CHAPTER

The effect of analytic and

experiential modes of
thought on moral judgment
11
Trevor Kvaran1, Shaun Nichols, Alan Sanfey
Department of Psychology, University of Arizona, Tucson, AZ, USA
Corresponding author. Tel.: þ1-520-3939900, Fax: 520-751-8668,
1

e-mail address: [email protected]

Abstract
According to dual-process theories, moral judgments are the result of two competing pro-
cesses: a fast, automatic, affect-driven process and a slow, deliberative, reason-based process.
Accordingly, these models make clear and testable predictions about the influence of each
system. Although a small number of studies have attempted to examine each process indepen-
dently in the context of moral judgment, no study has yet tried to experimentally manipulate
both processes within a single study. In this chapter, a well-established “mode-of-thought”
priming technique was used to place participants in either an experiential/emotional or analytic
mode while completing a task in which participants provide judgments about a series of moral
dilemmas. We predicted that individuals primed analytically would make more utilitarian
responses than control participants, while emotional priming would lead to less utilitarian
responses. Support was found for both of these predictions. Implications of these findings
for dual-process theories of moral judgment will be discussed.

Keywords
moral judgment, dual-process, moral psychology, mode of thought, utilitarianism

1 INTRODUCTION
Since at least the time of the ancient Greeks, philosophers have contemplated the psy-
chological mechanisms that underlie everyday moral judgment. However, it was not
until the twentieth century that psychologists began to empirically study the moral fac-
ulties of humans (Kohlberg, 1981; Piaget, 1932). Contemporary psychological
research has focused heavily on how individuals reason and make judgments about
complex moral dilemmas, that is, cases or situations in which two moral concerns
are brought into tension with each other, such that any action taken will lead to at least
one of the concerns being violated. For example, is it morally permitted to kill one
person in order to save five? These dilemmas have been used in philosophy to pit
Progress in Brain Research, Volume 202, ISSN 0079-6123, http://dx.doi.org/10.1016/B978-0-444-62604-2.00011-3
© 2013 Elsevier B.V. All rights reserved.
187
188 CHAPTER 11 Modes of thought and moral judgment

utilitarian and deontological concerns against each other (e.g., Thomson, 1976).
In broad terms, utilitarian judgments are those that maximize the overall benefits
and minimize the overall costs in a given situation (it is permissible to kill one to save
five), while deontological judgments favor individual rights and duties even if the con-
sequences are worse (killing one is not morally permitted, despite the loss of five).
In recent years, dual-process models, which argue that moral judgments are the
result of two competing neural systems, have become the focus of considerable re-
search. Although dual-process models have been prominent in social psychology for
quite some time (Chaiken and Trope, 1999), they were first proposed as a model of
moral judgment by Greene and colleagues (2001, 2004). According to the dual-
process model, certain moral dilemmas provoke responses from two separable,
though competing, psychological processes often referred to simply as System 1
and System 2 (Stanovich, 1999). System 1 is associated with fast, automatic,
affect-laden processing, with System 2 characterized by more conscious, deliberate,
and controlled reasoning.
Within the context of moral judgment, the dual-process account posits that, when
faced with certain moral dilemmas, two competing responses are generated: an af-
fective response from System 1 (in our example, an aversive reaction to the prospect
of taking a life), in addition to a deliberative cost-benefit response from System 2
(weighing the calculus of the lives lost and saved).1 This theory proposes that non-
utilitarian responses are generated when an aversive System 1 response overwhelms
cognitive mechanisms that allow for utilitarian-consistent reasoning. In contrast, di-
lemmas low in emotional salience are likely to provoke utilitarian judgments.
Initial support for the dual-process model came from fMRI studies (Greene et al.,
2001, 2004). Since then, a small number of behavioral studies have been used to sup-
port the model. Valdesolo and Desteno (2006) found that inducing positive mood
with a humorous video clip led to increased rates of utilitarianism in moral dilemmas.
This is argued as support for the dual-process model on the grounds that the positive
mood induction counteracted the aversive emotional response of System 1 and cre-
ated less conflict with the utilitarian System 2 response. In another study, Greene
et al. (2008) found that individuals took longer to make utilitarian, but not nonuti-
litarian, judgments while under cognitive load, suggesting that the cognitive pro-
cesses under load were causally relevant to utilitarian judgments in particular, as
predicted by dual-process theory.
Additional evidence for the dual-process model comes from a study examining
moral judgment in patients with lesions in ventromedial prefrontal cortex (VMPFC)
(Koenigs et al., 2007). Damage to this area is thought to lead to deficits in emotional
processing (Damasio, 1994), and the researchers found that the VMPFC patients

1
Although the dual-process model proposed by Greene and colleagues suggests competition between a
single affective response and a single cost-benefit response, this need not necessarily be the case. For
instance, in our example of considering the loss of one life as opposed to the loss of five lives, separate
affective and deliberative responses may be generated for each of these possibilities. See Moll et al.
(2008) for a detailed treatment of this important issue.
 2 Methods 189

were significantly more likely to make utilitarian moral judgments than controls.
Dual-process proponents have argued that this finding supports the model, as damp-
ened emotion processing should lead to less conflict between the two systems, thus
resulting in increased utilitarian judgments (Greene, 2007).
However, recent research on gender differences in moral judgment complicates
the dual-process model. Fumagalli and colleagues (2010) found that men made sig-
nificantly more utilitarian judgments than women, even after controlling for several
cultural and environmental factors. Additionally, Harenski et al. (2008) showed that,
while female moral judgments were predicted by activity in brain regions thought to
be particularly important for emotional processing, male moral judgments were pre-
dicted by activation in regions previously noted for involvement in recruiting cog-
nitive resources for difficult moral decisions (Greene et al., 2004).
These gender differences overlap considerably with the fMRI results of Greene
et al. (2001, 2004) and suggest that System 1 and System 2 may operate differently in
men and women. Men may rely relatively more on System 2, whereas System 1 may
be more prominent in women. This would explain the finding that males make sig-
nificantly more utilitarian judgments.
In this study, we address some still unanswered questions surrounding dual-
process theory and moral decision making. In particular, while support for the role
of System 1 has been relatively strong, evidence for System 2 processes has been less
obvious. Greene et al. (2008) found reaction time differences consistent with the
dual-process theory but were unable to show that manipulation of System 2 processes
actually altered the judgments themselves. Here, we primed participants toward
either an experiential (System 1) or analytic (System 2) mode of thought using a
method developed by Hsee and Rottenstreich (2004). We therefore predicted that,
overall, participants primed analytically would show increased rates of utilitarian
judgments relative to controls, while experientially primed individuals would show
decreased rates of utilitarianism. Additionally, and importantly, we further predicted
that men primed experientially would make significantly fewer utilitarian judgments
than controls but would not be affected by analytic priming, while women would
make significantly more utilitarian judgments when primed analytically but would
not be affected by experiential priming.

2 METHODS
2.1 Participants and procedure
One hundred and seventy-five participants (70% female, mean age ¼ 18.9,
SD ¼ 1.16) were recruited from the University of Arizona subject pool. All partici-
pants were undergraduates and received course credit for participating. Participants
were excluded if their average reaction time during either of two critical phases of the
experiment was greater than 2.5 standard deviations from the mean. This left a final
sample of 166 participants (70% female, mean age ¼ 18.84, SD ¼ 1.17), used for all
190 CHAPTER 11 Modes of thought and moral judgment

reported analyses. The experiment was presented to participants using E-Prime, a

computerized experimental design software package.
The experiment used a 3  3 mixed design, in which mode of thought was a
between-subject variable and dilemma type was a within-subject variable. Participants
were randomly assigned to one of three conditions: analytic mode of thought, experi-
ential mode of thought, or a control condition in which participants received no prim-
ing. All participants responded to a set of 30 dilemmas, comprising 10 personal, 10
impersonal, and 10 nonmoral dilemmas. Greene et al. (2001) proposed that moral
dilemmas can be usefully categorized as either personal or impersonal. Personal
dilemmas are those involving (a) serious bodily harm, (b) to one or more particular
individuals, where (c) this harm is not the result of deflecting an existing threat.
Dilemmas were categorized as impersonal if they did not meet these criteria. Personal
dilemmas then are those in which physical harm is brought to an individual as a means
of preventing some other harm, while impersonal dilemmas are those in which the
harm arises as an unintended consequence of an action. For example, an individual
choosing to push a stranger in the line of gunfire in order to save his daughter would
be an example of a personal dilemma, while choosing to pull one’s daughter out of the
way of a bullet, inadvertently leading to the bullet hitting a stranger standing behind her
would be an example of an impersonal dilemma (see Greene and Haidt, 2002 for a
discussion of the personal/impersonal distinction in moral judgment).

2.2 Mode of thought manipulation

To induce an analytic mode of thought, participants were asked to answer a short
math problem prior to the presentation of each dilemma (e.g., “If an object travels
at five feet per minute, then by your calculations how many feet will it travel in 360
seconds?”). For experiential priming, participants were shown a word before each
trial and asked to simply write down the feeling they most associate with this word
prior to the presentation of each dilemma (e.g., “When you hear the word crying what
do you feel? Please use one word to describe your predominant feeling”). All prim-
ing questions were presented to participants via computer and they responded in writ-
ing on an answer sheet (Hsee and Rottenstreich, 2004). Participants in the control
condition were presented with a screen showing a fixation cross for 10 s, in place
of any priming task.

2.3 Moral dilemma task

Thirty moral dilemmas were selected from a larger set of dilemmas used in several pre-
vious studies (Greene et al., 2001, 2004; Koenigs et al., 2007). Participants were pre-
sented with 10 personal, 10 impersonal, and 10 nonmoral dilemmas. In each
scenario, participants were faced with competing utilitarian and nonutilitarian interests
(though they were not labeled as such) and asked to make a judgment about what they
would do in each case. These judgments always came in the form of a binary yes/no
judgment, where a “yes” response corresponded to making a utilitarian judgment and
 2 Methods 191

a “no” response corresponded to a nonutilitarian judgment, a convention adopted to be

consistent with previous research (Greene et al., 2008; Koenigs et al., 2007) (see Fig. 1).

2.4 Analysis
Analyses were run using a 3 (prime type: analytic, experiential, control)  3
(dilemma type: personal, impersonal, nonmoral) mixed-factor analysis of covariance
(ANCOVA), with average reaction time to make a moral judgment included as a co-
variate. Reaction times were treated as a covariate because it has been suggested in
the literature that increased deliberation when making a moral judgment is likely to
be correlated with increased utilitarian judgments (Bartels, 2008). Rate of utilitarian
judgments was the primary dependent variable. Rates were computed by dividing the
total number of utilitarian judgments within a dilemma type by the total number of
dilemmas within that type.
Planned contrasts were conducted for both prime type and dilemma type. Two
contrasts were planned for prime type such that both experimental groups (analytic
and experiential) were independently compared to the control group. Two contrasts

Participant makes
judgment

¥ time ¥ time ¥ time 5s

If an object travels at You are at the wheel of a runaway Would you hit the
+
5 feet per minute, then trolley quickly approaching a fork in switch in order to
by your calculations the tracks. On the tracks extending to avoid the deaths
how many feet will it the left is a group of five railway of the five
travel in 360 s? workmen. On the tracks extending to workmen?
the right is a single railway workman.
If you do nothing the trolley will
proceed to the left, causing the deaths
of the five workmen. The only way to
avoid the deaths of these workmen is to
hit a switch on your dashboard that will
cause the trolley to proceed to the right,
causing the death of the single
workman.

Reveal prime Reveal dilemma Reveal dilemma Reveal fixation

question question
FIGURE 1
Timeline for a single trial in the analytic mode-of-thought condition. Each trial began with the
presentation of a prime question, which the participant responded to on an accompanying
paper answer sheet. After answering the prime question, the participant was presented with a
moral dilemma. Next the participant was presented with a question about the dilemma and
made a judgment about what they would do in the dilemma. The participant made her
judgment by pressing either “Q” for yes or “P” for no on the keyboard. After making a
judgment, a fixation point was presented for 5 s. The complete task consisted of 30 trials, with
order for both prime question and dilemma randomized across participants.
192 CHAPTER 11 Modes of thought and moral judgment

were also run for each dilemma type. The first contrast combined both personal and
impersonal moral dilemmas and compared them to nonmoral dilemmas. The second
contrast tested for differences between personal and impersonal moral dilemmas. All
analyses were conducted using SPSS (version 17) statistical software.

3 RESULTS
The analysis revealed a significant main effect of mode of thought (F(2, 160) ¼
3.227, p ¼ 0.042) as well as the predicted effect of the reaction time covariate
(F(1160) ¼ 4.479, p ¼ 0.036, see Fig. 2). Planned contrasts revealed significantly in-
creased rates of utilitarianism in the analytic group (M ¼ 0.63, SE ¼ 0.014) compared
to the control group (M ¼ 0.58, SE ¼ 0.014), p ¼ 0.048. No significant difference was
found between the experiential (M ¼ 0.58, SE ¼ 0.015) and control groups. The anal-
ysis also revealed a significant main effect of dilemma type (F(2, 320) ¼ 4.129,
p ¼ 0.017). Planned contrasts revealed a significant difference between moral
(M ¼ 0.556, SE ¼ 0.013) and nonmoral dilemmas (M ¼ 0.688, SE ¼ 0.007),
p ¼ 0.015. No significant difference was found between personal and impersonal di-
lemmas. No significant interaction between prime type and dilemma type was found.
Additional analyses were conducted to examine gender differences on the effect of
mode of thought on moral judgment. First, we examined whether males and females in
the control condition were significantly different from each other in their judgments.
A 2 (gender: male, female)  3 (dilemma type: personal, impersonal, nonmoral)
ANCOVA with average reaction time to make moral judgments included as a covar-
iate was used to test for gender differences. This analysis revealed a significant gender
difference (F(1, 56) ¼ 4.456, p ¼ 0.039) with males (M ¼ 0.637, SE ¼ 0.028) demon-
strating higher rates of utilitarianism than females (M ¼ 0.0.568, SE ¼ 0.016). No
interaction between gender and dilemma type was found.

0.75

0.7 *

0.65
Mean

0.6

0.55

0.5
Analytic Control Experiential
FIGURE 2
Mean rates of utilitarianism by mode-of-thought condition across all dilemma types. Error
bars represent standard error of the mean.
 4 Discussion 193

To explore this gender difference further, we ran two additional 3(prime type) 
x 3(dilemma type) ANCOVA analyses, one restricted to male participants and one
restricted to female participants, again including average reaction time to respond to
dilemmas as a covariate. Among male participants, there was a significant effect of
mode of thought (F(2, 48) ¼ 3.787, p ¼ 0.03) and the reaction time covariate was sig-
nificant (F(1, 48) ¼ 7.766, p < 0.008). Simple contrasts revealed that the experiential
group (M ¼ 0.558, SE ¼ 0.026) gave significantly less utilitarian responses than the
control group (M ¼ 0.634, SE ¼ 0.027), p ¼ 0.045. No significant effect was found
between analytic (M ¼ 0.654, SE ¼ 0.025) and control groups. No interaction effect
was found between prime type and dilemma type (see Fig. 3).
Among female participants, there was no significant effect of priming. However,
planned simple contrasts revealed a trend toward significance (p ¼ 0.054) between
analytic (M ¼ 0.614, SE ¼ 0.017) and control (M ¼ 0.569, SE ¼ 0.016) groups, with
analytic participants providing higher rates of utilitarian judgments than controls.
No such trend was observed between the experiential (M ¼ 0.589, SE ¼ 0.017) and
control groups (see Fig. 3).

4 DISCUSSION
Priming individuals toward an analytic mode of thought had a significant effect upon
moral judgments. More specifically, individuals in an analytic mode were signifi-
cantly more likely to make utilitarian judgments than individuals who were not
primed. This finding is particularly important because it is the first time, to our

0.8
0.75
0.7
0.65
Mean

0.6
0.55
0.5
0.45
0.4
Personal Impersonal Nonmoral Personal Impersonal Nonmoral

Male Female

Analytic Experiential Control

FIGURE 3
Rates of utilitarianism for males and females by mode-of-thought and dilemma-type
conditions. Error bars represent standard error of the mean.
194 CHAPTER 11 Modes of thought and moral judgment

knowledge, that a behavioral manipulation of System 2 processes has resulted in a

significant change in the rate of utilitarian judgments made in a moral dilemma task.
Previous research aimed at manipulating System 2 (Greene et al., 2008) found reac-
tion time differences consistent with the dual-process model but failed to show an
effect in the actual judgments made.
Our findings are also consistent with those of Koenigs et al. (2007), who found
that VMPFC lesion patients showed increased rates of utilitarian judgments. In that
study, the increased rates of utilitarianism were explained in terms of decreased input
from System 1, thought to be underlain in part by activity in VMPFC, which allowed
System 2 processes to be more heavily weighted in the patients’ moral judgments.
This study provides an important counterpoint to these results by showing that rates
of utilitarian judgment can also be increased by priming System 2 and additionally
that these effects can be demonstrated in healthy participants.
As noted previously, Harenski et al. (2008) provide support for the notion that,
while females rely more on emotional processing when making moral judgments,
males make preferential use of deliberative processes. One possible interpretation
of the gender differences seen in our data is that males may already be at, or at least
near, ceiling for any effects that can be induced via a subtle analytic prime such as the
one used here. This could explain why experiential, but not analytic, priming is ef-
fective in males. Similarly, females may be near ceiling for any effects likely to result
from the experiential primes used in this study. This interpretation fits the dual-
process model but suggests that moral judgment in females depends more on System 1
processes, while moral judgment in males is more influenced by System 2 proces-
sing. This interpretation also fits with the finding that males in general tended to
be more utilitarian than females in the control condition. This is, of course, just
one interpretation of the results and can be seen as potentially reinforcing negative
gender stereotypes. Because of this, these results should be considered with care.
Further studies are needed to clarify the precise role of gender differences within
a dual-process framework.
A caveat worth mentioning in these findings is that we did not observe an inter-
action between mode of thought and dilemma type, as would be predicted by previ-
ous work (Greene et al., 2001, 2004). Although moral dilemmas, particularly
personal moral dilemmas, showed greater differences between mode-of-thought
groups than nonmoral dilemmas, these differences did not result in a statistically sig-
nificant interaction. This lack of an interaction is perhaps not as surprising, given that
the priming manipulation used demonstrates effects in a wide variety of decision-
making domains. It does, however, raise the noteworthy question as to the extent
to which the processes underlying moral judgment are domain general or domain
specific. Further research aimed at investigating this question would provide impor-
tant new directions for the study of moral judgment.
In conclusion, this study provides the first evidence that manipulation of System 2
processes actually alter moral judgments, as predicted by the dual-process theory of
moral judgment. Participants who were primed to be in an analytic mode of thought
made significantly more utilitarian judgments than control participants. This is an
 References 195

important piece of new evidence in understanding moral decision making. Although

much support has been shown for the role of emotions in moral judgment (Valdesolo
and Desteno, 2006), and much of this support has been interpreted as consistent with
a dual-process system of moral judgment, there has been a surprising lack of evi-
dence for the deliberative System 2. Without data such as that presented here, it
is difficult, if not impossible, to distinguish between the dual-process model and
alternative models such as Haidt’s Social Intuitionism model of moral judgment
(Haidt, 2001). It is only through conclusive evidence that manipulation of both Sys-
tem 1 and System 2 alters moral judgment as predicted that the dual-process model
can be considered as a plausible account of the psychology of moral judgment.
Additionally, our data suggest that gender plays an important role in moral judg-
ment. This idea has a long history in moral psychology (Gilligan, 1977) but has
received relatively little attention in contemporary research (Fumagalli et al.,
2010, Harenski et al., 2008). Our data suggest that investigating gender differences
in moral judgment within the context of dual-process theory is an exciting and prom-
ising direction for future research, although one that should be approached carefully
so as not to reinforce unwarranted gender stereotypes.

References
Bartels, D.M., 2008. Principled moral sentiment and the flexibility of moral judgments and
decision making. Cognition 108, 381–417.
Chaiken, S., Trope, Y., 1999. Dual-Process Theories in Social Psychology. Guilford Press,
New York.
Damasio, A., 1994. Descartes’ Error: Emotion, Reason, and the Human Brain. Gosset/Putnam
Press, New York, NY.
Fumagalli, M., Ferrucci, R., Mameli, F., Marceglia, S., Mrakic-Sposta, S., et al., 2010. Gender-
related differences in moral judgment. Cogn. Process. 11, 219–226. http://dx.doi.org/
10.1007/s10339-009-0335-2.
Gilligan, C., 1977. In a different voice: women’s conceptions of self and of morality. Harv.
Educ. Rev. 47, 481–517.
Greene, J.D., 2007. Why are VMPFC patients more utilitarian?: a dual-process theory of moral
judgment explains. Trends Cogn. Sci. 11 (8), 322–323.
Greene, J., Haidt, J., 2002. How (and where) does moral judgment work? Trends Cogn. Sci.
6 (12), 517–523.
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, 2105–2108.
Greene, J.D., Nystrom, L.E., Engell, A.D., Darley, J.M., Cohen, J.D., 2004. The neural bases
of cognitive conflict and control in moral judgment. Neuron 44, 389–400.
Greene, J.D., Morelli, S.A., Lowenberg, K., Nystrom, L.E., Cohen, J.D., 2008. Cognitive load
selectively interferes with utilitarian moral judgment. Cognition 107, 1144–1154.
Haidt, J., 2001. The emotional dog and its rational tail: a social intuitionist approach to moral
judgment. Psychol. Rev. 108, 814–834.
Harenski, C., Antonenko, O., Shane, M., Kiehl, K., 2008. Gender differences in neural mech-
anisms underlying moral sensitivity. Soc. Cogn. Affect. Neurosci. 3, 313–321.
196 CHAPTER 11 Modes of thought and moral judgment

Hsee, C.K., Rottenstreich, Y., 2004. Music, pandas and muggers: on the affective psychology
of value. J. Exp. Psychol. Gen. 133, 23–30.
Koenigs, M., Young, L., Adolphs, R., Tranel, D., Cushman, F., et al., 2007. Damage to the
prefrontal cortex increases utilitarian moral judgments. Nature 446 (7138), 908–911.
Kohlberg, L., 1981. Essays on Moral Development, Volume 1: The Philosophy of Moral
Development. Harper Row, New York.
Moll, J., De Oliveira-Souza, R., Zahn, R., 2008. The neural basis of moral cognition: senti-
ments, concepts, and values. Ann. N. Y. Acad. Sci. 1124, 161–180.
Piaget, J., 1932. The moral judgment of the child. Harcourt, Brace Jovanovich, Oxford.
Stanovich, K.E., 1999. Who is Rational? Studies of Individual Differences in Reasoning.
Erlbaum, Mahwah, NJ.
Thomson, J., 1976. Killing, letting die, and the trolley problem. Monist 59, 204.
Valdesolo, P., DeSteno, D., 2006. Manipulations of emotional context shape moral judgment.
Psychol. Sci. 17, 476.