Jury Decision Making reflection

chengxihe

Ruva2007Effectofpretrialpublicity.pdf

Home >Psychology homework help >Jury Decision Making reflection

Effects of Pre-Trial Publicity and Jury Deliberation on Juror Bias and Source Memory Errorsy

CHRISTINE RUVA1,2*, CATHY McEVOY3

and JUDITH BECKER BRYANT1

1Department of Psychology, University of South Florida, USA 2Department of Psychology, University of South Florida,

Sarasota/Manatee campus, Sarasota, Florida, USA 3School of Aging Studies, University of South Florida, USA

SUMMARY

We examined the effects of exposure to pre-trial publicity (PTP) and jury deliberation on juror memory and decision making. Mock jurors either read news articles containing negative PTP or articles unrelated to the trial. They later viewed a videotaped murder trial, after which they either made collaborative group decisions about guilt or individual decisions. Finally, all participants independently attributed specific information as having been presented during the trial or in the news articles. Exposure to PTP significantly affected guilty verdicts, sentence length, perceptions of defendant credibility, and misattributions of PTP as having been presented as trial evidence. Jury deliberation had significant effects on jury verdicts, perceptions of defendant credibility, source memory for trial items, and confidence in source memory judgements, but did not affect sentences or critical source memory errors. Copyright # 2006 John Wiley & Sons, Ltd.

Today, in the perpetually accelerating information age, the courts have the difficult duty of

protecting a defendant’s Sixth Amendment right to a ‘speedy and public trial, by an

impartial jury of the State and district wherein the crime shall have been committed’

(United States Constitution) while protecting citizens’ First Amendment rights to free

speech (freedom of the press). Research has demonstrated that one aspect of free speech,

pre-trial publicity (PTP), can bias jury decision making (Dexter, Cutler, & Moran, 1992;

Kerr, Niedermeier, & Kaplan, 1999; Kramer, Kerr, & Carroll, 1990; Ogloff & Vidmar,

1994; Otto, Penrod, & Dexter, 1994). Content analyses of media sources have found

considerable PTP that includes information the American Bar Association has categorised

as potentially prejudicial (Imrich, Mullin, & Linz, 1995; Tankard, Middleton, & Rimmer,

1978). Because biased PTP is so pervasive, the issue of whether jurors can render a verdict

in a criminal case based solely on the evidence presented during trial is very important to

the functioning of our judicial system. Failure of the courts to protect the defendant’s right

APPLIED COGNITIVE PSYCHOLOGY Appl. Cognit. Psychol. 21: 45–67 (2007)

Published online 21 June 2006 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/acp.1254

*Correspondence to: C. Ruva, Department of Psychology, University of South Florida, Sarasota/Manatee campus, 5700 North Tamiami Trial, USS 805E, Sarasota, FL 34243, USA. E-mail: ruva@shell.cas.usf.edu yThis article was published online on 21 June 2006. An error was subsequently identified and corrected by an Erratum notice that was published online only on 30 August 2006; DOI: 10.1002/acp.1292. This printed version incorporates the amendments identified by the Erratum notice.

to a fair trial, due to PTP bias, has resulted in the reversal of a number of lower courts’

decisions (see Davis, 1986, and Imrich et al., 1995, for reviews).

The courts have several remedies that can be used to combat the biasing effects of

negative PTP1: judicial instruction, voir dire (i.e. jury selection), deliberation, continuance,

change of venue, and bench trial as opposed to jury trial. Mock-jury/juror research has

examined the effectiveness of the first four and found them to be ineffective in reducing

juror bias (Dexter et al., 1992; Kramer et al., 1990). The two court remedies that have not

been studied by experimenters, change of venue and bench trial, may be the most

successful for dealing with negative PTP because change of venuewill usually eliminate or

at least reduce the PTP, and the bench trial eliminates the jury. However, the criminal

defendant does not have a constitutional right to a bench trial (Deitz & Sissman, 1984) and

a change of venue is infrequently granted (Moran &Cutler, 1991). One reason given for the

infrequency of granting a change of venue is that the courts use a different definition of

prejudice than do social scientists (Moran & Cutler, 1991; Studebaker & Penrod, 1997).

The courts define a juror as being free from prejudice if he/she reports the ability to set

aside opinion and render a verdict based solely on the evidence presented at trial (Imrich

et al., 1995). The courts have indicated that social scientists do not have an adequate

understanding of how PTP influences the thought processes of prospective jurors (Moran &

Cutler, 1991). If we do not know how PTP imparts its biasing effect on jurors, the courts

will not be able to make educated decisions on the types of remedies that may be effective

in combating PTP effects.

One way that PTP may bias a jury’s verdict is through encoding processes in which jurors

selectively attend to trial information that is consistent with their preconceived notions based

on the PTP. Hope, Memon, and McGeorge (2004) found that exposure to negative PTP

causes pre-decisional distortion in which jurors evaluate pro-prosecution evidence more

favourably than its actual probative value. They suggest that negative PTP operates as a form

of confirmation bias resulting in a bias search and evaluation of trial evidence. Research

suggests that negative PTP causes jurors to form negative impression of the defendant (Otto

et al., 1994) and that these impressions can have an extremely detrimental effect on jury

verdicts (Dexter et al., 1992; Kramer et al., 1990; Otto et al., 1994).

PTP may also bias a jury’s verdict by influencing jurors’ memory for the trial and in turn

the jury’s decision. Research has consistently demonstrated that people’s memory for an

event can be significantly influenced by misinformation presented after an event (e.g.

Loftus & Palmer, 1974) or by misinformation presented before the to-be-remembered

event (reverse suggestibility, Lindsay & Johnson, 1989; Rantzen & Markham, 1992).

Findings using the reverse suggestibility paradigm suggest that one mechanism by which

PTP influences jurors’ memories for information presented during the trial may be source

monitoring (SM) errors (Johnson, Hashtroudi, & Lindsay, 1993; Lindsay, 1994;Weingardt,

Toland, & Loftus, 1994). Jurors may wrongly believe that the PTP they were exposed to

was part of the trial testimony because they are unable to distinguish between the sources of

these two types of information. These source misattributions can arise because recollecting

information about an event and the source of that information are believed to be two

1Although all types of PTP have the potential to have biasing effects on jury decision making, this paper focuses on negative PTP. Negative PTP is biasing against the defendant and would have the tendency to push jury verdicts in the direction of guilty. This type of PTP is of special concern to the justice system because of its desire to protect the defendant against Sixth Amendment violations.

46 C. Ruva et al.

separate cognitive acts (Johnson et al., 1993), and memory performance for event

information is generally better than memory for source information (Kelly, Carroll, &

Mazzoni, 2002).

Although individual jurors may succumb to memory errors and biases, courts

believe that groups (e.g. juries) are especially good at catching the memory errors

and balancing out the biases of their members. The supposedly superior ‘error catching’

ability of collaborating groups is one reason why our court system relies on juries

to make decisions about guilt (Bourgeois, Horowitz, Fosterlee, & Graphe, 1995;

Davis, Spitzer, Nagao, & Stasser, 1978). A number of studies support this view, in that

groups produce fewer memory errors (e.g. false memories, inferential errors

and confusional errors) than do individuals (Clark, Stephenson, & Kniveton, 1990; Clark,

Stephenson, & Rutter, 1986; Hartwick, Sheppard, & Davis, 1982; Hinz, 1990;

Stephenson, Abrams, Wagner, & Wade, 1986; Vollrath, Sheppard, Hinsz, & Davis,

1989). However in other studies, collaborating groups were found to be more susceptible

to memory errors than were individuals or nominal groups2 (Basden, Basden, Bryner, &

Thomas, 1997; Basden, Basden, Thomas, & Souphasith, 1998; Sheppard, 1980, as

cited in Clark et al., 1990; Vollrath et al., 1989). Comparison among studies is difficult

because of methodological differences. One problem with past research is that group

consensus on the accuracy of an item was required before it was recorded, making

comparisons between collaborative and nominal groups less meaningful. Findings from

research in which consensus on individual memory items are required are also difficult

to apply to jury decision making because jurors must come to a consensus on their final

decision (guilty vs. not guilty), but not on each fact or argument presented during

deliberations.

Very little research has been conducted to explore the effect of collaboration on memory

errors utilising a method that does not require consensus and that uses the nominal group

method. The existing research of this type indicates that collaboration, rather than reducing

memory errors, may actually increase them. For example, Basden and colleagues,

employing Deese, Roediger, and McDermott’s (1995) paradigm, have demonstrated that

collaborating groups produce as many or more false memories than do nominal groups

(Basden et al., 1997, 1998). Basden et al. (1997) attributed the collaborative groups’ higher

number of false recalls to members feeling some obligation to contribute and therefore

lowering their response thresholds.

In summary, questions regarding how PTP and group interaction affect memory for an

event and, ultimately, decision making are important ones for those interested in jury

decision making for a number of reasons. First, PTP is so pervasive that it would be hard, if

not impossible, to seat a jury that had not been exposed to PTP about the trial they will be

deciding. Because of this, it is important to investigate whether jurors can discriminate

between information presented at trial and information presented prior to trial. If jurors

cannot discriminate between these two sources of information this is extremely

problematic for our criminal justice system. Second, jury decision making and the

2Nominal groups consist of people who work individually on a task (e.g. memory test) and whose non-redundant output is pooled and used as a measure of nominal group performance. The nominal group method consists of comparing the non-redundant output of interacting individuals (collaborative groups) with the non-redundant output of an equal number of people who worked individually on the task (nominal groups) (Basden et al., 1997).

Effects of pre-trial publicity and jury deliberation 47

deliberation process occur in a group setting in which jurors engage in collaborative

cognition and form a collaborative memory for the trial facts. The jury’s verdict depends to

a large degree on jurors’ recall (or non-recall) of trial facts and arguments during the

deliberation process. Third, the effects of group collaboration on memory and decision

making is also of interest because some previous researchers attempted to apply results of

individuals’ behaviour to group phenomena. However, group norms (e.g. ‘majority rule’),

goals, and social processes (e.g. social loafing, diffusion of responsibility) are likely to

influence what information is remembered by group members both before and after

deliberation by influencing what information is attended to during trial presentation and

what information is discussed during deliberations.

To investigate the effects of negative PTP and collaboration on memory errors and

juror bias, participants in the present study read either negative PTP or unrelated news

articles and then acted as jurors to make either group or individual decisions about

the defendant’s guilt. Four to seven days following the PTP exposure, participants viewed a

videotaped criminal trial. Half then made a group decision about guilt (collaborating

jurors), and the other half made decisions about guilt on their own (nominal jurors).

After deciding on the guilt of the defendant, all participants were given a source

monitoring (SM) test, which they completed on their own. This test included facts

that appeared only in the trial, in PTP, in unrelated news articles, or in neither trial nor

the PTP.

It was expected that jurors exposed to negative PTP would be more likely to find the

defendant guilty and less likely to find him credible. It was expected that jurors who were

exposed to negative PTP would have good memory for information presented at trial and in

the PTP, but would be more likely to misattribute information presented only in the PTP to

the trial than would non-exposed jurors.

The PTP exposure effects were expected to interact with juror collaboration on the post-

deliberation measures of guilt and on credibility ratings. Jurors in the collaborating

exposed condition were expected to be significantly more likely, after deliberation,

to render a verdict of guilty, give higher guilt ratings, suggest longer sentences, and

perceive the defendant as being less credible than all other groups. The nominal

exposed condition was expected to show the second most bias, and the two non-exposed

groups were expected to show the least bias. Past research has shown that groups are

likely to polarise individual biases held by a majority of group members (Alderton &

Frey, 1983, 1986; Davis, 1992; Kerr et al., 1999; Kramer et al., 1990; Seibold &

Meyers, 1986). Jurors exposed to negative PTP were expected to be more likely than

non-exposed jurors to have pre-discussion preferences associated with viewing

the defendant as guilty, and group discussion should polarise these individual

preferences.

Collaboration was expected to influence jurors’ confidence in their SM judgements,

with collaborating jurors indicating higher levels of confidence in both their accurate

and inaccurate SM judgements than nominal jurors. Researchers have found that

collaboration increases people’s confidence in both their accurate and inaccurate memory

judgements (Clark et al., 1990; Stephenson & Wagner, 1989: Stephenson et al., 1986).

Stephenson and colleagues have labelled collaboration’s propensity to increase group

members’ confidence in their inaccurate memory judgements as the ‘misplaced

overconfidence effect’. Because of the inconsistent finding of collaborative memory

research, no prediction was made regarding the effect of collaboration on critical source

memory errors.

48 C. Ruva et al.

METHOD

Participants

The participants consisted of 558 university students (128 males and 430 females) who

received extra course credit for participating in the experiment.3 They ranged in age from

18 to 52 years (M¼ 20.6 years). Three hundred fifty four were Caucasian, 86 were African

American, 36 were Asian or Pacific Islander, 73 were Hispanic, 2 were Native Americans,

and 2 fell into the other category. Only six indicated that they had been jurors in an actual

criminal trial prior to participating in the experiment. There were 138 participants in both

of the collaborating groups, 140 in the nominal exposed groups and 142 in the nominal

non-exposed groups. Participants were randomly assigned to one of the collaboration

conditions (collaborative vs. nominal) at the beginning of phase 2. Chi-square analyses

examining the association between PTP exposure and collaboration for all of the

demographic variables assessed (i.e. age, sex, and race) revealed no evidence of a confound

(all p> 0.19).4

Design

This experiment utilised a 2 (PTP: exposed vs. non-exposed)� 2 (collaboration:

collaborating vs. nominal) between subjects hierarchical design. There were 25 groups

(juries) per condition. The exposed jurors received negative PTP about the defendant in the

stimulus trial, and the non-exposed jurors read unrelated crime articles. The nominal

groups were randomly generated after data collection was completed and consisted of five

to six individuals who worked alone on all experimental tasks. The collaborating groups

consisted of four to six people who deliberated as a jury and made a group decision

regarding guilt of the defendant. Of our 50 groups of collaborating jurors, 68% consisted of

6 jurors, 18% consisted of 5 jurors and 14% consisted of 4 jurors. Due to the fact that using

12-person juries would make most mock jury research prohibitive, most researchers

employ four- to six-person juries. Research examining the effects of jury size on outcomes

suggests that 6-person juries are more variable in their outcomes than 12-person juries, but

has established little else (Davis, Hulbert, Au, Chen, & Zarnoth (1997); Devine, Clayton,

Dunford, Seying, & Pryce, 2001).

Stimuli

Trial

The stimulus materials consisted of a real videotaped criminal trial (NJ v Bias) that has

been used in prior research (e.g. Pritchard &Keenan, 1999, 2003 and Hope et al., 2004) and

written PTP. The videotape depicting a trial of a man who was accused of murdering his

wife was edited to run approximately 30 minutes. The defendant pled not guilty and

3Research contrasting the decisions and judgements made by college students with those of community members at large has usually shown little if any difference (Bornstein, 1999; Cutler, Penrod, & Dexter, 1990; Goodman, Golding, Helgeson, Haith, & Michelli, 1987; Isquith, 1988, cited in Bottoms & Goodman, 1994; Lindsay, 1994; Yarmey & Jones, 1983). 4Data from an additional 70 participants were not included in the analyses because they failed to follow instructions or provided incomplete data (n¼ 22; 9 responded ‘new’ to the majority of SM items and 13 had excessive missing data) or did not participate in both phases of the experiment (n¼ 48).

Effects of pre-trial publicity and jury deliberation 49

claimed that his wife accidentally shot herself when he tried to prevent her from

committing suicide by shooting herself in the head. Pilot work based on techniques used by

Brian Cutler (personal communication, May 13, 2000) and his colleagues indicated that the

trial was ambiguous as to guilt and was perceived as being realistic and believable. A trial

that is ambiguous as to guilt is more likely to be open to biasing influences than one that is

unambiguous.

Pre-trial publicity

All participants were given media packets containing news articles about crimes that

were taken from a web-based archive for the Morning Call newspaper. This Allentown,

PA, newspaper was chosen because it published a large number of stories about the

NJ v Bias murder investigation and trial. Participants in the PTP-exposed conditions

received packets containing news stories that were modified from actual PTP from the NJ v

Bias trial. These news stories contained general information about the case (e.g. victim,

when and where the crime took place, description of the crime) as well as information that

was not presented at trial and that could have a biasing effect on juror verdicts (see

Appendix A)

Participants in the non-exposed PTP conditions received packets containing actual news

articles involving an unrelated crime in which a woman was accused of embezzling child

support funds. These articles were similar in composition to the news articles in the

exposed condition (i.e. both packets contained 9 separate news articles of approximately

the same length and consisted of 10 pages of text).

Measures

Verdicts, guilt ratings and sentence length

The participants were asked for their verdicts (not guilty, hung, or guilty), how confident

they were in these verdicts (guilt ratings), and to specify the length of the prison sentence

(from a given range of 30–45 years) they would suggest if they found the defendant guilty

(see Appendix B).

Source monitoring test

Participants completed a SM test in which they indicated whether a particular statement

appeared in the experiment either as part of the trial or in one of the articles they read, and if

so, the source of the item (i.e. trial, articles, or both the trial and articles). This test

contained information presented only in the trial, information provided only in the PTP,

information provided only in the unrelated articles, and information not provided in either

the PTP or the trial (see Appendix A). Each item received a SM judgement and a

confidence rating. For the confidence judgements, participants used a 7-point Likert scale

with 1 indicating that they were not at all confident and 7 indicating they were extremely/

completely confident about their SM judgements.

Credibility ratings

A number of characteristics relevant to the credibility of the defendant were assessed using

a 7-point Likert scale (see Appendix C). This scale has been used in past research and has

been found to have high internal consistency and convergent validity (Ruva & Bryant,

2004; Ruva, Bryant, & Brannick, 1996). The internal consistency of the scale was

50 C. Ruva et al.

determined using Cronbach’s alpha. The scale had an alpha of 0.87 (item total correlations

ranged from 0.32 to 0.72) demonstrating good internal consistency.

Procedure

This experiment consisted of two phases, which were 4–7 days apart (M¼ 4.02,

SD¼ 0.80). The vast majority of participants had a 4-day delay between phase 1 and phase

2 (n¼ 509), and Chi-square tests revealed no evidence of a confound between the length of

the delay and PTP exposure or collaboration, x2 (3, N¼ 558)¼ 5.19 and 5.96, respectively,

p> 0.11. In the first phase, participants were exposed to the PTP or unrelated articles, and

in the second phase they were exposed to the trial.

First phase

During phase 1, participants were run in groups of 16 or fewer, with all participants in a

given session assigned to the same PTP condition (exposed or non-exposed). They were

told that this was a two-part study examining the stability of emotional reactions to

different types of information and thus they would need to come back the following week

for the second phase of the experiment. This cover story was provided so that the

participants would not know they were part of an experiment involving jury research.

Participants received a packet containing personality tests and a demographic

questionnaire. Once these tests were completed, participants received packets containing

the news articles and were asked to read all of the articles thoughtfully. They were then

given 15 minutes to write down as much information as they could remember from the

news articles. After recall was completed they were asked to indicate their emotional

response to reading each piece of information.

Second phase

Approximately 4 days after exposure to the articles, participants viewed the videotaped

trial. They were told that they would watch a trial and then come to a group decision about

the defendant’s guilt. They were informed that during the first phase of the study they might

have read crime stories related to the trial that they were about to view. They were told that,

like actual jurors, they were not to use any of this prior information when making decisions

about the defendant’s guilt; that they must only use the evidence presented during the trial.

The trial was then presented via television and VCR and all participants in a given session,

regardless of whether they were in the collaborating or nominal condition, watched the trial

together. Immediately following the viewing of the trial, each juror was given a verdict

form asking for a verdict, guilt ratings, and sentence length (see Appendix B). Once this

task was completed, those in the nominal group were moved to a different room from that

of the collaborating group.

The collaborating jurors were told that they had one half hour to deliberate and come to a

group decision and that they should deliberate until they reached a unanimous verdict.

They were told that if such a verdict was not reached within one half hour the jury would be

considered hung, resulting in a mistrial. They were given a jury verdict form that included

space for a group verdict, guilt ratings, and sentence length. They were informed that they

would receive a warning 5 minutes prior to the end of the deliberation period and at that

time they should finish deliberations and decide on a verdict. They were told that, because

this was a murder trial, if they reached a unanimous verdict prior to the 5-minute warning

they were to continue to discuss the case to ensure that they had made the right decision.

Effects of pre-trial publicity and jury deliberation 51

The experimenter then left the room and did not return until it was time to give the 5-minute

warning. If the jury had not reached a unanimous decision at that time they were instructed

to try their best to do so within the next 5 minutes.

The nominal jurors were asked to individually recall in writing as many trial facts as they

could. They were given 15 minutes to complete this task. They were then asked to indicate,

in the space provided on the recall sheet, which trial facts had the greatest influence on their

decision and why. They were given 10 minutes to complete this task.5 The nominal jurors

then individually completed a verdict form. This verdict form was the same as the one

completed by collaborating jurors except that it did not provide the verdict option of hung.

Once the verdict forms were collected, jurors in both the collaborating and nominal

conditions each individually completed the SM test and indicated their confidence in each

answer (see Appendix A). Participants were then asked individually to report verdicts, guilt

ratings and sentences one final time as well as their ratings of the defendant’s credibility

(see Appendix C). A debriefing questionnaire was administered in order to assess whether

the participants believed the cover story from phase one of the experiment. The majority of

both PTP exposed (83%) and non-exposed jurors (84%) indicated that they did in fact

believe the cover story, x2 (2, N¼ 558)¼ 0.21, p¼ 0.65. When the same participants were

asked whether they had a suspicion that the research involved jury decision making, 36% of

the PTP exposed jurors and 18% of the non-exposed jurors indicated that they had been

suspicious, x2 (2, N¼ 558)¼ 21.89, p< 0.01. It is not clear why the exposed and non-

exposed jurors’ responses differed on this question. Chi-squares comparing verdict 1

frequencies as a function of participants’ responses to this question and F-test comparing

the critical SM errors of these groups revealed no significant difference between them, x2

(2, N¼ 558)¼ 0.003, p¼ 0.95 and F(1, 554)¼ 1.28, MSE¼ 0.002, p¼ 0.25. Those

participants who were suspicious of our cover story responded no differently than those

who indicated that they were not suspicious. Therefore, regardless of suspicion level, all

participants were included in all of the data analyses presented below.

RESULTS

For all analyses the alpha level for significance was set at 0.05. Unless otherwise specified,

2 (PTP exposure: exposed vs. non-exposed)� 2 (collaboration: collaborating vs. nominal)

between-subjects ANOVAs were performed on each guilt measure6. Because half of the

participants in the study acted as members of a jury, a hierarchical nested design with

participants nested within groups and groups nested within PTP exposure and collaboration

conditions were employed. The nested error term was used for all of the F tests for which

there were significant effects of groups nested within PTP exposure and collaboration

conditions. This resulted in the nested error term being used for all analyses except for

those involving the pre-deliberation guilt measures. For analyses using the nested error

5The recall task made the time between trial and source monitoring test equivalent across conditions. It also allowed nominal jurors to engage in an activity that would mimic the process that a juror might go through if he/ she had participated in a group deliberation process. Of course, this task does not involve the social factors that may be present in a group decision making situation, but it did force jurors to recall facts and provide justifications to support their verdicts, as individuals are called on to do in the group deliberation situation. 6ANOVAs were used for these analyses because we were interested in examining the interactive effects of collaboration and PTP on juror/jury verdicts. Also, hierarchical ANOVAs allowed us to test for significant effects of groups nested within PTP exposure and collaboration conditions and then use the more appropriate error term (nested error term) if this effect was significant. Chi square does not allow for these types of analyses. Finally, we viewed our verdict measure as measure as assessing degree of guilt ranging from not guilty to guilty.

52 C. Ruva et al.

term the level of analysis for the nominal juries is the groupmean. Effect sizes for ANOVAs

are reported as Cohen’s (1988).

The guilt measures, which included verdicts, guilt ratings and sentence length were

completed by the jurors three times. First they completed the measures individually just

prior to deliberation (pre-deliberation measures). Second they completed these measures

just after deliberation or recall of the testimony (post-deliberation measures). The

collaborating jurors did so in a group setting, and the nominal jurors did so individually.

Finally, jurors completed these measures individually after completing the SM test (post-

source monitoring measures). The post-deliberation guilt measures were the only measures

for which individuals in a jury provided the same scores. The only exception being juries

that could not come to a unanimous verdict. Those juries were scored as hung and guilt

ratings could vary among individual jurors.

Pre-deliberation measures

As predicted, jurors in the exposed PTP conditions were significantly more likely to vote

guilty than jurors in the non-exposed conditions (see the top panel of Table 1), F(1,

458)¼ 54.89, MSE¼ 0.86, p< 0.01, d¼ 0.63. The top panel of Table 2 provides the

frequency counts of pre-deliberation verdicts by PTP and collaboration conditions.

Also as predicted, the jurors in the exposed PTP conditions gave significantly higher

guilt ratings than did jurors in the non-exposed conditions (see the middle panel of Table 1),

F(1, 458)¼ 74.48, MSE¼ 2.51, p< 0.01, d¼ 0.74. Finally, jurors in the exposed

conditions who found the defendant guilty gave significantly longer sentences than did

non-exposed jurors who found the defendant guilty (see the bottom panel of Table 1), F(1,

207)¼ 17.86, MSE¼ 31.29, p< 0.01, d¼ 0.57.

There were no significant main or interactive effects of collaboration on any of the pre-

deliberation guilt measures (see top panel of Table 2 for frequency of pre-deliberation

verdicts by PTP and collaboration conditions).

Post-deliberation and post-source monitoring measures

Because the results for the post-deliberation and post-source monitoring measures were

similar these results are presented together. The post-deliberation results are listed first for

each analysis. As predicted, jurors exposed to PTP were significantly more likely to vote

guilty, Fs (1, 96)¼ 15.49 and 20.48, MSEs¼ 2.35 and 2.03, p< 0.01, d¼ 0.60 and 0.57,

give higher guilt ratings, Fs (1, 96)¼ 39.57 and 41.84, MSEs¼ 6.42 and 5.04, p< 0.01,

d¼ 0.75 and 0.80, and give longer sentences, F(1, 57)¼ 7.90 and 19.56,MSE¼ 59.57 and

41.28, p< 0.01, d¼ 0. 92 and 0.80, than were non-exposed jurors (see Table 1). There were

no significant main or interactive effects of collaboration on any of the post-deliberation or

post-source monitoring guilt measures (see the middle and bottom panels of Table 1 and 2).

Contrary to expectations, there were no significant interactive effects of PTP exposure

and collaboration for any of the guilt measures on the separate analyses for the three testing

times. In order to examine further whether jury deliberation had an effect on the guilt

measures, three-way (PTP exposure, collaboration and time of test) mixed-design

hierarchical ANOVAs were performed with the pre-deliberation and post-source

monitoring data on verdicts, guilt ratings and sentences as the within-subjects variable.

On the verdict measure, there was a significant effect of time of test and a significant

interaction between time of test and collaboration, F(1, 96)¼ 11.86 and 6.63,

Effects of pre-trial publicity and jury deliberation 53

T ab le

1 . M ea n s fo r th e g u il t m ea su re s

P re -d el ib er at io n

P o st -d el ib er at io n

P o st -s o u rc e m o n it o ri n g

E x p o se d

N o n -e x p o se d

M ea n

E x p o se d

N o n -e x p o se d

M ea n

E x p o se d

N o n -e x p o se d

M ea n

V er d ic ts a

C o ll ab o ra ti n g

2 .4 6 (0 .8 9 )b

1 .8 3 (0 .9 9 )

2 .1 5 (0 .9 9 )

2 .2 2 (0 .8 4 )

1 .7 2 (0 .7 2 )

1 .9 7 (0 .8 2 )

2 .2 5 (0 .9 7 )

1 .6 7 (0 .9 4 )

1 .9 5 (1 .0 0 )

N o m in al

2 .3 0 (0 .9 6 )

1 .7 4 (0 .9 7 )

2 .0 2 (1 .0 0 )

2 .2 7 (0 .9 7 )

1 .7 1 (0 .9 6 )

1 .9 9 (1 .0 0 )

2 .2 7 (0 .9 7 )

1 .7 1 (0 .9 6 )

1 .9 9 (1 .0 0 )

M ea n

2 .3 8 (0 .9 3 )

1 .7 8 (0 .9 8 )

2 .2 5 (0 .9 0 )

1 .7 2 (0 .8 5 )

2 .2 5 (0 .9 7 )

1 .6 9 (0 .9 5 )

G u il t ra ti n g sc

C o ll ab o ra ti n g

5 .2 9 (1 .6 5 )

4 .0 8 (1 .6 7 )

4 .6 9 (1 .7 7 )

5 .6 3 (1 .3 0 )

3 .9 6 (1 .2 7 )

4 .7 9 (1 .5 3 )

5 .4 5 (1 .4 4 )

4 .0 4 (1 .5 3 )

4 .7 5 (1 .6 4 )

N o m in al

5 .0 5 (1 .5 0 )

3 .8 9 (1 .6 1 )

4 .4 7 (1 .6 6 )

5 .0 7 (1 .6 1 )

3 .9 9 (1 .7 1 )

4 .5 4 (1 .7 4 )

5 .0 9 (1 .6 0 )

3 .9 9 (1 .6 8 )

4 .5 4 (1 .7 3 )

M ea n

5 .1 7 (1 .5 8 )

3 .9 8 (1 .6 4 )

5 .3 4 (1 .4 9 )

3 .9 7 (1 .5 0 )

5 .2 7 (1 .5 3 )

4 .0 1 (1 .6 1 )

S en te n ce

L en g th

C o ll ab o ra ti n g

4 0 .7 1 (5 .2 4 )

3 7 .7 4 (5 .7 5 )

3 9 .6 4 (5 .5 9 )

4 0 .6 4 (4 .8 9 )

3 6 .9 1 (5 .7 5 )

3 9 .7 2 (5 .3 3 )

4 1 .0 2 (5 .2 2 )

3 5 .8 0 (6 .0 1 )

3 9 .2 0 (6 .0 3 )

N o m in al

4 0 .3 7 (5 .4 2 )

3 7 .6 9 (5 .9 4 )

3 9 .4 0 (5 .7 4 )

4 0 .7 4 (5 .4 1 )

3 8 .1 0 (5 .6 4 )

3 9 .8 0 (5 .6 1 )

4 0 .7 4 (5 .4 1 )

3 8 .1 4 (5 .4 5 )

3 9 .8 1 (5 .5 5 )

M ea n

4 0 .5 5 (5 .3 1 )

3 7 .7 2 (5 .8 2 )

4 0 .7 0 (5 .1 8 )

3 7 .7 4 (5 .6 6 )

4 0 .8 8 (5 .3 0 )

3 7 .0 2 (5 .8 1 )

a T h e fo ll o w in g sc al e w as

u se d to

co d e v er d ic ts : 1 ¼ n o t g u il ty , 2 ¼ h u n g , 3 ¼ g u il ty . T h e h u n g o p ti o n w as

o n ly

av ai la b le

fo r ju ro rs

in th e co ll ab o ra ti v e co n d it io n s w h en

d ec id in g

v er d ic t 2 .

b S ta n d ar d d ev ia ti o n s o r st an d ar d er ro rs

(s en te n ce

le n g th

v ar ia b le ) ap p ea r in

p ar en th es es .

c T h e fo ll o w in g sc al e w as

u se d fo r g u il t ra ti n g s:

1 ¼ h ig h co n fi d en ce

n o t g u il ty , 4 ¼ u n su re

w h et h er

d ef en d an t is

g u il ty

o r n o t g u il ty , 7 ¼ h ig h co n fi d en ce

g u il ty .

d O n ly se n te n ce

d at a fr o m p ar ti ci p an ts w h o v o te d g u il ty ar e in cl u d ed

in th is ta b le b ec au se

th ey

ar e th e o n ly o n es

w h o m ad e th es e ra ti n g s. T h e se n te n ce

le n g th o p ti o n s ra n g ed

fr o m 3 0 to

4 5 y ea rs .

54 C. Ruva et al.

MSE¼ 0.285, respectively. Overall, jurors were less likely to find the defendant guilty on

the post-source monitoring measures than on the pre-deliberation measures (see the top

panel of Table 1 and the top and bottom panels of Table 2), but this effect was only apparent

for the jurors in the collaborating conditions. This finding suggests a leniency shift in which

deliberation may heighten awareness of the defendant protection norm and highlight the

reasonable doubt standard that acquitting the guilty is less serious an error than convicting

the innocent (Kerr, 1993). Regardless of time of test, PTP had a significant effect on

verdicts with jurors in the exposed conditions being significantly more likely to find the

defendant guilty than jurors in the non-exposed conditions, F(1, 96)¼ 30.65,MSE¼ 2.89,

p< 0.01. The interaction between time of test and PTP exposure was not significant, F(1,

96)¼ 0.30,MSE¼ 0.285, p¼ 0.51, suggesting that PTP exposure does not contribute to the

time of test effect.

Source monitoring test, critical source errors

The critical source monitoring errors of interest were those in which the participants

misattributed information contained only in the PTP as appearing in the trial or in both the

trial and the PTP (Zaragoza & Lane, 1994). Corrected error scores were calculated by

taking each participant’s initial error score and subtracting from this the proportion of new

items (those not presented in either the trial or the PTP) the participant identified as being

either in the trial or in both the trial and the PTP (see Table 3). Then two-variable (PTP

exposure� collaboration) between-subjects hierarchical ANOVAs were performed on the

corrected error scores in order to test the prediction that PTP exposure would have a

Table 2. Frequency counts and percentages for verdicts

Condition Not Guilty Hung Guilty Total

Pre-deliberation verdicts Collaboration� PTP Collaborating exposed 37 (27%)a NAb 101 (73%) 138 Collaborating non-exposed 81 (59%) NA 57 (41%) 138 Nominal exposed 50 (35%) NA 92 (65 %) 142 Nominal non-exposed 88 (63%) NA 52 (37%) 140

Total 256 (46%) NA 302 (54%) 558 (100%) Post-deliberation verdicts

Collaboration� PTP Collaborating Exposed 36 (26%) 35 (25%) 67 (49%) 138 Collaborating Non-exposed 60 (43%) 56 (41%) 22 (16%) 138 Nominal Exposed 52 (37%) NA 90 (63%) 142 Nominal Non-exposed 90 (64%) NA 50 (36%) 140

Total 238 (43%) 91 (16%) 229 (41%) 558 (100%) Post-source monitoring verdicts

Collaboration� PTP Collaborating exposed 52 (38%) NA 86 (62%) 138 Collaborating non-exposed 92 (67%) NA 46 (33%) 138 Nominal exposed 52 (37%) NA 90 (63%) 142 Nominal non-exposed 90 (64%) NA 50 (36%) 140

Total 286 (51%) NA 272 (49%) 558 (100%)

aRow percentages appear in parentheses beside their respective means. bOnly collaborating jurors could render a hung verdict (null verdict) during the group decision. Row percentages appear in parentheses beside their respective means.

Effects of pre-trial publicity and jury deliberation 55

significant effect on source monitoring errors. The nested error term was used for all

analyses. As can be seen in Table 3, the jurors whowere exposed to PTP made significantly

more critical source monitoring errors (M¼ 0.15, SD¼ 0.13) than did non-exposed jurors

(M¼ 0.10, SD¼ 0.12), F(1, 96)¼ 17.07,MSE¼ 0.02, p< 0.01, d¼ 0.40. That is, jurors in

the exposed conditions were significantly more likely than jurors in the non-exposed

conditions to attribute information presented only in the PTP to either the trial or both the

trial and the PTP. Critical source monitoring errors were also significantly related to

verdicts and guilt ratings, r(558)¼ 0.10 and 0.17, p< 0.05 and 0.01, respectively. These

associations are explored in the mediation analyses presented below. Neither the main

effect of collaboration nor the interaction of collaboration and PTP exposure were

significant, Fs(1, 96)¼ 0.49 and 0.015, MSE¼ 0.02, p> 0.60), respectively.

Source monitoring test, correct judgements of trial items

In order to examine differences among groups in the proportion of correct SM judgements

for trial items, a two-variable (PTP exposure� collaboration) between-subjects

hierarchical ANOVAwas performed on trial items (corrected for guessing: the proportion

of new items the participant identified as being in the trial was subtracted from the

participant’s initial correct judgement score) and the individual error term was used. Jurors

in the non-exposed conditions accurately identified significantly more of the trial items as

coming from the trial (M¼ 0.75, SD¼ 0.17) than did the exposed jurors (M¼ 0.63,

SD¼ 0.17), F(1, 457)¼ 76.56,MSE¼ 0.03, p< 0.05, d¼ 0.71. There was also a small but

significant main effect of collaboration, with collaborating jurors providing more accurate

Table 3. Mean proportions of source monitoring responses as a function of PTP exposure and collaboration

Response

Actual Source

Trial Pre-trial publicity New

Collaborating exposed jurors Trial 0.67 0.04 0.01 Pre-trial publicity 0.05 0.57 0.22 Both 0.15 0.14 0.01 New 0.14 0.25 0.76

Collaborating non-exposed jurors Trial 0.79 0.13 0.03 Pre-trial publicity 0.00 0.00 0.00 Both 0.00 0.00 0.00 New 0.21 0.87 0.97

Nominal exposed jurors Trial 0.63 0.04 0.03 Pre-trial publicity 0.05 0.56 0.18 Both 0.11 0.15 0.02 New 0.21 0.25 0.77

Nominal non-exposed jurors Trial 0.77 0.13 0.03 Pre-trial publicity 0.00 0.00 0.00 Both 0.00 0.00 0.00 New 0.22 0.87 0.96

56 C. Ruva et al.

responses (M¼ 0.70, SD¼ 0.16) than nominal jurors (M¼ 0.67, SD¼ 0.20), F(1,

457)¼ 6.36, MSE¼ 0.028, p< 0.01, d¼ 0.17. The interaction was not significant, F(1,

457)¼ 1.11, MSE¼ 0.028, p¼ 0.24.

Confidence in source monitoring judgements

In order to test the prediction that collaboration would have an effect on participants’

confidence in their SM judgements, two-variable (PTP exposure� collaboration) between-

subjects hierarchical ANOVAwere performed on the SM items. The nested error term was

used for all analyses. The means for the confidence ratings of participants’ SM judgements

are presented in Table 4.

As expected, jurors in the collaborating conditions were more confident than nominal

jurors about both their accurate and inaccurate SM judgements for the trial items, Fs(1,

96)¼ 5.15 and 5.44,MSEs¼ 0.57 and 1.65, p< 0.05, ds¼ 0.20 and 0.08, respectively. PTP

exposure did not have a significant effect on these confidence judgements, Fs(1, 96)¼ 3.03

and 4.75, MSEs¼ 0.57 and 1.65, respectively, p> 0.06. There was no significant effect of

collaboration or PTP exposure on confidence ratings for the critical SM errors, Fs(1,

96)¼ 1.07 and 1.43, MSE¼ 1.22, respectively, p> 0.18.

Credibility measures

To test the effects of PTP exposure and collaboration on credibility ratings of the defendant,

a two-variable (PTP exposure� collaboration) between-subjects hierarchical ANOVAwas

performed on credibility ratings of the defendant and the nested error term was used for all

analyses. As predicted, the jurors exposed to PTP perceived the defendant as less credible

(M¼ 40.83, SD¼ 10.58) than jurors in the non-exposed conditions (M¼ 49.57,

SD¼ 12.98), F(1, 96)¼ 41.48, MSE¼ 114.16, p< 0.01, d¼ 0.74. There was also a

significant effect of collaboration on perceived credibility of the defendant, F(1,

96)¼ 8.71, MSE¼ 114.16, p< 0.01, d¼ 0.32, with nominal jurors perceiving the

defendant as more credible (M¼ 47.13, SD¼ 12.31) than did collaborating jurors

(M¼ 43.20, SD¼ 12.62).

Mediation analyses

One of the main purposes for this research was to explore how PTP imparts its biasing

effect on jury decision making. It was hypothesise that both critical SM errors and

perceived credibility of the defendant may be relevant mechanisms. The above analyses

suggest a link between PTP exposure and critical PTP source memory errors and perceived

Table 4. Mean confidence ratings for correctly and incorrectly recognised trial and critical source memory errors

Response

Condition

Collaborating Nominal

Trial correct 6.38 (0.59) 6.25 (0.70) Trial incorrect 5.56 (1.19) 5.33 (1.28) Critical source errors 5.81 (0.94) 5.73 (1.06)

Note: Standard deviations are presented in parentheses following their respective means. Critical source errors include items that were incorrectly attributed to the trial or both the trial and PTP.

Effects of pre-trial publicity and jury deliberation 57

credibility of the defendant. They also suggest a link between PTP exposure and guilt

ratings. In order to explore these relationships both critical source memory errors and

perceived credibility of the defendant, we treated as mediation variables in two mediation

models of X!M!Y (where X is PTP, Y is post-source monitoring test guilt ratings andM is

the mediating variable) (Baron & Kenny, 1986; Shrout & Bolger, 2002). As Baron and

Kenny (1986) suggested, we constructed three regression equations for both models. For

the model with critical SM errors the first equation related guilt ratings to PTP (X!Y) and

was statistically significant, b¼ 1.25, t(556)¼ 9.44, p< 0.01. The second related critical

SM errors to PTP (X!M) and was statistically significant, b¼�0.05, t(556)¼ 4.6,

p< 0.01. The third equation related the guilt ratings to both PTP and critical SM errors and

indicated that both critical SM errors and PTP were significant, b¼ 1.28 and 1.19,

t(555)¼ 2.45 and 8.84, respectively, p< 0.05. The indirect effect of PTP on the guilt

ratings via the mediator (critical SM errors) was significant, Sobel (1982) test z¼ 2.17.

Therefore, the effect of PTP on guilt ratings does appear to be mediated by critical SM

errors.

Another mediation analysis was performed to test the model of X!M!Y (where X is

PTP, Y is post-source monitoring guilt ratings, and M is perceived credibility of the

defendant) (Baron & Kenny, 1986; Shrout & Bolger, 2002). The first regression equations

related guilt ratings to PTP (X!Y) and was statistically significant, b¼ 1.25, t(556)¼ 9.44,

p< 0.01. The second related perceived credibility of the defendant to PTP (X!M) and was

statistically significant, b¼�8.73, t(556)¼�8.71, p< 0.01. The third related the guilt

ratings to both PTP and critical SM errors and indicated that both critical SM errors and

PTP were significant, b¼�0.09 and 0.50, t(555)¼�20.07 and 4.67, respectively,

p< 0.01. The indirect effect of PTP on the guilt ratings via the mediator (perceived

credibility of the defendant) was significant, Sobel (1982) test z¼ 2.44. Overall, the

mediation analyses suggest that the effect of PTP on guilt ratings was mediated by both

perceived credibility of the defendant and critical SM errors.

GENERAL DISCUSSION

Consistent with previous research, the results of the experiment indicate that negative PTP

about a defendant can have an extremely biasing effect on juror decision making. They also

suggest that the effects of deliberation depend on how juror bias is measured (e.g. verdicts,

credibility ratings of the defendant, SM misattributions). For example, we found that jury

deliberation had no significant effect on post-deliberation measures of verdicts, guilt

ratings, or prison sentences. In contrast, the comparison of pre-deliberation and post-source

monitoring verdicts suggested that group deliberation attenuated the bias associated with

negative PTP. Such attenuation has been termed a ‘leniency shift’. Deliberations

presumably highlight the law’s reasonable doubt standard and the fact that the conviction

of the innocent should be avoided (Kerr, 1993). Although attenuation was found for

individual and not group verdicts, it does provide some reassurance for a criminal justice

system that relies on juries to eliminate or at least reduce individual biases.

Deliberation was also found to increase juror bias in that jurors who deliberated viewed

the defendant as less credible than jurors who did not regardless of PTP exposure. This is

consistent with past research that has shown that groups are likely to polarise individual

biases held by a majority of group members (Davis, 1992; Davis et al., 1978; Kerr et al.,

1999). Our research suggests that this type of bias is especially important because

58 C. Ruva et al.

perceived credibility of the defendant mediated the effect of PTP on guilt ratings and

explained nearly 40% of the variance in juror verdicts.

One purported benefit of utilising juries is that they are especially good at catching their

members’ memory errors. Does the deliberation process reduce memory errors? The

research presented here cannot directly answer this question because the jurors completed

the SM test individually. It can address whether prior collaboration, involving discussion of

information contained on an SM test, reduces the number of critical source monitoring

errors. We found no difference in the proportion of critical SM errors for jurors in the

collaborating and nominal conditions, but we did find collaborating jurors provided more

accurate SM responses than did nominal jurors. In contrast, Bourgeois et al. (1995) found

nominal jurors’ memory performance to be superior to that of collaborating jurors and

attributed this difference to social loafing. They varied the instructions between their

collaborating and nominal jurors informing the former that they would make a group

decision and the latter that they would make an individual one. All of the jurors in the

present study were told that they would deliberate as a group and then make a group

decision regarding the defendant’s guilt. Therefore, if such instructions promote social

loafing it would have been equivalent across all of our conditions and may be responsible

for the null effect of collaboration. We kept jury instructions consistent across the

conditions so as not to confound instructions with collaboration and to allow random

assignment of jurors to nominal and collaborative conditions.

It is hard to compare the current research to the collaborative memory studies in which

individuals work as a group to complete a memory test because our collaborating jurors did

not collaborate on the SM test. Instead, they collaborated to come to a group decision

regarding guilt and then completed the SM test individually. Thus, the individual memory

test assesses individual jurors’ memory performance after collaboration. This method is

more applicable to the criminal justice system than are typical methods employed to assess

collaborative memory (e.g. consensus, turn-taking, or no instructions on how to resolve

disputed group memories) for two reasons. First, real jurors are not required to come to a

consensus on each fact presented during deliberation: they only need to come to a

unanimous decision regarding guilt. Second, just because a fact is mentioned during

deliberations (as in the turn-taking strategy) does not mean that it now becomes part of the

jurors’ collective memory for what occurred during trial or that it will be used by jurors

when deciding on a verdict.

We also explored the effect of deliberation on juror confidence in their source memory

judgements. We found that deliberation did not have a significant effect on confidence

about critical source memory judgements, but it did significantly affect confidence ratings

about trial items with collaborating jurors indicating more confidence in both their accurate

and inaccurate source memory judgements. Why the confidence results differed between

critical source items and trial items is not clear. It could be that jurors spent the majority of

their deliberation time discussing trial items. After all, these items should have been highly

accessible because jurors viewed the trial just prior to deliberations. Our confidence

findings are consistent with past findings that collaboration increases people’s confidence

in both their accurate and inaccurate memory judgements (Clark et al., 1990; Stephenson &

Wagner, 1989: Stephenson et al., 1986). These findings are noteworthy because past

research has shown that juror participation in deliberations is related juror confidence

(Kassin & Wrightsman, 1988) and because juror confidence is not predictive of juror

memory (Pritchard & Keenan, 1999) the jurors who participate most in jury deliberations

are not necessarily the most accurate (Pritchard & Keenan, 2002). Thus, jury deliberations

Effects of pre-trial publicity and jury deliberation 59

can bolster jurors’ confidence in their inaccurate memories and thereby increase the

discussion of inaccurate information. Increased discussion makes it more likely that

inaccurate information will become part of the collaborative memory of the trial.

As noted above, contrary to our predictions collaboration did not have a significant effect

on our post-deliberation measures of guilt. One reason for the null effect of collaboration

on these measures may be the fact that we limited jury deliberations to 30 minutes, and

those juries that could not come to a unanimous verdict in this time period were considered

hung. Second, the method we employed also may have reduced the differences between

nominal and collaborating jury verdicts. Whereas the collaborating jurors deliberated as a

group, the nominal jurors deliberated individually, recalling facts from the trial and

indicating which facts had the greatest impact on their decisions and why. In traditional

mock-juror research, jurors are exposed to the trial and then immediately make decisions

about guilt. Note that our recall task made the time between trial and SM test equivalent

across conditions. It also allowed nominal jurors to engage in an activity that would mimic

the process that jurors might go through if they had participated in a group deliberation

process. That is, it forced nominal jurors to provide facts and justification to support their

verdicts, as individuals are called on to do in the group deliberation situation. Much

research in the area of jury decision making is criticised because it does not employ

interacting jurors who collaborate to come to a decision. The propensity to conduct

research on individuals rather than groups is most likely driven by the large numbers of

participants and logistical arrangements required for group designs. Further exploration

into the method we employed may provide a compromise between group and traditional

individual research and could be useful for exploratory studies. The recall test could be

changed to more closely mimic the deliberation process. For example, nominal jurors

could first be asked to provide facts/arguments that support their verdict and then to provide

facts/arguments that support the alternative verdict.

The second main objective of this research was to explore how PTP imparts its biasing

effect on jury decision making. One possible mechanism is by influencing jurors’ memory

for the trial and in turn the jurors’/jury’s decisions. Jurors who were exposed to negative

PTP could misattribute information contained in the PTP to evidence presented at trial, and

in the current study this appeared to be the case. PTP-exposed jurors were more likely than

non-exposed jurors to make these misattributions, and they were very confident when they

did so. The majority of the critical SM errors for the exposed jurors were due to attributing

this information inaccurately to both the trial and the PTP. The jurors did not forget that

they read about this information in the PTP, but they also believed that they heard it during

the trial. The belief that they had received the information from two separate sources

coupled with their high confidence in these SM errors may have made this information

extremely influential on their decisions. These critical SM errors were also found to

mediate the effect of PTP on guilt ratings and therefore may be amechanism through which

PTP imparts its biasing effects on juror decision making.

Although, critical SM errors did significantly mediate the effect of PTP on guilt ratings

the rate of critical SM errors for both exposed and non-exposed jurors was low (15% and

10%, respectively). These error rates are likely to be substantially lower for the mock jurors

in our study than they would be for actual jurors for a number of reasons. First, in the

experiment jurors were informed, just prior to viewing the trial, that they may have read

PTP related to the trial they were about to view and were admonished not to use this

information when making decisions about guilt. Past research suggests that these

instructions may increase jurors’ ability to discriminate between PTP items and trial items

60 C. Ruva et al.

(McDermott & Roediger, 1998; Neuschatz, Payne, Lampinen, & Toglia, 2001). Second,

the mock jurors in these experiments were only exposed to negative PTP about the trial on

one occasion. In contrast, real jurors involved in a trial in which there is moderate to

substantial negative PTP are likely to have been exposed to negative PTP about the case on

several occasions. Third, the trial used in these studies was much shorter (30 minutes) than

an actual trial would be. Some real trials last days, weeks, or even months. Also, with the

shorter trial the delay between learning a fact and having to identify its source was much

shorter. Past research has demonstrated that SM errors increase as the delay between

encoding and retrieval of a fact increases (Frost, Ingraham, &Wilson, 2002). Finally in real

trials, PTP and early trial testimony may be much closer in time than early trial testimony

and later trial testimony. Such close proximity of PTP and early testimony may make it

even more difficult for jurors to distinguish between PTP and facts presented early during

the trial, typically when the prosecution is presenting its case. All of these methodological

factors suggest that jurors in our study were under optimum conditions when they

completed the SM test. Nonetheless, those who were exposed to negative PTP were

significantly more likely to misattribute information contained in the negative PTP to the

trial. Future research should examine how timing of the trial and PTP affect the rate of

critical PTP source memory errors.

The fact that jurors who were exposed to negative PTP perceived the defendant as being

less credible than did non-exposed jurors suggests that negative PTP may have operated, at

least in part, by causing jurors to form a negative impression of the defendant. These

impressions then may have influenced jurors’ interpretations of trial evidence and hence

their verdicts (Otto et al., 1994), which is consistent with the finding that the credibility

ratings of the defendant explained nearly 40% of the variance in the jurors’ verdicts and

mediated the effect of PTP on guilt ratings. The results are also consistent with Pennington

and Hastie’s (1988, 1993) story model which suggests that jurors begin the decision-

making process by constructing a model or mental representation (a story) which is

comprised of both information presented at trial as well as information that the jurors bring

to trial (e.g. PTP, definitions of crime categories, knowledge about what makes a complete

story). These mental representations are then used as frameworks within which subsequent

trial evidence is interpreted. The negative PTP set up a framework that was consistent with

the prosecution’s story, making this story appear much more plausible and consistent than

the defense’s story. On the other hand, jurors who were not exposed to the negative PTP

were more likely to find both stories plausible and so were more likely to find the defendant

not guilty because of the reasonable doubt instructions that all of the jurors received.

Clearly, the courts have the difficult duty of maintaining a defendant’s Sixth Amendment

right to a fair trial without ‘infringing on the Constitutional guarantee of freedom of the press’

(Otto et al., 1994, p. 454). In a casewhere there is a lot of negative pre-trial publicity about the

defendant, it may be impossible for that defendant to get a fair trial. Even if jurors are

instructed not to use information contained in the PTP tomake decisions about guilt, theymay

be unable to do so because of source memory errors or negative impressions they have formed

about the defendant. If jurors are mistakenly using information provided in PTP to make

verdict decisions, then the defendant cannot receive a fair trial. Although the rate of critical

PTP sourcememory errors for exposed jurors was low, it does suggest that PTP can cause such

errors even under optimal experimental conditions. More importantly, such errors are related

to verdict decisions. More research is needed to examine the factors that affect jurors’ ability

to discriminate between information they receive prior to trial and facts presented at trial

(e.g. timing of the PTP; length of deliberations; type of case; amount of PTP; modality of

Effects of pre-trial publicity and jury deliberation 61

PTP—transcript, video, audiotape; and type of PTP—emotional vs. factual). Until

researchers can provide the courts with strong evidence of the mechanisms underlying the

PTP bias, the courts will continue to rely on questionably effective remedies such as judicial

instructions to ignore pre-trial publicity.

REFERENCES

Alderton, S. M., & Frey, L. R. (1983). Effects of reactions to arguments on group outcome: The case of group polarization. Central States Speech Journal, 34, 88–95.

Alderton, S. M., & Frey, L. R. (1986). Argumentation in small group decision-making. In R. Y. Hirokawa, & M. S. Poole (Eds.) Communication and group decision making (pp. 157–173). Beverly Hills, CA: Sage Publications.

Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Person- ality of Social Psychology, 51, 1173–1182.

Basden, B. H., Basden, D. R., Bryner, S., & Thomas, R. L., III. (1997). A comparison of group and individual remembering: Does collaboration disrupt retrieval strategies? Journal of Experimental Psychology: Learning, Memory, & Cognition, 23, 1176–1191.

Basden, B. H., Basden, D. R., Thomas, R. L., III., & Souphasith, S. (1998). Memory distortion in group recall. Current Psychology: Developmental, Learning, Personality, 16, 225–246.

Berlo, D. K., Lemert, J. B., & Mertz, R. J. (1970). Dimensions for evaluating the acceptability of message sources. Public Opinion Quarterly, 33, 563–578.

Bornstein, B. H. (1999). The ecological validity of jury simulations: Is the jury still out? Law and Human Behavior, 23, 75–92.

Bottoms, B. L., & Goodman, G. S. (1994). Perceptions of children’s credibility in sexual assault cases. Journal of Applied Social Psychology, 24, 702–732.

Bourgeois, M. J., Horowitz, I. A., Fosterlee, L., & Graphe, J. (1995). Nominal and interactive groups: Effects of preinstruction and deliberations on decisions and evidence recall in complex trials. Journal of Applied Psychology, 80, 58–67.

Brigham, J. C., & Spier, S. A. (1992). Opinions held by professionals who work with child witnesses. In H. Dent, & R. Flin (Eds.), Children as witnesses (pp. 93–111). Chichester, England: JohnWiley & Sons.

Clark, N. K., Stephenson, G.M., &Kniveton, B. H. (1990). Social remembering: Quantitative aspects of individual and collaborative remembering by police officers and students. British Journal of Psychology, 81, 73–94.

Clark, N. K., Stephenson, G. M., & Rutter, D. R. (1986). Memory for a complex social discourse: The analysis and prediction of individual and group remembering. Journal of Memory and Language, 25, 295–313.

Cohen, J. (1988). Statistical power analyses for the behavioral sciences. Hillsdale, NJ: Erlbaum. Cutler, B. L., Penrod, S. D., & Dexter, H. R. (1990). Juror sensitivity to eyewitness identification evidence. Law and Human Behavior, 14, 185–191.

Davis, R. W. (1986). Pretrial publicity, the timing of the trial, and mock jurors’ decision processes. Journal of Applied Social Psychology, 16, 590–607.

Davis, J. H. (1992). Some compelling intuitions about group consensus decisions, theoretical and empirical research, and interpersonal aggregation phenomena: Selected examples, 1950–1990. Organizational Behavior and Human Decision Processes, 52, 3–38.

Davis, J. H., Hulbert, L., Au, W. T., Chen, X., & Zarnoth, P. (1997). Effects of group size and procedural influence on consensus judgments of quantity: The examples of damage award and mock civil juries. Journal of Personality and Social Psychology, 57, 1000–1012.

Davis, J. H., Spitzer, C. E., Nagao, D. H., & Stasser, G. (1978). Bias in social decisions by individuals and groups: An example from mock juries. In H. Brandstatter, J. H. Davis, & H. Schuler (Eds.), Dynamics of group decisions (pp. 33–52). Beverly Hills, CA: Sage Publications.

Dexter, H. R., Cutler, B. L., & Moran, G. (1992). A test of voir dire as a remedy for the prejudicial effects of pretrial publicity. Journal of Applied Social Psychology, 22, 819–832.

62 C. Ruva et al.

Deitz, S. R., & Sissman, P. L. (1984). Investigating jury bias in a child molestation case. Behavioral Sciences & the Law, 2, 423–435.

Devine, D. J., Clayton, L. D., Dunford, B. B., Seying, R., & Pryce, J. (2001). Jury decision making: 45 years of empirical research on deliberating groups. Psychology, Public Policy, & Law, 7, 622–727.

Frost, P., Ingraham, M., & Wilson, B. (2002). Why misinformation is more likely to be recognised over time: A source monitoring account. Memory, 10, 179–185.

Goodman, G. S., Golding, J. M., Helgeson, V. S., Haith, M. H., & Michelli, J. (1987). When a child takes the stand. Law and Human Behavior, 11, 27–40.

Hartwick, J., Sheppard, B. H., & Davis, J. H. (1982). Group remembering: Research and implications. In R. A. Guzzo (Ed.), Improving decision making in organizations (pp. 41–72). London: Academic Press.

Hinz, V. B. (1990). Cognitive and consensus processes in group recognition memory performance. Journal of Personality and Social Psychology, 59, 705–718.

Hope, L., Memon, A., & McGeorge, P. (2004). Understanding pretrial publicity: Predecisional distorition of evidence by mock jurors. Journal of Experimental Psychology: Applied, 10, 111–119.

Imrich, D. J., Mullin, C., & Linz, D. (1995). Measuring the extent of prejudicial pretrial publicity in major American newspapers: A content analysis. Journal of Communication, 45(3), 94–117.

Johnson, M. K., Hashtroudi, S., & Lindsay, D. (1993). Source monitoring. Psychological Bulletin, 114, 3–28.

Kassin, S. M., & Wrightsman, L. S. (1988). The American jury on trial: Psychological perspectives. Bristol, PA: Taylor and Francis.

Kelly, A., Carroll, M., & Mazzoni, G. (2002). Metamemory and reality monitoring. Applied Cognitive Psychology, 16, 407–428.

Kerr, N. (1993). Stochastic models of juror decision making. In R. Hastie (Ed.), Inside the juror: The psychology of juror decision making (pp. 116–135). New York, NY: Cambridge University Press.

Kerr, N. L., Niedermeier, K. E., &Kaplan, M. F. (1999). Bias in jurors vs. bias in juries: New evidence from the SDS perspective. Organizational Behavior and Human Decision Processes, 80, 70–86.

Kramer, G. P., Kerr, N. L., & Carroll, J. S. (1990). Pretrial publicity, judicial remedies, and jury bias. Law and Human Behavior, 14(5), 409–438.

Leippe, M., Manion, A., & Romanczyk, A. (1992). Eyewitness persuasion: How and howwell do fact finders judge the accuracy of adults’ and children’s’ memory reports. Journal of Personality and Social Psychology, 63(2), 181–197.

Leippe, M. R., & Romanczyk, A. (1987). Children on the witness stand: A communication/ persuasion analysis of jurors’ reactions to child witnesses. In S. J. Ceci, M. P. Toglia, & D. F. Ross (Eds.), Children’s eyewitness memory (pp. 155–177). New York: Springer-Verlag.

Lindsay, D. S. (1994). Memory, source monitoring, and eyewitness testimony. In D. F. Ross, D. J. Read, &M. P. Toglia (Eds.), Adult eyewitness testimony: Current trends and developments (pp. 27– 55). New York: Cambridge University Press.

Lindsay, D. S., & Johnson, M. K. (1989). Reality monitoring and suggestibility: Children’s ability to discriminate among memories from different sources. In S. J. Ceci, D. F. Ross, & M. P. Toglia (Eds.), Perspectives on children’s testimony (pp. 92–121). New York: Springer-Verlag.

Loftus, E. F., & Palmer, J. C. (1974). Reconstruction of automobile destruction: And example of the interaction between language and memory. Journal of Verbal Learning and Behavior, 13, 585– 589.

McCroskey, J. C. (1966). Scales for the measurement of ethos. Speech Monographs, 33, 65–72. McCroskey, J. C., & Young, J. Y. (1981). Ethos and credibility: The construction and its measurement after three decades. Central States Speech Journal, 32, 24–34.

McDermott, K. B., & Roediger, H. L., III. (1998). Attempting to avoid illusory memories: Robust false recognition of associates persists under conditions of explicit warnings and immediate testing. Journal of Memory & Language, 39, 508–520.

Moran, G., & Cutler, B. L. (1991). The prejudicial impact of pretrial publicity. Journal of Applied Social Psychology, 21, 345–367.

Neuschatz, J. S., Payne, D. G., Lampinen, J. M., Toglia, M. P. (2001). Assessing the effectiveness of warnings and the phenomenological characteristics of false memories. Memory, 9, 53–71.

Nigro, G. N., Buckley, M. A., Hill, D. E., & Nelson, J. (1989). When juries ‘‘hear’’ children testify: The effects of eyewitness age and speech style on jurors’ perceptions of testimony. In S. J. Ceci, D. F. Ross, & M. P. Toglia (Eds.), Perspectives on children’s testimony (pp. 57–69). New York: Springer-Verlag.

Effects of pre-trial publicity and jury deliberation 63

Ogloff, J. R. P., & Vidmar, N. (1994). The impact of pretrial publicity on jurors: A study to compare the relative effects of television and print media in a child sex abuse case. Law and Human Behavior, 18, 507–525.

Otto, A. L., Penrod, S. D., & Dexter, H. R. (1994). The biasing impact of pretrial publicity on juror judgements. Law and Human Behavior, 18, 453–469.

Pennington, N., & Hastie, R. (1988). Explanation-based decision making: Effects of memory structure on judgment. Journal of Experimental Psychology: Learning, Memory, & Cognition, 14, 521–533.

Pennington, N., & Hastie, R. (1993). The story model for juror decision making. In R. Hastie (Ed.), Inside the juror: The psychology of juror decision-making (pp. 192–221). New York, NY: Cambridge University Press.

Pritchard, M. E., & Keenan, J. M. (1999). Memory monitoring in mock jurors. Journal of Applied Psychology, 84, 340–348.

Pritchard, M. E., & Keenan, J. M. (2002). Does jury deliberation really improve jurors’ memories? Applied Cognitive Psychology, 16, 589–601.

Rantzen, A., & Markham, R. (1992). The reversed eyewitness testimony design: More evidence for source monitoring. The Journal of General Psychology, 119(1), 37–43.

Ross, D. F., Dunning, D., Toglia, M. P., & Ceci, S. J. (1989). The child in the eyes of the jury: Age stereotypes, communication modality, and mock jurors’ perceptions of the child witness. In S. J. Ceci, D. F. Ross, & M. P. Toglia (Eds.), Perspectives on children’s testimony (pp. 37–56). New York: Springer-Verlag.

Ruva, C. L., & Bryant, J. B. (2004). The impact of age, speech style, and question form on perceptions of witness credibility and trial outcome. Journal of Applied Social Psychology, 34, 1919–1944.

Ruva, C. L., Bryant, J. B., & Brannick, M. T. (1996).Development of the adults’ general beliefs about children scale for research on witness credibility. Poster presented at the Conference on Human Development, Birmingham, AL.

Schmidt, C.W., & Brigham, J. C. (1996). Jurors’ perceptions of child victim-witnesses in a simulated sexual abuse trial. Law and Human Behavior, 20(6), 581–606.

Seibold, D. R., & Meyers, R. A. (1986). Communication and influence in group decision-making. In R. Y. Hirokawa, & M. S. Poole (Eds.) Communication and group decision making (pp. 133–155). Beverly Hills, CA: Sage Publications.

Shrout, P. E., & Bolger, N. (2002). Mediation in experimental and nonexperimental studies: New procedures and recommendations. Psychological Methods, 4, 422–445.

Sobel,M. E. (1982). Asymptotic confidence intervals for indirect effects in structural equationmodels. In S. Leinhardt (Ed.), Sociological methodology (pp. 290–312). San Francisco: Jossey-Bass.

Stephenson, G. M., Abrams, D., Wagner, W., & Wade, G. (1986). Partners in recall: Collaborative order in the recall of a police interrogation. British Journal of Social Psychology, 25, 341–343.

Stephenson, G. M., &Wagner, W. (1989). Origins of the misplaced confidence effect in collaborative recall. Applied Cognitive Psychology, 3, 227–236.

Studebaker, C. A., & Penrod, S. D. (1997). Pretrial publicity: The media, the law, and common sense. Psychology, Public Policy, & Law, 3, 428–460.

Tankard, J. W., Jr. Middleton, K., & Rimmer, T. (1978). Compliance with American Bar Association fair trial-free press guidelines. Journalism Quarterly, 464–468.

Vollrath, D. A., Sheppard, B. H., Hinsz, V. B., & Davis, J. H. (1989). Memory performance by decision-making groups and individuals. Organizational Behavior & Human Decision Processes, 43, 289–300.

Weingardt, K. R., Toland, H. K., & Loftus, E. F. (1994). Reports of suggested memories: Do people truly believe them. In D. F. Ross, D. J. Read, & M. P. Toglia (Eds.), Adult eyewitness testimony: Current trends and developments (pp. 3–27). New York: Cambridge University Press.

Wells, G. L., Turtle, J.W., & Luus, C. A. E. (1989). The perceived credibility of child witnesses:What happens when they use their own words? In S. J. Ceci, D. F. Ross, & M. P. Toglia (Eds.), Perspectives on children’s testimony (pp. 23–36). New York: Springer-Verlag.

Yarmey, A. D., & Jones, H. P. (1983). Is the psychology of eyewitness identification a matter of common sense? In S. M. A. Lloyd-Bostock, & B. R. Clifford (Eds.), Evaluating witness evidence (pp. 13–40). Chichester, England: John Wiley & Sons.

Zaragoza, M. S., & Lane, S. M. (1994). Source misattributions and the suggestibility of eyewitness memory. Journal of Experimental Psychology: Learning, Memory, & Cognition, 20, 934–945.

64 C. Ruva et al.

APPENDIX A: A SAMPLE OF ITEMS FROM

THE SOURCE MONITORING TEST

Facts presented only in the PTP

1. Lise did not know how to use guns and disliked them.

2. Daniel Bias has a bad temper.

3. Lise Bias’ body was found in the doorway of the couple’s bedroom.

4. The prosecutor questioned why Daniel Bias kept a loaded gun in the house if he

believed that his wife was suicidal.

5. Daniel Bias had been drinking alcohol on the night of his wife’s death.

6. Just prior to her death Lise had received a promotion at her work.

7. Lise Bias did not leave a suicide note.

8. On the day of her death Daniel and Lise had an argument about Lise buying new

clothes.

9. Daniel Bias was also charged with resisting arrest.

10. Daniel had wanted Lise to quit her job and start a family.

Facts presented only in the trial

1. Detective John Flynn failed to do any fingerprint testing on the gun that killed Lise

Bias.

2. On the day of her death Daniel and Lise had an argument about a piece of jewelry that

Lise had seen earlier that day.

3. During the last couple of nights before her death, Lise was very clingy. She demanded

that Friday and Saturday that Daniel go to bed when she went to bed.

4. Daniel Bias had his hand on his wife’s hand when the gun went off.

5. Prior to the night of her death, Lise Bias was brought to Warren County Hospital after

having an argument with her husband, in which she had threatened to shoot herself

with her husband’s gun.

6. If Lise Bias had shot herself the entry wound’s edges would not have been circular, but

rather, they would be star-shaped.

7. On the night of Lise’ death when Daniel first saw Lise with the gun in her hand he

believed that she was clowning around and was doing it for attention.

8. Lise Bias was standing by the mirror when the gun went off.

9. Daniel Bias had been to the shooting range on the day of his wife’s death.

10. The autopsy report of Dr. Isidore Mihalakis indicated that the doctor had performed

gentle rinsing on the hair that he clipped from Lise Bias’ head.

New Facts

1. Lise’ parents had been against her marrying Daniel Bias.

2. Lise was Daniel Bias’ second wife.

3. Daniel Bias refused to take a polygraph (lie detector) test.

4. Just prior to his wife’s death Daniel Bias was reported to be having an extramarital

affair.

Effects of pre-trial publicity and jury deliberation 65

Facts presented only in the Unrelated Crime Stories

1. Some cash child support payments from fathers were held for days, months or even

years before they were deposited into the account.

2. Renee Godshalk, a 23-year-old computer operator in the county domestic relations

office, was arrested for embezzling child support payments.

3. Godshalk took another $54 493 under another scheme in which she relied on a flaw in a

computer program to cover missing money.

4. One reason the shortages weren’t detected sooner was that domestic relations simply

wasn’t balancing its checkbook on time.

APPENDIX B: GUILT MEASURES

Individual/juror verdicts

� What is your verdict in this case?

______Guilty

______Not guilty

Group/jury verdicts

(This was only used to assess the second verdict by collaborative jurors.)

� What is your jury’s verdict in this case?

______Guilty

______Not guilty

______Hung (The jury was not able to reach a unanimous decision regarding the guilt of

the defendant)

Guilt ratings

� Please indicate how confident you are about your verdict.

Sentence Length

� If you found the defendant guilty, howmany years should he be sentenced to New Jersey

State Prison? Sentencing guidelines for the State of New Jersey give a permitted range of

I am certain that he is not guilty

I am pretty sure that he is not guilty

I’m not sure but I think he is not guilty

I’m unsure whether he is guilty of not guilty

I’m not sure but I think he is guilty

I am pretty sure that he is guilty

I’m certain that he is guilty

1 2 3 4 5 6 7

66 C. Ruva et al.

30–45 years. Therefore, you must stay within these guidelines when deciding on a

sentence.

__________ I did not find the defendant guilty

Note: The participants in this study resided in Florida. The sentence length was limited to

30–45 years to be consistent with Florida sentencing guidelines at the time of the

experiments inception.

APPENDIX C: CREDIBILITY RATINGS

The following questions were used to assess mock jurors’ perceptions of the defendant’s

credibility. Due to space limitations the response categories are not included.

1. What percentage of the questions do you believe the defendant answered correctly?

2. How confident did the defendant appear?

3. How biased was defendant’s testimony?

4. How intelligent did the defendant appear to be?

5. How likely is it that the defendant had ulterior motives for giving his testimony?

6. In your opinion, how accurate was the defendant’ testimony?

7. How honest did the defendant appear to be?

8. How credible did the defendant appear to be?

9. How believable was the defendant?

10. How detailed was the defendant’s testimony?

11. How consistent was the defendant in recounting the alleged events?

12. How nervous did the defendant appear to be?.

30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

Note: All items were score on a 7-point Likert-type scale with the exception of item 1,

which was converted from a percentage to a 10-point scale. Therefore, the maximum

possible perceived credibility score was 87.

Note: Item generation for the credibility scale involved a review of the witness

credibility literature in order to identify those variables most frequently associated with

witness credibility (Brigham & Spier, 1992; Goodman et al., 1987; Leippe, Manion, &

Romanczyk, 1992; Leippe & Romanczyk, 1987; Nigro, Buckley, Hill, & Nelson, 1989;

Ross, Dunning, Toglia, & Ceci, 1989; Schmidt & Brigham, 1996; Wells, Turtle, & Luus,

1989; Yarmey & Jones, 1983). Empirically validated measures of source credibility scales

were also consulted (Berlo, Lemert, & Mertz, 1970; McCroskey, 1966; McCroskey &

Young, 1981).

Effects of pre-trial publicity and jury deliberation 67