PreprintPDF Available

Justice Before Expediency: Robust Intuitive Concern for Rights Protection in Criminalization Decisions

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

Abstract and Figures

Is it better “that ten guilty persons escape than that one innocent suffer”? The notion that a false conviction is worse than a false acquittal is a deep-seated commitment in the theory of criminal law. In three experiments (total N = 1551), we examine this question through the phenomenon of proxy crime. Our studies manipulated the rate at which a series of hypothetical proxy crimes lead to falsely convicting the innocent and falsely acquitting those guilty of a target wrongdoing. Are people’s evaluations of such policies influenced by the proxy conduct’s specificity and/or sensitivity in predicting actual crime? Across several experimental conditions, participants’ approval of proxy criminalization depended primarily on its rate of false conviction, with false acquittal exerting a negligible effect. Furthermore, this tendency was unrelated to participants’ professed views about the purpose of criminal punishment, or with their probability literacy as measured on a prior training task. Though false acquittal rates had no effect on decisions to approve or reject a proposed proxy crime, when asked to justify their decisions, participants retrospectively supported their judgments by highlighting the proxy crime’s efficacy (or inefficacy) in combating crime. Taken together, these results reveal that people favor criminal policies that protect the rights of the innocent, but report comparable concern for their expediency in fighting crime. In this regard, our study documented a dissociation between the factors that shape intuitive attitudes toward criminalization and the reasons that people subsequently provide.
Content may be subject to copyright.
JUSTICE BEFORE EXPEDIENCY1
Justice Before Expediency:
Robust Intuitive Concern for Rights Protection in Criminalization Decisions
Abstract
Is it better “that ten guilty persons escape than that one innocent suffer”? The notion
that a false conviction is worse than a false acquittal is a deep-seated commitment in the theory
of criminal law. In three experiments (total N = 1551), we examine this question through the
phenomenon of proxy crime. Our studies manipulated the rate at which a series of hypothetical
proxy crimes lead to falsely convicting the innocent and falsely acquitting those guilty of a
target wrongdoing. Are people’s evaluations of such policies influenced by the proxy conduct’s
specificity and/or sensitivity in predicting actual crime? Across several experimental
conditions, participants’ approval of proxy criminalization depended primarily on its rate of
false conviction, with false acquittal exerting a negligible effect. Furthermore, this tendency
was unrelated to participants’ professed views about the purpose of criminal punishment, or
with their probability literacy as measured on a prior training task. Though false acquittal rates
had no effect on decisions to approve or reject a proposed proxy crime, when asked to justify
their decisions, participants retrospectively supported their judgments by highlighting the
proxy crime’s efficacy (or inefficacy) in combating crime. Taken together, these results reveal
that people favor criminal policies that protect the rights of the innocent, but report comparable
concern for their expediency in fighting crime. In this regard, our study documented a
dissociation between the factors that shape intuitive attitudes toward criminalization and the
reasons that people subsequently provide.
JUSTICE BEFORE EXPEDIENCY2
1. Introduction
It was a felony in 18th century England for a mother to conceal the birth of a child, as
such behavior indicated that she intended to kill, or had already killed, the child (Bentham,
1887). Similarly, many countries now punish citizens for merely traveling to areas under the
control of terrorist groups, on the presumption that these individuals are involved in acts of
terrorism (De Guttry et al., 2016).
Why criminalize these proxy conducts that are only probabilistically linked to criminal
acts, and not wrongful in themselves? One reason is that actual criminal misconduct can be
hard for law enforcement to observesince offenders are motivated to avoid punishment by
concealing their wrongdoing. Second, even when criminal conduct could in principle be
observed, it may be prohibitively costly (or even impossible) to prove it beyond a reasonable
doubt in order to secure criminal conviction. Proxy crimes offer a practical solution to this
problem.
On a theoretical plane, proxy crimes offer a window into the motives behind criminal
policy. A proxy crime targets a primary wrongdoing indirectly, hence it is both over-inclusive
(as some of the criminalized proxy conduct is not associated with the primary wrongoing) and
under-inclusive (as some instances of the primary wrongdoing are not associated with the
criminalized proxy conduct). The scope of that over- and under-inclusion is arguably the main
factor that should determine the normative assessment of a given proxy offense. On one hand,
proxy crimes can be evaluated on the basis of their effectiveness in curtailing crime: that is,
proxy offenses are acceptable insofar as they maximize sensitivity (i.e., they help convict many
actual criminals) without incurring too much costs on the innocent. This criterion is prominent
among consequentialist approaches to punishment and accompanies permissive attitudes
toward proxy crimes in the jurisprudential literature (Bystranowski & Mungan, 2022; Schauer,
2003, 2011; Teichman, 2017).
JUSTICE BEFORE EXPEDIENCY3
On the other hand, proxy crimes can be evaluated on the basis of their justice, or respect
for the innocent’s rights: that is, proxy crimes are acceptable insofar as they maximize
specificity and refrain from convicting citizens who are innocent of the primary wrongdoing.
These considerations are essential among non-consequentialist schools of thought, according
to which punishment is only justified by the defendant’s criminal actus reus (Duff et al., 2007;
Picinali, 2018)an act which cannot be replaced by mere correlates of crime (Alexander &
Ferzan, 2009; Duff, 1997). Doing so interferes with fundamental principles of retributivist
criminal justice, such as the presumption of innocence (Tomlin, 2013).
Research on punishment (Carlsmith, 2008; Carlsmith et al., 2002; Darley et al., 2000;
McFatter, 1982; Sharp & Otto, 1910) has already demonstrated that non-consequentialisti.e.,
primarily retributivistmotives guide people’s sentencing recommendations. Furthermore,
despite evincing retributivist motives, people decry retributivist rationales (aimed at punishing
the guilty for their past wrongs) and gravitate toward consequentialist justifications for
punishment (aimed at deterring future crime) instead (Carlsmith, 2008). This discrepancy
between participants’ tacit judgments and their professed convictions has been documented
pervasively in previous studies on moral cognition (Haidt, 2001).
In the present work, we examine the principles that guide people’s decisions about
criminalizationthat is, about whether various criminal policies should be adopted in the first
place. To pursue this question, we focus on proxy crimes (Bystranowski & Mungan, 2022;
McAdams, 2005), and evaluate the extent to which factors crucial to expediency (i.e.,
apprehending the guilty) and justice (i.e., sparing the innocent) shape people’s support for
proxy criminal policies. For instance, does people’s support for the aforementioned travel ban
depend on whether it helps to apprehend terrorists or on whether it preserves innocent travelers’
right to transit?
JUSTICE BEFORE EXPEDIENCY4
From the consequentialist perspective, the decision should rely on a calculus integrating
both the ban’s capacity to hinder terrorist operations and deter or intercept the terrorists
themselves, as well as its detriment to well-meaning travelers. If the former exceeds the latter,
there is reason to support the proxy crimeotherwise not (Teichman, 2017).
In contrast, non-consequentialism (Duff et al., 2007) argues that trade-offs between
expediency and protecting rights are fundamentally misguided. From a non-consequentialist
perspective, convicting the innocent violates a fundamental tenet of criminal law, and is
therefore wrong even if doing so would come with enormous benefits for a law’s expediency—
and, in turn, for social welfare. If innocent travelers are subject to criminal conviction, there is
reason to oppose the proxy crime–and only if they are not, could the proxy crime’s efficacy
constitute a reason in favor of its introduction. As such, the retributivist answer depends almost
solely on the ban’s detriment to the welfare of innocent travelers–and weakly, if at all, on its
value as a counterterrorist measure.
In our studies, we vary the rates of false conviction and false acquittal and investigate
whether participants’ attitudes toward proxy criminal statutes are influenced by these
manipulations. If participants display consequentialist attitudes toward criminalization, their
evaluation of proxy crimes ought to depend on both the false conviction and false acquittal
rates. If instead people demonstrate non-consequentialist principles, their approval of proxy
crimes ought to depend primarily on the rate of false conviction.
Whether such intuitive principles are introspectively recognized and expressly
endorsed are a separate question. Growing evidence under the guise of social intuitionism
(Haidt, 2001) highlights how the eliciting factors that shape people’s intuitive judgments bear
little relation to the justifying reasons people subsequently provide (Carlsmith et al., 2002;
Almagro et al., 2021). In the present research, we explore the possibility that attitudes toward
JUSTICE BEFORE EXPEDIENCY5
criminalization exhibit a similar discrepancy between the factors that shape intuitive
preferences regarding criminalization, and the reasoning that people provide post hoc.
Data, analysis scripts and materials are openly accessible on the Open Science
Framework at: https://osf.io/tn6j8/?view_only=ae18f93757ad4f68af55705e7803b773.
Statistical analyses were conducted in R version 4.1.2 using the lme4 and emmeans packages.
Our primary analysis are mixed-effects models with random intercepts of participant and
scenario, for which we report F tests using the Kenward-Roger approximation to the degrees
of freedom.
2. Study 1
In Study 1, we investigate the eliciting factors that shape people’s attitudes toward
proxy criminal policy. Participants evaluated a series of proposed proxy crimes devised in order
to address a target wrongdoing (see Table 1). We manipulated the rates of false conviction and
false acquittal for each proxy criminal policy, and evaluated their impact on people’s
evaluations. In addition, we examined people’s justifications for their evaluations of each proxy
crime through multi-item measures.
We ran Study 1 on two occasions, and here we report the results of the exploratory and
confirmatory runs of this experiment together. For the sake of concision, we employ an ad hoc
method to report both results consecutively. In braces, we report a series of two values: The
first corresponds to the exploratory result, and the second to the confirmatory result (e.g., p =
{.048, .198}). We also report 95% confidence intervals around the pooled estimate.
2.1. Methods
2.1.1. Participants
Data were collected through the Prolific Academic crowdsourcing service in May 2020
and March 2021. Participants were (i) native English speakers with (ii) at least a high
school/secondary education. For each iteration of the study, we recruited 400 participants
JUSTICE BEFORE EXPEDIENCY6
({229, 269} women, age: M = {34.3, 34.2} years, SD = {12.7, 13.2}). We did not apply any
exclusion criteria or exclude any observations.
The target sample size was decided on the basis of funding availability. A sample size
of 400 participants afforded excellent statistical power (95%) to detect a small-to-medium
correlation (r = .18; smaller than the average effect in social psychology, see Richard, Bond
Jr., & Stokes-Zoota, 2003) when α equals .05. Thus, our studies also had very good statistical
power (.952; or 90%) to reliably detect small-to-medium effects twice, i.e., in both exploratory
and confirmatory studies.
2.1.2. Materials
Participants viewed four short vignettes in a random order. The introduction to each
vignette described an unlawful activity occurring in a hypothetical community:
A dangerous cult is gaining popularity in the country of Zubrovka. Most of its followers
have settled outside state control, in the Western Desert, where they fight the
government army and from where they send assassins that spread fear across the
country. Although involvement in the illegal activities of the cult is a serious crime,
growing numbers of young people are joining the cult, as it is almost impossible to
prove that they are involved in criminal wrongdoing in the Western Desert.
Next, participants learned that lawmakers were considering whether to pass a proxy
criminal policy:
The Parliament of Zubrovka considers passing a new offense to make prosecution more
effective: The Anti-Cult Act. The Anti-Cult Act would make it a crime to enter Zubrovka
from the Western Desert. These days, most people leaving the Western Desert have
been involved in the illegal activities of the cult. Zubrovka used to be a popular getaway
for residents of the Western Desert, but travel in the region has been steadily declining
JUSTICE BEFORE EXPEDIENCY7
due to the risk posed by the cult. So, anyone traveling to Zubrovka from the Western
Desert is likely a member of the cult.
The proxy conduct either causally preceded the target offense (i.e., traveling to the
Western Desert) or resulted from the target offense (i.e., traveling from the Western Desert).
Table 1 lists each target offense and both corresponding proxy behaviors.
Table 1.
Stimuli: Primary offenses and corresponding proxy behaviors.
Target
offense
Precursor
Consequent
Poaching
Possession of swift bullets
Selling bear skin
Doping
Possession of a syringe
Taking a drug that masks the presence of
the illicit substance
Fraud
Possession of a marked deck of
cards
Possession of over $500 in cash while
being a vagrant
Terrorism
Possession of a bomb-
construction manual
Wearing head-covering in public after an
explosion
Cult *
Travelling to a terrorist-
controlled territory
Travelling from a terrorist-controlled
territory
Note. *: The Cult scenario was only employed in Study 2.
Note. *: The Cult scenario was only employed in Study 2.
Participants then learned of the proxy criminal policy’s specificity and sensitivity in
predicting the target wrongdoing, as well as the base rate of the target wrongdoing in the
community. These values were displayed visually in the form of a dot plot (see Figure 1),
purportedly part of a report commissioned by the lawmakers to learn more about the unlawful
activity. The visual display was presented on a separate screen, and participants could not
advance to the following section until 8 seconds had elapsed. In the visual display, criminals
JUSTICE BEFORE EXPEDIENCY8
and innocents were lateralized, as in Figure 1, and engaging in versus refraining from the proxy
conduct was color-coded.
Figure 1. Sample Display.
To generate these frequencies, we defined three parameters:
1) the base rate: the proportion of the entire population that engages in the primary
wrongdoing,
2) the false negative or false acquittal rate: the proportion of criminals who do not
engage in the proxy conduct, and
3) the false positive or false conviction rate: the proportion of innocents who do
engage in the proxy conduct.
These parameters were randomly sampled on each trial from uniform distributions. The
base rate was allowed to range from 10% to 30% of the entire population. The false acquittal
and false conviction rates were allowed to range between 1% and 19% (of the criminals and
innocents respectively). In other words, the frequency of false acquittals would equal
JUSTICE BEFORE EXPEDIENCY9
population size (which was held constant at 526) × base rate × false negative rate, rounded to
the nearest integer.
2.1.3. Procedure
After providing informed consent to take part in the study, participants completed the
Training Task, the Experimental Block, and the Individual Difference Measures in a fixed
order.
Training Task. Participants first saw two example displays showing the prevalence of
an undesirable activity (e.g., alcohol abuse) and its associated “symptom” in a community (e.g.,
having bloodshot eyes). They were then asked five true or false questions, each of which
required an inference from the example displays (e.g., ‘A person with bloodshot eyes is more
likely than not to abuse alcohol’). The purpose of this task was for participants to familiarize
with the visual display and practice their interpretation. We calculated the sum score of correct
responses on the training task (Ms = {3.43, 3.28}, SDs = {1.14, 1.19}) as an individual
difference measure of probability literacy.
Experimental Block. In a balanced incomplete block design, each participant was
assigned to a block in which they viewed two precursor and two consequent proxy criminal
policies. In each block, participants viewed either the Precursor or Consequent variant of each
of four offenses and judged the battery of four cases in a randomized order. In Appendix 1, we
report the effect of the precursor versus consequent manipulation on participants’ responses.
Each case was composed of a sequence of slides: the vignette (describing the primary
offense and the proposed proxy crime), the visual display, followed by the decision, and
justification slides.
On the decision slide, participants were asked whether the proxy criminal policy should
be passed on a 7-point Likert scale, anchored at 1: ‘Certainly not’ and 7: ‘Certainly’. Then, on
the justification slide, participants evaluated the proxy crime’s effectiveness (whether the proxy
JUSTICE BEFORE EXPEDIENCY10
crime helps apprehend offenders, prevents wrongdoing, and deters potential offenders) and
justice (whether it punishes/affects the lives of/is unfair toward innocent civilians) through a
total of six items. Each item was assessed on a 7-point agreement scale, anchored at 1: ‘Strongly
disagree’ and 7: ‘Strongly agree’.
Individual Difference Measures. After the experimental portion of the study,
participants completed two individual difference measures: the 6-item Punitiveness
Questionnaire (Kemme et al., 2014), and an adaptation of the retribution/deterrence motives
task introduced by Carlsmith and colleagues (2002). Finally, participants optionally provided
demographic information, such as their gender, age, educational attainment, religiosity, and
political orientation.
2.1.4. Measures
For each scenario and participant, we recorded the following three independent
measures: crime rate, false conviction, and false acquittal. The main response measure on each
trial was approval: whether or not the proxy crime should be adopted (7-point Likert scale).
Participants also completed a six-item assessment of the effectiveness and justice of
each proxy crime. We submitted these items to a maximum-likelihood factor analysis with
varimax rotation. Two factors exhibited SS loadings above one ({2.17, 2.05}, and {2.11, 2.04},
with the 3rd factor dropping to {0.20, 0.62}), and explained 71% and 64% of the variance in
evaluations, in the exploratory and confirmatory studies respectively. Every item loaded onto
a single factor, as hypothesized, with factor loadings > {.76, .72} (and failed to load onto the
other factor, with the absolute value of every factor loading < {.29, .25}). We therefore
calculated two additional measures per trial:
1. effectiveness, the three-item average (Cronbach’s α = {.89, .87}) of whether the proxy
crime effectively combats crime; and
JUSTICE BEFORE EXPEDIENCY11
2. justice, the three-item average (Cronbach’s α = {.87, .85}) of whether the proxy crime
safeguards the rights of innocent citizens.
2.2. Results
Summary statistics for approval, justice, and effectiveness are reported in Table 2. In
the aggregate, participants tended to approve of the proxy crimes (M = {4.94, 4.89}, Pooled
95% CI [4.53, 5.29]), and view them as effective (M = {5.11, 4.99}, Pooled 95% CI [4.79,
5.30]), but were divided as to whether they were just (M = {3.86, 3.84}, Pooled 95% CI [3.56,
4.14]).
Table 2.
Descriptive Statistics: Study 1.
Scenario
N
Approval
Justice
Effectiveness
M
(SD)
M
(SD)
M
(SD)
M
(SD)
M
(SD)
M
(SD)
Poaching
202
198
5.42
(1.46)
5.41
(1.63)
4.23
(1.30)
4.33
(1.31)
5.32
(1.14)
5.32
(1.17)
198
202
5.47
(1.63)
5.65
(1.50)
4.24
(1.46)
4.33
(1.34)
5.51
(1.24)
5.41
(1.15)
Doping
197
196
4.52
(1.70)
4.40
(1.70)
3.34
(1.51)
3.33
(1.41)
5.24
(1.15)
4.89
(1.30)
203
204
5.00
(1.55)
4.71
(1.60)
3.99
(1.59)
3.56
(1.40)
5.35
(1.15)
5.24
(1.20)
Fraud
198
202
4.91
(1.74)
4.90
(1.77)
3.86
(1.45)
4.00
(1.44)
5.07
(1.35)
5.07
(1.23)
202
198
4.13
(1.69)
4.17
(1.73)
3.36
(1.37)
3.34
(1.27)
4.64
(1.31)
4.66
(1.17)
JUSTICE BEFORE EXPEDIENCY12
Terrorism
203
204
5.50
(1.44)
5.34
(1.53)
4.25
(1.53)
4.16
(1.46)
5.27
(1.25)
4.91
(1.36)
197
196
4.56
(1.74)
4.52
(1.83)
3.61
(1.55)
3.67
(1.45)
4.47
(1.48)
4.42
(1.51)
In the analyses below, we examine whether participants’ approval judgments were
shaped by rates of false conviction and/or acquittal (while accounting for variation in the base
rate of crime). To this end, we enter approval ratings as the dependent measure in a mixed-
effects model, with random effects of participant and scenario, and rates of false conviction,
false acquittal, and the base rate as fixed effects. We then repeat this procedure with the
secondary dependent measures of perceived effectiveness and justice.
2.2.1. Effects of false conviction and acquittal
Approval of a proxy crime depended on the rate of false convictions (B = {-
5.06, -4.23}, Pooled 95% CI [-5.63, -3.69], t = {-7.32, -5.92}, ps < .001), but not on the rate of
false acquittals (B = {-0.43, 0.96}, Pooled 95% CI [-0.75, 1.18], t = {-0.62, 1.35}, p = {.53,
.18}), or on the base rate of crime (B ={1.16, -0.11}, Pooled 95% CI [-0.36, 1.42], t = {1.86, -
0.17}, p = {.06, .87}). See Figure 2.
JUSTICE BEFORE EXPEDIENCY13
Figure 2. Study 1: Standardized (i.e., z-scored) evaluations of the proxy criminal
policies by false conviction and acquittal rates, and by the base rate of crime. Each plot displays
(locally estimated) curve fit and mean values for each study (exploratory: crosses;
confirmatory: circles) in percent increments.
2.2.2. Effectiveness and justice evaluations
A corresponding model of justice evaluations indicated that false convictions
negatively predicted justice, B = {-7.19, -6.97}, Pooled 95% CI [-7.91, -6.30], t = {-12.20, -
12.09}, ps < .001. Meanwhile, the rate of false acquittals (B = {-0.40, 1.26}, Pooled 95% CI [-
0.41, 1.19], t = {-0.67, 2.21}, p = {.50, .027}) and variation in the base rate (B = {0.73, 0.79},
Pooled 95% CI [-0.00, 1.48], t = {1.36, 1.47}, p = {.17, .14}) had no impact on perceptions of
JUSTICE BEFORE EXPEDIENCY14
justice. Thus, the more a proxy crime falsely convicted innocent civilians, the more likely
participants were to view it as unjust.
Next, we examined the determinants of proxy crimes’ perceived effectiveness.
Effectiveness did not depend on the rate of false acquittals, B = {-0.98, -0.31}, Pooled 95% CI
[-1.38, 0.05], t = {-1.90, -0.60}, p = {.053, .55} or on the base rate of crime, B = {0.73, -0.15},
Pooled 95% CI [-0.39, 0.94], t = {1.57, -0.32}, p = {.12, .75}, as might have been expected.
Instead, reported effectiveness depended on the rate of false convictions, B = {-2.76, -2.82},
Pooled 95% CI [-3.53, -2.09], t = {-5.33, -5.36}, ps < .001. In other words, proxy crimes were
perceived as effective in apprehending criminals not when few criminals were acquitted, but
when few innocent civilians were convicted.
We therefore reasoned that reports of effectiveness served to justify participants’
decisions (i.e., to approve or reject the proxy criminal policy). If so, the influence of false
convictions on effectiveness may reflect an indirect effect via approval. Indeed, approval and
effectiveness were strongly related, B = {0.49, 0.41}, Pooled 95% CI [0.43, 0.47], t = {33.92,
26.27}, ps < .001. Furthermore, entering approval as an additional predictor in the model of
effectiveness substantially weakened the influence of false convictions (B = {-0.31, -1.14},
Pooled 95% CI [-1.36, -0.16], t = {-0.76, -2.51}, p = {.45, .012}).
2.3. Discussion
Study 1 provided evidence that false convictions shape attitudes toward criminal
policywhereas false acquittals have a limited effect. Participants were greatly concerned with
the risk of convicting innocent citizens, but inattentive to the risk of acquitting criminals.
Despite this fact, when asked to provide reasons in support of their prior decision,
supporters claimed that the policies were effective in combating crime while detractors claimed
that they were ineffective. As such, the perceived effectiveness of proxy crime was unrelated
to the proportion of criminals which it helped to apprehend, and depended primarily on the rate
JUSTICE BEFORE EXPEDIENCY15
at which innocent citizens were falsely convicted. In this regard, Study 1 points toward a
dissociation between the factors that shape participants’ judgments and their subsequent
justifications: Considerations of justice strongly impacted decisions regarding proxy criminal
statutes, whereas considerations of effectiveness were brought to bear to justify them.
3. Study 2
Study 1 documented a dominant role of false convictions in shaping
criminalization preferences. In Study 2, we examine the cognitive basis of this effect through
a manipulation of time pressure. First, we ask whether participants manifest an emphasis on
false conviction even when faced with limited time to decide. Such a result would suggest that
the non-consequentialist emphasis on false conviction (and neglect of false acquittal) is
relatively intuitive and automatic.
Second, we investigate whether reflection inductions weaken the effect of false
conviction and/or magnify the effect of false acquittal through various comparisons: (i)
between decisions under time pressure versus after a forced delay, and (ii) when evaluating the
proxy crime’s effectiveness and justice prior to issuing their judgment. Furthermore, we also
evaluate the impact of (iii) information presented in visual or numerical formon the
assumption that a numerical presentation would facilitate participants’ inferences about the
probative value of each proxy crime.
Finally, in Study 2 we addressed certain limitations in our design of Study 1:
Specifically, the visual display in Study 1 used a fixed lateralization and color scheme, which
could have influenced participants’ allocation of attention. In Study 2, we randomized the
lateralization and regimented the color scheme to rule out these effects.
JUSTICE BEFORE EXPEDIENCY16
3.1. Methods
3.1.1. Participants
751 participants (373 women, age: M = 36.2 years, SD = 12.4) were recruited through
the Prolific Academic crowdsourcing service in November 2021. Participants were (i) native
English speakers with (ii) at least a high school/secondary education. The target sample size
was decided on the basis of a rule of thumb: We sought to recruit at least 120 participants per
condition in a 3 x 2 between-subjects design (i.e., 720).
3.1.2. Materials
Participants viewed five short vignettes (see Table 1) in a random order. Each
participant was presented with two or three vignettes in the Precursor version, with the
remaining vignettes in the Consequent version.
As in Study 1, participants received information about the probative value of each proxy
conduct. In Study 2, we manipulated whether the information was displayed in visual form (as
in Study 1; with a uniform color scheme while randomizing lateralization across participants)
or numerical form. Participants in the numerical conditions received the information in the
form of a confusion matrix. For example, in the Cult scenario (where the target wrongdoing is
participation in a dangerous cult, and the proxy offense is traveling to/from the Western
Desert), participants might receive the following information:
There are 73 cult members and 448 non-members.
69 cult members travel to the Western Desert.
4 cult members don't travel to the Western Desert.
394 non-members don't travel to the Western Desert.
54 non-members travel to the Western Desert.
JUSTICE BEFORE EXPEDIENCY17
3.1.3. Procedure
In a 2 (Visual, Numerical) x 3 (Time Pressure, Forced Delay, Deliberation) between-
subjects design, participants were randomly assigned to one of six conditions. As in Study 1,
participants in every condition completed a Training Task, the Experimental Block, and the
Individual Difference Measures in a fixed order.
Training Task. The training task was composed of either visual displays or numerical
information, depending on condition assignment. In the Visual mode, the training task was just
as in Study 1. In the Numerical mode, participants were presented with corresponding
confusion matrices. We again defined probability literacy as the sum score of correct answers
to five true or false questions. Probability literacy was slightly higher in the Numerical mode
(M = 3.54, SD = 1.07) than the Visual mode (M = 3.39, SD = 1.09), Welch’s t(748) = 1.94, p =
.052, Cohen’s d = 0.14.
Experimental Block. Next, participants viewed five scenarios in a randomized order.
Participants in the Forced Delay condition followed the procedure as in Study 1, with the
vignette, display (visual or numerical according to condition), decision and justification slides
in sequence. The Time Pressure condition differed from the Forced Delay condition in that the
display slide was presented for 12 seconds and then automatically advanced. In the
Deliberation condition, the decision and justification questions were presented on the same
screen as the display, with the justification items before the decision question.
Individual Difference Measures. As in Study 1, participants completed the individual
difference measures and provided demographic information after the experimental portion of
the study.
JUSTICE BEFORE EXPEDIENCY18
3.1.4. Measures
As in Study 1, for each scenario and participant, we recorded the three independent
measures (crime rate, false conviction, false acquittal), and three dependent measures:
approval, effectiveness (Cronbach’s α = .89), and justice (Cronbach’s α = .87).
3.2. Results
Below we report a series of mixed-effects models with scenarios and
participants as crossed random effects. In the fixed effects portion of the model, we enter mode
of presentation, and two dummy-codes reflecting whether the nudge was present or absent,
whether time pressure was present or absent, and the two-way interactions between both
dummy variables and presentation mode.
3.2.1. Effects of condition and presentation mode
We observed no main effects of mode of presentation, nudge, time pressure, or any
two-way interaction effects in our primary model of approval, all Fs < 1.41, ps > .24. Pairwise
comparisons revealed no significant differences between conditions in the Numerical mode, ps
> .47, or the Visual mode, ps > .72. An intercept-only model revealed that the grand mean (M
= 4.97) significantly differed from the midpoint, t = 6.74, p < .001, indicating that participants
tended to approve of the proxy criminal policies across conditions.
A corresponding model of justice evaluations revealed a main effect of mode of
presentation, F(1, 745) = 6.24, p = .013, and no other effects, Fs < 0.66, ps > .42. This main effect
reflected higher justice evaluations in the Visual display mode (M = 4.00, 95% CI [3.63, 4.38])
than the Numerical summary mode (M = 3.83, 95% CI [3.46, 4.21]), t = 2.46, p = .014.
No corresponding effects arose in a model of effectiveness judgments, all Fs < 0.04, ps
> .84. An intercept-only model revealed that the grand mean (M = 5.15) significantly differed
from the midpoint, t = 11.95, p < .001, indicating that participants tended to view the proxy
criminal policies as efficient across conditions.
JUSTICE BEFORE EXPEDIENCY19
3.2.2. Effects of false conviction and acquittal
Next, we examined whether our continuous manipulations of the false positive (i.e.,
conviction) and false negative (i.e., acquittal) rates, as well as of the base rate of crime,
influenced participants’ evaluations. As in Study 1, we observed a strong main effect of false
convictions on approval, F(1,3435) = 158.2, p < .001. False convictions reduced approval of proxy
criminal policies across conditions, B = -5.48, 95% CI [-6.33, -4.63], t = -12.66, p < .001and
the effect was robust across presentation modes and conditions, -6.39 < Bs < -4.81, -6.12 < ts
< -4.45, ps < .001.
This time, however, we also observed effects of false acquittals, F(1, 3397) = 6.2, p = .012,
and overall crime (i.e., the base rate), F(1, 3411) = 4.69, p = .027. False acquittals reduced
approval of proxy criminal policies (B = -1.07, 95% CI [-1.92, -0.22], t = -2.46, p = .014) while
increases in the base rate of crime promoted approval, B = 0.87, 95% CI [0.11, 1.62], t = 2.24,
p = .025though both effects were small (see Figure 3).
Furthermore, the influence of false acquittal varied across conditionswhich was
reflected by a non-significant three-way interaction with time pressure and mode, F(1, 3408) =
3.81, p = .051. In the Numerical mode, time to reflect did not moderate the impact of false
acquittals on approval, B = 1.26, t = 0.83, p = .41. Furthermore, the simple effect of false
acquittals was non-significant in both the delayed (B = -0.44, 95% CI [-1.89, 1.00]) and speeded
(B = -1.71, 95% CI [-4.34, 0.92]) conditions. However, in the Visual mode, the forced delay
did strengthen the effect of false acquittals on approval, B = -2.86, t = -1.96, p = .0497.
Specifically, false acquittals reduced approval in the delayed condition (B = -2.26, 95% CI [-
3.76, -0.76]) but had no effect under time pressure (B = 0.60, 95% CI [-1.92, 3.11]).
3.2.3. Effectiveness and justice evaluations
A model of justice evaluations revealed effects of false conviction, F(1, 3397) =
327.9, p < .001, but also false acquittal, F(1, 3360) = 7.50, p = .006, and the base rate, F(1, 3374) =
JUSTICE BEFORE EXPEDIENCY20
7.73, p = .005. False convictions negatively predicted justice evaluations, B = -6.84, 95% CI [-
7.57, -6.10], t = -18.21, p < .001. This time, unlike Study 1, false acquittals also rendered proxy
crimes more unjust, B = -1.04, 95% CI [-1.77, -0.30], t = -2.76, p = .006. Additionally, increases
in the base rate of crime rendered proxy crimes more just, B = 0.92, 95% CI [0.26, 1.58], t =
2.75, p = .006. Importantly, false acquittals and crime rates had a weaker influence on
perceptions of justice than did false convictions.
A model of effectiveness evaluations revealed effects of both false conviction,
F(1, 3319) = 82.05, p < .001, and false acquittal, F(1, 3288) = 10.38, p = .001but not of the base
rate, F(1, 3299) = 2.09, p = .15. As in Study 1, false convictions negatively predicted effectiveness,
B = -2.96, 95% CI [-3.60, -2.33], t = -9.13, p < .001.
The main effect of false acquittals on effectiveness, though, was qualified by a three-
way interaction with presentation mode and time pressure, F(1, 3298) = 5.50, p = .019. This
interaction indicated that a forced delay moderated the effect of false acquittals on effectiveness
in the Visual mode, B = -2.39, t = -2.20, p = .025, such that the negative effect was present
following a forced delay (B = -1.54, 95% CI [-2.66, -0.42]) but absent under time pressure (B
= 0.84, 95% CI [-1.03, 2.73]). The forced delay had no corresponding influence in the
Numerical mode, B = 1.31, t = 1.14, p = .25where the trend was, if anything, weaker after a
forced delay (B = -1.01, 95% CI [-2.10, 0.06]) than under time pressure (B = -2.33, 95% CI [-
4.29, -0.36]).
JUSTICE BEFORE EXPEDIENCY21
Figure 3. Study 2: Standardized (i.e., z-scored) evaluations of the proxy criminal
policies by false conviction and acquittal rates, and by the base rate of crime. Each plot displays
(locally estimated) curve fit and mean values for each mode of presentation (Numerical:
crosses; Visual: circles) in percent increments.
Importantly, a forced delay did not moderate any of the corresponding effects of false
conviction on effectiveness, ps > .30, or justice evaluations, ps > .34, whether in the numerical
or visual modes. Thus, the emphasis on false conviction may be relatively automatic, whereas
directing attention toward false acquittal requires additional time and/or cognitive resources.
Study 2 replicated the influence of false convictions on effectiveness observed
in Study 1. So, once again, we assessed whether this effect could be indirect via approval.
JUSTICE BEFORE EXPEDIENCY22
Approval and effectiveness were strongly positively related, B = 0.52, 95% CI [0.50, 0.54], t
= 57.41, p < .001. Including approval in the model rendered the influence of false convictions
on effectiveness non-significant, B = -0.13, 95% CI [-0.61, 0.35], t = -0.53, p = .60. Appendix
2 reports further analyses of the proposed model linking the causes of approval to its
justification.
3.3. Discussion
Study 2 replicated and extended our primary findings. As in Study 1,
participants’ evaluations of proxy offenses depended primarily on the rate at which innocent
citizens would be falsely convicted. This pattern arose regardless of whether information about
the probative value of the proxy conduct was presented in numerical or visual form, and
whether participants had limited or unlimited time to interpret the evidence. Thus, the emphasis
on false conviction rates appears to reflect a spontaneous cognitive process.
The likelihood of acquitting offenders had no influence on participants’
decisions in Study 1. So, in Study 2, we asked whether inducing participants to reflect might
bring about greater concern for the rate of false acquittal. In the aggregate, Study 2 revealed
some concern for false acquittalthough these effects were modest when compared to the
impact of false conviction. Furthermore, rates of false acquittal reduced approval of proxy
criminal policies only under specific experimental conditionsnamely, when participants had
unlimited time to interpret the evidence and it was presented in visual form. All in all, Study 2
suggested that false acquittals can undermine support for proxy criminal policies, but this
relationship is fickle and demands greater cognitive effort.
Why a visual mode of presentation would inspire concern for false acquittals is unclear.
Responses to the training tasks suggested that, if anything, visual presentation of the evidence
hindered participants’ efforts to interpret the data relative to the numerical summary condition.
Therefore, ease of interpretation cannot easily explain the selective effect of false acquittals in
JUSTICE BEFORE EXPEDIENCY23
the visual presentation condition. Alternatively, it may be that a visual display enables the
valuation of false acquittals: People’s representation of the value of false acquittals may be
more responsive to visual changes (i.e., in location, shape and hue) than to equivalent changes
represented using numerals.
4. Individual Differences Analyses
In our studies, participants completed the following individual difference measures,
which we analyze in the aggregate below:
1. probability literacy, the sum score of correct responses in the training task;
2. punitive tendencies, the six-item average (Cronbach’s α = .85) of the Punitiveness
Questionnaire (Kemme et al., 2014);
3. punishment theory: a single item, self-reported, binary measure of which of the two
theories of punishment (retributivism or deterrence)
4. retribution motives: a single-item self-reported assessment of whether retributive
motives guided their judgments of proxy crimes, and
5. deterrence motives: a single-item self-reported assessment of whether deterrence
motives guided their judgments of proxy crimes.
The items assessing retribution and deterrence motives were adapted from work by
Carlsmith and colleagues (2002).
4.2. Analysis
In this section, we report a series of mega-analyses (total N = 1551) to examine
the effect of the individual difference measures. First, we conducted model comparisons to
ascertain whether the individual difference measures had a main effect on approval, justice and
effectiveness, and/or moderated the influence of false convictions, acquittals and crime rates
in each model. The models included random effects of participant and scenario nested within
study.
JUSTICE BEFORE EXPEDIENCY24
We then asked whether, relative to the baseline regression model with three fixed
effects (i.e., false conviction, false acquittal, and crime rate), model fit improved by entering
the individual difference measure as an additional predictor (i.e., in Model 1). Next, we asked
whether (Model 2) including the two-way interaction terms between the individual difference
measure and the three parameters improved model fit (compared to Model 1).
Where the model comparisons point to a significant effect, we followed up by reporting
the results of the corresponding regression model. The individual difference measures and the
dependent variables were standardized to one standard deviation, in order to produce readily
interpretable and comparable estimates of the magnitude of the effects.
4.3. Results
People claim to prefer the deterrence theory of punishment over retributivism in
abstract (in the binary choice question, 76% claim to prefer the former), but also when asked
to reflect on motives that guided their choices in the context of proxy crimes: People claim to
be guided more by deterrence motives than by retributive ones (see Table 2; t(2680) = 24.21,
p < .001, Cohen’s d = 0.87).
Increased punitiveness is associated with elevated levels of both retributive and
deterrence motives; the latter are also positively correlated with each other. Higher probabilistic
literacy, in turn, is associated with lower punitiveness and retributive motives (see Table 2).
JUSTICE BEFORE EXPEDIENCY25
Table 3.
Correlation Table (N = 1551).
M (SD)
(1)
(2)
(3)
(4)
(1) Punitiveness
5.36
(1.09)
-
(2) Retributivism
4.95
(1.39)
.44
[.40, .48]
-
(3) Deterrence
5.98
(0.92)
.25
[.20, .29]
.11
[.06, .16]
-
(4) Probability Literacy
3.42
(1.12)
-.14
[-.18,-.09]
-.08
[-.13, -.04]
-.05 #
[-.10, .00]
-
Note. #: non-significant (p > .05). All other p values below .001.
4.3.1. Punitiveness
Adding punitiveness improved all three models’ fit to the data, χ2s > 30.6, ps < .001,
whereas adding the interaction terms did not, χ2s < 4.52, ps > .21. Overall, punitive individuals
were more likely to support proxy crimes (B = 0.23), and to consider them just (B = 0.10) and
effective (B = 0.24), all ps < .001. This pattern held regardless of the proxy crimes’ specificity
and sensitivity in relation to the target offense.
4.3.2. Retribution and deterrence motives
Adding punishment motives also improved the models’ fit, χ2s > 58.8, ps < .001.
Endorsement of deterrence theory was associated with favorable views of proxy criminal
policies (approval: B = 0.21, justice: B = 0.12, effectiveness: B = 0.25), all ps < .001. To a lesser
JUSTICE BEFORE EXPEDIENCY26
extent, endorsement of retributivism was also tied to greater approval (B = 0.12) and
effectiveness (B = 0.13, both ps < .001), but not justice (B = 0.02, p = .21).
This time, including the interaction terms marginally improved the fit of the
effectiveness model, χ2 = 11.38, p = .077, but not the approval or justice models, χ2s < 6.10, ps
> .41. Deterrence motives moderated the influence of false acquittal on effectiveness, B = -
0.38, 95% CI [-0.75, -0.02], t = -2.07, p = .039. Among participants with elevated (+1SD) or
average (Mean) reported level of deterrence motivation, false acquittals reduced perceived
effectiveness (Mean: B = -0.64; +1SD: B = -1.03, both ps < .001), while no such effect arose
among participants with low (-1SD) level, B = -0.26, t = -0.98, p = .33.
4.3.3. Probability literacy
Model comparisons indicated that entering probability literacy did not improve model
fit, χ2s < 1.35, ps > .24. In contrast, adding the interaction terms improved the fit of the justice
model, χ2 = 11.38, p < .001. A trend was also observed for the approval model, χ2 = 7.46, p =
.058, but not the effectiveness model, χ2 = 1.52, p = .68.
Probability literacy moderated the effect of false conviction on justice (B = -0.86, p <
.001) and approval (B = -0.47, p = .014). In other words, the effect of false conviction was
stronger among participants with higher scores on the training task (than among those with
lower scores).
JUSTICE BEFORE EXPEDIENCY27
Table 4. Individual Difference Analyses (N = 1551).
Approval
Justice
Efficiency
B
t
p
B
t
p
B
t
p
False Conviction
-5.08
[-5.72, -4.44]
-15.51
< .001
-6.97
[-7.52, -6.42]
-24.83
< .001
-2.91
[-3.39, -2.43]
-11.90
< .001
False Acquittal
-0.34
[-0.98, 0.30]
-1.03
.30
-0.30
[-0.85, 0.25]
-1.07
.28
-0.76
[-1.24, -0.28]
-3.12
.002
Base Rate
0.64
[0.06, 1.21]
2.15
.032
0.80
[0.31, 1.30]
3.16
.002
0.30
[-0.13, 0.73]
1.37
.17
Punitiveness
0.24
[0.19, 0.29]
9.66
< .001
0.09
[0.04, 0.14]
3.63
< .001
0.18
[0.14, 0.23]
8.76
< .001
Probability
Literacy
0.06
[0.02, 0.10]
2.70
.007
0.03
[-0.01, 0.07]
1.45
.15
0.07
[0.04, 0.11]
3.91
< .001
Retribution
0.07
[0.03, 0.10]
3.50
< .001
-0.00
[-0.04, 0.03]
-0.21
.84
0.07
[0.04, 0.010]
4.40
< .001
Deterrence
0.34
[0.28, 0.39]
12.60
< .001
0.18
[0.13, 0.23]
6.92
< .001
0.33
[0.28, 0.37]
14.49
< .001
Punitiveness×F.
Conviction
0.34
[-0.34, 1.01]
0.98
.33
0.13
[-0.45, 0.71]
0.43
.67
0.01
[-0.49, 0.51]
0.04
.97
Punitiveness×F.
Acquittal
0.56
[-0.11, 1.23]
1.63
.10
0.38
[-0.19, 0.95]
1.30
.20
0.26
[-0.24, 0.76]
1.04
.30
Punitiveness×Base
Rate
-0.17
[-0.78, 0.45]
-0.53
.60
-0.28
[-0.81, 0.24]
-1.07
.29
0.10
[-0.35, 0.56]
0.45
.65
Literacy×F.
Conviction
-0.76
[-1.33, -0.19]
-2.61
.009
-1.19
[-1.68, -0.71]
-4.78
< .001
-0.08
[-0.51, 0.34]
-0.38
.70
Literacy×F.
Acquittal
-0.31
[-0.89, 0.27]
-1.60
.29
-0.04
[-0.53, 0.46]
-0.14
.89
-0.24
[-0.67, 0.19]
-1.08
.28
Literacy×Base
Rate
0.07
[-0.46, 0.59]
0.24
.81
0.34
[-0.11, 0.79]
1.49
.14
0.06
[-0.34, 0.45]
0.27
.78
Retribution×F.
Conviction
-0.01
[-0.52, 0.50]
-0.04
.97
0.07
[-0.37, 0.50]
0.30
.77
-0.06
[-0.44, 0.32]
-0.33
.74
JUSTICE BEFORE EXPEDIENCY28
Deterrence×F.
Conviction
-0.12
[-0.85, 0.61]
-0.32
.75
-0.69
[-1.32, -0.07]
-2.17
.030
0.02
[-0.52, 0.57]
0.57
.94
Retribution×F.
Acquittal
-0.10
[-0.62, 0.42]
-0.39
.70
-0.17
[-0.62, 0.27]
-0.77
.44
0.20
[-0191, 0.59]
1.01
.31
Deterrence×F.
Acquittal
-0.47
[-1.20, 0.26]
-1.26
.21
-0.52
[-1.15, 0.10]
-1.65
.10
-0.60
[-1.14, -0.06]
-2.16
.031
Retribution×Base
Rate
0.01
[-0.46, 0.48]
0.04
.97
0.12
[-0.28, 0.51]
0.56
.57
-0.32
[-0.67, 0.03]
-1.81
.071
Deterrence×Base
Rate
-0.01
[ -0.66, 0.63]
-0.04
.97
0.10
[ -0.46, 0.65]
0.33
.74
0.24
[ -0.24, 0.73]
0.99
.32
4.4. Discussion
Attitudes toward punishment influenced participants’ evaluations of proxy criminal
policies. As expected, punitive individuals demonstrated a more favorable attitude toward
proxy crimes–which did not depend on the proxy crimes’ specificity and sensitivity. Then,
when examining participants’ theories of punishment, we corroborated a pattern previously
observed within legal scholarship: Specifically, endorsement of deterrence theory was
associated with more favorable views of proxy crime than was endorsement of retributivism.
Probability literacy, as measured on the training task, was unrelated to overall approval
of proxy crimes. Higher scores on the training task were, however, tied to a greater emphasis
on false convictions. In other words, decisions to approve or oppose proxy criminal policies
were more closely related to false conviction rates among those participants who scored highly
on the training task (than among those with lower scores).
JUSTICE BEFORE EXPEDIENCY29
5. General Discussion
In debates concerning the principles that should guide criminalization, two
broad camps have dominated the theoretical landscape. On one hand, consequentialists argue
that new criminal offenses may be rightfully introduced as long as their benefits, primarily,
their effectiveness in combating crime, outweigh their social costs. In contrast, non-
consequentialists advocate certain categorical constraints on the legitimate scope of
criminalizationone of which is non-infringement on the rights of the innocent. Does people’s
reasoning about criminalization illustrate the commitments of either theoretical perspective?
To answer this question, we conducted a series of experiments probing people’s
reactions toward a series of proxy criminal statutes. Because proxy crimes target wrongdoing
indirectly, and incur in both false positives (when they wrongfully convict innocent people)
and false negatives (when they falsely acquit criminals), they provide a unique window into
the way concerns about crime reduction and protection of the innocent guide people’s
evaluations of criminal policy, dissociating the impact of each.
In this regard, our studies revealed a striking discrepancy in laypeople’s
thinking about criminal law. When introspecting about the motives driving their
criminalization decisions, participants alleged both a concern with tackling crime as well as
with safeguarding the innocent’s rights. Yet, attitudes toward concrete criminal policies were
influenced primarily, perhaps even solely, by their risk of convicting innocent civiliansin line
with a non-consequentialist position. Furthermore, this effect of wrongful convictions was
remarkably robust to the imposition of time pressure favoring an intuitive cognitive style,
variation in the mode of presentation (i.e., whether the data were presented visually or
numerically), and arose regardless of people’s probability literacy and beliefs about the purpose
of legal punishment.
JUSTICE BEFORE EXPEDIENCY30
According to a consequentialist school of thought, the rate of false acquittals and the
incidence (or base rate) of crime ought to influence evaluations of criminal policies. Yet these
parameters mattered little if at allexcept under specific conditions promoting cognitive
reflection. In sum, returning to our opening question, our studies documented a predominant
and univocal commitment to the principle that it is better “that ten guilty persons escape than
that one innocent suffer” (Blackstone, 1765). This intuitive commitment arose among
individuals with diverse explicit views on the function of criminal policy, under various
experimental conditions, and even though participants tended retrospectively to adduce a
greater focus on the statutes’ capacity to combat and deter crime.
Taken together, these results highlight a dissociation between the factors that
shape intuitive attitudes toward criminalization and the reasoning that people retrospectively
offer. As such, our present studies dovetail with a growing body of research on legal decision-
making (Carlsmith et al., 2002; Struchiner et al., 2020; Costa et al., 2019), documenting a
recurring discrepancy between people’s intuitive legal judgment and their explicit avowals.
References
Alexander, L., & Ferzan, K. K. (2009). Crime and Culpability: A Theory of Criminal
Law. Cambridge University Press.
Bentham, J. (1887). Theory of legislation (R. Hildreth, Trans.). Trübner & Company.
Bystranowski, P., & Mungan, M. C. (2022). Proxy Crimes. American Criminal Law
Review, 59(1), 1-38.
Carlsmith, K. M. (2008). On Justifying Punishment: The Discrepancy Between Words
and Actions. Social Justice Research, 21(2), 119137. https://doi.org/10.1007/s11211-008-
0068-x
JUSTICE BEFORE EXPEDIENCY31
Carlsmith, K. M., Darley, J. M., & Robinson, P. H. (2002). Why do we punish?
Deterrence and just deserts as motives for punishment. Journal of Personality and Social
Psychology, 83(2), 284.
Costa, L. L. F., Esteves, A. B. D., Kreimer, R., Struchiner, N., & Hannikainen, I. (2019).
Gender stereotypes underlie child custody decisions. European Journal of Social Psychology,
49(3), 548-559.
Darley, J. M., Carlsmith, K. M., & Robinson, P. H. (2000). Incapacitation and just
deserts as motives for punishment. Law and Human Behavior, 24(6), 659683.
De Guttry, A., Capone, F., & Paulussen, C. (2016). Foreign fighters under international
law and beyond. Springer.
Duff, R. A. (1997). Criminal Attempts. Oxford University Press.
Duff, R. A., Farmer, L., Marshall, S., & Tadros, V. (2007). The Trial on Trial: Volume
3. Towards a Normative Theory of the Criminal Trial. Hart Publishing.
Haidt, J. (2001). The emotional dog and its rational tail: A social intuitionist approach
to moral judgment. Psychological Review, 108(4), 814.
Kemme, S., Hanslmaier, M., & Pfeiffer, C. (2014). Experience of parental corporal
punishment in childhood and adolescence and its effect on punitiveness. Journal of Family
Violence, 29(2), 129142.
McFatter, R. M. (1982). Purposes of punishment: Effects of utilities of criminal
sanctions on perceived appropriateness. Journal of Applied Psychology, 67(3), 255.
Picinali, F. (2018). Can the Reasonable Doubt Standard be Justified? A Reconstructed
Dialogue. Canadian Journal of Law & Jurisprudence, 31(2), Article 2.
Rosseel, Y. (2012). Lavaan: An R package for structural equation modeling and more.
Version 0.512 (BETA). Journal of Statistical Software, 48(2), 136.
Schauer, F. (2003). Profiles, probabilities, and stereotypes. Harvard University Press.
JUSTICE BEFORE EXPEDIENCY32
Schauer, F. (2011). Bentham on Presumed Offences. Utilitas, 23(4), 363379.
Sharp, F. C., & Otto, M. C. (1910). A Study of the Popular Attitude Towards
Retributive Punishment. International Journal of Ethics, 20(3), 341357. JSTOR.
Struchiner, N., Almeida, G. F. C. F., & Hannikainen, I. R. (2020). Legal decision-
making and the abstract/concrete paradox. Cognition, 205, 104421.
Teichman, D. (2017). Convicting With Reasonable Doubt: An Evidentiary Theory of
Criminal Law. Notre Dame Law Review, 93(2), 757810.
Tomlin, P. (2013). Extending the golden thread? Criminalisation and the presumption
of innocence. Journal of Political Philosophy, 21(1), 4466.
JUSTICE BEFORE EXPEDIENCY33
Open Practices Statement
The experiments were not formally preregistered. Materials, data, and scripts have been
made publicly available on the Open Science Framework at:
https://osf.io/tn6j8/?view_only=ae18f93757ad4f68af55705e7803b773
ResearchGate has not been able to resolve any citations for this publication.
Article
Full-text available
The family, as the primary instance of socialization, plays a key role in nurturing values and attitudes. Based on this notion, this paper looks at how parental corporal punishment in childhood and adolescence, as an expression of a strict, authoritarian upbringing, can influence punitiveness later in life. The results of a representative German sample using multivariate analyses show that individuals who were physically punished or abused by their parents during childhood or adolescence are more punitive than non-victims of parental violence. Based on these findings, the question of whether changing parenting styles might have implications at the macro-level of punitiveness is addressed.
Article
The justification of the reasonable doubt standard has been hotly debated in recent years. Deontologists—including retributivists—have generally defended the standard, whilst consequentialists have generally argued for a lower standard of proof. Captivating arguments have been produced from both sides. The paper narrates this debate through a dialogue between ideal representatives of these different camps. In doing so, it recasts—and, hopefully, improves—some of the arguments presented thus far. Then, the paper introduces a new participant in the debate, the Intermediary. The Intermediary is under the impression that the debate has reached an impasse, due to fundamental moral disagreements between the parties involved. Therefore, she presents them with a challenge: to find a common ground that allows the parties to justify to each other the choice of a standard of proof, notwithstanding their different basic moral commitments. The Intermediary takes up this challenge, and provides a justification for the reasonable doubt standard based on the value of respecting defendants and on rules of instrumental rationality.
Article
Research on moral judgment has been dominated by rationalist models, in which moral judgment is thought to be caused by moral reasoning. The author gives 4 reasons for considering the hypothesis that moral reasoning does not cause moral judgment; rather, moral reasoning is usually a post hoc construction, generated after a judgment has been reached. The social intuitionist model is presented as an alternative to rationalist models. The model is a social model in that it deemphasizes the private reasoning done by individuals and emphasizes instead the importance of social and cultural influences. The model is an intuitionist model in that it states that moral judgment is generally the result of quick, automatic evaluations (intuitions). The model is more consistent than rationalist models with recent findings in social, cultural, evolutionary, and biological psychology, as well as in anthropology and primatology.
Article
This Article presents an evidentiary theory of substantive criminal law according to which sanctions are distributed in proportion to the strength of the evidence mounted against the defendant. It highlights the potential advantages associated with grading penalties in proportion to the probability of wrongdoing and situates this claim within both consequentialist and deontological theories of punishment. Building on this analysis, the Article reviews the doctrinal tools used to achieve the goal of evidentiary grading of sanctions and shows that key factors in criminal law are geared towards dealing with evidentiary uncertainty. Finally, the Article explores the underlying logic of the evidentiary structure of criminal law and argues that this structure can be justified on psychological, economic, and expressive grounds.
Book
This book offers various perspectives, with an international legal focus, on an important and underexplored topic, which has recently gained momentum: the issue of foreign fighters. It provides an overview of challenges, pays considerable attention to the status of foreign fighters, and addresses numerous approaches, both at the supranational and national level, on how to tackle this problem. Outstanding experts in the field – lawyers, historians and political scientists – contributed to the present volume, providing the reader with a multitude of views concerning this multifaceted phenomenon. Particular attention is paid to its implications in light of the armed conflicts currently taking place in Syria and Iraq. Andrea de Guttry is a Full Professor of International Law at the Scuola Superiore Sant’Anna, Pisa, Italy. Francesca Capone is a Research Fellow in Public International Law at the Scuola Superiore Sant’Anna. Christophe Paulussen is a Senior Researcher at the T.M.C. Asser Instituut in The Hague, the Netherlands, and a Research Fellow at the International Centre for Counter-Terrorism – The Hague.
Article
This book presents a comprehensive overview of what the criminal law would look like if organized around the principle that those who deserve punishment should receive punishment commensurate with, but no greater than, that which they deserve. Larry Alexander and Kimberly Kessler Ferzan argue that desert is a function of the actor’s culpability and that culpability is a function of the risks of harm to protected interests that the actor believes he is imposing and his reasons for acting in the face of those risks. The authors deny that resultant harms, as well as unperceived risks, affect the actor’s desert. They thus reject punishment for inadvertent negligence as well as for intentions or preparatory acts that are not risky. Alexander and Ferzan discuss the reasons for imposing risks that negate or mitigate culpability, the individuation of crimes, and omissions. They conclude with a discussion of rules versus standards in criminal law and offer a description of the shape of criminal law in the event that the authors’ conceptualization is put into practice.
Article
It seems to be widely accepted that the presumption of innocence, and the attendant standard of 'beyond reasonable doubt' properly apply in the courtroom as a procedural principle directly grounded in the moral imperative to avoid punishing those who should not be punished. In this article I argue that if this is correct, then we ought be as careful about what we criminalise, as we are about who we punish, since people can be wrongfully punished by criminalisation errors as well as by conviction errors.