1 Introduction
Jury research shows that verdicts in criminal trials vary according to the strength of the evidence presented by the prosecution and the defence (Devine, Clayton, Dunford, Seying, & Pryce, 2001). But because ‘hard evidence’ is usually lacking in sexual assault trials, it can be expected that jurors will rely on extra-legal factors (myths, biases and misconceptions) in deciding a case with ambiguous evidence such as a word-against-word case.
Studies of the legal and extra-legal variables that affect convictions in sexual assault cases go back to the 1960s and 1970s with findings that the degree of ‘bad character’ evidence about a complainant influences jurors’ perceptions of her responsibility for the defendant’s sexual conduct, which, in turn, may be associated with more lenient sentences for the defendant (Feldman-Summers & Lindner, 1976).
In their classic investigation of jury decision-making, Kalven and Zeisel (1966: 252–253) found that juries were four times more likely to convict in a sexual assault trial if the defendant was a stranger, if there was more than one assailant, or if the assault involved violence. Although the authors did not specifically investigate the impact of variables associated with the complainant, they found that the degree of agreement between judges and juries in cases with one of the above three factors was 88%, compared to 40% in cases without one of these factors. Kalven and Zeisel (1966) hypothesised that the discrepancy between judge/jury agreement could be attributed to juries’ focus on the behaviour of complainants, specifically the extent to which complainants were perceived to be responsible for their victimisation.
Nonetheless, sexual assault cases are not easy for judges to preside over. A study of a sample of US judges found that cases ‘involving sexual offenses were rated by the judges as more difficult’ compared to other criminal cases (Bumby & Maddox, 1999: 311). Specifically, 63.4% found sexual assault cases and 76.2% found CSA cases more or much more difficult from a legal/technical standpoint; 87.8% found sexual assault cases and 78.6% found CSA cases more or much more difficult from a personal/emotional standpoint, while 87.8% found both types of cases more or much difficult from a public scrutiny/pressure point of view.
Various reasons were given for these greater difficulties including, the ‘rules of evidence make convictions more difficult and tend to promote hiding the truth’; ‘ambiguity in evidence’; ‘one-on-one crime’; ‘lack of corroborating evidence/witness’; ‘victim won’t testify’; ‘reluctance of juries to convict on testimony of only one witness’; ‘assessing credibility of witness’; ‘age of child witness’; and ‘ease of defendant’s injection of reasonable doubt into circumstances normally surrounding such a crime’ (Bumby & Maddox, 1999: 312–313). As we will see below, these difficulties still characterise sexual assault trials two decades later.
2 Factors that Predict Outcomes in Sexual Assault Trials: Physical Evidence
Where the evidence is weak or ambiguous in a sexual assault trial, it can be expected that jurors will have expectations of corroborating evidence, such as medical and/or forensic evidence. For example, 57% of mock jurors agreed with the following statement, ‘A physical examination by a doctor will almost always show whether a child has been sexually abused’ in studies by Quas, Thompson, and Clarke-Stewart (2005) and Kovera and Borgida (1997).
Jurors’ expectations of medical evidence have also been reported in simulated adult sexual assault cases. In a study of 233 jurors who participated in 27 simulated trials, Ellison and Munro (2009b: 376) found a ‘widespread belief that genital trauma is an inevitable outcome of rape’ contrary to a medical report in the trials which noted that the absence of genital injury ‘was neither consistent nor inconsistent with rape’ and instructions to jurors that ‘the absence of physical injury did not rebut an allegation of rape’.
In a study involving 18 simulated sexual assault trials, Taylor and Joudo (2005) reported that, during deliberation, mock jurors attempted to ‘fill in the gaps’ in the evidence by speculating not only about the complainant’s behaviour but also discussed the fact that there was no DNA evidence, and no medical evidence of injury (Taylor & Joudo, 2005: 61).
These studies raise the question, does the presence or absence of medical and/or forensic evidence affect outcomes in sexual assault cases? The literature that documents the impact of such evidence on the likelihood of prosecution and on outcomes in sexual assault trials reveals contradictory findings.
A few studies have found no association between medical evidence of sexual assault and the decision to prosecute or convict (Hagemann, Stene, Myhre, Ormstad, & Schei, 2011; Hansen, Mikkelsen, Sabroe, & Charles, 2010; Saint-Martin, Bouyssy, Jacquet, & O’Byrne, 2007), although ‘differences in the legal systems in European countries [may] account for the lack of correlation between medical evidence and legal outcomes’ (Cossins, Jayakody, Norris, & Parkinson, 2016: 931).
By comparison, the presence of positive medical evidence was the strongest predictor of conviction in actual CSA cases as reported by Bradshaw and Marks (1990). Similarly, Palusci, Cox, Cyrus, Heartwell, Vandervort, and Pott (1999) found that a positive medical finding was more likely to predict convictions in CSA cases compared to other factors such as behavioural symptoms displayed by the child complainant.
Out of a sample of 1547 cases of adult and child rape in South Africa, Jewkes, Christofides, Vetten, Jina, Sigsworth, and Lotts (2009) found that both non-genital and genital injuries were associated with the likelihood of CSA cases being prosecuted, compared to adult rape cases. In adult cases, a conviction was more likely if evidence of injury was presented by the prosecution, irrespective of the type of injury. Thus, convictions in the absence of injuries were found to be relatively low in South African courts, that is, 9% in adult sexual assault cases and 33% in CSA cases. By comparison, the presence of DNA evidence had no impact on outcomes in child or adult rape cases.
Similarly, McGregor, Du Mont, and Myhr (2002: 644) found that the extent of injuries documented by a medical examination (mild, moderate or severe) in 193 adult sexual assault cases had a significant positive association with the likelihood of charges being filed, while ‘the only variable found to be associated with conviction was a clinical injury extent score of severe’.
‘documentation of injury was the strongest predictor of a positive legal outcome. In just under half (44%) of those studies that examined the presence of general physical injuries, a significant association was found’1;
the defendant’s prior convictions and/or charges;
multiple charges against the defendant;
penetration;
use of a weapon; and
corroborating evidence from witnesses.
whether her character or reputation were perceived as being negative … and her behaviours before (e.g., she engaged in “risk-taking” activities such as walking alone at night), during (e.g., she did not resist the assailant) and following an assault (e.g., she did not report the assault promptly to the police). (Du Mont & White, 2007: 49)
The composite profile of the victim least likely to see her assailant convicted is of an older, poorer woman, whose character or reputation is perceived negatively as she is deemed to have been sexually promiscuous, a sex worker, has a psychiatric history, drug abuse problem and/or a criminal record. (Du Mont & White, 2007: 30; references omitted)
- i.
in half the studies reviewed, generally, ‘forensic medical evidence was associated with positive a legal outcome’, although ‘a range of other evidentiary and credibility factors also have a role to play’.
- ii.
Evidence of severe or serious injury was ‘most often associated with a positive legal outcome’ although other ‘biological evidence do[es] not appear to have any impact on legal outcomes’.
- iii.
No single factor was found to predict ‘case progression or positive legal outcomes’ because of ‘complex interactions between numerous factors’, including forensic evidence.
stereotypical beliefs around what constitutes ‘real’ rape … and who is a ‘deserving’ victim …, rather than the availability of forensic or medical evidence availability, continue to have a profound and significant [e]ffect on decision-making processes and case outcome for sexual offences.
the defendant was a stranger;
the defendant was non-Caucasian;
physical or verbal expression of non-consent by the victim;
victim injury;
use of force by the defendant;
evidence of some threat, force, injury or weapon use;
additional evidence linking the defendant to the offence.
While no single factor predicted prosecution, multivariate analyses found that ‘cases were significantly more likely to proceed if the defendants used force and the victims actively expressed non-consent’, both of which ‘help to establish the mental element of the offence; that is, that the defendant knew the victim was not consenting, but intended to have sex regardless’ (Lievore, 2005: 5).
prosecutors who pre-empt the views of the jury on the basis of various sociodemographic, personal or other extralegal variables may exclude particular types of victims from justice. They may then fail to challenge and expand the bounds of convictability by dealing with actual juries to learn how to win believable but risky cases. (Lievore, 2005: 6)
While jurors might expect supporting medical or forensic evidence, it is uncommon for a sexual assault to result in demonstrable injury to the ano-genital area of adults and children (Cossins et al., 2016; Du Mont & White, 2007; Gray-Eurom, Seaberg, &Wears, 2002; Hansen et al., 2010; Heger, Ticson, Velasquez, & Bernier, 2002; Jina et al., 2015; Saint-Martin et al., 2007), or to result in the collection of forensic evidence.
The typicality of delayed complaint also prevents the collection of medical evidence while, in adult sexual assault cases, DNA evidence has no probative value in cases in which a defendant admits he had sex with the complainant but denies that it was non-consensual.
In a review of the literature on medical examinations involving suspected CSA, Cossins et al. (2016: 926) reported that ‘there is a narrow window of time [around 72-96 hours] in which physical evidence may be recovered from the body of a pre-pubertal child’, such as blood or semen, while injuries are generally absent due to the elasticity of the ano-genital tissues or because they are difficult to accurately detect (Cossins et al., 2016). As well, the ‘competence among medical, law enforcement, forensic science, and legal professionals who deal with victims and their cases of sexual assault often negatively impact the integrity of medico-legal findings’, as does active disbelief of a victim’s allegations (Du Mont & White, 2013: 1231, 1232). Thus, both jurors’ expectations of injury and prosecutors’ filtering of cases based on lack of injury do not match the reality of most sexual assaults.
3 Factors that Predict Outcomes in Sexual Assault Trials: Other Corroborating Evidence
While the above studies were mostly limited to examining the impact of physical evidence on case outcomes in sexual assault trials, it is likely that a variety of legal factors are implicated in jury outcomes in these trials. In fact, mock jury studies predict that the overall strength of the prosecution’s case is a significant factor in case outcomes (Bottoms et al., 2014). For example, in a large-scale, simulated jury study of CSA cases involving jury deliberation, when corroboration was provided by two other victims about being sexually abused by the same defendant, conviction rates increased significantly compared to a trial with no corroborating witnesses, irrespective of whether the corroborating evidence came from victim-witnesses, or victim-complainants in a joint trial (Goodman-Delahunty, Cossins, & Martschuk, 2016).
Studies show that evidence strength is also associated with convictions in actual sexual assault cases. When LaFree (1980) analysed 124 forcible rape cases from a large American city that were adjudicated between 1970 and 1975, he found that as ‘the amount of prosecution evidence increased, the likelihood of guilty verdicts increased’, with several factors having a greater impact on verdicts. The defendant’s criminal record of sex offences, the seriousness of his prior sex offences, the number of offenders, and prompt report by the complainant all predicted guilty verdicts (LaFree, 1980: 843). In addition to the presence of bodily and/or genital trauma, Gray-Eurom, Seaberg and Wears’ (2002) analysis of 801 sexual assault examinations over a 2-year period found that the victim’s age (less than 18 years) and the use of a weapon by the assailant were also significantly associated with successful prosecutions.
- i.
age of child at trial;
- ii.
presence of similar fact evidence2;
- iii.
recent complaint evidence3;
- iv.
presence of penile penetration;
- v.
eyewitness evidence;
- vi.
two or more complainants in one trial;
- vii.
more than four charges;
- viii.
partial acknowledgement by the defendant;
- ix.
positive medical evidence;
- x.
presence of a venereal disease, or
- xi.
DNA evidence of the defendant in forensic samples.
The dependent variable was at least one guilty verdict for a sex offence.
The study found that the more evidential factors present in a CSA trial, the more likely it was for the jury to deliver a guilty verdict. Thus, ‘the presence of any three or more of the nine evidential variables was significantly predictive of conviction’ (Blackwell & Seymour, 2014: 573), and demonstrated the predictive value of overall evidence strength. Out of the above nine variables, the three that were most predictive of at least one guilty verdict were: the presence of similar fact evidence, eyewitness evidence and positive medical or DNA evidence.
This sample of cases also confirmed the fact that most CSA cases do not involve evidence to corroborate a complainant’s allegations with only 10.9% of cases featuring similar fact evidence; 28.5% featuring medical evidence (although in two-thirds of those cases the medical evidence was equivocal); 10.2% featuring DNA evidence and 12.4% featuring eyewitness evidence (Blackwell & Seymour, 2014: 572; see also De Jong & Rose, 1991; Lewis, Klettke, & Day, 2014; Saint-Martin et al., 2007).
The paucity of corroborating evidence in a typical CSA trial means that in most cases, the credibility of the child complainant will be crucial to the jury verdict. Indeed, most charges in the 137-case sample resulted in acquittals (62.5%), confirming the relatively low conviction rate of sex offences seen in other studies (Daly & Bouhours, 2010; Fitzgerald, 2006; see review by Quadara, Fileborn, & Parkinson, 2013; see Chapter 1). This suggests that ‘jurors require a high threshold of corroborative evidence to convict’ (Blackwell & Seymour, 2014: 573),4 a finding that supports the leniency effect observed in criminal trials, generally speaking.
Although Blackwell and Seymour (2014: 573) concluded that ‘jurors’ verdicts were dominated by evidential issues’, they were unable to account for any variance in guilty verdicts that may have been attributed to extra-legal variables such as jurors’ biases and misconceptions, nor the differential impact of legal and extra-legal factors in strong and weak cases. In other words, Blackwell and Seymour (2014) did not assess the extent to which the evidence in the 137 CSA trials was filtered through any CSA misconceptions held by jurors and jurors’ expectations about how child complainants should behave. As a result, the omission of the effects of extra-legal factors may lead to inflated measures of the effects of legal, evidentiary variables.
where the victims displayed some observable behavioral response to the alleged abuse, a guilty verdict was almost seven times more likely. (Lewis et al., 2014: 437)
These behaviours included: ‘difficulty sleeping/nightmares, inappropriate displays of affection with parents, reluctance to undress, reluctance to bathe, displaying adult sexual behavior, withdrawing, problems at school, … promiscuity, depression, and suicidal ideation/attempts’. (Lewis et al., 2014: 435)
Counter-intuitively, the presence of medical evidence had no impact on guilty verdicts while the behaviour of ‘acting out’ was the only significant behavioural predictor of an acquittal, ‘suggest[ing] that when child victims showed evidence of destructive behavior or contrary to what is expected …, a jury was more likely to discredit their account of the alleged abuse and therefore return a not guilty verdict’ (Lewis et al., 2014: 438).
Nonetheless, small sample sizes for each factor in the study (medical evidence plus eleven different behavioural responses) suggest that further research is needed to confirm the findings of Lewis et al. (2014).
4 The Impact of Extra-Legal Factors on Outcomes in Sexual Assault Trials: Victim, Defendant and Juror Characteristics
In a large meta-analysis of jury studies to determine the impact of a range of juror, victim and defendant factors on outcomes in criminal trials, Devine and Caughlin (2014: 122) found that prior criminal record was one of two defendant characteristics that had a significant impact on guilty verdicts. However, ‘the overall effect of defendant criminal record on guilty judgments … was only modest’ and likely to depend on other variables such as the degree of similarity between the prior charges and the current charge.
By contrast, in a sexual assault trial, evidence of a defendant’s prior criminal record is not likely to be a typical feature because of the extent to which sexual assault is an under-reported crime,5 as discussed in Chapter 1. Because the focus of the typical adult sexual assault trial is whether or not the complainant consented, credibility assessments of the complainant mediate jurors’ verdicts. In CSA cases, the fact that most children do not report the abuse to authorities for months or years means that the collection of medical evidence or inculpatory DNA evidence to prove the alleged sexual conduct is impossible. Even if there is prompt reporting or discovery of the abuse, medical examinations of children rarely show physical evidence of sexual abuse (Cossins et al., 2016; Johnson, 2004), while DNA evidence may not be available because of the type of sexual assault. Thus, the focus of the CSA trial is also on complainant credibility.
Because of the lack of so-called ‘hard evidence’ of a defendant’s alleged sexual acts, ‘jurors will turn to the only other information they have – their own expectations and beliefs’ (Cossins, 2013: 95). This suggests that extra-legal variables, such as jurors’ biases and misconceptions and/or defendant and complainant characteristics will have a significant impact on trial outcomes in sexual assault trials.
In an empirical study of jury decision-making in 18 simulated adult sexual assault trials,6 Taylor and Joudo (2005) found a significant relationship between juror demographics, credibility and verdict, in that credibility ratings of the complainant were significantly associated with juror gender, socioeconomic status (SES) and jurors’ political attitudes. In fact, jurors’ pre-existing attitudes about sexual assault influenced their judgements about the credibility of the complainant and the guilt of the accused more than the facts of the case and the manner in which the evidence was given. Men ‘tended to hold less favourable attitudes towards rape victims in general’ and were significantly less likely to perceive the complainant as credible compared to women (Taylor & Joudo, 2005: 32).
Overall, the juror factors that were significantly associated with less favourable attitudes towards victims included being male, having a lower income and being politically conservative. In turn, these factors were linked to the belief that the defendant was not guilty. Those jurors who favoured a guilty verdict were significantly more likely to believe the defendant was guilty both before and after jury deliberation, to hold more positive attitudes towards victims, to believe the complainant to be highly credible, and to have low empathy for the defendant and higher empathy for the complainant (Taylor & Joudo, 2005: 56–57).
Although a majority of mock jurors (74%) favoured a not guilty verdict, half of this group believed the defendant was guilty, but did not believe the prosecution had proved their case beyond reasonable doubt (Taylor & Joudo, 2005: 48), a finding that is consistent with the high threshold that the standard of proof represents for laypeople and the leniency effect associated with jurors’ verdicts.
The results of this study help explain the relatively low conviction rates in sexual assault trials in Australia and E&W, as discussed in Chapter 1. More mock jurors may have voted guilty and some of the hung juries may have become unanimous7 if the jurors had been exposed to educational information. Such information may have challenged the type of myths and misconceptions that some mock jurors relied on when assessing the truth of the complainant’s testimony since there was a ‘high degree to which many jurors believed many of the myths which surround rape in general’. In particular, mock jurors had ‘strong expectations about how a “real” victim would behave before, during and after an alleged sexual assault’ which influenced their assessment of the complainant’s credibility during deliberations (Taylor & Joudo, 2005: 59).
Distrust and contempt for the unchaste female accuser was formalised into a set of legal rules unique to rape cases. The most prominent rule allowed the use at trial of evidence of the complainant’s unchaste conduct. These rules combined to shift the usual focus of a criminal trial from an inquiry into the conduct of the offender to that of the moral worth of the complainant.8 (Galvin, 1986)
Although progressive judges have recognised that a complainant has ‘her victimisation measured against the current rape mythologies’,9 unfortunately ‘many judges show a conventional wisdom about human behaviour which may represent the limitations of their background’.10 Researchers have also recognised that the typical lack of eyewitness and forensic evidence in sexual assault trials means that a word-against-word trial is inherently ‘unstable’ (LaFree, 1980), thus inviting jurors to rely on other, external sources of information. This suggests that verdicts may depend less on the evidence and more on whether the complainant is considered to be worthy of the law’s protection.
While there is no ‘general propensity for female jurors to be harsher on criminal defendants than male jurors’ (Devine & Caughlin, 2014: 122), Devine and Caughlin’s (2014) recent meta-analysis of 11 focal characteristics of defendants and jurors found that juror gender has a significant impact on verdicts in both adult and child sexual assault trials. Female jurors are more likely to convict in those cases, particularly where the complainant is a child, compared to all other types of criminal trials (Devine & Caughlin, 2014: 122; see also Bottoms, Golding, Stevenson, Wiley, & Yozwiak, 2007).
Does this gender effect arise because men are more likely to adhere to rape myths and CSA misconceptions compared to women, or as a result of women’s greater empathy for victims, or both? For example, studies with both student and community samples have found gender differences in empathy for victims, the believability of children as witnesses and attitudes towards CSA (Bottoms, 1993; Bottoms et al., 2014). Most recently, Bottoms et al. (2014: 789) found that ‘women compared to men were more empathic toward child victims, more opposed to adult/child sex, more pro-women, … more inclined to believe children generally and … made more pro-victim judgments in hypothetical abuse cases’.
Victim age is also a key predictor of mock jurors’ verdicts in CSA cases since age has been shown to affect ‘perceptions of competence and trustworthiness’ and hence credibility (Bottoms et al., 2007: 519; see also Bottoms et al., 2014). Generally, younger victims are perceived as more trustworthy compared to older children and teenagers, possibly because they are ‘sexually naïve, lacking the knowledge and cognitive capacity to fabricate sexual encounters’ (Bottoms et al., 2007: 519) and are unlikely to be the subject of allegations of drunkenness and sexual promiscuity during defence cross-examination. It is possible that age is a key factor in explaining the data in Chapter 1 that showed that the conviction rate for child sex offences compared to the conviction rate for adult sex offences was higher in NSW over an eight-year period.
‘the age of the child, the actions of the child after the fact, the language used in the child’s statement’, and ‘he is totally guilty, as I don’t believe a 6 year old would make it up and a 6 year old would not fantasise about those sorts of things’. Thus, mock jurors seemed to perceive a child of 6 years as less responsible and more vulnerable, trustworthy, and sexually naïve. (ibid.: 51)
perceptions of competence and trustworthiness are not straightforward because this age group exhibits attributes of a child and emerging adult (i.e., sexual maturity). Thus, as the effect of age became more ambiguous, evidence and context of assault became more important. (ibid.: 51)
Thus, it appears that the mock jurors in this study linked age and sexual maturity with victim blaming and victim responsibility. While the 15-year-old complainant was perceived as ‘more competent to recall the assault’, she was also perceived ‘as less trustworthy, less vulnerable, more responsible, and more likely to lie’ (ibid.: 52), as well as having a greater ability to resist the abuse.
A meta-analysis by Whatley (1996) examined the degree of responsibility that observers impose on a female rape victim by investigating common variables associated with victim blame. The study found that the victim’s clothing revealingness, and character significantly affected the likelihood that observers would blame a victim of sexual assault. As well, a victim with ‘questionable’ character is held to be more responsible than a more ‘respectable’ victim (Whatley, 1996: 90), as is the victim who is acquainted with the perpetrator (compared to a stranger).
In a meta-analysis of studies that examined the impact of victim behaviour and demeanour in CSA cases, Bottoms et al. (2007: 521) found that male mock jurors were more likely to attribute responsibility to victims who did not resist. This finding suggests that resistance is a gendered concept, with men judging females according to their expectations of how a victim ought to behave when faced with a threat. As well, actual jurors’ verdicts were correlated with ‘the child’s facial expressions, gestures and movements, eye contact, nervousness, and manner of speaking’.
This suggests that jurors look for emotions that are expected of a sexual abuse victim, which are then used to make credibility assessments, as discussed in the next section.
5 The Impact of the Emotionality of the Victim/Complainant on Case Outcomes
When it comes to victim demeanour, various studies have shown that laypeople have particular expectations about emotional salience and CSA. Generally speaking, laypeople do not expect a neutral or non-emotional response from an abused child (Bederian-Gardner & Goldfarb, 2014; Wessel, Magnussen, & Melinder, 2013) and expect that children will commonly exhibit emotions such as fear and sadness (Bederian-Gardner & Goldfarb, 2014; Myers, Redlich, Goodman, Prizmich, & Imwinkelried, 1999; Regan & Baker, 1998), with emotional display being associated with increased guilty verdicts compared to calm emotional displays (Regan & Baker, 1998; see also Golding, Fryman, Marsil, & Yozwiak, 2003).
The same expectations appear to be true for adult complainants. For example, Ellison and Munro (2009a: 206–212) presented nine different video versions of a simulated sexual assault trial (in which the complainant’s evidence was varied) to 233 mock jurors to investigate ‘the influence upon juror deliberations of key behavioural cues involving lack of resistance, delayed reporting and calm complainant demeanour’. Because the evidence supporting the occurrence of the sexual assault was equivocal, jurors spent a large part of their deliberations discussing the complainant’s emotional responses, as well as her ‘“appropriate” socio-sexual behaviour’ before, during and after the alleged sexual assault to determine her veracity.
The factors that weighed in favour of a credible story were: the complainant’s distress, her emotional state in court and her immediate complaint to the police. The factors that weighed against a credible story included the complainant’s failure to resist or physically struggle, lack of significant injury, a three-day delay in complaint, and the complainant’s calm, unemotional delivery of her evidence.
Ellison and Munro (2009a: 212, 214) concluded that ‘mock jurors appeared to have little understanding of the psychological effects … that could influence a rape complainant’s demeanour in court’, while jurors’ assumptions about ‘the instinct to fight back, the compulsion to report immediately and the inability to control one’s … emotions … clearly influenced’ their deliberations and verdicts.
In a subsequent study, the authors examined whether or not expert testimony or an expansive judicial direction would address the popular misconceptions observed in mock jurors’ deliberations. In simulated trials where the complainant displayed a calm, unemotional demeanour, delayed her complaint by three days or showed lack of resistance, jurors who received educational guidance ‘were far more likely to offer thoughts on what could possibly account for the complainant’s lack of emotionality’ and were ‘significantly more likely to state that they were untroubled’ by the delayed complaint (Ellison & Munro, 2009b: 368, 370). However, no empirical analysis was carried out to determine the impact of the expert educational guidance on jury deliberations or case outcomes.
In a more recent study, when laypeople and professional child protection workers watched mock police interviews with child actors who disclosed physical abuse, Wessel et al. (2013) found that the emotions displayed by the children significantly affected the credibility ratings of the child and culpability ratings of the perpetrator given by both laypeople and professionals. The highest credibility ratings were for the sad, emotional expression ‘followed by the neutral expression and a drop in perceived credibility when the witness displayed more positive emotional expression’ (Wessel et al., 2013: 614).
Those who perceived the child to be more emotional were more likely to render guilty verdicts, view the child as credible, and view the defendant as less credible. Moreover, … these trends were especially robust when participants perceived the child who was shown as calm while testifying as being highly emotional. … Participants who perceived the child to be highly emotional were most likely to vote guilty. (Cooper, Quas, & Cleveland, 2014: 823)
By contrast, Bederian-Gardner and Goldfarb (2014: 840) found that both victim age and victim gender affected laypeople’s expectations of the emotions of a sexually abused child. In particular, laypeople responded according to gender stereotypes in that they ‘expected to see more sadness and fear from female than from male victims’. As well, the age of the child affected perceptions in that participants ‘expected fear to be expressed more than sadness regardless of gender’ for younger victims (aged 5 years) but expected anger to be the emotion displayed by the older, 13-year-old, male victim.
Other studies have also found that mock jurors, non-empanelled jurors called up for jury duty and actual jurors expect sexually abused children to display strong emotional reactions (Cossins, Goodman-Delahunty, & O‘Brien, 2009; Goodman-Delahunty, Martschuk, & Cossins, 2017; McCauliff & Kovera, 2012). Indeed, the subjective nature of these expectations was recently demonstrated by Bederian-Gardner, Goldfarb and Goodman (2017) who found that the degree of empathy for participants predicted laypeople’s assessments of a CSA victim’s feelings, which, in turn, predicted ratings of perceived believability.
Jurors’ perceptions of emotion are important to consider in relation to some of the techniques commonly used to reduce a child’s stress while testifying in open court, including screens, videotaped testimony, and closed circuit television …. These techniques, although often effective at reducing children’s distress, may also lead to a child witness acting in a manner inconsistent with the expectations of jurors.
The authors suggest that education is needed to inform jurors about the ‘variability in children’s general emotional reactions’ when giving evidence, and ‘jurors’ own potential biases about emotional displays in sexual abuse cases’ (Cooper et al., 2014: 825) so that invalid considerations are not used to unfairly measure a child’s credibility.
Mock jurors and observers also expect an adult complainant to be emotional when giving evidence (Taylor & Joudo, 2005). A number of studies show that an unemotional sexual assault victim is perceived as being significantly less credible compared to an emotional victim (Bollingmo, Wessel, Eilertsen, & Magnussen, 2008; Bollingmo, Wessel, Sandhold, Eilertsen, & Magnussen, 2009; Kaufman, Drevland, Wessel, Overskeid, & Magnussen, 2003; Klippenstine & Schuller, 2012; Lens, Doorn, Pemberton, & Bogaerts, 2014; Vrij & Fischer, 1997; Wessel, Drevland, Eilertsen, & Magnussen, 2006).
These effects have been found to be ‘quite robust and turn up … across the complete age range of adults eligible for jury duty’, as well as with experienced police investigators and judges (Bollingmo et al., 2009: 62). In addition, consistency in emotional response over time is important to people’s perceptions of the believability of an allegation of sexual assault (Klippenstine & Schuller, 2012). It is also likely that these beliefs will influence jurors’ perceptions of what is, and is not, a reasonable doubt even when they believe that the sexual assault occurred. In other words, ‘[t]he more ambiguous the available evidence, the greater the influence of general beliefs and sentiments’ (Schuller et al., 2010: 760).
A related topic that has been investigated by Pickel and Gentry (2017) is the influence of observers’ expectations of psychological harm on credibility assessments of rape complainants. Although it might be expected that evidence of high levels of psychological harm would result in more positive credibility assessments and guilty verdicts, research shows that this is too simple a conclusion. In the study by Pickel and Gentry (2017), undergraduate students were assigned to one of two simulated rape scenarios: a prototypical case in which the defendant was a stranger with the attack occurring late at night in a secluded area or a non-prototypical case in which the defendant was an acquaintance, with the attack occurring in the early evening in a private residence. The two scenarios were further manipulated by including either that the complainant was suffering from PTSD or from mild anxiety. Pickel and Gentry (2017) found that a diagnosis of PTSD was more congruent with jurors’ expectations of harm in the prototypical scenario compared with the non-prototypical scenario. By contrast, in the non-prototypical case, jurors expected the complainant to be suffering from mild anxiety rather than PTSD.
- i.
in the prototypical condition than the non-prototypical condition;
- ii.
in the prototypical condition when the complainant was suffering PTSD compared to when she was suffering mild anxiety;
- iii.
in the non-prototypical condition when the complainant was suffering mild anxiety compared to when she was suffering PTSD.
these differences were not related to the presence of a weapon, the victim’s credibility, the severity of her physical injuries, her fear of being physically injured or killed, the extent to which she communicated that she did not want to have sexual intercourse, or the degree to which the perpetrator threatened to kill her. (Pickel & Gentry, 2017: 268)
Thus, ‘the effect of harm on ratings of the victim’s credibility is mediated by the extent to which the harm is in line with mock jurors’ expectations’ which were related to the prototypicality of the rape (Pickel & Gentry, 2017: 266). These findings suggest that credibility ratings and verdicts are not only assessed according to a complainant’s behaviour before and during the sexual assault but also according to laypeople’s expectations of post-rape trauma. However, these expectations do not necessarily reflect the findings from the literature on trauma, as discussed in Chapter 11.
6 The Influence of Victim Gender, Victim Race and Other Extra-Legal Factors on Case Outcomes
In a mock jury study involving 896 community volunteers aged 19 years or over, Feild (1979: 279) reported that verdicts were influenced by the race of the victim and the defendant, the physical attractiveness of the complainant, and the complainant’s sexual history as well as her behaviour with the defendant prior to the alleged offence.
Although LaFree (1980) reported that as the strength of the prosecution’s case increased, the likelihood of guilty verdicts also increased, he also found that as the time taken for the complainant to report increased, the chances of a guilty verdict decreased while the complainant’s moral character and conduct (such as hitchhiking, at a bar by herself, out late at night) also predicted acquittals. Nonetheless, ‘the single most important predictor [of a conviction] was the victim’s race’ in that men on trial for sexually assaulting black women were significantly less likely than other men of being convicted (LaFree, 1980: 846). This finding suggests that minority women are perceived to be less worthy of the law’s protection.
Overall, men who sexually assaulted (i) black women, (ii) women who engaged in ‘risky’ behaviour, (iii) women outside their homes, (iv) women who delayed their complaint and (v) women who knew or were related to the defendant, were less likely to be convicted compared to other men. LaFree (1980: 848) concluded that these outcomes indicate that both authorities and fact-finders are influenced by their ‘perceptions of the victim’s behaviour and characteristics’, as well as the characteristics of the defendant, all of which is mediated by cultural conceptions of rape.
LaFree, Reskin and Visher (1985) conducted post-trial interviews with 331 jurors in 38 forcible sexual assault trials between 1978 and 1980 where the complainant was at least 12 years old to examine the relationship between the following independent variables: (i) jurors’ characteristics and attitudes; (ii) complainants’ characteristics and behaviour; (iii) defendants’ characteristics and behaviours; (iv) evidence and case characteristics. The dependent variable was jurors’ pre-deliberation verdicts.
Neither the use of a weapon or victim injury significantly affected jurors’ verdicts in cases where the defendant either denied the sexual act or said the sexual act was consensual. By contrast, jurors were influenced by the complainant’s character and were ‘less likely to believe in a defendant’s guilt when the victim had reportedly engaged in sex outside marriage, drank or used drugs, or had been acquainted with the defendant’. In particular, jurors who had conservative sex-role attitudes were more likely to find the defendant innocent when the complainant used drugs or alcohol, compared to more liberal-minded jurors (LaFree et al., 1985: 397–399).
Jurors were also influenced by defendant characteristics in that they were more likely to vote guilty if he had few social ties, presented a negative courtroom appearance or had a previous criminal record.
Using the same dataset as LaFree et al. (1985), Reskin and Visher (1986) studied the impact of four types of corroborating evidence (a recovered weapon; victim injury, other physical evidence and eyewitness evidence) and extra-legal factors (defendant attractiveness; defendant’s employment status; negative comments about the complainant’s moral character; perceptions of victim carelessness) on jurors’ pre-deliberation verdicts. All four types of corroborative evidence ‘jointly explained almost one-fifth of the variation in jurors’ judgments about the defendant’s guilt’, although a recovered weapon was the most influential (Reskin & Visher, 1986: 429).
While jurors did take account of evidence that corroborated the complainant’s testimony, they ‘selectively interpreted’ that evidence (Reskin & Visher, 1986: 436; emphasis added), giving more weight to evidence of force and seriousness of the sexual assault and less to eyewitness testimony.11
After controlling for the effects of corroborating evidence, attractive or employed defendants were less likely to be perceived as guilty, while jurors were more likely to believe in the defendant’s innocence if the complainant had not, in their opinion, ‘exercised sufficient caution or was of poor moral character’. In fact, these two variables about the complainant ‘exerted the strongest independent effects on jurors’ predeliberation verdicts’ (Reskin & Visher, 1986: 431).
Overall, Reskin and Visher (1986: 434) found that their results supported Kalven and Zeisel’s (1966) liberation hypothesis. As discussed in Chapter 2, the liberation effect is believed to arise in cases with ambiguous or weak evidence, such that juries’ verdicts are more susceptible to the influence of extra-legal or non-evidentiary factors which, in turn, supports the existence of a leniency effect in sexual assault trials.
In strong sexual assault cases (defined as having two or more types of corroborative evidence), jurors were more likely to believe in the defendant’s guilt, compared to weak cases (those that had none or only one type of corroborating evidence). Out of the 38 sexual assault trials studied, ‘91 percent of the strong cases resulted in guilty verdicts in contrast to 53 percent of the weak cases’ (Reskin & Visher, 1986: 434).
The strength of a sexual assault case also affected whether jurors considered extra-legal factors in coming to a decision. In cases where there were at least two types of corroborating evidence, jurors tended to ignore the complainant’s and defendant’s personal characteristics, although jurors’ opinions about whether the complainant had failed to exercise sufficient caution still significantly affected assessments of the defendant’s guilt in strong cases.
By contrast, in weak cases, jurors’ verdicts were influenced by their own values, that is, their attitudes towards crime, as well as the characteristics of complainants and defendants. Defendants were more likely to be found innocent if the complainant had behaved carelessly or was judged to have been of low moral character. Thus, Reskin and Visher’s (1986) study confirms the fact that the effects of evidence strength decrease when extra-legal factors are taken into account in statistical analyses, such that omitting the effects of these factors is likely to lead to inflated measures of the effects of legal factors on trial outcomes.
7 Summary
Generally speaking, verdicts in criminal trials, including sexual assault trials, vary according to the strength of the case evidence. While the above studies show that various legal factors predict verdicts in sexual assault trials (the defendant’s prior criminal record, number of offenders, use of a weapon, medical evidence of victim injury, including ano-genital injury, other physical evidence, such as DNA, corroborating evidence from witnesses, prompt report to police and young age of complainant), they also tell us that jurors are influenced by various extra-legal factors that moderate the effects of legal factors. One of these is the relationship between victim and defendant so that jurors are more likely to convict when the parties are strangers to each other.
Studies on the influence of extra-legal factors have found that mock jury verdicts are influenced by juror demographics (gender and SES), as well as negative perceptions of the complainant’s character or reputation (her sexual history, so-called ‘risky’ behaviours such as style of dress and her alcohol/drug use), the complainant’s SES, her degree of resistance, and her emotional displays while giving evidence. Similarly, studies involving actual jurors have found that a complainant’s negative moral character and ‘risky’ behaviour predicted acquittals, although a juror’s gender and conservative or liberal value system influences their interpretation of a complainant’s character. In addition, complainants’ post-behavioural responses to sexual abuse influence verdicts in CSA trials.
Jurors have also been influenced by defendant characteristics, such as employment and SES, as well as the race and attractiveness of both the complainant and defendant and emotional displays by the complainant. It should be noted, however, that Devine and Caughlin (2014) concluded from their meta-analysis of juror and defendant characteristics and judgements of guilt that defendant race had limited impact in criminal trials, generally, while defendant attractiveness had no impact.
However, female jurors are ‘notably more conviction prone’ in sexual assault cases compared to other types of criminal cases, such that case type ‘moderate[s] the relationship between juror gender and judgments of defendant culpability’ (Devine & Caughlin, 2014: 118). Indeed, there is evidence to support Kalven and Zeisel’s (1966) liberation hypothesis, discussed in Chapter 2, that juries’ verdicts are more susceptible to the influence of extra-legal or non-evidentiary factors in cases with ambiguous or weak evidence, that is, with little or no corroborating evidence. It appears that the leniency effect observed in criminal trials will be exacerbated in sexual assault trials with ambiguous evidence and by the extent to which jurors rely, consciously or unconsciously, on one or more of the extra-legal factors discussed above which have been found to exert strong, independent effects on credibility ratings and verdict.
Key studies tell us that the match or mismatch between complainant and gender/victim stereotypes will influence jurors’ credibility perceptions and verdicts. For most jurors in a sexual assault trial, it can be expected that ‘many of their questions [will] remain … unanswered’ and they will try ‘to fill in the gaps … by drawing inferences about what may or may not have happened’ (Taylor & Joudo, 2005: 61). Without corroborating evidence, it appears that jury deliberations create a vignette of consensual sex based on jurors’ own assumptions and beliefs which then determines how they interpret the circumstances of the case at hand and the complainant’s evidence (Taylor & Joudo, 2005).
These results suggest that reforms to the sexual assault trial need to focus on ways to counter the negative credibility perceptions of complainants which are based on both gender and victim stereotypes, since ‘people may use their overall judgement of the gender stereotypicality of the complainant to anchor’ other expectations about victim behaviour, such as emotionality (Schuller et al., 2010: 776). These issues are discussed in more detail in Chapter 5.
In the next chapter, I explore jurors’ and laypeople’s perceptions of guilt and responsibility in the context of drug and/or alcohol use by the complainant, given that a majority of adult sexual assaults involve consumption of these substances by either the complainant, the defendant or both.