ArticlePDF Available

Juror Sensitivity to False Confession Risk Factors: Dispositional vs. Situational Attributions for a Confession

Authors:

Abstract and Figures

Research on jurors’ perceptions of confession evidence suggests that jurors may not be sensitive to factors that can influence the reliability of a confession. Jurors’ decisions tend not to be influenced by situational pressures to confess, which suggests that jurors commit the correspondence bias when evaluating a confession. One method to potentially increase sensitivity and counteract the correspondence bias is by highlighting a motivation other than guilt for the defendant’s confession. We conducted three experiments to evaluate jurors’ sensitivity to false confession risk factors. Participants read a trial transcript that varied the presence of false confession risk factors within an interrogation. Some participants also read testimony that presented an alternative motivation for the confession (expert testimony, Experiments 1 & 3; defendant testimony, Experiment 2). Across three experiments, participants were generally able to distinguish between interrogation practices that can produce a false confession, regardless of the presence or absence of expert or defendant testimony. Experiment 3 explored whether participants’ attributions for the confessor’s motivation were affected by interrogative pressure and expert testimony, and whether these attributions affected verdicts. Participants’ reluctance to convict when false confession risk factors were present was associated with situational, rather than dispositional, attributions regarding the defendant’s motivation to confess. It is possible that increased knowledge is responsible for participants’ improved sensitivity to false confession risk factors.
Content may be subject to copyright.
Running head: SENSITIVITY 1
Juror Sensitivity to False Confession Risk Factors:
Dispositional vs. Situational Attributions for a Confession
Skye A. Woestehoff
University of Texas at El Paso
Christian A. Meissner
Iowa State University
In Press at Law & Human Behavior
Author Note
Skye A. Woestehoff, Department of Psychology, University of Texas at El Paso;
Christian A. Meissner, Department of Psychology, Iowa State University.
This research was supported in part by a student grant-in-aid from the American
Psychology-Law Society.
Correspondence concerning this article should be addressed to Skye A. Woestehoff,
Department of Psychology, University of Texas at El Paso, 500 W University Ave, El Paso, TX
79902. Email: sawoestehoff@miners.utep.edu
SENSITIVITY 2
Abstract
Research on jurors’ perceptions of confession evidence suggests that jurors may not be sensitive
to factors that can influence the reliability of a confession. Jurors’ decisions tend not to be
influenced by situational pressures to confess, which suggests that jurors commit the
correspondence bias when evaluating a confession. One method to potentially increase
sensitivity and counteract the correspondence bias is by highlighting a motivation other than
guilt for the defendant’s confession. We conducted three experiments to evaluate jurors’
sensitivity to false confession risk factors. Participants read a trial transcript that varied the
presence of false confession risk factors within an interrogation. Some participants also read
testimony that presented an alternative motivation for the confession (expert testimony,
Experiments 1 & 3; defendant testimony, Experiment 2). Across three experiments, participants
were generally able to distinguish between interrogation practices that can produce a false
confession, regardless of the presence or absence of expert or defendant testimony. Experiment 3
explored whether participants’ attributions for the confessor’s motivation were affected by
interrogative pressure and expert testimony, and whether these attributions affected verdicts.
Participants’ reluctance to convict when false confession risk factors were present was associated
with situational, rather than dispositional, attributions regarding the defendant’s motivation to
confess. It is possible that increased knowledge is responsible for participants’ improved
sensitivity to false confession risk factors.
Keywords: confession, interrogation, juries, expert testimony, attribution
SENSITIVITY 3
Juror Sensitivity to False Confession Risk Factors:
Dispositional vs. Situational Attributions for a Confession
Trials are generally considered a search for the truth. To determine what occurred, jurors
should consider the reliability of the evidence in an unbiased manner and arrive at a decision
regarding the defendant’s guilt (Miller & Boster, 1977). Unfortunately, the search for truth may
be inhibited by flawed evidence. DNA testing has revealed that jurors have convicted innocent
individuals in cases involving evidence such as mistaken eyewitness identifications, faulty
forensic evidence, and false confessions (Garrett, 2011).
One type of evidence that jurors may evaluate is the defendant’s confession; however,
confessions are not always reliable. Innocent suspects are more likely to confess if the
interrogator offers an explicit promise of leniency or uses minimization techniques by
downplaying the seriousness of what happened (Russano, Meissner, Narchet, & Kassin, 2005).
An innocent suspect may also confess if the interrogator exaggerates the strength of the evidence
via maximization techniques (Horgan, Russano, Meissner, & Evans, 2012), such as claiming to
having non-existent evidence (Kassin & Kiechel, 1996; Perillo & Kassin, 2011). Unfortunately,
institutional safeguards may not prevent these false confessions from occurring or from being
presented at trial (see Kassin et al., 2010). This may be problematic given the persuasiveness of a
confession (Kassin & Neumann, 1997) and the difficulty jurors have recognizing false
confessions; false confessors who brought their cases to trial were convicted 73 to 81% of the
time (Drizin & Leo, 2004; Leo & Ofshe, 1998). In the current research, we explored the extent to
which testimony from an expert or the defendant might improve participants’ sensitivity to false
confession risk factors.
Juror Sensitivity to False Confession Risk Factors
SENSITIVITY 4
Although some scholars argue that the issue is more complex (see Martire & Kemp,
2011), a lack of sensitivity about the evidence could contribute, at least in part, to wrongful
convictions. Sensitivity relates to a juror’s capability to discern differences in evidence quality
(knowledge) and a juror’s ability to use this knowledge when rendering a verdict (integration;
Cutler, Penrod, & Dexter, 1989; Leippe, 1995; Leippe & Eisenstadt, 2009). Sensitivity should
lead jurors to recognize poor quality evidence and render more appropriate verdicts. Conversely,
a lack of sensitivity could lead jurors to unquestionably accept poor quality evidence as proof of
a defendant’s guilt (Cutler, Dexter, & Penrod, 1989).
Knowledge. The first component of sensitivity relates to knowledge about factors that
influence evidence reliability (Cutler, Penrod, et al., 1989). Scholars have expressed concerns
that jurors lack knowledge about interrogations and false confessions (Leo, 2008), and such
concerns appear justified. Although people are aware that torture or the threat of torture can lead
to false confessions (Henkel, Coffman, & Dailey, 2008; Leo & Liu, 2009), people sometimes fail
to recognize that psychologically manipulative tactics can as well. People believe, for example,
that implications of leniency and the presentation of false evidence are unlikely to produce false
confessions (Leo & Liu, 2009), although this is not the case (Kassin & Kiechel, 1996; Russano et
al., 2005).
Integration. The second component of sensitivity relates to integration, or the ability to
adjust one’s verdict to account for variations in evidence quality (Cutler, Penrod, et al., 1989).
Unfortunately, people often convict a confessor regardless of how the confession was elicited,
even when they acknowledge the situational pressures to confess. People view some
interrogation techniques as coercive (Leo & Liu, 2009) and the resulting confession as less
voluntary (Kassin & McNall, 1991; Kassin & Wrightsman, 1980, 1981), yet they tend to convict
SENSITIVITY 5
confessors regardless of the interrogation techniques used (see also Woody & Forrest, 2009;
Woody, Forrest, & Yendra, 2014).
The available research suggests that jurors may not be sensitive to false confession risk
factors. Jurors are generally unknowledgeable about why a suspect might falsely confess (Leo &
Liu, 2009) and when jurors do recognize coercion in interrogations, they tend to discount the risk
factors and rely on the confession when rendering a verdict (Kassin & Sukel, 1997). These
findings are not confined to simulated trials; innocent confessors have a high likelihood of being
convicted if they bring their cases to trial (see Drizin & Leo, 2004; Leo & Ofshe, 1998).
Improving Sensitivity to Confession Evidence
One explanation for jurors’ insensitivity to false confession risk factors is the
correspondence bias – the tendency to believe that a person’s behavior reflects an internal,
dispositional state even when situational factors may have influenced the behavior (Gawronski,
2004; Gilbert & Malone, 1995). One reason the correspondence bias occurs is because people
may not know that situations can influence behavior (Gawronski, 2004; Gilbert & Malone,
1995). With regards to a confession, jurors may trust a confession elicited with psychologically
manipulative techniques if they are unaware that those techniques can produce false confessions.
It may not be the case that jurors are completely unaware of the situational influences on
a confession, however. Instead, jurors may recognize, but discount, the influence of interrogation
techniques when determining a confessor’s guilt. The correspondence bias may thus occur if
people believe that situational factors do not influence behavior (Gilbert & Malone, 1995),
perhaps because the behavior is perceived as diagnostic of a disposition (Gawronski, 2004). A
behavior is considered diagnostic if the behavior would not occur without the presence of the
corresponding disposition (Gawronski, 2004). When a behavior is diagnostic, situational
SENSITIVITY 6
pressures are deemed irrelevant and discounted because people assume that situational pressures
would only affect those with the disposition. Thus, a dispositional inference is made, regardless
of whether situational pressures are present (Gawronski, 2004; Reeder, 1993; Reeder & Brewer,
1979).
The available research suggests that confessions may be seen as diagnostic of guilt.
Jurors tend to believe that confessors are guilty (Henkel et al., 2008) and jurors convict
confessors despite recognizing situational pressures to confess (Kassin & Sukel, 1997). This
suggests that jurors appreciate, but discount, the influence of situational pressures. Jurors seem to
believe a confession is the product of an internal drive (i.e., guilt) rather than external pressure
from an interrogator, even when they are cognizant of interrogative pressure.
One potential method to counteract the belief that interrogation techniques do not lead to
false confessions is through the perceived motivation of the confessor. People are less likely to
demonstrate the correspondence bias when an alternative motivation for a behavior exists (Fein,
Hilton, & Miller, 1990). Accentuating the motivation for a false confession might encourage
people to believe that the defendant confessed for a reason other than guilt (cf. Palmer, Button,
Barnett, & Brewer, 2014). In the current research, we sought to highlight the motivation for a
false confession via expert testimony (Experiments 1 & 3) and the defendant’s testimony
(Experiment 2). We also explored participants’ beliefs about the defendant’s motivation to
confess (Experiment 3).
Expert testimony. Jurors’ insensitivity to false confession risk factors may be due to a
lack of knowledge about the influence of situational factors on a confession (Leo & Liu, 2009;
Gawronski, 2004; Gilbert & Malone, 1995). Expert testimony can improve knowledge for some
factors related to eyewitness accuracy (Cutler, Penrod, et al., 1989; Fox & Walters, 1986), child
SENSITIVITY 7
sexual abuse (Buck, London, & Wright, 2011; Gabora, Spanos, & Joab, 1993; Goodman-
Delahunty, Cossins, & O’Brien, 2011), and hearsay evidence (Nuñez, Gray, & Buck, 2012).
Regarding confession evidence, people who hear expert testimony rate some psychologically
manipulative interrogation techniques as more coercive (Blandón-Gitlin, Sperry, & Leo, 2011;
Woody & Forrest, 2009) and more likely to elicit false confessions (Gomes, Stenstrom, &
Calvillo, 2014). These findings suggest that expert testimony could improve knowledge about
the situational pressures to confess, potentially elucidating the motivations underlying a false
confession.
While knowledge is an important component of sensitivity, jurors must ultimately
integrate their knowledge when rendering a verdict (Cutler, Penrod, et al., 1989). Jurors may
view situational pressures to confess as irrelevant if they believe a confession is diagnostic of
guilt (Gawronski, 2004; Reeder, 1993; Reeder & Brewer, 1979), which could explain why
jurors’ knowledge tends not to influence verdicts (Kassin & Sukel, 1997). Simply educating
jurors about why a coerced confession is unreliable may be insufficient to influence verdicts
(Kassin & Wrightsman, 1981); jurors may need to perceive the information as relevant in order
to use it.
It is possible that the manner in which expert testimony is presented could influence
perceived relevance. Jurors may ignore an expert’s testimony if the testimony is perceived as
irrelevant, as may be the case with general expert testimony (Bar-Hillel, 1980; Borgida &
Brekke, 1981) where the expert discusses psychological theories or research related to the
evidence (Leippe & Eisenstadt, 2009). In contrast, case-specific expert testimony may be
perceived as more relevant and thus more likely to be used by jurors (Bar-Hillel, 1980; Borgida
& Brekke, 1981; Leippe & Eisenstadt, 2009). Case-specific expert testimony is where the expert
SENSITIVITY 8
connects his or her testimony to the evidence, such as by noting the presence of risk factors
(Leippe & Eisenstadt, 2009) or by responding to a hypothetical scenario (Brekke & Borgida,
1988).
Case-specific expert testimony may be more likely than general expert testimony to lead
jurors to apply the expert’s testimony when it is appropriate to do so (i.e., sensitivity). Mock
jurors exposed to case-specific expert testimony have demonstrated greater sensitivity in cases
involving child sexual abuse (Kovera, Gresham, Borgida, Gray, & Regan, 1997) and eyewitness
testimony (Cutler, Penrod, et al., 1989; Geiselman et al., 2002, Exp. 1; Geiselman & Mendez,
2005; Phillips, 2001) than jurors not exposed to expert testimony. Although general expert
testimony has also led to sensitivity in some experiments (Buck et al., 2011; Cutler, Penrod, et
al., 1989; Devenport, Stinson, Cutler, & Kravitz, 2002; Laub, 2010; Wells & Wright, 1983 as
cited in Wells, 1986), such testimony appears more likely to result in no effect or desensitization
(Devenport & Cutler, 2004; Geiselman et al., 2002, Exp. 1; Kovera et al., 1997; Laimon, 2005,
Exp. 2; Lindsay, 1994, Exp. 2 & 5; Martire & Kemp, 2009).
Whether general or case-specific expert testimony about false confessions can facilitate
the integration of knowledge and lead to sensitivity has not yet been examined. Although expert
testimony can influence jurors’ verdicts (Blandón-Gitlin et al., 2011; Gomes et al., 2014; Woody
& Forrest, 2009), to the best of our knowledge no published studies involving a confession from
the defendant have been designed to test for sensitivity, and none have compared general and
case-specific expert testimony. A test for sensitivity is possible if evidence quality is manipulated
(Cutler, Penrod, et al., 1989) via the presence or absence of risk factors that could undermine the
credibility of the evidence (Martire & Kemp, 2011).
SENSITIVITY 9
Research suggests that (case-specific) expert testimony could lead to sensitivity. Jurors
may commit the correspondence bias and be insensitive to false confession risk factors because
they are unknowledgeable or view situational pressures to confess as irrelevant (Gawronski,
2004; Gilbert & Malone, 1995). General and case-specific expert testimony can educate jurors
about false confession risk factors, thus providing an alternative motivation for the confession.
However, case-specific expert testimony is likely to also lead jurors to integrate this knowledge
at trial, given that the expert’s information would be more explicitly relevant to the case.
Defendant testimony. In addition to expert testimony, another way to provide an
alternate motivation for the confession is via defendant testimony about the interrogation. If a
confession is seen as diagnostic of guilt (Henkel et al., 2008), jurors may view situational factors
as irrelevant when rendering a verdict (Kassin & Sukel, 1997). However, if defendants discuss
the role of situational factors in producing their false confessions (an alternative motivation),
jurors may recognize the importance of attending to these factors.
People seek explanations for others’ behavior (cf. Malle, Guglielmo, & Monroe, 2014)
and the defendant’s testimony may have an advantage over expert testimony in this regard.
People expect the defendant to testify and they perceive the defendant’s silence as evidence of
guilt (Antonio & Arone, 2005). Additionally, jurors view defendant testimony as more important
to their final decisions than expert testimony (Bridgeman & Marlowe, 1979). Defendant
testimony may be perceived as more important than expert testimony because experts testify
about how individuals behave in general, but the legal system is concerned with explaining a
specific defendant’s behavior (cf. Costanzo & Krauss, 2012). Although case-specific expert
testimony could bridge the gap between abstract research and the individual on trial, testimony
from the defendant may be seen by jurors as particularly relevant to their decisions. While the
SENSITIVITY 10
defendant’s testimony has been presented in a few prior experiments (Gomes et al., 2014;
Henkel, 2008; Kassin & Sukel, 1997), such testimony is more often not present. Despite a lack of
research on whether a defendant’s testimony about the interrogation might influence verdicts,
research on the correspondence bias (Fein et al., 1990) suggests that it may by providing jurors
with a motivation for the defendant’s confession other than guilt (see also Palmer et al., 2014).
We hypothesized that defendant testimony would make participants aware of an alternative
motivation to confess and assist participants in applying this knowledge at trial.
Overview of the Current Studies
Research thus far has suggested that jurors are not sensitive to false confession risk
factors because they lack knowledge about the role of interrogation techniques in eliciting false
confessions (Leo & Liu, 2009) and they fail to apply the limited knowledge they do have (Kassin
& Sukel, 1997). We conducted three experiments to examine the impact of expert and defendant
testimony on juror sensitivity to false confession risk factors. Participants read a trial transcript
adapted from previous research (Henkel, 2008; Kassin & Sukel, 1997) where the defendant
confessed under low, medium, or high pressure, or not at all. In Experiments 1 and 3,
participants read general, case-specific, or no expert testimony. In Experiment 2, the presence of
the defendant’s testimony about the interrogation and the confession was manipulated. In
Experiment 3 we also investigated whether participants believed the defendant’s confession was
dispositionally or situationally motivated, and whether these attributions influenced verdicts.
Experiment 1
We hypothesized that general and case-specific expert testimony would improve
knowledge about false confession risk factors, relative to no expert testimony. We also
hypothesized that case-specific expert testimony would lead to sensitivity; participants would be
SENSITIVITY 11
more likely to convict when the interrogation did not contain false confession risk factors (low-
pressure interrogation) and more likely to acquit when the interrogation did contain false
confession risk factors (medium- and high-pressure interrogations), compared to participants
who did not receive expert testimony. In contrast, we hypothesized that general expert testimony
would result in either a skepticism effect (a reduction in convictions across interrogations) or no
effect, compared to no expert testimony.
Method
Participants. Six hundred and ninety-eight students at a southwestern university and
community members from Amazon Mechanical Turk completed the study. Students (N = 397)
completed the experiment in exchange for course credit. Students were excluded for taking
longer than 2 hours to complete the experiment (n = 19), for failing an attention check question
(n = 2), or because they were not citizens of the United States (n = 24). The final student sample
was 352. Students were predominately female (65.34%) and Hispanic/Latino (83.24%), ages 18
to 50 (M = 21.07, SE = 0.27).
Community members (N = 301) were recruited via Mechanical Turk and were
compensated $1.50 for their time. IP addresses were restricted to the United States. Community
members were excluded for taking longer than 2 hours to complete the experiment (n = 14), for
failing an attention check question (n = 3), or because they were not citizens of the United States
(n = 3). The final community sample was 281. Community members were predominately female
(69.04%) and White/Caucasian (75.45%), ages 18 to 72 (M = 36.98, SE = 0.76).
Design. A 3 (interrogative pressure: low, medium, high) x 3 (expert testimony: none,
general, case-specific) + 1 (no confession control) between-participants design was used. The
effect of sample (student vs. community) was also assessed.
SENSITIVITY 12
Materials. Participants were randomly assigned to read one of ten murder trial scenarios
adapted from previous research (Kassin & Sukel, 1997). The defendant confessed to the crime
during a low-, medium-, or high-pressure interrogation. Participants also read general, case-
specific, or no expert testimony. The no-confession control condition excluded the confession
and expert testimony. Except for the confession, the evidence was circumstantial (see Kassin &
Sukel, 1997).
Interrogation. Interrogation techniques were additive such that higher pressure
interrogations included techniques used in lower pressure interrogations as well as additional
techniques. When a confession was present, the prosecuting attorney asserted in opening
statements that an innocent person would never confess to a crime and claimed the only reason
the defendant confessed is because he was guilty. The defense attorney argued that the defendant
was innocent and was in shock and consumed with grief when he confessed.
Low-pressure interrogation. The interrogator testified that he encouraged the defendant
to tell the truth. The defendant confessed and said he was sorry for what he had done. On cross-
examination the interrogator testified that the defendant did not reveal the location of the murder
weapon or divulge any information that only the perpetrator would know. The suspect had been
emotionally distraught and retracted his confession after the interrogation ended, claiming he
was in shock and not thinking clearly when he confessed.
The defendant testified that he had been upset and in shock. He did not know why he
confessed and he denied committing the murders. On cross-examination he testified that he never
stopped the interrogation, asked for a lawyer, or left the interrogation room.
SENSITIVITY 13
In closing arguments, the defense attorney argued that the defendant was not in his right
mind when he confessed. The only details in the defendant’s confession related to how the
victims were killed; information that anyone at the crime scene could have known.
Medium-pressure interrogation. After testifying about the low-pressure interrogation
tactics, the interrogator testified that he told the defendant he had the ability to wipe the slate
clean and that he would put it in his report if the suspect said he was sorry for what he had done.
The interrogator also told the defendant that the crime was not planned, the victims were to
blame for what happened, the crime could have been worse, and that anyone would have reacted
in the same way. The interrogator then claimed to have DNA evidence that did not exist.
The defendant also testified that he wanted to go home but the interrogator questioned
him relentlessly. He testified that he felt trapped, thought the only way he would be able to leave
was if he confessed, and thought the DNA evidence would prove his innocence.
The defense attorney reiterated that the defendant believed he had no other choice but to
confess.
High-pressure interrogation. After testifying about the low- and medium-pressure
interrogation tactics, the interrogator testified that he told the defendant he would receive the
death penalty if he did not confess. The defendant confessed after 7 hours of interrogation. On
cross-examination, the interrogator testified that he had waved his gun around, interrogated the
defendant all night, and denied the defendant breaks on multiple occasions.
The defendant also testified that he was exhausted and believed he would die unless he
confessed to the crime.
SENSITIVITY 14
The defense attorney added that the defendant had been forced to stay in the interrogation
room all night, even when he explicitly asked for breaks, and that the defendant had been scared
for his life because he was threatened with a gun and the death penalty.
The high-pressure interrogation included elements from Kassin and Sukel’s (1997) high-
pressure interrogation and elements from the interrogations of the Norfolk Four. Although many
of the factors present in this interrogation could render the confession involuntary (cf. Kassin et
al., 2010), this type of interrogation may still appear in court, as was the case with the Norfolk
Four. We included this interrogation to determine whether participants could recognize an
egregious interrogation, above and beyond psychologically manipulative techniques.
Expert testimony. Participants who read a confession received general, case-specific, or
no expert testimony.
General expert testimony. The general expert testified about false confessions, their
prevalence, and research on false confessions. The expert explained that an innocent suspect is
more likely to confess if the interrogator implies leniency, threatens the suspect, or lies about
evidence; if the interrogation is lengthy; or if the suspect is sleep-deprived, under stress, or
depressed. The expert noted that it is difficult to determine the veracity of a confession because
false confessions may include details about the crime or statements of remorse; however, one
marker of a true confession is whether the confession leads to the discovery of new evidence.
On cross-examination the expert admitted that some innocent people would not confess
in response to the aforementioned factors, and that a confession that does not lead to the
discovery of new evidence is not necessarily a false confession. Experimental research on false
confessions uses college students as suspects and interrogators, and participants in these
SENSITIVITY 15
experiments are not accused of actual crimes. The expert had only testified for the defense and
was being paid $200 per hour.
Case-specific expert testimony. In addition to the testimony from the general expert, the
case-specific expert added a response to a hypothetical scenario that was similar to the
interrogation. In the low-pressure interrogation, the expert testified that the hypothetical
suspect’s psychological state could make the suspect more susceptible to pressure from the
interrogator. The expert noted that the confession did not contain new details about the crime. In
the medium-pressure interrogation, the expert added that the hypothetical suspect would be more
likely to falsely confess because the interrogator implied leniency and lied about evidence. In the
high-pressure interrogation, the expert added that the hypothetical suspect might falsely confess
because of the interrogation’s length and the threat of the death penalty.
On cross-examination, the case-specific expert testified about the risk factors that were
not present during the hypothetical interrogation. In the low-pressure interrogation, the expert
testified that the hypothetical interrogator did not imply leniency, lie about evidence, or threaten
the suspect with harsher punishment, and the interrogation did not appear to be extraordinarily
long. In the medium-pressure interrogation, the expert testified that the hypothetical interrogation
did not include a threat and did not appear to be lengthy. In the high-pressure interrogation, the
expert’s testimony was identical to the general expert condition, given that all risk factors were
present.
Questionnaire. To ensure that participants were paying attention, participants answered
factual questions about the case throughout the trial, including how many victims there were,
whether the murder weapon was found, how one of the victims was killed, and whether the
private investigator found evidence that one of the victims was having an affair. Importantly,
SENSITIVITY 16
none of these questions referenced the interrogation. After the trial, participants read the
definition for first-degree felony murder and reasonable doubt. Participants then rendered a
dichotomous verdict and answered questions about the case, including the likelihood the
defendant committed the crime (0-100%), perceptions of the interrogation and the confession on
7-point scales, and knowledge about false confession risk factors. The questionnaire also
included several questions to determine whether participants were answering questions randomly
(e.g., “What do you breathe?” where the correct answer was “Air”). Finally, participants
answered a series of demographics questions.
Participants answered six questions regarding their perceptions of the interrogation and
the confession: the pressure placed on the defendant to confess, the likelihood the defendant’s
confession was false, how voluntary his confession was, how incriminating the confession was,
how much the confession influenced participants’ verdicts, and how justified the interrogator’s
actions were during the interrogation. A principal component factor analysis was used to extract
a factor score representing participants’ perceptions of the interrogation and confession, with all
loadings 0.67 and α = .83.
Knowledge of false confession risk factors was assessed by asking participants to list
reasons for why an innocent suspect would falsely confess to a crime. Participants’ responses
were coded based upon 21 predetermined reasons for a false confession. These reasons were
grouped into four categories: motive, which focused on the suspect’s perceptions of the
interrogation and his or her motivation to confess (e.g., to end the interrogation); interrogation
techniques, which focused on the interrogator’s actions (e.g., minimization); disposition, which
focused on the suspect’s state (e.g., cognitive state); and types of false confessions (e.g.,
voluntary).
SENSITIVITY 17
Two trained researchers who were blind to condition independently coded the data.
Coder 1 coded all of the data and Coder 2 coded a random sample of the data (20%; n = 128) to
ensure reliability. Reliability was determined for each false confession reason; reasons were
eliminated if reliability was poor (κ < 0.70). Coder 1’s scores were used for all analyses.
Inter-rater reliability was acceptable for all but one reason (κs > 0.82). The remaining
reason, “interrogative promise”, did not have acceptable inter-rater reliability (κ = 0.32) and was
excluded from all analyses. Participants’ false confession reasons were summed to create one
knowledge score for each participant, with higher scores indicating more false confession risk
factors mentioned.
Procedure. Participants completed the experiment online. An audio version of the trial
played while participants read the transcript. Participants could not proceed to the next portion of
the trial until the audio had finished. The audio consisted of one male narrator who was
instructed to be as consistent as possible between conditions. The no confession control lasted 25
minutes, while the low-, medium-, and high-pressure interrogations lasted, on average, 36, 39,
and 41 minutes, respectively. On average, the general and case-specific expert testimony
conditions took 41 and 43 minutes, respectively.
Results
Manipulation checks. Participants correctly answered 96.74% (SE = 0.01) of the factual
questions about the case. Community members had higher accuracy (M = 98.52%, SE = .01) than
did students (M = 95.32%, SE = .01), F(1, 631) = 20.84, p < .001, d = 0.37, 95% CI [0.36, 0.37].
Fifty-one participants incorrectly classified the presence of a confession or expert testimony; the
majority (n = 41) believed an expert had testified when he had not. Excluding these participants
did not meaningfully alter the results; therefore, these participants were included in the analyses.
SENSITIVITY 18
We used an ANOVA to evaluate whether our manipulations affected participants’
perceptions of pressure. The main effects of interrogative pressure was significant, F(2, 559) =
415.34, p < .001, η
p
² = .60. Participants in the low-pressure interrogation (M = 3.39, SE = 0.13)
believed that the defendant had been placed under less pressure to confess than did participants
in the medium- (M = 6.29, SE = 0.08) and high-pressure (M = 6.82, SE = 0.04) interrogations,
t(380) = 18.79, p < .001, d = 1.93, 95% CI [1.78, 2.08] and t(374) = 24.32, p < .001, d = 2.52,
95% CI [2.38, 2.66], respectively. Additionally, participants believed the defendant was under
less pressure to confess in the medium-pressure interrogation than in the high-pressure
interrogation, t(376) = 5.91, p < .001, d = 0.61, 95% CI [0.52, 0.70].
There was also a main effect of expert testimony, F(2, 559) = 7.26, p = .001, η
p
² = .03,
which was qualified by an interaction, F(4, 559) = 4.84, p = .001, η
p
² = .03. The interaction
suggested that expert testimony only affected perceptions of pressure in the low-pressure
interrogation. Participants exposed to general expert testimony thought there was more pressure
in the low-pressure interrogation (M = 4.09, SE = 0.23) than participants exposed to case-specific
(M = 2.94, SE = 0.21) or no expert testimony (M = 3.13, SE = 0.22), t(125) = 3.66, p < .001, d =
0.65, 95% CI [0.34, 0.96] and t(126) = 3.00, p = .003, d = 0.53, 95% CI [0.22, 0.85],
respectively.
We used an ANOVA to assess the effect of our manipulations on participants’
knowledge. The interaction between expert testimony and interrogative pressure was not
significant, F(4, 559) = 0.23, p = .922, η
p
² < .01. Consistent with our hypothesis, however, there
was a main effect of expert testimony, F(2, 565) = 10.49, p < .001, η
p
² = .04. Participants who
received general (M = 3.04, SE = 0.11) or case-specific (M = 3.22, SE = 0.12) expert testimony
mentioned more false confession risk factors than did participants who did not receive expert
SENSITIVITY 19
testimony (M = 2.52, SE = 0.10), t(381) = 3.57, p < .001, d = 0.35, 95% CI [0.20, 0.50] and
t(374) = 4.41, p < .001, d = 0.46, 95% CI [0.30, 0.61], respectively. Knowledge did not differ
between general or case-specific expert testimony, t(375) = 1.14, p = .256, d = 0.11, 95% CI [-
0.05, 0.27].
There was also a main effect of interrogative pressure, F(2, 559) = 6.68, p = .001, η
p
² =
.02. Participants mentioned more false confession risk factors in the high-pressure interrogation
(M = 3.24, SE = 0.11) than in the medium- (M = 2.86, SE = 0.11) or low-pressure (M = 2.68, SE
= 0.11) interrogations, t(376) = 2.41, p = .017, d = 0.25, 95% CI [0.09, 0.40] and t(374) = 3.57, p
< .001, d = 0.37, 95% CI [0.21, 0.52], respectively.
Sample differences. There were no significant interactions between sample type and
interrogative pressure or expert testimony, ps .088, η
p
²s .01. Several main effects of sample
type did emerge, however. Students mentioned fewer false confession risk factors than did
community members (Ms = 2.58 and 3.36, SEs = 0.08 and 0.09), F(1, 550) = 40.02, p < .001, d =
0.52, 95% CI [0.40, 0.65]. Students were also more likely to convict the defendant than were
community members (50% and 30.24%), χ² (1, N = 633) = 17.62, p < .001, V = .17, 95% CI [.09,
.25].
Given that sample did not interact with our independent variables, all subsequent
analyses collapsed across sample type.
Juror integration. Verdict. We hypothesized that case-specific expert testimony would
result in a sensitivity effect, demonstrated by an interaction between expert testimony and
interrogative pressure. A 3 (interrogation pressure: low, medium, high) x 3 (expert testimony:
general, case-specific, none) hierarchical log-linear analysis revealed no significant effects of
expert testimony on participants’ verdicts. Expert testimony did not lead to a main effect (i.e.,
SENSITIVITY 20
expert-induced skepticism), χ² (2, N = 568) = 1.19, p = .552, V = .05, 95% CI [.00, .08];
convictions did not differ between the no expert (43.98%), general expert (43.23%), and case-
specific expert (38.92%) conditions (see Table 1). Expert testimony also did not interact with
interrogative pressure (i.e., expert-induced sensitivity), χ² (4, N = 568) = 1.13, p = .890, V = .05,
95% CI [.00, .06].
There was a significant main effect of interrogative pressure, however, χ² (2, N = 568) =
10.82, p = .004, V = .14, 95% CI [.03, .15], suggesting that participants were able to discriminate
between the three interrogative pressure conditions independent of expert testimony. Convictions
were higher in the low-pressure interrogation (51.58%) than in the medium- (38.54%) and high-
pressure (36.02%) interrogations, χ² (1, N = 382) = 6.56, p = .010, V = .13, 95% CI [.03, .23] and
χ² (1, N = 376) = 9.24, p = .002, V = .16, 95% CI [.06, .26], respectively. The medium- and high-
pressure interrogations did not differ, χ² (1, N = 378) = 0.26, p = .612, V = .03, 95% CI [.00, .13].
Additional analyses demonstrated that convictions were higher in the low-pressure interrogation
than in the no confession control (33.85%), χ² (1, N = 255) = 6.11, p = .013, V = .16, 95% CI
[.03, .28]. There were no differences between the no confession control and the medium- or
high-pressure interrogations, χ² (1, N = 257) = 0.46, p = .499, V = .04, 95% CI [.00, .16] and χ²
(1, N = 251) = 0.10, p = .752, V = .02, 95% CI [.00, .13], respectively.
Likelihood of commission. We analyzed likelihood of commission with a 3 (interrogative
pressure: low, medium, high) x 3 (expert testimony: general, case-specific, none) ANOVA.
Consistent with verdict, there was no main effect or interaction involving expert testimony, F(2,
559) = 1.58, p = .207, η
p
² = .01 and F(4, 559) = 1.25, p = .289, η
p
² = .01. Participants’ beliefs
about the likelihood that the defendant committed the crime did not differ between the general
SENSITIVITY 21
(M = 56.84%, SE = 2.31), case-specific (M = 59.24%, SE = 2.36), and no expert (M = 62.63%,
SE = 2.32) conditions (see Table 1).
However, there was a main effect of interrogative pressure, F(2, 559) = 12.27, p < .001,
η
p
² = .04. Participants believed it was more likely that the defendant had committed the crime in
the low-pressure interrogation (M = 68.72%, SE = 2.14) than in the medium- (M = 56.78%, SE =
2.48) or high-pressure (M = 53.14%, SE = 2.37) interrogations, t(380) = 3.65, p < .001, d = 0.37,
95% CI [0.34, 0.41] and t(374) = 4.87, p < .001, d = .50, 95% CI [0.47, 0.54], respectively. The
medium- and high-pressure interrogations did not differ, t(376) = 1.06, p = .290, d = 0.11, 95%
CI [0.08, 0.14]. Additional analyses demonstrated that likelihood of commission estimates were
higher in the low-pressure interrogation than in the no confession control (M = 55.49%, SE =
3.94), t(253) = 3.06, p = .002, d = 0.44, 95% CI [0.41, 0.48]. There were no differences between
the no confession control and the medium- or high-pressure interrogations, t(255) = 0.27, p =
.791, d = 0.04, 95% CI [-0.003, 0.08] and t(249) = 0.51, p = .612, d = 0.07, 95% CI [0.03, 0.11],
respectively.
Direct and indirect effects. We conducted a mediational path analysis with interrogative
pressure and expert testimony as the independent variables, knowledge and the perception of the
interrogation factor score as mediators, and verdict as the dependent variable. Given that no
differences emerged between the two types of expert testimony (general or case-specific), expert
testimony was collapsed into present/absent. Similarly, given the pattern of effects in verdict, the
medium- and high-pressure interrogations were combined and compared to the low-pressure
interrogation. In the model, perception of the interrogation was predicted by interrogative
pressure, expert testimony, and knowledge. Knowledge was predicted by expert testimony.
Verdict was predicted by interrogative pressure, expert testimony, perception of the
SENSITIVITY 22
interrogation, and knowledge. Model fit was acceptable (CFI = .987, RMSEA = .105, χ²/df =
7.27, p = .007) and the model explained 53% of the variance in verdict.
Figure 1 displays the significant paths observed in the model. We observed significant
direct (β = .24, 95% CI [.15, .33], p < .001) and indirect (β = -.41, 95% CI [-.47, -.34], p < .001)
effects of interrogative pressure on verdict. The indirect effect showed that the medium- and
high-pressure interrogations led to more negative perceptions of the interrogation (β = -.52, 95%
CI [-.58, -.47], p < .001), which in turn predicted a reduction in convictions (β = .78, 95% CI
[.71, .85], p < .001).
The direct effect of expert testimony on verdict was non-significant, β = .02, 95% CI [-
.07, .11], p = .662. However, expert testimony indirectly affected verdict through knowledge (β =
-.03, 95% CI [-.05, -.01], p = .003) and through knowledge and perceptions of the interrogation
(β = -.02, 95% CI [-.04, -.01], p = .001). Expert testimony increased knowledge (β = .19, 95% CI
[.10, .27], p < .001), which in turn led to fewer convictions (β = -.16, 95% CI [-.24, -.08], p <
.001). Greater knowledge also led to more negative perceptions of the interrogation (β = -.16,
95% CI [-.22, -.10], p < .001); negative perceptions in turn reduced convictions.
Finally, knowledge directly and indirectly (β = -.13, 95% CI [-.17, -.08], p < .001) led to
a reduction in convictions. Participants with greater knowledge had more negative perceptions of
the interrogation and therefore were less likely to convict.
We also conducted the path model without collapsing interrogative pressure and expert
testimony. The key difference between models lies in the indirect effect of expert testimony on
verdict. There were no indirect effects involving the comparison between general and no expert
testimony, βs -.01, ps .248. However, compared to no expert testimony, case-specific expert
testimony indirectly led to fewer convictions through knowledge (β = -.03, 95% CI [-.05, -.01], p
SENSITIVITY 23
= .009) and through knowledge and perceptions of the interrogation (β = -.02, 95% CI [-.03, -
.01], p = .006).
Discussion
We hypothesized that expert testimony would counteract the correspondence bias by
educating jurors about interrogation techniques, and that case-specific expert testimony in
particular would help jurors apply this knowledge. While both general and case-specific expert
testimony did improve knowledge about false confession risk factors, expert testimony did not
directly influence participants’ verdicts. A path analysis, however, showed that expert testimony
indirectly affected verdict through false confession knowledge and perceptions of the
interrogation, though the effects were small. Assessment of this expert testimony effect
suggested that it was driven by case-specific expert testimony.
Contrary to prior research (Kassin & Sukel, 1997), participants were generally sensitive
to false confession risk factors independent of expert testimony. Participants convicted the
defendant more often when false confession risk factors were absent (the low-pressure
interrogation) than when false confession risk factors were present (the medium- and high-
pressure interrogations). Verdicts did not differ between the medium-pressure interrogation, the
high-pressure interrogation, and the no confession control, suggesting that the majority of
participants were not influenced by the confession when it had been elicited with false
confession risk factors. Generally, participants did not appear to commit the correspondence bias
or view a confession as diagnostic of guilt.
The null effect of expert testimony is unlikely to be due to low power. A power analysis
indicated we needed 259 participants to detect a medium effect size; Experiment 1 included 633
participants. The null effects are also not likely due to a lack of variability in our dichotomous
SENSITIVITY 24
variable, verdict, given that we found the same pattern of results for our continuous variable,
likelihood of commission.
People are less likely to commit the correspondence bias and more likely to respond to
situational factors when they are aware of an alternative motivation for an actor’s behavior (Fein
et al., 1990). In Experiment 1, we sought to provide participants with an alternative motivation
for the confession via expert testimony. We had believed that participants would commit the
correspondence bias and view a confession as diagnostic of guilt in the absence of such
testimony, regardless of interrogation pressure. In Experiment 2, we presented this alternative
motivation for the confession via defendant testimony. We hypothesized that the presence of
defendant testimony would increase participants’ knowledge and lead participants to respond to
differences in interrogative pressure, relative to conditions without defendant testimony.
Experiment 2
Method
Participants. Five hundred and seven students from a southwestern university and
community members from Mechanical Turk participated in the experiment. Students (N = 298)
participated for course credit. Students were excluded for taking longer than 2 hours (n = 5), for
failing an attention check question (n = 2), or because they were not citizens of the United States
(n = 22). The final student sample was 269. Students were mostly female (72.86%) and
Hispanic/Latino (81.41%), ages 18 to 51 (M = 20.42, SE = 0.29).
Community members (N = 209) were recruited from Mechanical Turk and were
compensated $1.50 for their participation. IP addresses were restricted to the United States.
Community members were excluded for taking longer than 2 hours (n = 3), for failing an
attention check question (n = 4), because they were not citizens of the United States (n = 3), or
SENSITIVITY 25
because they had participated in Experiment 1 (n = 1). The final community sample was 198.
Community members were mostly female (62.63%) and White/Caucasian (72.22%), ages 18 to
76 (M = 37.24, SE = 0.92).
Design. A 3 (interrogative pressure: low, medium, high) x 2 (defendant testimony:
present, absent) + 1 (no confession control) between-participants design was used. The effect of
sample (student, community) was also assessed.
Materials. Participants were randomly assigned to read one of seven trial scenarios.
Scenarios with defendant testimony were identical to the scenarios without expert testimony in
Experiment 1. In the no defendant scenarios, all mentions of the defendant’s perceptions of the
interrogation and any factors that could cast doubt on the confession were removed. The defense
attorney did not provide a justification for why the suspect confessed. The interrogator did not
testify that the defendant retracted his confession after the interrogation ended. The defendant did
not testify about the interrogation or his confession; however, the defendant still denied
committing the murders.
Questionnaire. The questionnaire was identical to Experiment 1. Participants answered
six questions regarding their perceptions of the interrogation and the confession. A principal
component factor analysis was used to extract a factor score representing participants’
perceptions, with all loadings 0.65 and α = .85.
False confession knowledge was coded using the same coding scheme from Experiment
1. Coder 1 coded all the data and Coder 2 coded a subset of the data (20%, n = 94). Inter-rater
reliability was acceptable for all but two reasons (κs > 0.80). The two ratings that did not have
acceptable inter-rater reliability, “interrogative leniency” and “minimization” (κ 0.65), were
SENSITIVITY 26
excluded from analyses. Participants’ false confession reasons were summed to create one score
for each participant.
Procedure. Audio from Experiment 1 was edited to remove defendant testimony from
the appropriate conditions. All participants completed the experiment online. Procedures were
identical to Experiment 1. The trial lasted, on average, 33 minutes with the defendant’s
testimony and 29 minutes without.
Results
Manipulation checks. Participants correctly answered 97.38% (SE = 0.01) of the factual
questions about the case.
Community members had higher accuracy (M = 98.23%, SE = .01) than
did students (M = 96.75%, SE = .01), F(1, 465) = 3.94, p = .048, d = 0.19, 95% CI [0.18, 0.19].
Eighteen participants misidentified the presence or absence of the confession. Excluding these
participants did not meaningfully alter the results.
We used an ANOVA to analyze whether our manipulations affected the perceived
pressure to confess. There was a main effect of interrogative pressure, F(2, 393) = 299.61, p <
.001, η
p
² = .60. Participants indicated the defendant was under less pressure to confess in the
low-pressure interrogation (M = 3.20, SE = 0.11) than in the medium- (M = 6.34, SE = 0.12) and
high-pressure (M = 6.91, SE = 0.12) interrogations, t(264) = 15.71, p < .001, d = 1.94, 95% CI
[1.75, 2.14] and t(269) = 21.47, p < .001, d = 2.62, 95% CI [2.45, 2.79], respectively.
Participants also indicated the defendant was under less pressure in the medium-pressure
interrogation than in the high-pressure interrogation, t(259) = 5.39, p < .001, d = 0.67, 95% CI
[0.56, 0.77]. The main effect of defendant testimony and the interaction were not significant,
F(1, 393) = 0.55, p = .458, d = 0.07, 98% CI [-0.14, 0.27] and F(2, 393) = 1.63, p = .197, η
p
² =
.01, respectively.
SENSITIVITY 27
We conducted an ANOVA to assess whether our manipulations affected knowledge
about false confession risk factors. There was a main effect of defendant testimony, F(1, 392) =
5.94, p = .015, d = 0.24, 95% CI [0.10, 0.38]. Contrary to our hypothesis, participants mentioned
fewer false confession risk factors when the defendant’s testimony was present (M = 2.31 SE =
0.10) than absent (M = 2.64, SE = 0.09).
There was also a main effect of interrogative pressure on knowledge, F(2, 392) = 9.68, p
< .001, η
p
² = .05. Participants in the high-pressure interrogation (M = 2.89, SE = 0.12) mentioned
more false confession risk factors than participants in the medium- (M = 2.35, SE = 0.12) or low-
pressure (M = 2.19, SE = 0.11) interrogations, t(258) = 3.12, p = .002, d = 0.38, 95% CI [0.21,
0.55] and t(269) = 4.21, p < .001, d = 0.51, 95% CI [0.34, 0.67], respectively. The interaction
was not significant, F(2, 392) = 1.00, p = .369, η
p
² = .01.
Sample differences. There were no significant interactions between sample type and the
independent variables (interrogative pressure and defendant testimony), ps .312, η
p
²s .01,
although there were some main effects. Students mentioned fewer false confession risk factors
than did community members (Ms = 2.21 and 2.84, SEs = 0.09 and 0.11), F(1, 386) = 20.50, p <
.001, d = 0.45, 95% CI [0.31, 0.59]. Additionally, students were more likely to convict the
defendant than were community members (58% and 40.40%), χ² (1, N = 467) = 18.22, p < .001,
V = .20, 95% CI [.10, .29].
Given that sample did not interact with the independent variables, subsequent analyses
collapsed across sample type.
Juror integration. Verdict. We hypothesized that defendant testimony would lead to a
sensitivity effect. We conducted a 3 (interrogative pressure: low, medium, high) x 2 (defendant
testimony: present, absent) hierarchical loglinear analysis on participants’ verdicts. Defendant
SENSITIVITY 28
testimony did not interact with interrogative pressure (i.e., testimony-induced sensitivity), χ² (2,
N = 399) = 2.06, p = .358, V = .07, 95% CI [.00, .16] or produce a main effect on verdict (i.e.,
skepticism), χ² (1, N = 399) = 0.75, p = .385, V = .04, 95% CI [.00, .14]. Convictions did not
differ based on the presence (53.57%) or absence (49.26%) of defendant testimony (see Table 2).
Consistent with Experiment 1, there was a main effect of interrogative pressure, χ² (2, N
= 399) = 9.71, p = .008, V = .16, 95% CI [.03, .18], suggesting that the majority of participants
were capable of discriminating between the interrogative pressure conditions independent of
defendant testimony. Convictions were higher in the low-pressure interrogation (60.15%) than in
the high-pressure interrogation (41.35%), χ² (1, N = 271) = 9.57, p = .002, V = .19, 95% CI [.07,
.31]. The medium-pressure interrogation (52.34%) did not differ from the low- or high-pressure
interrogations, χ² (1, N = 266) = 1.64, p = .200, V = .08, 95% CI [.00, .20] and χ² (1, N = 261) =
3.17, p = .075, V = .11, 95% CI [.00, .22], respectively. Additional analyses demonstrated that
convictions were higher in the low-pressure interrogation than the no confession control
(45.59%), χ² (1, N = 206) = 3.91, p = .048, V = .14, 95% CI [.00, .27]. The no-confession control
did not differ from the medium- or high-pressure interrogations, χ² (1, N = 196) = 0.81, p = .368,
V = .06, 95% CI [.00, .20] and χ² (1, N = 201) = 0.33, p = .566, V = .04, 95% CI [.00, .18],
respectively.
Likelihood of commission. We analyzed likelihood of commission with a 3 (interrogative
pressure: low, medium, high) x 2 (defendant testimony: present, absent) ANOVA. Similar to
verdict, the main effect and interaction involving defendant testimony were non-significant, F(1,
393) = 0.41, p = .523, d = 0.06, 95% CI [0.03, 0.09] and F(2, 393) = 0.35, p = .702, η
p
² = .002.
Participants’ estimations of the likelihood that the defendant had committed the crime were not
SENSITIVITY 29
affected by the presence (M = 66.24%, SE = 2.21) or absence (M = 64.25%, SE = 2.25) of
defendant testimony (see Table 2).
The main effect of interrogative pressure, however, was significant, F(2, 393) = 4.49, p =
.012, η
p
² = .02. Participants believed it was more likely that the defendant had committed the
crime in the low-pressure interrogation (M = 70.43%, SE = 2.27), compared to the high-pressure
interrogation (M = 59.12%, SE = 2.95), t(269) = 3.05, p = .003, d = 0.37, 95% CI [0.34, 0.41].
The medium-pressure interrogation (M = 65.98%, SE = 2.89) did not differ from the low- or
high-pressure interrogations, t(264) = 1.22, p = .224, d = 0.15, 95% CI [0.11, 0.19] and t(259) =
1.66, p = .098, d = 0.21, 95% CI [0.17, 0.25], respectively. Additional analyses demonstrated
that the no confession control (M = 65.97%, SE = 3.29) did not differ from the low-, medium-, or
high-pressure interrogations, t(204) = 1.12, p = .264, d = 0.17, 95% CI [0.13, 0.20], t(194) <
0.01, p = .999, d < 0.01 , 95% CI [-0.04, 0.04], and t(199) = 1.44, p = .150, d = 0.22, 95% CI
[0.17, 0.26], respectively.
Direct and indirect effects. We conducted a mediational path analysis to examine the
direct and indirect effects of our interrogative pressure and defendant testimony on participants’
verdicts. The model was identical to Experiment 1 except that defendant testimony replaced
expert testimony as an independent variable. Model fit was acceptable (CFI = .978, RMSEA =
.134, χ²/df = 8.15, p = .004) and the overall model explained 64% of the variance in verdict.
Significant paths in the model are displayed in Figure 2. We observed direct (β = .23,
95% CI [.12, .33], p < .001) and indirect (β = -.38, 95% CI [-.46, -.30], p < .001) effects of
interrogative pressure on verdict. The medium- and high-pressure interrogations led to more
negative perceptions of the interrogation (β = -.48, 95% CI [-.55, -.41], p < .001), which in turn
led to a reduction in guilty verdicts (β = .79, 95% CI [.71, .88], p < .001).
SENSITIVITY 30
The direct effect of defendant testimony on verdict was non-significant (β = -.02, 95% CI
[-.11, .07], p = .690). However, defendant testimony indirectly produced small effects on verdict
by affecting knowledge (β = .02, 95% CI [.001, .05], p = .044) and by affecting knowledge and
perceptions of the interrogation (β = .03, 95% CI [.01, .06], p = .020). Defendant testimony
reduced knowledge (β = -.12, 95% CI [-.22, -.02], p = .017), which increased convictions (β = -
.20, 95% CI [-.30, -.10], p < .001). The reduction in knowledge via defendant testimony also led
to more positive perceptions of the interrogation (β = -.31, 95% CI [.38, .24], p < .001), leading
to more convictions.
The direct (β = -.20, 95% CI [-.30, -.10], p < .001) and indirect (β = -.25, 95% CI [-.31, -
.19], p < .001) effects of knowledge on verdict were also significant. Greater knowledge led to
more negative perceptions of the interrogation, and thus fewer convictions.
We observed the same pattern of effects with the model that did not collapse interrogative
pressure.
Discussion
Our hypotheses that defendant testimony would improve knowledge and lead to
sensitivity were not supported. Participants mentioned fewer false confession risk factors when
exposed to defendant testimony and there was no direct effect of defendant testimony on
participants’ verdicts. Consistent with Experiment 1, participants were generally sensitive to
variations in interrogative pressure, independent of defendant testimony. A path analysis again
demonstrated that the medium- and high-pressure interrogations led to a reduction in convictions
in part by influencing participants’ perceptions of the interrogation.
Participants generally appeared to recognize – and respond appropriately to – differences
in interrogative pressure, suggesting that they did not commit the correspondence bias. However,
SENSITIVITY 31
previous experiments have shown that people can recognize coercion in an interrogation (Kassin
& Sukel, 1997) or discount a confession (Henkel, 2008) and still believe the confessor to be
guilty. Perceptions of the interrogation and verdict cannot directly answer whether participants in
higher-pressure interrogations discounted the confession because they believed the defendant
confessed for a situational, rather than dispositional, motivation.
We conducted Experiment 3 to directly assess whether participants’ attributions for the
defendant’s motivation to confess were affected by interrogative pressure. We hypothesized that,
consistent with Experiments 1 and 2, there would be a main effect of interrogative pressure for
participants’ verdicts, but no direct effect of expert testimony. We further hypothesized that
participants in the medium- and high-pressure interrogations would believe the defendant
confessed because of a situational motivation, which would predict not-guilty verdicts. We
hypothesized that participants in the low-pressure interrogation would believe the defendant
confessed because of a dispositional motivation, which would predict guilty verdicts.
In Experiment 3 we also addressed a limitation of Experiments 1 and 2 by removing the
additive nature of the interrogation techniques, given that previous experiments did not use such
an additive manipulation (Kassin & Sukel, 1997). Because there were no interactions involving
sample type in Experiments 1 and 2, we only collected data from a student sample.
Experiment 3
Method
Participants. Students from a southwestern university participated in the experiment (N
= 389). Participants were excluded for taking longer than 2 hours to complete the experiment (n
= 9), for failing an attention check question (n = 3), or because they were not citizens of the
SENSITIVITY 32
United States (n = 24). The final sample size was 353. Participants were primarily female (62%)
and Hispanic/Latino (86.69%), ages 18 to 48 (M = 20.47, SE = 0.18).
Design. A 3 (interrogative pressure: low, medium, high) x 3 (expert testimony: general,
case-specific, none) + 1 (no confession control) between-participants design was used.
Materials. Participants were randomly assigned to 1 of 10 conditions. The materials were
identical to Experiment 1 except interrogation techniques were no longer additive. In the low-
pressure interrogation, the interrogator encouraged the defendant to tell the truth. In the medium-
pressure interrogation, the interrogator minimized the suspect’s culpability for the crime and
falsely claimed to have evidence. In the high-pressure interrogation, the interrogator waved his
gun around, threatened the defendant, and denied him breaks throughout the 7-hour
interrogation. The defendant and the defense attorney presented reasons for the defendant’s
confession that were consistent with the interrogation.
Questionnaire. The questionnaire was nearly identical to Experiments 1 and 2. One new
question was added that asked participants why they believed the defendant confessed.
Participants’ responses were coded as dispositional and/or situational motivations. A
dispositional motivation was coded as present if the defendant confessed because he committed
the crime. A situational motivation was coded as present if the defendant confessed because of
the interrogator’s behavior or as a result of the interrogator’s behavior (e.g., to avoid the death
penalty). Coder 1 coded all of the data and Coder 2 coded a subset of the data (20%, n = 70).
Interrater reliability was acceptable (κ .86).
Participants answered the same six perceptions questions as in Experiments 1 and 2. The
scale had good reliability, α = .78, and a principal component factor analysis was used to extract
a factor score (loadings .60).
SENSITIVITY 33
Participants also answered the same knowledge question as in Experiments 1 and 2.
Coder 1 coded all of the data while Coder 2 coded a subset of the data (20%, n = 70). Interrater
reliability was acceptable for the majority of reasons (κs > 0.79). One reason, “torture”, did not
have acceptable reliability (κ = 0.66) and was excluded from analyses. Participants’ responses
were summed to create one knowledge score for each participant.
Procedure. The audio from Experiment 1 was edited to remove the additive interrogation
techniques. The low-, medium-, and high-pressure interrogations lasted an average of 36, 38, and
37 minutes, respectively. The general and case-specific expert testimony conditions lasted an
average of 39 and 42 minutes, respectively. Procedures were identical to those in Experiments 1
and 2.
Results
Manipulation checks. Participants correctly answered 93.50% (SE = .01) of the factual
questions about the trial. Several participants misidentified the presence of a confession (n = 19)
or expert testimony (n = 34). Excluding these participants did not meaningfully alter the results;
therefore, all participants were included in the final analyses.
We used an ANOVA to determine whether our manipulations affected perceptions of the
pressure to confess. There was a main effect of interrogative pressure, F(2, 306) = 124.05, p <
.001, η
p
² = .45. Participants believed the defendant had been placed under more pressure to
confess in the medium- (M = 6.12, SE = 0.12) and high-pressure (M = 6.56, SE = 0.08)
interrogations, compared to the low-pressure interrogation (M = 3.68, SE = 0.20), t(209) =
10.65, p < .001, d = 1.47, 95% CI [1.25, 1.70] and t(204) = 13.31, p < .001, d = 1.87, 95% CI
[1.65, 2.08], respectively. Participants also believed the defendant had been placed under more
SENSITIVITY 34
pressure to confess in the high-pressure interrogation than in the medium-pressure interrogation,
t(211) = 3.01, p = .003, d = 0.41, 95% CI [0.27, 0.56].
The main effect of expert testimony was not significant, F(2, 306) = 2.05, p = .130, η
p
² =
.01, but there was an interaction between expert testimony and interrogative pressure, F(4, 306)
= 3.50, p = .008, η
p
² = .04. Expert testimony only affected perceptions of pressure in the low-
pressure interrogation. Participants who received general expert testimony thought there was
more pressure in the low-pressure interrogation (M = 4.40, SE = 0.30), compared to the no expert
testimony condition (M = 3.06, SE = 0.38), t(66) = 2.77, p = .007, d = 0.68, 95% CI [0.22, 1.15].
We conducted an ANOVA to examine whether our manipulations affected knowledge.
There was a main effect for expert testimony, F(2, 303) = 7.19, p = .001, η
p
² = .05. Participants
who did not receive expert testimony (M = 2.19, SE = 0.13) mentioned fewer false confession
risk factors than participants who received general (M = 2.77, SE = 0.13) or case-specific (M =
2.83, SE = 0.14) expert testimony, t(205) = 3.09, p = .002, d = 0.43, 95% CI [0.25, 0.61] and
t(206) = 3.35, p = .001, d = 0.47, 95% CI [0.28, 0.65], respectively. General and case-specific
expert testimony did not differ, t(207) = 0.32, p = .753, d = 0.04, 95% CI [-0.23, 0.14].
There was also a main effect of interrogative pressure on knowledge, F(2, 303) = 3.04, p
= .049, η
p
² = .02, though the interaction was not significant, F(4, 303) = 1.17, p = .325, η
p
² = .02.
People in the high-pressure interrogation (M = 2.87, SE = 0.14) mentioned more false confession
risk factors than people in the medium-pressure (M = 2.44, SE = 0.13) and low-pressure (M =
2.49, SE = 0.13) interrogations, t(208) = 2.18, p = .030, d = 0.30, 95% CI [0.11, 0.49] and t(204)
= 1.97, p = .050, d = 0.28, 95% CI [0.09, 0.46].
Juror integration. Verdict. We used a hierarchical loglinear analysis to examine the
effects of interrogative pressure and expert testimony on verdict. Consistent with our prior
SENSITIVITY 35
experiments, there was a main effect of interrogative pressure, χ² (2, N = 315) = 6.77, p = .034, V
= .15, 95% CI [.00, .18]. Participants were more likely to convict the defendant in the low-
pressure interrogation (56.86%) than in the medium- (41.28%) or high-pressure (42.31%)
interrogations, χ² (1, N = 211) = 5.12, p = .024, V = .16, 95% CI [.01, .29] and χ² (1, N = 206) =
4.36, p = .037, V = .15, 95% CI [.00, .28], respectively. Convictions did not differ between the
medium- or high-pressure interrogations, χ² (1, N = 213) = 0.02, p = .880, V = .01, 95% CI [.00,
.06]. Additional analyses showed that convictions in the no confession control condition
(42.10%) were not different from the low-, medium-, or high-pressure interrogations, χ² (1, N =
140) = 2.42, p = .120, V = .13, 95% CI [.00, .30], χ² (1, N = 147) = 0.01, p = .930, V = .01, 95%
CI [.00, .05], and χ² (1, N = 142) < 0.01, p = .983, V = .002, 95% CI [.00, .05], respectively (see
Table 3).
There was also a main effect of expert testimony, χ² (2, N = 315) = 10.56, p = .005, V =
.18, 95% CI [.04, .20], although the interaction was not significant, χ² (8, N = 315) = 3.03, p =
.932, V = .09, 95% CI [.00, .07]. Expert testimony led to skepticism; participants exposed to both
general (37.14%) and case-specific (44.34%) expert testimony were less likely to convict than
participants who did not receive expert testimony (58.65%), χ² (1, N = 209) = 9.69, p = .002, V =
.22, 95% CI [.08, .35] and χ² (1, N = 210) = 4.31, p = .038, V = .14, 95% CI [.00, .28],
respectively. Convictions did not differ between general and case-specific expert testimony, χ²
(1, N = 211) = 1.13, p = .287, V = .07, 95% CI [.00, .21].
Likelihood of commission. We evaluated the effect of interrogative pressure and expert
testimony on likelihood of commission with an ANOVA. Similar to verdict, there was a main
effect of interrogative pressure, F(2, 306) = 7.97, p < .001, η
p
² = .05. Participants believed it was
more likely the defendant had committed the crime in the low-pressure interrogation (M =
SENSITIVITY 36
68.28%, SE = 2.86) than in the medium- (M = 59.65%, SE = 3.11) or high-pressure (M =
51.01%, SE = 3.23) interrogations, t(209) = 2.03, p = .043, d = 0.28, 95% CI [0.24, 0.32] and
t(204) = 3.99, p < .001, d = 0.56, 95% CI [0.52, 0.60], respectively. The medium- and high-
pressure interrogations did not differ, t(211) = 1.93, p = .055, d = 0.27, 95% CI [0.22, 0.31].
Additional analyses showed that participants thought it was more likely the defendant had
committed the crime in the low-pressure interrogation, compared to the no confession control (M
= 55.18%, SE = 5.30), t(138) = 2.30, p = .023, d = 0.44, 95% CI [0.39, 0.49]. Likelihood of
commission did not differ between the no confession control and the medium- or high-pressure
interrogations, t(145) = 0.73, p = .467, d = 0.14, 95% CI [0.09, 0.19] and t(140) = 0.67, p = .504,
d = 0.13, 95% CI [0.07, 0.18], respectively (see Table 3).
The ANOVA also revealed a main effect of expert testimony, F(2, 306) = 4.78, p = .009 ,
η
p
² = .03. Participants believed it was more likely the defendant had committed the crime when
they did not receive expert testimony (M = 67.27%, SE = 3.00) compared to general (M =
55.66%, SE = 3.19) or case-specific (M = 55.95%, SE = 3.12) expert testimony, t(207) = 2.65, p
= .009, d = 0.37, 95% CI [0.33, 0.41] and t(208) = 2.62, p = .010, d = 0.36, 95% CI [0.32, 0.41],
respectively. Likelihood of commission did not differ between general and case-specific expert
testimony, t(209) = 0.07, p = .947, d = .01, 95% CI [-0.05, 0.03].
Direct and indirect effects: Verdict. We conducted a mediational path analysis with
interrogative pressure and expert testimony as predictors, knowledge and perceptions of the
interrogation as mediators, and verdict as the outcome variable. The model was identical to
Experiment 1. Model fit was good (CFI = 1.000, RMSEA = .000, χ²/df = 0.93, p = .335) and the
model explained 52% of variance in verdict. Significant paths are displayed in Figure 3.
SENSITIVITY 37
Consistent with prior experiments, interrogative pressure directly (β = .21 p = .002, 95%
CI [.08, .34]) and indirectly (β = -.39, p < .001, 95% CI [-.48, -.30]) influenced verdicts. Higher-
pressure interrogations led to more negative perceptions of the interrogation (β = -.53, 95% CI [-
.61, -.46], p < .001), in turn leading to fewer convictions (β = .74, 95% CI [.62, .85], p < .001).
Although expert testimony did not directly affect verdict (β = -.06, 95% CI [-.17, .05], p
= .301), expert testimony did produce a significant indirect effect on verdict through perceptions
of the interrogation (β = -.10, 95% CI [-.17, -.03], p = .003), knowledge (β = -.03, 95% CI [-.06, -
.002], p = .040), and perceptions of the interrogation and knowledge (β = -.03, 95% CI [-.05, -
.01], p = .005). Expert testimony led to more negative perceptions of the interrogation (β = -.14,
95% CI [-.23, -.05], p = .003), which led to fewer convictions. Expert testimony also improved
knowledge (β = 0.21, 95% CI [.10, .32], p < .001), which led to fewer convictions (β = -.15, 95%
CI [-.27, -.03], p = .015). The increase in knowledge from expert testimony also affected
perceptions of the interrogation, leading to fewer convictions (β = -.20, 95% CI [-.29, -.12], p <
.001).
Finally, knowledge directly and indirectly led to fewer convictions via perceptions of the
interrogation (β = -.15, p < .001, 95% CI [-.22, -.08]). Participants who were more
knowledgeable viewed the interrogation more negatively, and thus were less likely to convict the
defendant.
The pattern of results was similar for the model that did not collapse expert testimony and
interrogative pressure.
Direct and indirect effects: Confession motivation. We conducted a mediational path
analysis to examine whether participants’ attributions for the defendant’s motivation to confess
predicted verdict. Interrogative pressure (low- vs. medium- and high-pressure) and expert
SENSITIVITY 38
testimony (present vs. absent) were included as predictors. Dispositional and situational
confession motivations were included as mediators. Verdict was the outcome variable. Model fit
was acceptable (CFI = .900, RMSEA = .421, χ²/df = 56.62, p < .001). The model explained 99%
of the variance in verdict. Significant paths are displayed in Figure 4.
Interrogative pressure influenced verdict directly (β = .41, 95% CI [.25, .58], p < .001), as
well as indirectly through dispositional motivation (β = -.24, 95% CI [-.36, -.12], p < .001) and
through situational motivation (β = -.35, 95% CI [-.45, -.25], p < .001). Participants in the low-
pressure interrogation were more likely to mention a dispositional motivation for confession (β =
-.29, 95% CI [-.42, -.16], p < .001), which in turn led to more guilty verdicts (β = .83, 95% CI
[.73, .92], p < .001). However, participants in the medium- and high-pressure interrogations were
more likely to mention a situational motivation for confession (β = .53, 95% CI [.43, .63], p <
.001), which then led to fewer guilty verdicts (β = -.66, 95% CI [-.78, -.55], p < .001).
Expert testimony did not affect verdict directly (β = .02, 95% CI [-.15, .18], p = .850) or
indirectly through dispositional motivation (β = -.11, 95% CI [-.23, .003], p = .057). However,
expert testimony did indirectly lead to fewer convictions by affecting situational motivation to
confess (β = -.11, 95% CI [-.20, -.02], p = .014). Participants exposed to expert testimony were
more likely to mention a situational motivation for confession (β = .17, 95% CI [.04, .30], p =
.010), which in turn led to a reduction in convictions.
The same pattern of results was observed in the uncollapsed model for the indirect effect
of interrogative pressure on verdict via situational motivation. However, the indirect effect on
verdict via dispositional motivation was significant for the low- vs. high-pressure comparison (β
= -.21, 95% CI [-.34, -.07], p = .002), but not the low- vs. medium-pressure comparison (β = -
.09, 95% CI [-.23, .04], p = .165). Additionally, the indirect effect of expert testimony on verdict
SENSITIVITY 39
via situational motivation was significant for case-specific expert testimony (β = -.15, 95% CI [-
.27, -.03], p = .017), but not general expert testimony (β = -.002, 95% CI [-.11, .11], p = .972).
Discussion
Consistent with Experiments 1 and 2, participants’ verdicts and estimates of the
likelihood the defendant committed the crime were affected by interrogative pressure.
Experiment 3 also confirmed our hypothesis that interrogative pressure influences participants’
attributions for the defendant’s confession. Participants in the low-pressure interrogation
condition were more likely to convict because they attributed a dispositional motivation for the
confession (i.e., guilt). However, participants in the medium- and high-pressure interrogation
conditions were more likely to acquit because they attributed a situational motivation for the
confession (i.e., the interrogator’s behavior or a consequence of the interrogator’s behavior).
Consistent with our prior experiments, participants did not need to be explicitly provided with an
alternative motivation for the confession in order to avoid committing the correspondence bias.
Generally, participants were able to recognize that a confession could be induced by situational
factors, and participants adjusted their decisions accordingly when those factors were present.
Unlike Experiment 1, expert testimony led to a skepticism effect: participants exposed to
expert testimony were less likely to convict the defendant and believed it was less likely that he
had committed the crime, compared to no expert testimony. Similar to interrogative pressure,
expert testimony did provide a motive other than guilt for the confession; participants exposed to
expert testimony were more likely to believe the defendant confessed for a situational reason.
Assessment of this expert testimony effect suggested that it was driven by case-specific expert
testimony. However, expert testimony led participants to distrust the confession overall, rather
than enhancing participants’ pre-existing sensitivity to false confession risk factors. These results
SENSITIVITY 40
are consistent with scholars’ findings that expert testimony is likely to lead to skepticism, rather
than sensitivity, particularly when the other evidence in the case is circumstantial (Leippe, 2005;
Leippe & Eisenstadt, 2009).
General Discussion
The purpose of the current research was to evaluate sensitivity to false confession risk
factors. We hypothesized that participants’ previous inability to respond to differences in
interrogations was due to the correspondence bias (Gawronski, 2004; Gilbert & Malone, 1995) –
that participants either did not know about the influence of situational factors on false
confessions or believed a confession was diagnostic of guilt. We expected that participants
would ignore situational pressures to confess and believe the confessor to be guilty unless
participants were provided with an alternative motivation for the confession (Fein et al., 1990).
It appeared that participants did not demonstrate the correspondence bias when
evaluating a confession that was elicited via varying interrogative pressure. Across three
experiments, participants’ verdicts and likelihood of commission estimates demonstrated that
participants were generally sensitive to false confession risk factors, regardless of expert or
defendant testimony. One instance of the correspondence bias is when people recognize, but
discount, the influence of situational factors (Gawronski, 2004; Gilbert & Malone, 1995). In the
current experiments, participants appeared not only to recognize situational pressures to confess
when evaluating an interrogation – they also applied this evaluation to their verdicts. Participants
in higher pressure interrogations appeared less likely to convict because they viewed the
interrogation more negatively, and because they believed the defendant’s confession was
motivated by situational pressures, rather than guilt. We should note here that our findings do not
suggest that all participants were immune to the correspondence bias, however. Rather, our
SENSITIVITY 41
experiments indicate that participants were generally able to recognize variations in interrogative
pressure, and that participants adjusted their estimates of likelihood of commission and their
determinations of guilt accordingly.
Both expert and defendant testimony were largely ineffective. Expert testimony produced
only a small indirect effect on verdict (Exp 1) or led participants to be skeptical of the
confession, regardless of risk factors (Exp 3). Assessments of this effect suggested that it was
driven by case-specific expert testimony. Defendant testimony similarly did not influence
sensitivity (Exp 2). Although (case-specific) expert testimony and defendant testimony presented
a motivation for the confession other than guilt, neither was necessary to enable participants to
respond to situational factors that might produce a false confession.
(in)Sensitivity in Prior Research
The effects of interrogative pressure in our research are inconsistent with some prior
findings despite the fact that we used the same trial transcript as Henkel (2008) and Kassin and
Sukel (1997), with only minor variations. Although participants in prior research did discount the
confession if it had been elicited via an explicit threat (Kassin & Wrightsman, 1980, 1981) or if
the suspect confessed due to concerns about his medical disorder (Henkel, 2008), participants did
not discount the confession if it had been elicited with implied or explicit promises (Kassin &
McNall, 1991; Kassin & Wrightsman, 1980, 1981) or more coercive techniques (Henkel, 2008;
Kassin & Sukel, 1997).
Other experiments have found that people can recognize and be sensitive to interrogative
pressure, however (Horgan et al., 2012; Shpurik, 2003). People believe that a false confession is
more likely when the interrogator downplays the consequences of confessing, provides the
suspect with face-saving excuses, and minimizes the seriousness of the offense (Horgan et al.,
SENSITIVITY 42
2012). The results of our research extend those findings by demonstrating that participants’
verdicts, likelihood of commission estimates, and attributions for the confession were also
influenced by these techniques. Researchers have also recently found juror sensitivity to other
variations in confession evidence, such as inconsistencies between the confession and case facts
(Palmer et al., 2014) and incentive for an informant to provide a secondhand confession (Maeder
& Pica, 2014). Similar to the current research, Maeder and Pica’s (2014) findings were
inconsistent with prior studies (Neuschatz, Lawson, Swanner, Meissner, & Neuschatz, 2008;
Neuschatz et al., 2012), despite using similar stimuli and samples.
The recent sensitivity effects do not seem to be due to changes in methodology, given the
similarities in trial scenarios in studies that did (Maeder & Pica, 2014) and did not (Neuschatz et
al., 2008) find sensitivity. Sensitivity also does not appear to be due to the presence of certain
interrogation techniques; for example, minimization techniques have been present in studies that
did (Horgan et al., 2012) and did not (Kassin & McNall, 1991) find sensitivity. The shift in
sensitivity may instead be related to improved knowledge and integration of this knowledge.
Maeder and Pica (2014) suggest that their results may reflect jurors’ increased knowledge about
how an informant may lie in exchange for an incentive (cf. Swanner, Beike, & Cole, 2010); an
increase in knowledge could similarly account for increased sensitivity to false confession risk
factors.
The Role of Knowledge
Knowledge is an important part of sensitivity (Cutler, Penrod, et al., 1989). People
perceive interrogation techniques differently when they learn that these techniques can lead
innocent people to confess (Blandón-Gitlin et al., 2011; Forrest et al., 2012); it is possible that
participants in our research were sensitive to false confession risk factors because of improved
SENSITIVITY 43
knowledge about interrogations and false confessions. Portrayals of eyewitnesses in crime
dramas has increased since the 1980s (Desmarais, Price, & Read, 2008) and lay knowledge about
factors that influence eyewitness (mis)identification has also grown over time (Desmarais &
Read, 2011). The effects we observed in our research may be due to a similar increase in media
exposure regarding false confessions. False confessions have become a recent topic in the media,
including fictional portrayals (Grisham, 2011) and documentaries such as The Confessions
(2010), The Central Park Five (2012), and Making a Murderer (2015).
It is not just an increase in knowledge that could account for improved sensitivity to false
confession risk factors, but improved integration as well. Participants in previous experiments
have recognized coercion in interrogations, yet still convicted the confessor (Kassin & Sukel,
1997). In our research, participants were able to apply their recognition of coercion in
interrogations when they decided whether to convict the defendant. Perhaps participants in
previous experiments viewed a confession as diagnostic of guilt and assumed that only guilty
people confess. Participants may therefore have believed that situational factors were irrelevant
when determining a confessor’s guilt. It may be that people no longer perceive a confession as
diagnostic of guilt, however. People are less likely to believe a confession is a strong indicator of
guilt when they have greater media exposure to false confessions and wrongful convictions
(Henkel et al., 2008). Perhaps media exposure has affected the perceived diagnosticity of a
confession and thus improved participants’ integration of their knowledge when evaluating a
confession.
Limitations and Future Directions
While the current findings may run counter to that of previous research, the effects are
rather robust. We collected large samples of student and community members in an effort to
SENSITIVITY 44
increase the generalizability of our findings. Although the experiments were conducted online,
several steps were taken to ensure high data quality. An audio version of the testimony played in
all experiments as participants read the trial and participants could not progress to the next part
of the trial until the audio had finished. Additionally, participants were excluded if they failed at
least one attention check question, took longer than 2 hours to complete the experiment, or were
not citizens of the United States. Participants also demonstrated high accuracy for the factual
questions about the trial, indicating that they attended to the trial.
Our community samples were recruited from Mechanical Turk, which has many benefits.
Mechanical Turk participants are often internally motivated to participate in experiments
(Buhrmester, Kwang, & Gosling, 2011) and are demographically more diverse than college
students (Buhrmester et al., 2011; Crump, McDonnell, & Gureckis, 2013). Additionally, the
quality of data from Mechanical Turk is often consistent with data from laboratory experiments
(Buhrmester et al., 2011; Crump et al., 2013), but only when the participants are from native
English-speaking countries (Feitosa, Joseph, & Newman, 2015). To ensure data quality, we
restricted IP addresses to the United States and excluded all participants who were not citizens of
the United States.
However, there are some limitations in our research. Data were only collected from
individual jurors; it would be beneficial to extend the current research to deliberating juries to
improve ecological validity (see Diamond, 1997). Another limitation is that the evidence other
than the confession was circumstantial. Jurors may be more likely to trust a confession,
regardless of how it is elicited, if other evidence corroborates the defendant’s guilt.
One additional question that remains unanswered is whether participants are sensitive to
actual true and false confessions. Many participants in our research were sensitive to false
SENSITIVITY 45
confession risk factors, but it is not known whether these participants could have accurately
distinguished between true and false confessions. Discrimination accuracy is important to
investigate because it directly tests whether participants would convict guilty suspects and acquit
innocent ones (Martire & Kemp, 2011). The results of our research can only suggest whether
participants would make accurate decisions by assuming that the absence or presence of false
confession risk factors perfectly correlates with true and false confessions, but this is unlikely to
be the case (Martire & Kemp, 2011). Some false confessions occur in the absence of false
confession risk factors and true confessions can occur in the presence of false confession risk
factors (Russano et al., 2005). To further complicate the issue, false confessions appear credible
when they contain details about the crime and statements of motive (Appleby, Hasel, & Kassin,
2013) and people are likely to believe that a confession is true regardless of whether or not this is
actually the case (Kassin, Meissner, & Norwick, 2005). Therefore, it may be more difficult for
jurors to discriminate between actual true and false confessions than it is to discriminate between
interrogations with and without false confession risk factors. Additional research exploring
discrimination accuracy is needed to address this limitation in our research.
Conclusions
Researchers have previously expressed concerns about jurors’ inability to respond to
variations in interrogations (e.g., Kassin & Sukel, 1997). The majority of participants in the
current research did not demonstrate this inability. Participants were likely to trust the confession
and convict the defendant when the interrogator implored the defendant to tell the truth (low-
pressure interrogation). However, many participants discounted the confession when the
interrogator minimized the suspect’s responsibility for the crime and presented false evidence
(the medium-pressure interrogation), and when the interrogator behaved aggressively, threatened
SENSITIVITY 46
the suspect with the death penalty, and interrogated him throughout night without breaks (the
high-pressure interrogation). The current data suggest that people may be better at evaluating an
interrogation and the resulting confession than previously thought, perhaps because knowledge
about false confessions has grown over time.
SENSITIVITY 47
References
Antonio, M. E., & Arone, N. E. (2005). Damned if they do, damned if they don’t: Jurors’
reaction to defendant testimony or silence during a capital trial. Judicature, 89, 60-66.
Appleby, S. C., Hasel, L. E., & Kassin, S. M. (2013). Police-induced confessions: An empirical
analysis of their content and impact. Psychology, Crime & Law, 19, 111-128. doi:
10.1080/1068316X.2011.613389
Bar-Hillel, M. (1980). The base-rate fallacy in probability judgments. Acta Psychologica, 44,
211-233. doi: 10.1016/0001-6918(80)90046-3
Blandón-Gitlin, I., Sperry, K., & Leo, R. (2011). Jurors believe interrogation tactics are not
likely to elicit false confessions: Will expert testimony inform them otherwise?
Psychology, Crime & Law, 17, 239-260. doi: 10.1080/10683160903113699
Borgida, E., & Brekke, N. (1981). The base rate fallacy in attribution and prediction. In J. H.
Harvey, W. J. Ickes, & R. F. Kidd (Eds.), New directions in attribution research (Volume
3) (pp. 63-95). Hillsdale, N. J.: Lawrence Erlbaum.
Brekke, N., & Borgida, E. (1988). Expert psychological testimony in rape trials: A social-
cognitive analysis. Journal of Personality and Social Psychology, 55, 372-386. doi:
10.1037/0022-3514.55.3.372
Bridgeman, D. L., & Marlowe, D. (1979). Jury decision making: An empirical study based on
actual felony trials. Journal of Applied Psychology, 64, 91-98. doi: 10.1037/h0078046
Buck, J. A., London, K., & Wright, D. B. (2011). Expert testimony regarding child witnesses:
Does it sensitize jurors to forensic interview quality? Law and Human Behavior, 35, 152-
164. doi: 10.1007/s10979-010-9228-2
SENSITIVITY 48
Buhrmester, M., Kwang, T., & Gosling, S. D. (2011). Amazon’s Mechanical Turk: A new source
of inexpensive, yet high-quality data? Perspectives on Psychological Science, 6, 3-5. doi:
10.1177/1745691610393980
Costanzo, M., & Krauss, D. (2012). Forensic and legal psychology: Psychological science
applied to the law (pp. 1-28). New York: Worth.
Crump, M. J. C., McDonnell, J. V., & Gureckis, T. M. (2013). Evaluating Amazon’s Mechanical
Turk as a tool for experimental behavioral research. PLoS ONE, 8, e57410. doi:
10.1371/journal.pone.0057410
Cutler, B. L., Dexter, H. R., & Penrod, S. D. (1989). Expert testimony and jury decision making:
An empirical analysis. Behavioral Sciences and the Law, 7, 215-225.
Cutler, B. L., Penrod, S. D., & Dexter, H. R. (1989). The eyewitness, the expert psychologist,
and the jury. Law and Human Behavior, 13, 311-332. doi: 10.1007/BF01067032
Desmarais, S. L., Price, H. L., & Read, J. D. (2008). ‘Objection, your honor! Television is not
the relevant authority.’ Crime drama portrayals of eyewitness issues. Psychology, Crime
& Law, 14, 225–243. doi: 10.1080/10683160701652583
Desmarais, S. L., & Read, J. D. (2011). After 30 years, what do we know about what jurors
know? A meta-analytic review of lay knowledge regarding eyewitness factors. Law and
Human Behavior, 35, 200-210. doi: 10.1007/s10979-010-9232-6
Devenport, J. L., & Cutler, B. L. (2004). Impact of defense-only and opposing eyewitness
experts on juror judgments. Law and Human Behavior, 28, 569-576. doi:
10.1023/B:LAHU.0000046434.39181.07
Devenport, J. L., Stinson, V., Cutler, B. L., & Kravitz, D. A. (2002). How effective are the cross-
examination and expert testimony safeguards? Jurors’ perceptions of the suggestiveness
SENSITIVITY 49
and fairness of biased lineup procedures. Journal of Applied Psychology, 87, 1042-1054.
doi: 10.1037/0021-9010.87.6.1042
Diamond, S. S. (1997). Illuminations and shadows from jury simulations. Law and Human
Behavior, 21, 561–571. doi: 10.1023/A:1024831908377
Drizin, S. A., & Leo, R. A. (2004). The problem of false confessions in the post-DNA world.
North Carolina Law Review, 82, 891-1007.
Fein, S., Hilton, J. L., & Miller, D. T. (1990). Suspicion of ulterior motivation and the
correspondence bias. Journal of Personality and Social Psychology, 58, 753–764. doi:
10.1037/0022-3514.58.5.753
Feitosa, J., Joseph, D. L., & Newman, D. A. (2015). Crowdsourcing and personality
measurement equivalence: A warning about countries whose primary language is not
English. Personality and Individual Differences, 75, 47-52. doi:
10.1016/j.paid.2014.11.017
Forrest, K. D., Woody, W. D., Brady, S. E., Batterman, K. C., Stastny, B. J., & Bruns, J. A.
(2012). False-evidence ploys and interrogations: Mock jurors’ perceptions of false-
evidence ploy type, deception, coercion, and justification. Behavioral Sciences and the
Law, 30, 342-364. doi: 10.1002/bsl.1999
Fox, S. G., & Walters, H. A. (1986). The impact of general versus specific expert testimony and
eyewitness confidence upon mock juror judgment. Law and Human Behavior, 10, 215-
228. doi: 10.1007/BF01046211
Gabora, N. J., Spanos, N. P., & Joab, A. (1993). The effects of complainant age and expert
psychological testimony in a simulated child sexual abuse trial. Law and Human
Behavior, 17, 103-119. doi: 10.1007/BF01044540
SENSITIVITY 50
Garrett, B. L. (2011). Convicting the innocent: Where criminal prosecutions go wrong.
Cambridge, MA: Harvard University Press.
Gawronski, B. (2004). Theory-based bias correction in dispositional inference: The fundamental
attribution error is dead, long live the correspondence bias. European Review of Social
Psychology, 15, 183-217. doi: 10.1080/10463280440000026
Geiselman, R. E., & Mendez, B. A. (2005). Assistance to the fact finder: Eyewitness expert
testimony versus attorneys’ closing arguments. American Journal of Forensic
Psychology, 23, 5-16.
Geiselman, R. E., Putman, C., Korte, R., Shahriary, M., Jachimowicz, G., & Irzhevsky, V.
(2002). Eyewitness expert testimony and juror decisions. American Journal of Forensic
Psychology, 20, 21-36.
Gilbert, D. T., & Malone, P. S. (1995). The correspondence bias. Psychological Bulletin, 117,
21-38. doi: 10.1037/0033-2909.117.1.21
Gomes, D. M., Stenstrom, D. M., & Calvillo, D. P. (2014). Examining the judicial decision to
substitute credibility instructions for expert testimony on confessions. Legal and
Criminological Psychology. doi: 10.1111/lcrp.12068
Goodman-Delahunty, J., Cossins, A., & O’Brien, K. (2011). A comparison of expert evidence
and judicial directions to counter misconceptions in child sexual abuse trials. Australian
& New Zealand Journal of Criminology, 44, 196-217. doi: 10.1177/0004865811405140
Grisham, J. (2011). The confession. New York: Dell.
Henkel, L. A. (2008). Jurors’ reactions to recanted confessions: Do the defendant’s personal and
dispositional characteristics play a role? Psychology, Crime & Law, 14, 565 – 578. doi:
10.1080/10683160801995247
SENSITIVITY 51
Henkel, L. A., Coffman, K. A. J., & Dailey, E. M. (2008). A survey of people’s attitudes and
beliefs about false confessions. Behavioral Sciences and the Law, 26, 555-584. doi:
10.1002/bsl.826
Horgan, A. J., Russano, M. B., Meissner, C. A., & Evans, J. R. (2012). Minimization and
maximization techniques: Assessing the perceived consequences of confessing and
confession diagnosticity. Psychology, Crime, & Law, 18, 65-78. doi:
10.1080/1068316X.2011.561801
Kassin, S. M., Drizin, S. A., Grisso, T., Gudjonsson, G. H., Leo, R. A., & Redlich, A. D. (2010).
Police-induced confessions: Risk factors and recommendations. Law and Human
Behavior, 34, 3-38. doi: 10.1007/s10979-009-9188-6
Kassin, S. M., & Kiechel, K. L. (1996). The social psychology of false confessions: Compliance,
internalization, and confabulation. Psychological Science, 7, 125-128. doi:
10.1111/j.1467-9280.1996.tb00344.x
Kassin, S. M., & McNall, K. (1991). Police interrogations and confessions: Communicating
promises and threats by pragmatic implication. Law and Human Behavior, 15, 233-251.
doi: 10.1007/BF01061711
Kassin, S. M., & Neumann, K. (1997). On the power of confession evidence: An experimental
test of the fundamental difference hypothesis. Law and Human Behavior, 21, 469-484.
doi: 10.1023/A:1024871622490
Kassin, S. M., Meissner, C. A., & Norwick, R. J. (2005). “I’d know a false confession if I saw
one”: A comparative study of college students and police investigators. Law and Human
Behavior, 29, 211-227. doi: 10.1007/s10979-005-2416-9
SENSITIVITY 52
Kassin, S. M., & Sukel, H. (1997). Coerced confessions and the jury: An experimental test of the
“harmless error” rule. Law and Human Behavior, 21, 27-46. doi:
10.1023/A:1024814009769
Kassin, S. M., & Wrightsman, L. S. (1980). Prior confessions and mock juror verdicts. Journal
of Applied Social Psychology, 10, 133-146. doi: 10.1111/j.1559-1816.1980.tb00698.x
Kassin, S. M., & Wrightsman, L. S. (1981). Coerced confessions, judicial instruction, and mock
juror verdicts. Journal of Applied Social Psychology, 11, 489-506. doi: 10.1111/j.1559-
1816.1981.tb00838.x
Kovera, M. B., Gresham, A. W., Borgida, E., Gray, E., Regan, P. C. (1997). Does expert
psychological testimony inform or influence juror decision making? A social cognitive
analysis. Journal of Applied Psychology, 82, 178-191. doi: 10.1037/0021-9010.82.1.178
Laimon, R. L. (2005). Effects of expert information on individuals’ perceptions of children’s
eyewitness reports. Unpublished doctoral dissertation, Central Michigan University,
Mount Pleasant, MI.
Laub, C. E. (2010). Can earwitness limitations be overcome by the court system? Strategies to
help mock jurors appreciate the limitations of earwitness testimony. Unpublished
doctoral dissertation, University of Nebraska-Lincoln, Lincoln, NE.
Leippe, M. R. (1995). The case for expert testimony about eyewitness memory. Psychology,
Public Policy, and Law, 1, 909-959. doi: 10.1037/1076-8971.1.4.909
Leippe, M. R. & Eisenstadt, D. (2009). The influence of eyewitness expert testimony on juror's
beliefs and judgments. In B. L. Cutler (Ed.) Expert Testimony on the Psychology of
Eyewitness Identification. New York- Oxford.
SENSITIVITY 53
Leo, R. A. (2008). Police interrogation and American justice. Cambridge, MA: Harvard
University Press.
Leo, R. A., & Liu, B. (2009). What do potential jurors know about police interrogation
techniques and false confessions? Behavioral Sciences and the Law, 27, 381-399. doi:
10.1002/bsl.872
Leo, R. A., & Ofshe, R. J. (1998). The consequences of false confessions: Deprivations of
liberty and miscarriages of justice in the age of psychological interrogation. Criminal
Law and Criminology, 88, 429-496.
Lindsay, R. C. L. (1994). Expectations of eyewitness performance: Jurors' verdicts do not follow
from their beliefs. In D. F. Ross, J. D. Read, & M. P. Toglia (Eds.), Adult eyewitness
testimony: Current trends and developments (pp. 362-384). New York: Cambridge
University Press.
Maeder, E. M., & Pica, E. (2014). Secondary confessions: The influence (or lack thereof) of
incentive size and scientific expert testimony on jurors’ perceptions of informant
testimony. Law and Human Behavior. doi: 10.1037/lhb0000106
Malle, B. F., Guglielmo, S., & Monroe, A. E. (2014). A theory of blame. Psychological Inquiry,
25, 147-186. doi: 10.1080/1047840X.2014.877340
Martire, K. A., & Kemp, R. I. (2009). The impact of eyewitness expert evidence and judicial
instruction on juror ability to evaluate eyewitness testimony. Law and Human Behavior,
33, 225-236. doi: 10.1007/s10979-008-9134-z
Martire, K. A., & Kemp, R. I. (2011). Can experts help jurors to evaluate eyewitness evidence?
A review of eyewitness expert effects. Legal and Criminological Psychology, 16, 24-36.
doi: 10.1348/135532509X477225
SENSITIVITY 54
Miller, G. R., & Boster, F. J. (1977). Three images of a trial: Their implications for
psychological research. In B. D. Sales (Ed.), Psychology in the legal process (pp. 19-38).
New York: Spectrum.
Neuschatz, J. S., Lawson, D. S., Swanner, J. K., Meissner, C. A., & Neuschatz, J. S. (2008). The
effects of accomplice witnesses and jailhouse informants on jury decision making. Law
and Human Behavior, 32, 137-149. doi: 10.1007/s10979-007-9100-1
Neuschatz, J. S., Wilkinson, M. L., Goodsell, C. A., Wetmore, S. A., Quinlivan, D. S., & Jones,
N. J. (2012). Secondary confessions, expert testimony, and unreliable testimony. Journal
of Police and Criminal Psychology, 27, 179-192. doi: 10.1007/s11896-012-9102-x
Nuñez, N., Gray, J., & Buck, J. A. (2012). Educative expert testimony: A one-two punch can
affect jurors’ decisions. Journal of Applied Social Psychology, 42, 535-559. doi:
10.1111/j.1559-1816.2011.00782.x
Palmer, M. A., Button, L., Barnett, E., & Brewer, N. (2014). Inconsistencies undermine the
credibility of confession evidence. Legal and Criminological Psychology. doi:
10.1111/lcrp.12048
Perillo, J. T., & Kassin, S. M. (2011). Inside interrogation: The lie, the bluff, and false
confessions. Law and Human Behavior, 35, 327-337. doi: 10.1007/s10979-010-9244-2
Phillips, M. (2001). The National Institute of Justice guide for eyewitness evidence: Can it
improve juror decision-making? Unpublished doctoral dissertation, Florida International
University, Miami, FL.
Reeder, G. D. (1993). Trait-behavior relations and dispositional inference. Personality and
Social Psychology Bulletin, 19, 586-593. doi: 10.1177/0146167293195010
SENSITIVITY 55
Reeder, G. D., & Brewer, M. B. (1979). A schematic model of dispositional attribution in
interpersonal perception. Psychological Review, 86, 61-79. doi: 10.1037/0033-
295X.86.1.61
Russano, M. B., Meissner, C. A., Narchet, F. M., & Kassin, S. M. (2005). Investigating true and
false confessions within a novel experimental paradigm. Psychological Science, 16, 481–
486. doi: 10.1111/j.0956-7976.2005.01560.x
Shpurik, M. K. (2003). Perceptions of alibi evidence as a function of the strength of
incriminating evidence. Unpublished doctoral dissertation, Florida International
University, Miami, Florida.
Swanner, J. K., Beike, D. R., & Cole, A. T. (2010). Snitching, lies and computer crashes: An
experimental investigation of secondary confessions. Law and Human Behavior, 34, 53-
65. doi: 10.1007/s10979-008-9173-5
Wells, G. L. (1986). Expert psychological testimony: Empirical and conceptual analyses of
effects. Law and Human Behavior, 10, 83-95. doi: 10.1007/BF01044560
Woody, W. D., & Forrest, K. D. (2009). Effects of false-evidence ploys and expert testimony on
jurors’ verdicts, recommended sentences, and perceptions of confession evidence.
Behavioral Sciences and the Law, 27, 333-360. doi: 10.1002/bsl.865
Woody, W. D., Forrest, K. D., & Yendra, S. (2014). Comparing the effects of explicit and
implicit false-evidence ploys on mock jurors’ verdicts, sentencing recommendations, and
perceptions of police interrogation. Psychology, Crime & Law, 20, 603-617. doi:
10.1080/1068316X.2013.804922
SENSITIVITY 56
Table 1
Percentage of Convictions and Likelihood of Commission as a Function of Interrogative
Pressure and Expert Testimony in Experiment 1
Interrogation
Expert testimony
Convictions (n)
Likelihood of
Commission (SE)
None
None
33.85% (65)
55.49% (3.94)
Low-pressure
None
57.14% (63)
76.62% (3.43)
General
49.23% (65)
63.40% (3.93)
Case-specific
48.38% (62)
66.27% (3.54)
Medium-pressure
None
37.88% (66)
54.71% (4.46)
General
42.86% (63)
55.60% (4.42)
Case-specific
34.92% (63)
60.11% (3.98)
High-pressure
None
37.10% (62)
56.56% (4.12)
General
37.50% (64)
51.51% (4.05)
Case-specific
33.33% (60)
51.33% (4.20)
SENSITIVITY 57
Table 2
Percentage of Convictions and Likelihood of Commission as a Function of Interrogative
Pressure and Defendant Testimony in Experiment 2
Interrogation
Defendant
Testimony
Convictions (n)
Likelihood of
Commission (SE)
None
None
45.59% (68)
65.97% (3.29)
Low-pressure
Defendant
66.67% (69)
72.06% (3.16)
No defendant
53.62% (69)
68.80% (3.28)
Medium-pressure
Defendant
54.10% (61)
68.30% (4.13)
No defendant
50.75% (67)
63.86% (4.06)
High-pressure
Defendant
39.39% (66)
58.27% (4.05)
No defendant
43.28% (67)
59.96% (4.31)
SENSITIVITY 58
Table 3
Percentage of Convictions and Likelihood of Commission as a Function of Interrogative
Pressure and Expert Testimony in Experiment 3
Interrogation
Expert testimony
Convictions (n)
Likelihood of
Commission (SE)
None
None
42.11% (38)
55.18% (5.30)
Low-pressure
None
72.72% (33)
78.49% (4.32)
General
45.71% (35)
61.83% (5.41)
Case-specific
52.94% (34)
65% (4.67)
Medium-pressure
None
54.05% (37)
66.73% (5.19)
General
25.71% (35)
53.29% (5.39)
Case-specific
43.24% (37)
58.60% (5.47)
High-pressure
None
50% (34)
56.97% (5.45)
General
40% (35)
51.86% (5.78)
Case-specific
37.14% (35)
44.37% (5.49)
SENSITIVITY 59
Figure 1. The significant effects of interrogative pressure, expert testimony, and knowledge of
false confession risk factors on participants’ verdicts in Experiment 1.
SENSITIVITY 60
Figure 2. The significant effects of interrogative pressure, defendant testimony, and knowledge
of false confession risk factors on participants’ verdicts in Experiment 2.
SENSITIVITY 61
Figure 3. The significant effects of interrogative pressure, expert testimony, and knowledge of
false confession risk factors on participants’ verdicts in Experiment 3.
SENSITIVITY 62
Figure 4. The significant effects of interrogative pressure and expert testimony on perceptions of
the defendant’s motivation to confess and on verdict in Experiment 3.
... This model sought to replicate the effects of perceptions of the interrogation that have been shown to mediate the influence of interrogation tactics on key outcomes (see Woestehoff & Meissner, 2016). We assessed noncoercive (direct questioning) versus coercive (psychologically manipulative, coercive) interrogation tactics and the reason-listing task (vs. ...
... Trial Transcript. We adapted a trial transcript used in previous research (Kassin & Sukel, 1997;Woestehoff & Meissner, 2016). The transcript described the trial of an adult male defendant who was accused of murdering his wife and neighbor, with the defense claiming that the evidence was all circumstantial. ...
... The study was administered online. After providing informed consent, participants read the transcript while listening to an audio version of the transcript (the same audio was used by Woestehoff & Meissner, 2016, with new audio from the same narrator for additions to this experiment). In the trial transcript, participants read testimony from the detective and the defendant that described the interrogation as being direct questioning, psychologically manipulative, or coercive. ...
Article
Full-text available
Objective: Prior research suggests that jurors may commit the fundamental attribution error when evaluating confession evidence (i.e., failing to recognize the situational pressures inherent to coercive interrogations) and exhibit belief perseverance when presented with expert testimony or judicial instructions seeking to remediate juror knowledge. Given mixed findings regarding the use of safeguards that might assist jurors in rendering appropriate decisions, the current research examined the effectiveness of reason elaboration instructions. Hypotheses: We hypothesized that instructing mock jurors to engage in reason elaboration (Experiments 1, 2, and 4: list reasons; Experiment 3: make an initial judgment and then list reasons for the opposite of their initial belief) for why an individual might confess may help them to become more sensitive to situational and dispositional confession risk factors. We expected that reason elaboration instructions would lead to fewer convictions when a coercive interrogation was presented, but not in cases in which a non-coercive interrogation was presented (i.e., a sensitivity effect). Method: Across four experiments, jury-eligible participants (N = 1,319) read a murder trial transcript and then responded to items measuring perceived interrogation coerciveness, defendant vulnerability, and verdict decision. We manipulated interrogation approach (noncoercive vs. coercive) and reason listing for a true and/or false confession. Results: Across all four experiments, mock jurors demonstrated appropriate knowledge of false confession risk factors, and there was no interactive effect of our reason elaboration task with interrogation condition. Conclusion: Reason elaboration does not appear to be an effective safeguard for debiasing and improving sensitivity in jurors’ evaluations of confession evidence. Jurors appeared relatively proficient in distinguishing between coercive and noncoercive interrogation tactics. Future research should assess alternative approaches that can leverage mock jurors’ knowledge of appropriate risk factors and further improve their decision making.
... Yet only a minority of survey respondents (43%) were aware that interrogators can legally lie to suspects to coerce a confession (e.g., claim to have evidence of guilt that they do not really have; Chojnacki et al., 2008). Although jurors have historically been unlikely to believe that people would confess to a crime they did not commit (Henkel et al., 2008), more recent research suggests that jurors are increasingly aware of factors that contribute to false confessions (Woestehoff & Meissner, 2016). 1 Consistent with psychological and legal frameworks, coercion can be defined as "persuasive techniques that limit the suspect's autonomy by manipulating the perceived costs and benefits of possible courses of action and/or depleting the suspect's motivation or ability to resist acceding to the investigators' demands" (Kaplan et al., 2019, p. 6). ...
... Yet other evidence suggests that jurors can discount coerced confessions (Jones et al., 2021) and that media exposure to cases involving false confessions has sensitized prospective jurors to factors that contribute to unreliable confession evidence (Costanzo et al., 2010;Woestehoff & Meissner, 2016). Thus, it is unclear to what extent jurors will make various types of attributions about a coerced confession during jury deliberation. ...
... On the one hand, a good deal of evidence suggests that jurors are often insensitive to coercive elements of a confession and would in turn be likely to render internal attributions regarding confession evidence (i.e., attributing the confession to inherent guilt; Kassin et al., 2010). On the other hand, growing public awareness of factors that contribute to false confessions stemming from media influences has sensitized jurors to coercive elements of an interrogation (e.g., Henkel et al., 2008;Woestehoff & Meissner, 2016). Given conflicting evidence and the novelty of our methodological approach, this research question was exploratory. ...
Article
Full-text available
Objective: Because confessions are sometimes unreliable, it is important to understand how jurors evaluate confession evidence. We conducted a content analysis testing an attribution theory model for mock jurors' discussion of coerced confession evidence in determining verdicts. Hypotheses: We tested exploratory hypotheses regarding mock jurors' discussion of attributions and elements of the confession. We expected that jurors' prodefense statements, external attributions (attributing the confession to coercion), and uncontrollable attributions (attributing the confession to defendant naivety) would predict more prodefense than proprosecution case judgments. We also expected that being male, politically conservative, and in support of the death penalty would predict proprosecution statements and internal attributions, which in turn would predict guilty verdicts. Method: Mock jurors (N = 253, Mage = 47 years; 65% women; 88% White, 10% Black, 1% Hispanic, 1% listed "other") read a murder trial synopsis, watched an actual coerced false confession, completed case judgments, and deliberated in juries of up to 12 members. We videotaped, transcribed, and reliably coded deliberations. Results: Most mock jurors (53%) rendered a guilty verdict. Participants made more prodefense than proprosecution statements, more external than internal attributions, and more internal than uncontrollable attributions. Participants infrequently mentioned various elements of the interrogation (police coercion, contamination, promises of leniency, interrogation length) and psychological consequences for the defendant. Proprosecution statements and internal attributions predicted proprosecution case judgments. Women made more prodefense and external attribution statements than men, which in turn predicted diminished guilt. Political conservatives and death penalty proponents made more proprosecution statements and internal attributions than their counterparts, respectively, which in turn predicted greater guilt. Conclusions: Some jurors identified coercive elements of a false confession and rendered external attributions for a defendant's false confession (attributing the confession to the coercive interrogation) during deliberation. However, many jurors made internal attributions, attributing a defendant's false confession to his guilt-attributions that predicted juror and jury inclinations to convict an innocent defendant. (PsycInfo Database Record (c) 2023 APA, all rights reserved).
... true) confessions (Mindthoff et al., 2018), and experimental work shows that potential jurors are less likely to convict defendants who confessed during a coercive (vs. non-coercive) interrogation (Jones et al., 2021;Woestehoff & Meissner, 2016); however, coerced confessions (vs. no confession evidence) persistently enhance perceptions of guilt (e.g., Jones & Penrod, 2018;Kassin & McNall, 1991;Schneider & Sauerland, 2023). ...
... Despite these skepticism (or null) findings when measured by verdicts, expert testimony does appear to impart valuable knowledge to jurors. Woestehoff and Meissner (2016;Study 1) found that although expert testimony did not directly affect conviction rates, it did indirectly influence jurors' verdict decisions (albeit to a small extent) by improving jurors' knowledge of false confession risk factors. Such a finding raises the question of whether the expert testimony skepticism effect stems not from an inherent issue with expert testimony itself, but rather with jurors' inability to effectively integrate the knowledge they obtain from expert testimony (i.e., inability to appropriately apply false confession knowledge when reaching a verdict decision). ...
Article
Full-text available
Objective: Over the past 4 decades, discrepant research findings have emerged in the juror–confession literature, prompting the need for a systematic review and meta-analysis that assesses the effect of confession evidence (coerced or noncoerced) on conviction rates and the efficacy of trial safeguards. Hypotheses: We did not predict any directional hypotheses. Some studies show increased convictions when a confession is present (vs. not), regardless of whether that confession was coerced; other studies demonstrate that jurors are able to discount coerced confessions. Studies have also demonstrated sensitivity effects (safeguards aided jurors in making appropriate decisions), skepticism effects (safeguards led jurors to indiscriminately disregard confession evidence), or null effects with regard to expert testimony and jury instructions. Method: We identified 83 independent samples (N = 24,860) that met our meta-analytic inclusion criteria. Using extracted Hedges’ g effect sizes, we conducted both network meta-analysis and metaregression to address key research questions. Results: Coerced and noncoerced confessions (vs. no confession) increased convictions (network gs = 0.34 and 0.70, respectively), yet coerced (vs. noncoerced) confessions reduced convictions (network g = −0.36). When jury instructions were employed (vs. not), convictions in coerced confession cases were reduced (this difference did not emerge for noncoerced confessions; a sensitivity effect). Expert testimony, however, reduced conviction likelihood regardless of whether a confession was coerced (a skepticism effect). Conclusion: Confession evidence is persuasive, and although jurors appear to recognize the detrimental effect of coercive interrogation methods on confession reliability, they do not fully discount unreliable confessions. Educational safeguards are therefore needed, but more research is encouraged to identify the most effective forms of jury instructions and expert testimony. One potential reform could be in the interrogation room itself, as science-based interviewing approaches could provide jurors with more reliable defendant statement evidence that assists them in reaching appropriate verdict decisions.
... If it is true that people will assume that all confessions are an accurate reflection of guilt (Leo & Davis, 2010), then it stands to reason that this overriding belief will come with them into the courtroom, and they will view confession evidence through the lens of assumed truthfulness. However, the last decade has seen the rise of true-crime media focused on wrongful convictions, which may be changing juror sensitivity to risk factors for falsely confessing (Mindthoff et al., 2018;Woestehoff & Meissner, 2016). If juror attitudes towards coerced confessions are changing, they may no longer reflect the longstanding assertion that people believe that nothing short of torture would result in an innocent person confessing to a crime they did not commit (Leo & Ofshe, 1998). ...
... This effect was not seen in those who had low belief in coerced confessions. The divergent direction of judgements of guilt between the low and high belief in coerced confession groups may be further evidence that jurors are able to consider that some situations could give rise to false confessions, and that the myth of psychological interrogation (Leo & Ofshe, 1998) might be less widespread due to education about wrongful convictions (Woestehoff & Meissner, 2016). For example, if a juror has been exposed to the idea that false confessions are a real phenomenon, they might perceive the inconsistencies as an indication that the suspect is unfamiliar with the facts of the crime due to innocence. ...
... There is some evidence that the public has awareness about the false confession phenomenon. For example, Woestehoff and Meissner (2016) found participants in their three experimental studies to be sensitive to false confessions risk factors. Specifically, participants were able to recognise interrogation practices that can produce a false confession and showed reluctance to convict a confessor when a situational false confession risk factorspecifically interrogative pressurewas present. ...
Article
Published survey data on perceptions of interrogations and false confessions have come from Western European and North American countries and have tended to focus on jury-eligible citizens’ perceptions. The present study examined perceptions of police officers from the Royal Malaysia Police. Fifty-nine police officers reported their overall perceptions of confessions, perceptions of personal factors contributing to false confessions, and self-perceived likelihood of false confession. Findings reveal that the Malaysian police did not seem to be cognizant of the fact that false confessions can happen and the contributing risk factors. This emphasises the need to raise awareness in the country – both from the top-down and bottom-up.
... Recent research has indicated that juror sensitivity to coercion in interrogations is improving and their skepticism regarding false confessions is declining (Jones et al., 2021;Mindthoff et al., 2018;Woestehoff & Meissner, 2016). Some attribute this shift to increased awareness of false confessions through mass media (Pastewski, 2023). ...
Article
A confession made to the police is a crucial factor to address in mounting a competent criminal defense. Expert witnesses specializing in confession evidence are therefore an important resource for defense attorneys. Despite a strong base of empirical research on false confessions, there exist no formal guidelines for confession experts on how to most effectively proffer that social science. Through structured interviews with 20 confession experts, we probed their processes, recommendations, and the most common challenges they encounter. Some overlap of opinion was found among the confession experts we interviewed; based on that concord and our professional experience we offer some recommendations for expert witnesses. Overall, however, there was a seeming lack of consensus among confession experts regarding the processes they use and how to present their findings to the courts. These results suggest the need for greater collaboration among confession experts to identify best practices and standardize the profession.
... Hasel & Kassin, 2009). Mock jurors may be able to recognize the potential risk of false confessions (Mindthoff et al., 2018;Woestehoff & Meissner, 2016), as they are less likely to convict juvenile defendants when the police interrogators are viewed unfairly (Redlich et al., 2008) or when the juvenile confessed under coercion compared to when they voluntarily confess (Najdowski et al., 2009). However, laypeople do not always grasp the risk associated with youth status in interrogations (Henkel et al., 2008;Mindthoff et al., 2018) and have even rendered guilty verdicts at high rates for juvenile false confessors (e.g. ...
... Past studies have shown that it emerges empirically in attributions for many social problems, such as crime and carceral inequality (Grasmick & McGill, 1994;Pickett & Ryon, 2017;. Studies have also found that the distinction is useful for understanding diverse issues, such as jurors' perceptions of false confessions (Woestehoff & Meissner, 2016), civilians' perceptions of COVID-19 risk (Dunning et al., 2020), beliefs about the role of genetics in crime , and perceptions of school-aged peers' reputations (Waas & Honer, 1990). ...
Article
Why have public reactions to police misconduct been so polarized, and why have opposing social movements emerged in response? This study explores attributions of police misconduct, using a myriad of possible attributions and a population-matched national sample ( N = 700), to extend our understanding of the perceived causes of police misconduct and who holds which attributions, focusing on race, racial attitudes, and political ideology. We find that attributions could be divided into (a) multifaceted attributions—the belief that misconduct has multiple causes; and (b) excusatory attributions—the belief that misconduct is caused by factors external to police officers and agencies. Endorsement of these attributions stems from racial and political attitudes, with mediation analyses finding that race plays an indirect role in endorsing attributions of police misconduct. As such, efforts to address police misconduct face not only a political power struggle but also a racially attitudinal one.
... Jurors rarely discount confession evidence (Kassin & Wrightsman, 1981), even when the confession was coerced by police (Kassin & Sukel, 1997;Shaked-Schroer et al., 2015), made under duress (Henkel et al., 2008), was disputed (Blandon-Gitlin et al., 2010), or was contradicted by exculpatory evidence (Appleby & Kassin, 2016;Wallace & Kassin, 2011). As such, many individuals who falsely confess are convicted at trial (Appleby et al., 2011;Drizin & Leo, 2004), likely because jurors struggle to imagine why an innocent person would confess (Appleby et al., 2011;Blandon-Gitlin et al., 2010; but see Mindthoff et al., 2018;Woestehoff & Meissner, 2016). Yet, false confessions are not a new or rare phenomenon (Kassin, 2012); to-date, false confessions have contributed to 29% of the 375 wrongful convictions documented by The Innocence Project (Quinoz, 2022). ...
Article
Full-text available
Objectives This study examines how confession (versus eyewitness) evidence and guilt status impacts mock defendants’ plea decisions and perceptions of their probability of conviction (PoC) and the strength of evidence (SoE), key elements of the shadow-of-the-trial model. Methods In a simulated mock-theft scenario, adult participants (n = 239) were randomly assigned to a guilt status (guilty/innocent) and evidence-type (confession/eyewitness) condition. They were offered a plea, and perceptions of PoC and SoE were measured. Results As predicted, confession evidence led to higher rates of pleas than eyewitness evidence. Guilty participants were also more likely to accept the plea than innocent participants. However, evidence type did not impact perceptions of PoC or SoE, though guilt status did. Conclusions Our findings empirically support the impact that confessions—true or false—and guilt status have on plea decision-making. We discuss implications of these results for the shadow-of-the-trial model and the cumulative-disadvantage framework.
Article
Full-text available
Mock jurors (N = 800) viewed a videotaped trial that included information about a lineup identification procedure. Suggestiveness of the eyewitness identification procedure varied in terms of foil, instruction, and presentation biases. Expert testimony regarding the factors that influence lineup suggestiveness was also manipulated. Criteria included juror ratings of lineup suggestiveness and fairness, ratings of defendant culpability, and verdicts. Jurors were sensitive to foil bias but only minimally sensitive to instruction and presentation biases. Expert testimony enhanced juror sensitivity only to instruction bias. These results have implications for the effectiveness of cross-examination and expert testimony as safeguards against erroneous convictions resulting from mistaken identifications.
Chapter
Adult Eyewitness Testimony: Current Trends and Developments provides an overview of empirical research on eyewitness testimony and identification accuracy, covering both theory and application. The volume is organized to address three important issues. First, what are the cognitive, social and physical factors that influence the accuracy of eyewitness reports? Second, how should lineups be constructed and verbal testimony be taken to improve the chances of obtaining accurate information? And third, whose testimony should be believed? Are there differences between accurate and inaccurate witnesses, and can jurors make such a distinction? Adult Eyewitness Testimony: Current Trends and Developments is crucial reading for memory researchers, as well as police officers, judges, lawyers and other members of the judicial system.
Article
Hundreds of research papers have been written concerning the potential inaccuracy of eyewitness recollections (1-3) and several legal remedies have been proposed including eyewitness expert testimony (4). However, relatively little research has been carried out to evaluate the effectiveness of eyewitness expert testimony on juror decision making. Experiment I compared the effects of both general and specific expert testimony on juror verdicts and rationales in a mock trial scenario presenting a single eyewitness to an armed robbery and murder. Only specific expert testimony, tying the principles of eyewitness psychology to the case evidence affected juror decisions. Experiment 2 evaluated the effectiveness of presenting these specifics through attorney closing arguments. Results indicated that the adversarial closing arguments led jurors to be skeptical of the eyewitness evidence, and to thereby base their decisions on other evidence. Implications for presenting eyewitness expert testimony in a court of law are discussed.
Article
Mistaken identity has been cited often as one of the leading causes of wrongful convictions in criminal trials. Several legal remedies have been proposed including eyewitness expert testimony (1). Geiselman et al. (2) found expert testimony to improve jurors' discrimination between good and poor eyewitnessing conditions, as described in a mock trial scenario, but inclusion of attorneys' adversarial closing arguments largely eliminated the enhanced discrimination. In the present experiment, judge's instructions were included prior to the closing arguments to explain that the attorneys' arguments are not evidence. This remedy partially restored the enhanced discrimination achieved with the expert testimony alone. Without expert testimony, the closing arguments (framed by the judge's instructions) were found to nearly eliminate all juror discrimination of the eyewitness evidence in their verdicts. Implications for presenting eyewitness expert testimony in a trial are discussed.
Article
This chapter examines the influence of eyewitness expert testimony on jurors' beliefs and judgments in trials that include eyewitnesses. Most surveys indicate that prospective jurors overly believe confident eyewitnesses and have misconceptions about memory. Expert testimony aims to correct these shortcomings by educating jurors about the psychology of eyewitness reporting and factors that affect reports. Two dozen trial simulation experiments over thirty years, taken together, indicate that expert testimony produces modest gains in juror knowledge, which may improve sensitivity to evidence quality, but often do not. Expert testimony also tends to increase skepticism about prosecution eyewitnesses and, in turn, increase acquittals. Appropriately, skepticism is typically limited to cases involving central but questionable eyewitness evidence, andsalient and memorable expert testimony. Overall, eyewitness expert testimony has limited influence. Expert testimony is identified as a form of persuasive communication, and recipient, source, message, and context factors that create obstacles to persuasion are reviewed.
Article
The purpose of Study 1 was to look at how participants perceived identification testimony as a function of modality (i.e., eyewitness vs. earwitness), witnessing conditions (i.e., good vs. poor), and whether trial safeguards (i.e., expert testimony, jury instructions, closing arguments) properly sensitized participants to witnessing factors that influence witness accuracy. A total of 426 undergraduate psychology students read 1 of 17 trial transcripts, completed a pre-trial and post-trial questionnaire concerning factors that affect identification accuracy, rendered a verdict, and rated the importance of witnessing factors and the credibility of each witness. Participants were aware of factors that affect the reliability of an identification, but only used that knowledge when rendering a verdict if explicitly told of those factors (e.g., through expert testimony). Eyewitness testimony was treated the same as earwitness testimony, despite being told of higher inaccuracies with the latter. When deciding guilt, participants relied heavily on the credibility of witnesses and the victim’s confidence in her identification, which is a poor predictor of identification accuracy. The factors that affect identification accuracy did not predict the verdict. ^ The purpose of Study 2 was to complete a comprehensive legal analysis to determine the type of issues arising from earwitness testimony. A total of 226 cases were identified. The breakdown of those cases revealed that 86 (38%) dealt with procedural issues (e.g., authentication, self-incrimination), 117 (52%) with reliability issues and 35 (15%) with the use of a safeguard. In 11 (5%) cases the earwitness testimony was ruled inadmissible, and expert testimony or cautionary jury instructions were only allowed in 5 (2%) cases. This is troubling due to the high reliance jurors place on an identification. The courts claim that jurors are adequately suited to evaluate the evidence properly. If witness memory is common sense to jurors, and problematic witness testimony is detected by juries during cross-examination and closing arguments, then it is difficult to understand how misidentifications are the leading cause of wrongful convictions in the US. Unless information is presented through other legal actors, such as the introduction of expert testimony, jurors will have no reasonable basis for making their judgments.