coronavirus-related deaths (New York Times,
2020). We hypothesized that regional variation in Covid-19 prevalence could affect
psychological realism, and therefore operationalized county-level case rates as a first realism
proxy.
Another feature that may affect the realism of triage dilemmas is participants’ individual
susceptibility to Covid-19 hospitalization and death. An elderly, at-risk person concerned with
their own risk of developing severe symptomatology, or an essential worker witnessing multiple
coronavirus-related deaths every day as part of their profession might view triage dilemmas in a
more realistic light than average. To evaluate this feature, which we labeled proximity, several
items in our survey assessed the degree to which people were personally affected by the
pandemic.
Thus, whereas prevalence constituted an objective measure of participants’ exposure to
Covid-19 (in virtue of their geographic location), proximity was a self-reported measure of
TRIAGE DILEMMAS 9
9
participants’ subjective affectedness (owing to factors such as occupational hazard, health status,
and so on).
Finally, we also explored whether the pandemic induced a general change in people’s
utilitarian inclinations. To examine this question, we compared participants’ judgments of
traditional sacrificial dilemmas during the pandemic (in May 2020) to a dataset collected one yea
prior. A potential period effect in responses, however, must be interpreted with caution. Given
that, at the height of the pandemic, triage dilemmas received ample attention in mainstream
media, and were at times discussed in conjunction with classic philosophical thought
experiments (e.g., The Atlantic, 2020), a period effect could also indicate a perceived increase in
realism and relevance of traditional dilemmas.
Overview
Our studies were conducted with approval from the Ethics Commission of the University
of Zurich. We first report attitudes toward a series of critical care dilemmas across high and low
prevalence states (Study 1). Then, in Studies 2 and 3, we pursue two potential explanations for
the emergence of heightened utilitarian tendencies in response to triage dilemmas involving
coronavirus patients: a period effect (Study 2) and a realism effect (Study 3). After reporting on
the three studies, we turn to individual difference analyses which help to elucidate variation in
response to triage versus traditional dilemmas.
Data and materials are available on the Open Science Framework at
osf.io/dpsq9/?view_only=54a7c150e03d4d78819b2954cee3a240. We disclose all measures,
manipulations, and exclusions in the studies. Participants were adults residing in the United
States recruited on Amazon Mechanical Turk (mturk.com). To maximize data quality, we
screened out participants who failed a two-part attention check at the start of the experiment.
TRIAGE DILEMMAS 10
10
This attention check resulted in 453 exclusions in Study 1 (33% exclusion rate), 149 exclusions
in Study 2 (14% exclusion rate), and 78 exclusions in Study 3 (14% exclusion rate). The method
by which we determined the sample size for each study is reported in the corresponding ‘Power
Analysis’ subsections.
Frequentist and Bayesian analyses were conducted using R version 3.6.2 and the car (Fox
& Weisberg, 2019) and BayesFactor (Morey & Rouder, 2018) packages. Pairwise comparisons
of simple and marginal effects were calculated with the emmeans (Lenth, 2019) package,
applying Tukey correction for multiple comparisons (notated as pTukey throughout the Results
sections).
Study 1
In Study 1, we surveyed participants’ attitudes toward critical care triage at the height of
the Covid-19 pandemic. The core questions in Study 1 were: Do people typically favor or oppose
utilitarian standards in ICU triage? Do such attitudes toward ICU triage depend upon the
magnitude of the Covid-19 threat? To answer our second question, we evaluated the effects of
geographic variation in the health burden of Covid-19 (as measured by daily, county-level
updates of case and mortality rates) by sampling from high and low prevalence regions of the
United States, and also the effects of subjective beliefs about the virus’s severity.
Methods
Materials
For Study 1, we devised three different critical care dilemmas: Redirect, Withdraw, and
Jeopardize. Each dilemma involves the sacrifice of a human life in order to save a larger number
of lives. In Redirect, an oxygen tank is on its way to a hospital where a single coronavirus patient
requires urgent treatment. The decision must be made whether to redirect the oxygen tank to a
TRIAGE DILEMMAS 11
11
second ICU nearby where five, similarly critical, coronavirus patients recently checked in.
Meanwhile, in Withdraw, an oxygen tank is already in use to treat a single coronavirus patient at
a hospital ICU. This time, the decision must be made whether to disconnect the oxygen tank and
drive it to a second ICU to save five coronavirus patients. Thus, the Redirect/Withdraw pair
serves as a test of the so-called ‘equivalence thesis’: i.e., the notion that withholding and
withdrawing treatment are equivalent from a moral point of view (Sulmasy & Sugarman, 1994).
Finally, in Jeopardize, a frail essential worker can be sent to pick up an oxygen tank and drive it
to an ICU where five coronavirus patients have just checked in. However, in doing so, the
essential worker would likely contract coronavirus and die.
In every scenario, we stipulated thatwith near-certaintythe patients would survive if
they received supplementary oxygen, and would die otherwise. We also stated that the patients
could not be transported (due to their critical condition), that a decision had to be made
imminently, and that there were no further alternatives (i.e., additional oxygen tanks, alternative
treatments, etc.).
Participants
Between April 17th and April 21st 2020, we recruited 906 crowdworkers (53% women;
age: M = 38.9, SD = 13.0, range = 18 to 75). To maximize the ability to detect effects of
prevalence, we recruited participants from five higher prevalence states (> 10 confirmed cases
per 10,000 residents; Florida, Georgia, Michigan, New York, and Pennsylvania), and four lower
prevalence states (< 10 confirmed cases per 10,000 residents; Arizona, Kentucky, North
Carolina, and Texas). Since transmission rates and Covid-19 prevalence were highest in densely
populated areas (i.e. metropolitan areas), and these areas also tend to be more Democratic
TRIAGE DILEMMAS 12
12
(Wilkinson, 2019), we sought to balance the voting trends of high and low prevalence groups of
states (see Table 1).
Table 1
By-State Sample Sizes, Covid-19 Statistics and Voting Trends
Prevalence
State
N
Cases 1
Dem.:Rep.
Ratio 2
High
New York
122
114
52:29
Florida
85
48.5
42:39
Michigan
105
29.2
45:38
Pennsylvania
106
22.1
46:41
Georgia
98
14.7
42:40
Low
Kentucky
68
6.0
41:45
Texas
69
5.8
38:41
Arizona
76
5.8
40:42
North Carolina
168
5.2
44:39
Note. Cases and death rates are per 10000 residents on April 17th 2020 (the first day of data
collection). 1: compiled from the New York Times (2020) GitHub repository. 2: Gallup (2017)
poll data.
Procedure
In a 3 × 1 between-subjects design, participants were asked to consider one of three
critical care dilemmas: Withdraw, Redirect, Driver. Participants were asked to rate the triage
decision through five assessments:
1) Prescription: whether the agent (1) “should” or (0) “should not” perform the
utilitarian action.
TRIAGE DILEMMAS 13
13
2) Permissibility: whether it is morally permissible for the agent to:
a) do nothing: 1 = “Yes”; 0 = “No”,
b) perform the utilitarian action: 1 = “Yes”; 0 = “No”.
In our studies, we compute the normalized average of both dichotomous items after
reverse scoring the first item (i.e., relative permissibility: 0 = Only Inaction
Permissible, 0.5 = Both/Neither Permissible, 1 = Only Action Permissible).
3) Blameworthiness: whether it is morally blameworthy for the agent to:
a) do nothing: 1 = “Not at all”; 7 = “Very much”,
b) perform the utilitarian action: = “Not at all”; 7 = “Very much”.
We compute the difference score, blame for action blame for inaction, (i.e., relative
blame, ranging from -6: most utilitarian, 0: equal blame, to 6: most deontological).
Following the triage dilemmas, participants completed a battery of individual difference
measures. The results of these individual difference analyses are reported in Section 5.
Secondary Data
To devise a measure of local prevalence of Covid-19, we divide the recorded number of
cases and deaths in each US county by its estimated population. County-level cases and deaths
were obtained from the New York Times GitHub repository (github.com/nytimes/covid-19-
data), and population estimates were obtained from the US Census Bureau’s Population
Estimates Program (www.census.gov/programs-surveys/popest.html).
Power Analysis
We set α to .05, and power to .80, in order to conduct a sensitivity power analysis for
logistic regression. Assuming a baseline probability of .6 (
x = 0 = .6), a sample size of 906
TRIAGE DILEMMAS 14
14
participants provided sufficient statistical power to detect effects with an odds ratio (OR) ≥ 1.21
(and ≤ 0.83).
Results
Scenario Differences
A 2 test revealed that prescriptive judgments differed by scenario, 2(2, 903) = 74.73, p <
.001, Cramer's V = .27, 95% CI [.20, .34]. Participants were more likely to save five patients by
redirecting the oxygen tank (
= .87, 95% CI [.83, .91]) than by disconnecting the oxygen tank (
= .60, 95% CI [.55, .66]), OR = 4.46, = .31, z = 7.24, or fatally infecting an essential worker (
= .62, 95% CI [.56, .67]), OR = 4.10, = .29, z = 7.00, both psTukey < .001 (see Figure 1). Still, it
is noteworthy that, by comparison to the uniform distribution, participants tended to favor even
instrumental sacrificewhether of a patient (in Withdraw) or an essential worker (in Jeopardize)
in order to save a larger number of lives.
Permissibility and blame judgments also differed by scenario, permissibility: 2(2, 903) =
24.59, Cramer's V = .16, 95% CI [.11, .20]; blame: F(2, 902) = 35.77, p2 = .08, both ps < .001.
Redirecting an oxygen tank was viewed as more permissible (
= .69, 95% CI [.64, .74]) than
disconnecting an oxygen tank (
= .55, 95% CI [.49, .60]), rpb = .21, 95% CI [.13, .28], z = 3.69,
or placing an essential worker at risk (
= .51, 95% CI [.64, .74]), rpb = .25, 95% CI [.17, .32], z
= 4.65, both psTukey < .001. Redirecting an oxygen tank was also judged less blameworthy ( = -
0.68, 95% CI [-0.95, -0.40]) than either disconnecting an oxygen tank ( = 0.45, 95% CI [0.16,
0.74]), Cohen's d = -0.48, 95% CI [-0.64, -0.31], t = -5.49, or infecting an essential worker ( =
0.97, 95% CI [0.70, 1.25]), Cohen's d = -0.67, 95% CI [-0.83, -0.51], t = -8.28, both psTukey <
.001.
TRIAGE DILEMMAS 15
15
Covid-19 Prevalence
County-level daily cases predicted participants’ self-reported exposure to Covid-19, OR
= 1.85, 95% CI [1.10, 3.10], z = 2.33, p = .020. Participants who reported having had a close
friend or family member hospitalized due to Covid-19 were more likely to reside in counties
with more Covid-19 casesserving as validation of our prevalence assessments.
However, in multiple regression models, county-level Covid-19 prevalence had no effect
on people’s moral attitude toward critical care dilemmas (prevalence: all 2s < 1.40, all ps > .25;
prevalence×scenario: all 2s < 3.50, all ps > .15). Since prevalence is greater in population-dense
areas, and greater population density is associated with a Democratic political orientation, we re-
ran the previous analyses entering participants’ reported voting intentions. These supplementary
analyses revealed qualitatively similar results: Prevalence of Covid-19 did not influence moral
attitudes on any dependent measure (prevalence: all 2s < 1.70, all ps > .15; prevalence×scenario:
all 2s < 3.05, all ps > .20). Figure 1 displays prescriptive judgments by scenario and voting
intention, separately for low- and high- prevalence states.
Covid-19 Threat Severity
We included a two-item measure of Covid-19 severity beliefs (= .73)
described in further detail in the ‘Individual Difference Analyses section. This measure revealed
a weak effect of perceived severity on moral attitudes toward critical care dilemmas: In a
multiple logistic and linear regressions, viewing Covid-19 as a severe threat was associated with
support for utilitarian triage decisions (OR = 1.14, z = 2.00, p = .045), and with judging that
utilitarian action in triage contexts is more permissible (OR = 1.18, t = 2.18, p = .030) and
marginally less blameworthy (B = -0.12, t = -1.69, p = .073, p2 = .003) than inaction.
Figure 1
TRIAGE DILEMMAS 16
16
Prescriptive Judgments in Triage Dilemmas by Prevalence and Voting Intention
Discussion
Participants reported relatively favorable attitudes toward utilitarian resolutions of critical
care dilemmas (see also Wilkinson et al., 2020). The clearest pattern emerged for collateral
sacrifice (i.e., in the Redirect scenario): Participants prescribed utilitarian action, and judged it
relatively more permissible and less blameworthy than inaction. For instrumental sacrifice,
participants continued to prescribe utilitarian action (in the Withdraw and Jeopardize scenarios),
despite rating it somewhat more blameworthy than inaction (and comparably permissible). These
attitudes were unrelated to local prevalence of Covid-19, though they appeared to be weakly
associated with beliefs about coronavirus severity.
TRIAGE DILEMMAS 17
17
Study 2
The observation of broad support for utilitarian interventions in Study 1 raises the
possibility that the Covid-19 pandemic has precipitated a shift in people’s moral attitudes
regarding human life. In Study 2, we examine whether the onset of the Covid-19 pandemic
increased support for utilitarian sacrifice more broadly. To evaluate this hypothesis, we ran a
two-wave longitudinal study contrasting people’s views about utilitarian sacrifice prior to versus
during the Covid-19 pandemic.
Methods
Materials
In Study 2, we drew on a diverse set of traditional moral dilemmas, adapted from
previous studies in moral psychology (e.g., Koenigs et al., 2007). Four scenarios involved
instrumental sacrifice (i.e., death as a means to the greater good; as in Withdraw), and four
scenarios involved collateral sacrifice (i.e., death as a side effect of pursuing the greater good; as
in Redirect). Importantly, none of the dilemmas employed in Study 2 alluded to the Covid-19
pandemic, or described critical care scenarios.
Participants
905 crowdworkers (48% women; age: M = 40.5, SD = 13.6, range = 18 to 81) were
recruited at two time points one year apart: May of 2019 (pre-pandemic n = 395), and May of
2020 (mid-pandemic n = 513).
Procedure
In a 2 (period: pre, mid) × 2 (condition: instrumental, collateral) between-subjects design,
participants were asked to consider a traditional moral dilemma involving either instrumental or
collateral sacrifice. They then made prescriptive and permissibility judgments (but not
TRIAGE DILEMMAS 18
18
blameworthiness judgments), as in Study 1. Following the moral dilemma, participants
completed a battery of individual difference measuresthe results of which are reported in the
Section titled ‘Individual Difference Analyses’.
Power Analysis
We set to .05, and to .20 (power to 80%), in order to conduct a sensitivity power
analysis for logistic regression. A sample size of 905 participants (and assuming
pre = .60)
provided sufficient statistical power to detect effects with an odds ratio (OR) ≥ 1.21 (and ORs ≤
0.83).
Results
We entered period, condition and the two-way interaction between period and condition
in mixed-effects regression models (with scenario as a random effect). Through model selection,
we observed no main effect of period either on prescriptive judgments, 2(df = 1) = 0.00, p = .97,
or on permissibility judgments, F(1, 896) = 0.22, p = .73. In other words, the marginal effects of
period on prescriptive judgments (
pre = .62, 95% CI [.48, .75]; vs.
mid = .63, 95% CI [.49, .75]),
OR = 1.02, z = 0.11, p = .91, and permissibility judgments (pre = -0.16, 95% CI [-1.20, 0.87]; vs.
mid = -0.10, 95% CI [-1.13, 0.93]), B = 0.06, z = 0.06, p = .95, were non-significant.
Furthermore, a Bayesian analysis provided strong evidence for the absence of period effects
(prescriptive: BF01 = 13.84; permissibility: BF01 = 12.92).
We also found no period×condition interaction effect on relative permissibility
judgments, F(1, 896) = 0.72, p = .63though we observed a marginally significant trend for
prescriptive judgments, 2(df = 1) = 2.95, OR = 1.67, p = .086 (see Figure 2). The simple effects
of period on prescriptive judgments were statistically non-significant (collateral: OR = 0.76, z =
TRIAGE DILEMMAS 19
19
-1.27, pTukey = .20; instrumental: OR = 1.27, z = 1.19, pTukey = .23), and absent from a Bayesian
perspective (collateral: BF01 = 9.54; instrumental: BF01 = 8.95).
Propensity Score Matching
Stay-at-home orders, and shifts in childcare duties and employment status, could have
brought about changes in crowdworkers’ demographic profile from May 2019 to May 2020. To
evaluate this concern, we first estimate propensity scores (Rosenbaum & Rubin, 1983) for each
participant: i.e., the predicted probability of participating in the study during the pandemic rather
than before. Indeed, a multiple logistic regression (with age, gender, educational attainment and
native language status) revealed significant differences in the demographic profile of our two
samples, omnibus ²(df = 4) = 17.86, p = .001. Specifically, whereas age (OR = 1.01, z = 1.67, p
= .095), gender (OR = 0.83, z = -1.35, p = .18), and native language (OR = 0.59, z = -1.16, p =
.25) differences were small and non-significant, educational attainment was a clear predictor of
participation during the pandemic (vs. before), OR = 1.18, z = 3.36, p = .001.
A nearest-neighbor matching algorithm yielded 395 matched pairs, which showed good
balance on all four covariates: age (OR = 1.00, z = 0.54, p = .59), gender (OR = 1.07, z = 0.44, p
= .66), native language (OR = 1.25, z = 0.40, p = .69), and educational attainment (OR = 1.05, z
= 0.97, p = .33). In this matched dataset, the main effect of period remained non-significant
(prescription: 2[df = 1] = 0.00, OR = 1.00, p = .99, BF01 = 11.05; permissibility: F[1, 781] =
0.21, p = .64, BF01 = 12.36), as did its interaction with condition (prescription: 2[df = 1] = 1.98,
OR = 1.55, p = .16, BF01 = 3.31; permissibility: F[1, 781] = 0.51, p = .47, BF01 = 7.31).
Discussion
The results of Study 2 suggest that the Covid-19 pandemic has not affected moral
attitudes toward utilitarian sacrifice in general (see also Francis, & McNabb, 2021). Therefore,
TRIAGE DILEMMAS 20
20
elevated utilitarian tendencies in Study 1 cannot straightforwardly be attributed to a period effect.
Next, we turn to the realism hypothesis.
Study 3
To investigate the realism hypothesis, we conducted a third study during the Covid-19
pandemic. According to this hypothesis, the psychological salience of Covid-19 could shift
moral attitudes toward scenarios describing triage of coronavirus patients. Specifically,
psychological realism may render triage dilemmas more engaging than hypothetical dilemmas
and, as a result, elicit different cognitive processes (see also Francis et al., 2016 2017; Patil et al.,
2014). Such a result would imply certain limits on the ecological validity of traditional moral
dilemma research (Baumann et al. 2014).
Methods
Materials
For this study, we devised matched pairs of vignettes involving either instrumental or
collateral sacrifice. In each pair, one vignette (taken from Study 1: i.e., Redirect or Withdraw)
involved triage of critical care resources to treat Covid-19 patients, and another involved a
tightly-matched, control dilemma about a house fire. More specifically, in the Collateral
condition, a fire truck is rerouted from one smaller house fire to a larger house fire to save a
larger number of lives. In the Instrumental condition, a fire truck ceases to extinguish a smaller
house fire in order to save a larger number of lives in a larger, nearby fire.
Participants
A minimum sample size of 400 (or 100 per group) was decided before any data collection
or analysis. 487 crowdworkers (44% women; age: M = 43.5, SD = 14.3, range = 21 to 79) were
recruited in May of 2020.
TRIAGE DILEMMAS 21
21
Procedure
In a 2 within- (realism: triage, control) × 2 between- (condition: personal, impersonal) ×
2 between- (order: triage-first, control-first) subjects design, participants were randomly
assigned to one of four groups. In every group, participants viewed both a triage and a control
(i.e., house fire) dilemma, and we manipulated the order in which the dilemmas appeared, as well
as whether they involved instrumental or collateral sacrifice of a human life.
Power Analysis
We set to .05, and to .20 (power to 80%), in order to conduct a sensitivity power
analysis for logistic regression. A sample size of 487 participants (and setting
control = .60)
provided sufficient statistical power to detect effects with an odds ratio (OR) ≥ 1.30 (and ORs ≤
0.77).
Results
First-Trial Analysis
In a logistic regression model, we entered realism, condition and the two-way interaction
between realism and condition as predictors of prescriptive judgments. A model comparison
revealed effects of condition, 2(df = 1) = 15.74, OR = 2.19, z = 3.92, and of the
realism×condition interaction, 2(df = 1) = 16.92, OR = 5.49, z = 4.00, both ps < .001. No main
effect of realism was observed, 2(df = 1) = 2.22, OR = 1.34, z = 1.48, p = .14.
Figure 2 displays the simple effects in the realism×condition interaction: Participants in
the Redirect condition were more likely to prescribe utilitarian action in a triage context (
.87,
95% CI [.80, .92]) than in a control context (
.65, 95% CI [.57, .73]), OR = 3.64, z = 3.88,
pTukey < .001which corresponds to decisive evidence in Bayesian terms (BF10 = 371.53). The
difference in the Withdraw condition was not statistically significant (
triage = .53, 95% CI [.44,
TRIAGE DILEMMAS 22
22
.63]; vs.
control = .64, 95% CI [.55, .71]), OR = 0.63, t = -1.55, pTukey = .12which constitutes
only anecdotal evidence in favor of the null (BF01 = 2.25).
This pattern of results was mirrored in the ratings of blame: A two-way ANOVA
revealed effects of condition, F(1, 479) = 9.49, p = .002, p2 = .02, realism, F(1, 479) = 4.44, p =
.036, p2 = .01, and of the realism×condition interaction, F(1, 479) = 17.78, p < .001, p2 = .05.
Participants in the Redirect condition assigned less blame for utilitarian action in a triage context
(-1.03, 95% CI [-1.41, -0.65]) than in a control context (0.17, 95% CI [-0-19, 0.54]), t = 4.46,
pTukey < .001, Cohen’s d = -0.64, 95% CI [-0.91, -0.38]. The difference in the Withdraw
condition was not statistically significant (triage: 0.42, 95% CI [0.02, 0.81; vs. control: 0.00,
95% CI [-0.37, 0.37]), t = -1.53, pTukey = .13, Cohen’s d = 0.18, 95% CI [-0.08, 0.44].
Relative permissibility judgments revealed a different pattern: We observed more modest
effects of condition, 2(df = 1) = 5.41, OR = 1.08, p = .020, and of the realism×condition
interaction, 2(df = 1) = 7.18, OR = 1.16, p = .007. No main effect of realism was observed, 2(df
= 1) = 0.24, OR = 1.02, p = .62. This time, utilitarian action resulting in collateral death was not
judged more permissible in a triage context (
.66, 95% CI [.60, .73]) than in a control context
(
.60, 95% CI [.54, .66]), t = 1.53, pTukey = .13, rpb = .10, 95% CI [-.02, .22]. However,
utilitarian action in the Instrumental condition was deemed less permissible in a triage context
(.50, 95% CI [.44, .57]) than in a control context (.60, 95% CI [.55, .66]), t = -2.26, pTukey = .024,
rpb = -.10, 95% CI [-.22, -.03].
Figure 2
Prescriptive Judgments in Studies 1, 2 and 3. Heightened Endorsement of Collateral Sacrifice
(Left) is Explained by an Effect of Realism (Right), not of Period (Center)
TRIAGE DILEMMAS 23
23
Order Effect
Through model comparison, we observed a two-way interaction between condition and
order for every dependent measure (prescription: 2[df = 1] = 13.20, OR = 3.88, p < .001;
permissibility: 2[df = 1] = 4.49, OR = 1.95, p = .034; blame: F(1, 480) = 14.84, p < .001, p2 =
.03). At a broad level, judgments of Collateral sacrifice depended on the order in which the
dilemmas were presented (see Figure 3): When the triage dilemma was presented first,
participants were more likely to report that agents should perform the utilitarian action across
contexts (triage: z = 3.65, OR = 12.23, pTukey < .001; control: z = 2.79, OR = 5.64, pTukey = .005),
that inaction was blameworthy in both contexts (triage: t = -3.20, Cohen's d = -0.45, 95% CI [-
0.71, -0.20], pTukey = .001; control: t = -4.61, Cohen's d = -0.65, 95% CI [-0.95, -0.39], pTukey <
.001), and that utilitarian action in triage contexts was permissible (triage: z = 2.48, OR = 1.96,
pTukey = .013; whereas control: z =1.27, OR = 1.29, pTukey = .21).
TRIAGE DILEMMAS 24
24
Figure 3
Order Effects in Study 3. Marginal Effect of Triage-First Order by Condition.
Note. *: The signs of the relative blame effects have been reversed in this figure. A positive effect
indicates heightened utilitarianism in every panel.
Discussion
The results of Study 3 suggested that heightened utilitarianism in Study 1 (also observed
in Wilkinson et al., 2020) may have arisen from the psychological realism of ICU triage
dilemmas amidst the Covid-19 pandemic. Participants demonstrated more utilitarian attitudes
when judging dilemmas involving critical care triage than when considering structurally
analogous, control dilemmasat least in circumstances demanding collateral sacrifice. This
effect then shaped participants’ moral attitudes in the subsequent trial, resulting in elevated
utilitarian attitudes toward control dilemmas when these were preceded by psychologically
realistic, triage dilemmas. In sum, we obtained evidence for the absence of a period effect in
Study 2 (through Bayesian analyses), but evidence in favor of a realism effect in Study 3.
TRIAGE DILEMMAS 25
25
Individual Difference Analyses
Attitudes towards the Covid-19 pandemic have been shown to vary considerably with
political ideology. Survey evidence reveals significant differences across party lines on
numerous measures: the belief that “powerful people intentionally planned the Covid-19
outbreak” (Republican: 34% v. Democrats 18%, Pew Research, 2020a), and that Covid-19
constitutes a major public health threat (Republicans: 46% v. Democrats: 85%, Pew Research,
2020c), as well as the fear of unwittingly spreading the virus (Republicans: 45% v. Democrats:
77%) or contracting it (Republicans: 35% v. Democrats: 64%; Pew Research, 2020b).
Given the pronounced differences in partisan views about the severity and origin of
Covid-19, as well as in compliance with public health recommendations (Gollwitzer et al.,
2020), one might also expect disagreement between liberals and conservatives concerning
critical care triage. In our studies, we sought to replicate evidence of political polarization
concerning the origin and severity of Covid-19, and to ask whether polarization extends to the
ethics of critical care triage (Hunter, 1991).
Additionally, to deepen our understanding of the role of experimental realism in moral
affect, we administered individual difference measures of empathy (Davis, 1983) and utilitarian
motivation (Kahane et al., 2018). Here we report aggregate analyses examining their associations
with moral attitudes toward triage dilemmas.
Measures
Interpersonal Reactivity Index
Participants in Studies 1, 2 and 3 completed the empathic concern (Cronbach’s αs = .87,
.87, .86) and perspective-taking (Cronbach’s αs = .83, .86, .84) subscales of the Interpersonal
Reactivity Index (Davis, 1983). Scores on the empathic concern subscale have been previously
TRIAGE DILEMMAS 26
26
associated with weaker utilitarian tendencies on hypothetical moral dilemmas (Gleichgerrcht &
Young, 2013; also Hannikainen et al., 2018).
Oxford Utilitarianism Scale
Participants in Studies 1, 2 and 3 completed the nine-item Oxford Utilitarianism Scale
(Kahane et al., 2018), composed of two subscales: instrumental harm (Cronbach’s αs = .77, .81,
.77) and impartial beneficence (Cronbach’s αs = .76, .81, .79). Support for utilitarian trade-offs
in hypothetical moral dilemmas has been associated primarily with scores on the instrumental
harm subscale (Kahane et al., 2018).
Social and Economic Conservatism Scale
Participants in Studies 1 and 3 completed the twelve-item Social and Economic
Conservatism Scale (Everett, 2013; Cronbach’s αs = .87, .88). Participants reported their feelings
toward twelve sociomoral issues that distinguish liberal from conservative ideology (e.g.,
patriotism, religion, welfare benefits [reversed], or gun ownership), on a scale anchored at 1:
Negative’ to 7: ‘Positive’.
Coronavirus Threat Severity
To assess participants’ beliefs about the severity of coronavirus, participants were asked
to answer two questions in Studies 1 and 3 (Cronbach’s αs = .73, .81): (1) “What is your general
attitude towards the current coronavirus situation?”, ranging from 1: ‘People are completely
exaggerating the danger’ to 7: ‘This is a very serious situation’; (2) “Do you think a lockdown as
it is carried out in many countries is warranted?”, ranging from 1: ‘No, I completely disagree
with the lockdown’ to 7: ‘Yes, I completely agree with the lockdown’.
TRIAGE DILEMMAS 27
27
Coronavirus Conspiracy Beliefs
We also introduced a novel five-item assessment of conspiratorial thinking about
coronavirus (Cronbach’s s = .93, .93), adapted from the Generic Conspiracist Beliefs Scale
(Brotherton, French, & Pickering, 2013). For each item (e.g., “The spread of coronavirus has
been the result of the deliberate, concealed efforts of some governments and scientists”),
participants reported their level of agreement from 1: ‘Strongly disagree’ to 7: ‘Strongly agree’.
Power Analyses
We set to .05, and to .20 (power to 80%), in order to conduct sensitivity power
analyses for pairwise correlation.
Traditional Dilemmas. 509 participants (in Study 2 mid-pandemic wave) judged a
traditional dilemma and completed the individual difference measures. This sample provided
sufficient statistical power (80%, with = .05) to detect correlations of |r| ≥ .12.
Triage Dilemmas. 1132 participants judged a triage dilemma (as first trial; in Studies 1
and 3) and completed the individual difference measures. This sample provided enough
statistical power (80%, with = .05) to detect correlations of |r| ≥ .08.
Results
Empathy in Triage versus Traditional Dilemmas
We partially reproduced an association between empathic concern and deontological
judgment (Gleichgerrcht & Young, 2013) in the context of standard hypothetical dilemmas:
Greater empathic concern predicted reduced permissibility of utilitarian action (relative to
inaction), r(509) = -.09, 95% CI [-.18, -.01], p = .034. No corresponding effect emerged for
prescriptive judgments or with the perspective-taking subscale (see Table 1).
TRIAGE DILEMMAS 28
28
In triage dilemmas, we also found an association between empathic concern and
permissibility judgmentsthough in the opposite direction, r(1132) = .09, 95% CI [.03, .14], p
= .003 (see Table 2). In other words, empathic individuals were more likely to view utilitarian
triage as permissible, but utilitarian action in hypothetical dilemmas as impermissible.
Furthermore, the effect of empathic concern on permissibility judgments was moderated
by type of dilemma (empathic concern×dilemma type: 2[df = 1] = 12.28, OR = 1.89, p = .001).
Partialing out the effect of scenario in a multiple regression, confirmed that the marginal slope of
empathic concern in traditional dilemmas was negative (as in prior research; OR = 0.66, partial r
= -.12) but positive in triage dilemmas (OR = 1.25, partial r = .06)and these effects
significantly differed, z = 3.52, p < .001. Still, generally speaking, the role of self-reported affect
appeared to be fickle in our studies.
Table 2
Pairwise Correlation Between IRI Subscales and Moral Judgments in Traditional (Above
Diagonal) and Triage (Below Diagonal) Dilemmas
Ns = 1132 (below); 509 (above).
(1)
(2)
(3)
(4)
(1) Empathic concern
-
.64 ***
[.59, .69]
-.01
[-.07, .066]
-.09 *
[-.18, -.01]
(2) Perspective-taking
.60 ***
[.56, .63]
-
.05
[-.04, .14]
-.01
[-.09, .08]
(3) Prescriptive judgment
.04
[-.02, .10]
.05
[-.01, .10]
-
.60 ***
[.56, .64]
(4) Relative permissibility
.09 ***
[.03, .14]
.07 *
[.01, .13]
.63 ***
[.59, .66]
-
(5) Relative blame
-.03
[-.09, .03]
-.03
[-.09, .03]
-.38 ***
[-.43, -.33]
-.40 ***
[-.45, -.35]
TRIAGE DILEMMAS 29
29
Note. *: p < .05; **: p < .01; ***: p < .005. Bold cells indicate significant correlations between
IRI subscales and moral attitudes.
Utilitarian Beliefs in Triage versus Traditional Dilemmas
Both facets of utilitarianism correlated with judgments of both triage and hypothetical
dilemmas (see Table 3).
Table 3
Pairwise Correlation Between OUS Subscales and Moral Judgments in Traditional (Above
Diagonal) and Triage (Below Diagonal) Dilemmas
Ns = 1132 (below); 509 (above).
(1)
(2)
(3)
(4)
(1) Instrumental harm
-
.37 ***
[.29, .44]
.35 ***
[.27, .42]
.40 ***
[.32, .47]
(2) Impartial beneficence
.29 ***
[.24, .34]
-
.10 *
[.01, .18]
.12 **
[.03, .20]
(3) Prescriptive judgment
.26 ***
[.21, .31]
.18 ***
[.12, .23]
-
.60 ***
[.56, .64]
(4) Relative permissibility
.14 ***
[.08, .20]
.10 ***
[.04, .15]
.63 ***
[.59, .66]
-
(5) Relative blame
-.16 ***
[-.22, -.11]
-.11 ***
[-.16, -.05]
-.38 ***
[-.43, -.33]
-.40 ***
[-.45, -.35]
Note. *: p < .05; **: p < .01; ***: p < .005. Bold cells indicate significant correlations between
OUS subscales and moral attitudes.
A series of regression models uncovered significant differences in the role of utilitarian
beliefs across dilemma types: First, instrumental harm and impartial beneficence differentially
predicted prescriptive judgments across dilemma types (IH×dilemma type: 2[df = 1] = 5.12, OR
= 1.28, p =.024; IB×dilemma type: 2[df = 1] = 8.30, OR = 1.37, p = .004). Specifically,
instrumental harm beliefs better predicted judgments about traditional dilemmas (OR = 1.96,
TRIAGE DILEMMAS 30
30
partial rpb = .37) than about triage dilemmas (OR = 1.53, partial rpb = .22), z = -2.26, p = .024. In
turn, impartial beneficence beliefs predicted judgments about triage dilemmas (OR = 1.25, partial
rpb = .10), but did not predict judgments of traditional dilemmas (OR = 0.91, partial rpb = -.05)
and these coefficients statistically differed from each other, z = 2.88, p = .004.
Second, utilitarian beliefs differentially predicted permissibility judgments across
dilemma types (IH×dilemma type: 2[df = 1] = 13.92, OR = 1.45, p < .001; IB×dilemma type:
2[df = 1] = 3.00, OR = 0.84, p = .083). Instrumental harm beliefs better predicted judgments of
permissibility in traditional dilemmas (OR = 1.74, partial r = .42) than in triage dilemmas (OR =
1.17, partial r = .14)and these effects statistically differed from each other, z = -5.68, p < .001.
Meanwhile, impartial beneficence predicted permissibility judgments for triage dilemmas (OR =
1.28, partial r = .14) thought not traditional dilemmas (OR = 1.08, partial r = .04); and the
difference between these marginal trends approached significance, z = 1.73, p = .083.
Third, impartial beneficence independently predicted judgments of relative blame in
triage dilemmas (B = -0.01, t = -2.22, p = .026, partial r = .07), even after controlling for the
effect of instrumental harm beliefs (B = -0.02, t = -4.19, p < .001, partial r = .12). We did not
include the blame dependent measures in Study 2, precluding the comparison across dilemma
types.
Collectively, these results show that views about triage dilemmas depend upon both
instrumental harm beliefs and impartial beneficencewhereas attitudes toward traditional
dilemmas are shaped primarily (or even solely) by beliefs about instrumental harm.
Ideological Disagreements on Triage Dilemmas
The 12-item SECS scale discriminated between Democrats (M = 4.13, SD = 0.93),
Republicans (M = 5.64, SD = 0.78) and Independents (M = 4.69, SD = 0.99), all pairwise ps <
TRIAGE DILEMMAS 31
31
.001. We observed clear associations between political ideology and general attitudes toward
Covid-19: Conservatives were more likely to hold conspiratorial beliefs about Covid-19, r(1132)
= .20, 95% CI [.14, .26], while liberals were more likely to treat Covid-19 as a serious threat,
r(1130) = -.31, 95% CI [-.36, -.26], both ps < .001.
Despite replicating ideological disagreements concerning the truth about Covid-19, we
observed few moral disagreements between liberals and conservatives (see Table 4): No
differences arose in either the Jeopardize or Reroute scenarios; meanwhile, in the Withdraw
scenario, conservatives were more likely to blame the utilitarian agent than were liberals (see
also Luke & Gawronski, 2021; Hannikainen, Miller & Cushman, 2017)but no parallel
differences emerged for permissibility or prescriptive judgments. Bayesian model comparisons
confirmed that liberals and conservatives revealed robust disagreement concerning the truth
about Covid-19 (BF10s > 1011), but tended to agree in their moral attitudes toward triage
dilemmas (BF01s > 5; except regarding permissibility, which was inconclusive BF10 < 3).
Table 4
Ideological Differences in Moral Attitudes, Conspiratorial Thinking and Perceived Severity
Withdraw
Jeopardize
Redirect
Main effect of
SECS score
(BF)
Ns
386
309
433
1128 (pooled)
(1) Prescriptive judgment
-.09
[-.18, .02]
.00
[-.11, .11]
-.07
[-.16, .03]
5.92
(null)
(2) Relative permissibility
-.02
[-.12, .08]
.02
[-.10, .13]
-.01
[-.11, .08]
15.22
(null)
(3) Relative blame
-.19 ***
[-.28, -.09]
-.06
[-.17, .05]
-.00
[-.10, .09]
2.21
(alt.)
TRIAGE DILEMMAS 32
32
(4) Corona-threat
-.24 ***
[-.33, -.14]
-.36 ***
[.-45, -.26]
-.33 ***
[.-42, -.25]
8.31 × 1027
(alt.)
(5) Corona-conspiracy
.14 **
[-.28, -.09]
.29 ***
[.18, .39]
.19 ***
[.10, .28]
5.99 × 1011
(alt.)
Note. *: p < .05; **: p < .01; ***: p < .005. null: in support of null model; alt.: in support of
alternative model. Bold cells indicate significant correlations with SECS scores.
Discussion
First, we found limited effects of empathy, which could stem from the self-reported
nature of our measures. Importantly, the effect of empathic concernwhere presentreversed
from traditional, trolley-type contexts to psychologically realistic, triage contexts: i.e.,
underlying deontological judgments in traditional dilemmas, but utilitarian judgments in triage
dilemmas.
Second, attitudes toward traditional dilemmas were linked to participants’ abstract
attitudes toward instrumental harm (in line with Kahane et al., 2015 2018), whereas attitudes
toward dilemmas involving critical care were also tied to participants’ tendency toward impartial
beneficence. Thus, for psychologically realistic dilemmas, prosocial attitudes including affective
empathy and impartial beneficence can also be predictors of greater utilitarian tendencies (in
contrast to previous work on trolley-type dilemmas; see e.g., Bartels & Pizarro, 2011;
Gleichgerrcht & Young, 2013).
Third, we replicated the association between political ideology and beliefs about the
origin and severity of Covid-19revealing substantial polarization. Yet, in their moral attitudes
toward critical care triage, liberals and conservatives appeared to be largely in agreement.
General Discussion
Our study documented heightened utilitarianism in diverse locations throughout the
United States in reaction to dilemmas describing ICU triage scenarios. These attitudes appeared
TRIAGE DILEMMAS 33
33
not to depend on the prevalence of Covid-19 confirmed positives; but, rather, arose
homogeneously across high and low prevalence regions (Study 1). We then reasoned that this
effect could stem from a period effect, i.e., heightened utilitarianism following the onset of the
Covid-19 pandemicyet our evidence spoke against period-driven change in moral judgment
(Study 2). Alternatively, the effect could be due to the psychological realism of critical care
dilemmas. Our next follow-up (Study 3) lent support to this interpretation, demonstrating a
selective tendency to favor utilitarian resolutions of triage dilemmas, relative to tightly-matched
dilemmas unrelated to the Covid-19 pandemic.
As such, our findings dovetail with prior evidence showing that promoting experimental
realism amplifiesrather than dampensutilitarian tendencies (Patil et al., 2014; Francis et al.,
2016, 2017, see also Baumann et al. 2012). This result cannot easily be explained by theories that
link utilitarian tendencies to sub-clinical deficits in affect (Bartels & Pizarro, 2011). Our
moderation analyses pointed toward a distinct explanation: First, affective engagement was
linked to utilitarian responses in triage dilemmas (but deontological responses to traditional
dilemmas). Thus, experimental realism could heighten utilitarian responses, by activating an
approach motivation (i.e., the concern with saving lives), which is otherwise absent in traditional
dilemma contexts. Adopting the two-dimensional model of utilitarianism provided convergent
evidence: As in previous research (Kahane et al., 2015), when considering traditional sacrificial
dilemmas, utilitarian responses depended primarily on the belief that instrumental harm is
permissible. In contrast, utilitarian triage decisions were independently linked to both dimensions
of utilitarianism: Beliefs about the permissibility of instrumental harm continued to predict
utilitarian attitudes toward triage, but so did attitudes of impartial beneficence.
TRIAGE DILEMMAS 34
34
Implications
Our study adds to growing concerns that the literature on trolley dilemmas has generated
a series of misconceptions about moral cognition in general (Baumann et al. 2012), and
utilitarian ethics in particular. First, predominant experimental paradigms may have downplayed
the extent to which people endorse utilitarian outcomes in real life (see also Francis et al., 2017;
Patil et al., 2014). Relatedly, existing work has associated utilitarian ethics to an antisocial
personality (Bartels & Pizarro, 2011), and yet our present studies demonstrate that utilitarian
attitudes can also arise among prosocial individualsat least when considering moral dilemmas
that afford psychological and experimental realism.
On a more promising note, our study reveals thatwhile the origin and severity of
Covid-19 are a matter of profound political divisionat least some ethical questions that arise
for the allocation of public health resources are not. Liberals and conservatives appear to agree,
in large part, on the morality of utilitarian resolutions to dilemmas arising in the context of
critical care triage.
Limitations and Future Research
First, our studies recruited large samples in order to maximize statistical power to observe
average-sized effects. As such, even very small individual differences (e.g., the effects of
empathy and perspective-taking on moral attitudes) reached the threshold of statistical
significance. However, statistical significance does not entail practical significance; so readers
should exercise caution and consider both statistical significance and effect size to gauge the
relevance of each of our findings.
Second, although we obtained strong evidence that psychological realism exerts an
influence on moral judgment, our studies provided limited insight into the mechanism inducing
TRIAGE DILEMMAS 35
35
this effect. Moderation analyses in Section 5 yielded tentative evidence that prosocial
dispositions may explain people’s heightened utilitarianism in critical care dilemmas:
Potentially, coronavirus patients elicit attitudes of impartial beneficence to a greater extent than
do the nondescript victims of traditional moral dilemmas. However, confirmatory research is
needed to establish the credibility of this particular explanation.
Third, given rapid change in the circumstances surrounding the Covid-19 pandemic, we
cannot confidently describe any constraints on the generality of our findings. Plausibly, as
Covid-19 continues to ‘normalize’, its psychological salience—the very feature at the core of our
findingsmight dwindle. If so, a deeper understanding of the present phenomenon may require
conceptual adaptations of the present methods to novel contexts.
Fourth, our conclusions are based strictly on online convenience samples of US
participants. Research comparing such samples to nationally representative samples underscores
that, while demographic differences are pronounced, the results of social science studies
generally replicate across sampling methodswith few exceptions (Chandler, Rosenzweig,
Moss, Robinson, & Litman, 2019). Furthermore, studies of utilitarian sacrifice have documented
a remarkable degree of universality (Awad et al., 2020). Relatedly, the present tendency toward
heightened utilitarianism on triage dilemmas emerged also in a nationally representative sample
(Wilkinson et al., 2020).
Conclusion
Moral psychologists have long debated whether, and under what conditions, people favor
utilitarian approaches to the sacrifice of human lives for the greater good. The early stages of the
Covid-19 pandemic provided a unique context to explore this issue in circumstances of
unprecedented realism and practical relevance: What should frontline ICU professionals do when
TRIAGE DILEMMAS 36
36
faced with the decision whether to sacrifice one patient and save a larger number of lives?
Despite flagrant political polarization surrounding the origin and severity of Covid-19, liberals
and conservatives alike supported utilitarian approaches to the rationing of scarce medical
resources during the pandemic. This result could not have easily been predicted by previous
research on trolley-type sacrificial dilemmas, perhaps because these unrealistic scenarios fail to
muster the same cognitive processes, such as impartial beneficence, that govern utilitarian
reasoning about pressing, real-world issues. As a result, our evidence provides grounds for
optimism about the multilateral development of public health guidelines to govern triage and
allocation of critical care resources.
TRIAGE DILEMMAS 37
37
References
Antoniou, R., Romero-Kornblum, H., Young, J. C., You, M., Kramer, J., & Chiong, W. (ms). No
Utilitarians in a Pandemic? Shifts in Moral Reasoning during the COVID-19 Global
Health Crisis. https://doi.org/10.31234/osf.io/yjn3u
Aronson, E., Wilson, T. D., & Brewer, M. B. (1998). Experimentation in social psychology. In
G. Lindsay & E. Aronson (Eds.), The Handbook of Social Psychology (4th edn, Vol. 1,
pp. 99142). Boston, MA: McGraw-Hill.
Arora, C., Savulescu, J., Maslen, H. et al. (2016). The Intensive Care Lifeboat: a survey of lay
attitudes to rationing dilemmas in neonatal intensive care. BMC Med Ethics 17, 69
(2016). https://doi.org/10.1186/s12910-016-0152-y
Awad, E., Dsouza, S., Shariff, A., Rahwan, I., & Bonnefon, J. F. (2020). Universals and
variations in moral decisions made in 42 countries by 70,000 participants. Proceedings of
the National Academy of Sciences, 117(5), 2332-2337.
Bauman, C. W., McGraw, A. P., Bartels, D. M., & Warren, C. (2014). Revisiting external
validity: Concerns about trolley problems and other sacrificial dilemmas in moral
psychology. Social and Personality Psychology Compass, 8(9), 536-554.
Bartels, D. M., & Pizarro, D. A. (2011). The mismeasure of morals: Antisocial personality traits
predict utilitarian responses to moral dilemmas. Cognition, 121(1), 154-161.
Bartels, D. M., Bauman, C. W., Cushman, F. A., Pizarro, D. A., & McGraw, A. P. (2015). Moral
judgment and decision making. The Wiley Blackwell handbook of judgment and decision
making, 63, 478-515.
TRIAGE DILEMMAS 38
38
Bostyn, D. H., Sevenhant, S., & Roets, A. (2018). Of mice, men, and trolleys: Hypothetical
judgment versus real-life behavior in trolley-style moral dilemmas. Psychological
science, 29(7), 1084-1093.
Brotherton, R., French, C. C., & Pickering, A. D. (2013). Measuring belief in conspiracy
theories: The generic conspiracist beliefs scale. Frontiers in psychology, 4, 279.
Buckwalter, W., & Peterson, A. (2020). Public attitudes toward allocating scarce resources in the
COVID-19 pandemic. PloS one, 15(11), e0240651.
Campbell, D. T. (1957). Factors relevant to validity of experiments in social settings.
Psychological Bulletin, 54, 297312.
Chandler, J., Rosenzweig, C., Moss, A. J., Robinson, J., & Litman, L. (2019). Online panels in
social science research: Expanding sampling methods beyond Mechanical Turk. Behavior
Research Methods, 51(5), 2022-2038.
Christensen, J. F., & Gomila, A. (2012). Moral dilemmas in cognitive neuroscience of moral
decision-making: A principled review. Neuroscience & Biobehavioral Reviews, 36(4),
1249-1264.
Cole Wright, J., Cullum, J., & Schwab, N. (2008). The cognitive and affective dimensions of
moral conviction: Implications for attitudinal and behavioral measures of interpersonal
tolerance. Personality and Social Psychology Bulletin, 34(11), 1461-1476.
Converse, P. E. (2006). The nature of belief systems in mass publics (1964). Critical review,
18(1-3), 1-74.
Conway, P., & Gawronski, B. (2013). Deontological and utilitarian inclinations in moral decision
making: a process dissociation approach. Journal of personality and social psychology,
104(2), 216.
TRIAGE DILEMMAS 39
39
Conway, P., Goldstein-Greenwood, J., Polacek, D., & Greene, J. D. (2018). Sacrificial utilitarian
judgments do reflect concern for the greater good: Clarification via process dissociation
and the judgments of philosophers. Cognition, 179, 241-265.
Cushman, F., Young, L., & Hauser, M. (2006). The role of conscious reasoning and intuition in
moral judgment: Testing three principles of harm. Psychological science, 17(12), 1082-
1089.
Emanuel, E. J., Persad, G., Upshur, R., Thome, B., Parker, M., Glickman, A., ... & Phillips, J. P.
(2020). Fair allocation of scarce medical resources in the time of Covid-19. New England
Journal of Medicine (382), 2049-2055.
Everett, J. A. C. (2013). The 12 item social and economic conservatism scale (SECS). PloS one,
8(12), e82131.
Fox, J., & Weisberg, S. (2019). An R Companion to Applied Regression, Third edition. Sage:
Thousand Oaks, CA.
Francis, K. B., Howard, C., Howard, I. S., Gummerum, M., Ganis, G., Anderson, G., & Terbeck,
S. (2016). Virtual morality: Transitioning from moral judgment to moral action?. PloS
one, 11(10), e0164374.
Francis, K., & McNabb, C. B. (2021). Moral Decision-Making during COVID-19: Moral
judgments, moralisation, and everyday behaviour. https://doi.org/10.31219/osf.io/jvfds.
Francis, K. B., Terbeck, S., Briazu, R. A., Haines, A., Gummerum, M., Ganis, G., & Howard, I.
S. (2017). Simulating moral actions: An investigation of personal force in virtual moral
dilemmas. Scientific Reports, 7(1), 1-11.
Foot, P. (1967). The problem of abortion and the doctrine of the double effect in virtues and
vices. Oxford Review, 5, 515.
TRIAGE DILEMMAS 40
40
Gleichgerrcht, E., & Young, L. (2013). Low levels of empathic concern predict utilitarian moral
judgment. PloS one, 8(4), e60418.
Graham, J., Haidt, J., & Nosek, B. A. (2009). Liberals and conservatives rely on different sets of
moral foundations. Journal of personality and social psychology, 96(5), 1029.
Greene, J. D., Sommerville, R. B., Nystrom, L. E., Darley, J. M., & Cohen, J. D. (2001). An
fMRI investigation of emotional engagement in moral judgment. Science, 293(5537),
2105-2108.
Greene, J. D. (2007). The secret joke of Kant’s soul. In W. Sinnott-Armstrong (Ed.), Moral
Psychology, Vol. 3: The Neuroscience of Morality: Emotion, Disease, and Development
(pp. 35117). Cambridge, MA: MIT Press.
Greene, J. D., Cushman, F. A., Stewart, L. E., Lowenberg, K., Nystrom, L. E., & Cohen, J. D.
(2009). Pushing moral buttons: The interaction between personal force and intention in
moral judgment. Cognition, 111(3), 364-371.
Hannikainen, I. R., Machery, E., & Cushman, F. A. (2018). Is utilitarian sacrifice becoming
more morally permissible?. Cognition, 170, 95-101.
Hannikainen, I. R., Miller, R. M., & Cushman, F. A. (2017). Act versus impact: Conservatives
and liberals exhibit different structural emphases in moral judgment. Ratio, 30(4), 462-
493.
Hunter, J. D. (1991). Culture wars: The struggle to define America. New York: Basic Books.
Kahane, G., Everett, J. A. C., Earp, B. D., Farias, M., & Savulescu, J. (2015). ‘Utilitarian’
judgments in sacrificial moral dilemmas do not reflect impartial concern for the greater
good. Cognition, 134, 193209.
TRIAGE DILEMMAS 41
41
Kahane, G., Everett, J. A., Earp, B. D., Caviola, L., Faber, N. S., Crockett, M. J., & Savulescu, J.
(2018). Beyond sacrificial harm: A two-dimensional model of utilitarian psychology.
Psychological Review, 125(2), 131.
Koenigs, M., Young, L., Adolphs, R., Tranel, D., Cushman, F., Hauser, M., & Damasio, A.
(2007). Damage to the prefrontal cortex increases utilitarian moral judgements. Nature,
446(7138), 908-911.
Koenigs, M., Kruepke, M., Zeier, J., & Newman, J. P. (2012). Utilitarian moral judgment in
psychopathy. Social cognitive and affective neuroscience, 7(6), 708-714.
Lenth, R. (2019). emmeans: Estimated Marginal Means, aka Least-Squares Means. R package
version 1.4.3.01. https://CRAN.R-project.org/package=emmeans
Luke, D. M., & Gawronski, B. (2021). Political Ideology and Moral Dilemma Judgments: An
Analysis Using the CNI Model. Personality and Social Psychology Bulletin,
0146167220987990.
Mikhail, J. (2007). Universal moral grammar: Theory, evidence, and the future. Trends in
Cognitive Sciences, 11, 143152.
Mikhail, J. (2009). Moral grammar and intuitive jurisprudence: A formal model of unconscious
moral and legal knowledge. In D. M. Bartels, C. W. Bauman, L. J. Skitka, & D. L. Medin
(Eds.), Moral Judgment and Decision Making: The Psychology of Learning and
Motivation (Vol. 50, pp 27100). San Diego: Elsevier.
New York Times. (2020). Coronavirus (Covid-19) Data in the United States.
https://github.com/nytimes/covid-19-data
Patil, I. (2015). Trait psychopathy and utilitarian moral judgement: The mediating role of action
aversion. Journal of Cognitive Psychology, 27(3), 349-366.
TRIAGE DILEMMAS 42
42
Patil, I., Cogoni, C., Zangrando, N., Chittaro, L., & Silani, G. (2014). Affective basis of
judgment-behavior discrepancy in virtual experiences of moral dilemmas. Social
neuroscience, 9(1), 94-107.
Patil, I., & Silani, G. (2014). Reduced empathic concern leads to utilitarian moral judgments in
trait alexithymia. Frontiers in psychology, 5, 501.
Pew Research Center (June 2020a), “Three Months In, Many Americans See Exaggeration,
Conspiracy Theories and Partisanship in COVID- 19 News.”
Pew Research Center (June 2020b), “Republicans, Democrats Move Even Further Apart in
Coronavirus Concerns”.
Pew Research Center (July 2020c), “Republicans remain far less likely than Democrats to view
COVID-19 as a major threat to public health”.
Reeskens, T., Roosma, F., & Wanders, E. (2021). The perceived deservingness of COVID-19
healthcare in the Netherlands: a conjoint experiment on priority access to intensive care
and vaccination. BMC public health, 21(1), 1-8.
Romero-Rivas, C., & Rodriguez-Cuadrado, S. (2021). The Psychological Impact of the COVID-
19 Pandemic Affected Decision-Making Processes. The Spanish journal of
psychology, 24.
Rozin, P., Lowery, L., Imada, S., & Haidt, J. (1999). The CAD triad hypothesis: a mapping
between three moral emotions (contempt, anger, disgust) and three moral codes
(community, autonomy, divinity). Journal of personality and social psychology, 76(4),
574.
TRIAGE DILEMMAS 43
43
Skitka, L. J., Bauman, C. W., & Sargis, E. G. (2005). Moral conviction: Another contributor to
attitude strength or something more?. Journal of personality and social psychology,
88(6), 895.
Tenbrunsel, A. E., & Messick, D. M. (1999). Sanctioning systems, decision frames, and
cooperation. Administrative Science Quarterly, 44(4), 684-707.
The Atlantic. (2020). The Extraordinary Decisions Facing Italian Doctors.
https://www.theatlantic.com/ideas/archive/2020/03/who-gets-hospital-bed/607807/
Thomson, J. J. (1976). Killing, letting die, and the trolley problem. The Monist, 59, 204217.
Thomson, J. J. (1985). The trolley problem. Yale Law Journal, 94, 13951415.
Wilkinson, D., Zohny, H., Kappes, A., Sinnott-Armstrong, W., & Savulescu, J. (2020). Which
factors should be included in triage? An online survey of the attitudes of the UK general
public to pandemic triage dilemmas. BMJ Open, 10:e045593. doi: 10.1136/bmjopen-
2020-045593.
Wilkinson, W. (2019). The Density Divide: Urbanization, Polarization, and Populist Backlash.
Niskanen Center Research Paper.
ResearchGate has not been able to resolve any citations for this publication.
Article
Full-text available
A sample of 641 participants were presented with four decision-making tasks during the first stages of the COVID–19 lockdown in Spain: The dictator game, framing problems, utilitarian/deontological and altruistic/egoistic moral dilemmas. Participants also completed questionnaires on mental health status and experiences related to the COVID–19 pandemic. We used boosted regression trees (an advanced form of regression analysis based on machine learning) to model relationships between responses to the questionnaires and decision-making tasks. Results showed that the psychological impact of the COVID–19 pandemic predicted participants’ responses to the framing problems and utilitarian/deontological and altruistic/egoistic moral dilemmas (but not to the dictator game). More concretely, the more psychological impact participants suffered, the more they were willing to choose the safest response in the framing problems, and the more deontological/altruistic were their responses to moral dilemmas. These results suggest that the psychological impact of the COVID–19 pandemic might prompt automatic processes.
Article
Full-text available
Background Amidst the COVID-19 pandemic, governments, health experts, and ethicists have proposed guidelines about ICU triage and priority access to a vaccine. To increase political legitimacy and accountability, public support is important. This study examines what criteria beyond medical need are deemed important to be perceived of priority COVID-19 healthcare access. Method Two conjoint experiments about priority over ICU treatment and early COVID-19 vaccination were implemented in a probability-based sample of 1461 respondents representative of the Netherlands. Respondents were asked who should receive treatment out of two fictitious healthcare claimants that differed in in age, weight, complying with corona policy measures, and occupation, all randomly assigned. Average marginal coefficient effects are estimated to assess the relative importance of the attributes; attributes were interacted with relevant respondent characteristics to find whether consensus exists in this relative ranking. Results The Dutch penalize those not complying with coronavirus policy measures, and the obese, but prioritize those employed in ‘crucial’ sectors. For these conditions, there is consensus among the population. For age, young people are prioritized for ICU treatment, while the middle-aged are given priority over a vaccine, with younger respondents favoring healthcare for elderly claimants, while older respondents favor support for young cohorts. Conclusion People who have no control over their social risk and are able to reciprocate to society are considered as more deserving of priority of COVID-19 healthcare. Our findings provide fair support for the implemented ethical guidelines about ICU-treatment and COVID-19 vaccines.
Article
Full-text available
Many real-world dilemmas involve disagreement about whether decisions should follow moral norms in an unconditional manner (deontology) or be based on the consequences for the greater good (utilitarianism). To examine how political ideology may account for some of these disagreements, the current research used a formal modeling approach to investigate associations between political ideology and (1) sensitivity to consequences, (2) sensitivity to moral norms, and (3) general preference for inaction versus action in responses to moral dilemmas. Across three studies (N = 996) with samples from the United States (Studies 1 and 3) and the United Kingdom (Study 2), conservatives were less influenced by overall consequences for the greater good in comparison to liberals. Political ideology was not significantly associated with sensitivity to moral norms and general action tendencies. The findings provide more nuanced insights into how political ideology may contribute to disagreements on real-world moral dilemmas.
Article
Full-text available
Objective As cases of COVID-19 infections surge, concerns have renewed about intensive care units (ICUs) being overwhelmed and the need for specific triage protocols over winter. This study aimed to help inform triage guidance by exploring the views of lay people about factors to include in triage decisions. Design, setting and participants Online survey between 29th of May and 22nd of June 2020 based on hypothetical triage dilemmas. Participants recruited from existing market research panels, representative of the UK general population. Scenarios were presented in which a single ventilator is available, and two patients require ICU admission and ventilation. Patients differed in one of: chance of survival, life expectancy, age, expected length of treatment, disability and degree of frailty. Respondents were given the option of choosing one patient to treat or tossing a coin to decide. Results Seven hundred and sixty-three participated. A majority of respondents prioritised patients who would have a higher chance of survival (72%–93%), longer life expectancy (78%–83%), required shorter duration of treatment (88%–94%), were younger (71%–79%) or had a lesser degree of frailty (60%–69%, all p<0.001). Where there was a small difference between two patients, a larger proportion elected to toss a coin to decide which patient to treat. A majority (58%–86%) were prepared to withdraw treatment from a patient in intensive care who had a lower chance of survival than another patient currently presenting with COVID-19. Respondents also indicated a willingness to give higher priority to healthcare workers and to patients with young children. Conclusion Members of the UK general public potentially support a broadly utilitarian approach to ICU triage in the face of overwhelming need. Survey respondents endorsed the relevance of patient factors currently included in triage guidance, but also factors not currently included. They supported the permissibility of reallocating treatment in a pandemic.
Article
Full-text available
The general public is subject to triage policies that allocate scarce lifesaving resources during the COVID-19 pandemic, one of the worst public health emergencies in the past 100 years. However, public attitudes toward ethical principles underlying triage policies used during this pandemic are not well understood. Three experiments (preregistered; online samples; N = 1,868; U.S. residents) assessed attitudes toward ethical principles underlying triage policies. The experiments evaluated assessments of utilitarian, egalitarian, prioritizing the worst-off, and social usefulness principles in conditions arising during the COVID-19 pandemic, involving resource scarcity, resource reallocation, and bias in resource allocation toward at-risk groups, such as the elderly or people of color. We found that participants agreed with allocation motivated by utilitarian principles and prioritizing the worst-off during initial distribution of resources and disagreed with allocation motivated by egalitarian and social usefulness principles. At reallocation, participants agreed with giving priority to those patients who received the resources first. Lastly, support for utilitarian allocation varied when saving the greatest number of lives resulted in disadvantage for at-risk or historically marginalized groups. Specifically, participants expressed higher levels of agreement with policies that shifted away from maximizing benefits to one that assigned the same priority to members of different groups if this mitigated disadvantage for people of color. Understanding these attitudes can contribute to developing triage policies, increase trust in health systems, and assist physicians in achieving their goals of patient care during the COVID-19 pandemic.
Preprint
Full-text available
The COVID-19 pandemic continues to pose significant health, economic, and social challenges. Given that many of these challenges have moral relevance, the present study investigates whether the COVID-19 pandemic is influencing moral decision-making and whether moralisation of behaviours specific to the crisis predict adherence to government-recommended behaviours. While proto-utilitarian tendencies and utilitarian moral judgments remain unchanged during the pandemic, individuals have moralised non-compliant behaviours associated with the pandemic such as failing to physically distance from others. Importantly, our findings show that this moralisation predicts individual compliance with government-recommended behaviours and as such, provides significant implications for public messaging strategies during the pandemic.
Article
Full-text available
When do people find it acceptable to sacrifice one life to save many? Cross-cultural studies suggested a complex pattern of universals and variations in the way people approach this question, but data were often based on small samples from a small number of countries outside of the Western world. Here we analyze responses to three sacrificial dilemmas by 70,000 participants in 10 languages and 42 countries. In every country, the three dilemmas displayed the same qualitative ordering of sacrifice acceptability, suggesting that this ordering is best explained by basic cognitive processes, rather than cultural norms. The quantitative acceptability of each sacrifice, though, showed substantial country-level variations. We show that low relational mobility (where people are more cautious about not alienating their current social partners) is strongly associated with the rejection of sacrifices for the greater good (especially for Eastern countries), which may be explained by the signaling value of this rejection. We make our dataset fully available as a public resource for researchers studying universals and variations in human morality: all the data and code used in this article can be downloaded at https://bit.ly/2Y7Brr9.
Article
Full-text available
Amazon Mechanical Turk (MTurk) is widely used by behavioral scientists to recruit research participants. MTurk offers advantages over traditional student subject pools, but it also has important limitations. In particular, the MTurk population is small and potentially overused, and some groups of interest to behavioral scientists are underrepresented and difficult to recruit. Here we examined whether online research panels can avoid these limitations. Specifically, we compared sample composition, data quality (measured by effect sizes, internal reliability, and attention checks), and the non-naivete of participants recruited from MTurk and Prime Panels—an aggregate of online research panels. Prime Panels participants were more diverse in age, family composition, religiosity, education, and political attitudes. Prime Panels participants also reported less exposure to classic protocols and produced larger effect sizes, but only after screening out several participants who failed a screening task. We conclude that online research panels offer a unique opportunity for research, yet one with some important trade-offs.
Preprint
The COVID-19 pandemic poses many real-world moral dilemmas, which can pit the needs and rights of the many against the needs and rights of the few. We investigated the influence of this contemporary global crisis on moral judgments in older adults, who are at greatest personal risk from the pandemic. We hypothesized that during this pandemic, individuals would give fewer utilitarian responses to hypothetical dilemmas, accompanied by higher levels of confidence and emotion elicitation. Our pre-registered analysis (https://osf.io/g2wtp) involved two waves of data collection, before (2014) and during (2020) the COVID-19 pandemic, regarding three categories of moral dilemmas (personal rights, agent-centered permissions, and special obligations). While utilitarian responses considered across all categories of dilemma did not differ, participants during the 2020 wave gave fewer utilitarian responses to dilemmas involving personal rights; that is, they were less willing to violate the personal rights of others to produce the best overall outcomes.