Children's perceived emotional behavior at disclosure and prosecutors' evaluations.

G Model CHIABU-2738; No. of Pages 12

ARTICLE IN PRESS Child Abuse & Neglect xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Child Abuse & Neglect

Children’s perceived emotional behavior at disclosure and prosecutors’ evaluations Paola Castelli a,b,∗ , Gail S. Goodman b a b

John Cabot University, Rome, Italy University of California, Davis, USA

a r t i c l e

i n f o

Article history: Received 5 September 2012 Received in revised form 20 February 2014 Accepted 24 February 2014 Available online xxx

Keywords: Child witness Emotion Child sexual abuse Abuse disclosure Prosecution

a b s t r a c t The present study investigated the perceived emotional behavior of alleged child victims when disclosing sexual abuse in a forensic interview. It also addressed whether the perceived emotional behavior influenced prosecutors’ evaluations of children’s potential as witnesses and prosecutors’ recommendations to press charges. Ninety-eight videotapes of forensic interviews with alleged child sexual abuse victims (4- to 17-year-olds) were coded for behavioral indicators of emotions. Case file information and district attorney evaluations were also coded. Results indicated that children were not generally perceived as being emotional (e.g., sad) during disclosure. However, the perceived intensity of expressed emotions was greater when children disclosed the alleged abuse compared to when they discussed more neutral topics in rapport building. Greater perceived emotional withdrawal by children at disclosure was associated with more negative evaluations of child witnesses by prosecutors. Moreover, children’s emotional behaviors, as noted by prosecutors, were among the predictors of prosecutors’ recommendations to file charges. Practical implications are discussed. © 2014 Elsevier Ltd. All rights reserved.

Introduction “Shanley’s accuser broke down on the stand while detailing years of alleged abuse” (The Boston Channel, 2005). “A man. . . tearfully testified that Shanley would pull him from catechism classes. . . His voice cracking, a hand over his face, the man also said Shanley would wait for him in the bathroom with the lights off” (The Washington Post, 2005). (Defense lawyer) “Mondano asked the judge to declare a mistrial, contending the man’s emotional outbursts during his testimony would taint jurors and prejudice them against his client” (Associated Press, 2005). These statements, taken from press coverage of the trial of Paul Shanley (one of the central figures in the Boston Archdiocese’s clergy sex abuse scandal), seem to support two common beliefs people hold regarding abuse victims: abuse disclosure should be accompanied by negative emotion (e.g., sadness) and such emotional behavior can be used to gauge the veracity of an abuse allegation. Indeed, research indicates that child victims of sexual abuse are deemed more credible when displaying distress as opposed to neutral demeanor while testifying (Goodman et al., 1992; Regan & Baker, 1998). Recent empirical evidence, however, highlights that children’s reports of abuse are not typically accompanied by visible emotional behaviors, at least when the disclosure occurs in the context of investigative interviews (Katz et al., 2012; Sayfan, Mitchell, Goodman, Eisen, & Qin, 2008; Wood, Orsak, Murphy, & Cross, 1996).

∗ Corresponding author. http://dx.doi.org/10.1016/j.chiabu.2014.02.010 0145-2134/© 2014 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Castelli, P., & Goodman, G.S. Children’s perceived emotional behavior at disclosure and prosecutors’ evaluations. Child Abuse & Neglect (2014), http://dx.doi.org/10.1016/j.chiabu.2014.02.010


2

ARTICLE IN PRESS P. Castelli, G.S. Goodman / Child Abuse & Neglect xxx (2014) xxx–xxx

Prosecuting decisions are likely influenced by children’s performance (e.g., if and how children disclose) in forensic interview contexts, given that these interviews are often seen as previews of children’s performance at trial and also given that videotapes of the interviews may be presented at hearings or at trials, in addition to or in place of children’s live testimony (Crawford v. Washington, 2004; Lyon & Dente, 2012; McWilliams et al., 2014). Therefore, it is important to explore children’s perceived emotional behavior during a forensic interview, and whether children’s perceived emotional behavior may affect prosecutors’ evaluations and decisions. Emotional Behavior During Disclosure of Abuse In a study of the effects of court involvement on alleged child sexual abuse victims, Goodman et al. (1992) reported that 65% of the children who testified in court exhibited some form of distress, such as crying. The few extant studies addressing children’s out-of-court disclosure of abuse, however, paint a different picture (Katz et al., 2012; Sayfan et al., 2008; Wood et al., 1996). Wood et al. (1996), for example, coded the behaviors displayed by children during forensic interviews in suspected child sexual abuse cases. Contrary to common belief, and to the children in Goodman et al.’s (1992) study, the majority of children displayed relaxed, neutral behaviors or relatively flat affect during the interview rather than sadness or shame. Sayfan et al. (2008) also coded maltreated children’s expressed emotion in forensic interviews and again found that the children displayed mainly neutral affect when disclosing alleged abuse. Additionally, Sayfan et al. (2008) reported that neither child age nor gender significantly predicted negative affect at disclosure. However, abuse frequency was significantly related to displayed affect, such that the greater the number of previous abuse allegations, the less likely children were to appear upset. Non-verbal emotion indices (i.e., stress, physical disengagement, positive and negative emotions) were seldom observed by Katz et al. (2012) in their sample of alleged abuse victims. However, when the investigative interview phases were each considered separately, stress indicators appeared to be more prominent in the rapport building and substantive phase of the interview compared to the introductory phase. Similarly, positive emotions were more prominent in the introductory compared to the rapport-building phase, and more prominent in the rapport building compared to the substantive phase of the interview. These findings suggest that interview phase is an important factor to consider when addressing children’s perceived affect when disclosing abuse. The evidence thus far suggests that children’s demeanor when disclosing abuse in forensic interviews is perceived as relatively neutral, especially when reporting frequent maltreatment (Sayfan et al., 2008). It is worth noting, however, that methodological limitations may have precluded detection of emotional behaviors and their predictors in the reviewed studies. Wood et al. (1996), for example, did not explore whether frequency and severity of disclosed abuse could have potentially influenced emotional display. Although the effect of these variables was addressed by Sayfan et al. (2008), these authors only obtained a general “negative affect” rating; thus, subtle differences in emotional behavior might have gone unnoticed. Katz et al. (2012) used a more sensitive measure of emotion behavior by time coding for specific behavioral signs and facial expression, and comparing them across interview phases, which may increase the probability of detecting emotional behavior. However, because the videotape sound was muted during coding, important cues to distress, as indicated in the children’s voices, may have been missed. It is also of interest to determine whether the type and intensity of children’s emotional behaviors when disclosing alleged abuse is noted by prosecutors, who likely look for emotional signs when evaluating potential witnesses (e.g., because prosecutors may try to anticipate jurors’ reactions). Further research on alleged victims’ emotional behavior during forensic interview is thus needed. Emotional Behavior, Victim Credibility, and Prosecution Understanding whether disclosure in forensic interviews is accompanied by emotional behavior is an important issue because research suggests that emotional behaviors are often considered signs of victim credibility (e.g., Cooper, 2012; Kaufmann, Drevland, Wessel, Overskeid, & Magnussen, 2003; Nadler & Rose, 2003; Vrij & Fischer, 1996). Indeed, victim demeanor (e.g., nervousness, embarrassment, crying) while testifying was one of four factors most likely to be mentioned by jurors in child sexual abuse trials as influencing their decision to believe the child victim witness (Myers, Redlich, Goodman, Prizmich, & Imwinkelried, 1999). Similarly, in a mock jury trial for battery against a child (Goodman et al., 2006), the majority of mock jurors rated child witness emotion, facial expression, body language, and nervousness as being important factors in their decision to believe the child, regardless of whether they were presented with the child’s live testimony or with the videotape of the child’s forensic interview. Whereas the link between emotion display and credibility has been extensively studied with adult alleged victims (Baldry, 1996; Bollingmo, Wessel, Sandvold, Eilertsen, & Magnussen, 2009; Bothwell & Jalil, 1992; Kaufmann et al., 2003; Winkel & Koppelaar, 1991), only a few experimental studies have addressed the effects of emotional expression on the credibility of child victims. Regan and Baker (1998) reported that mock jurors expected a 6-year-old child to cry and show fear and confusion upon confronting a defendant in court, and they were more likely to convict the defendant when they read about an incest trial in which a young child cried upon first seeing the defendant in the courtroom. More recently, Cooper (2012) addressed the effects of emotional display (as portrayed in a drawing) on child victims’ credibility in a mock child sexual abuse trial. Of interest, Cooper found that child victim emotional demeanor (i.e., calm vs. crying) as objectively Please cite this article in press as: Castelli, P., & Goodman, G.S. Children’s perceived emotional behavior at disclosure and prosecutors’ evaluations. Child Abuse & Neglect (2014), http://dx.doi.org/10.1016/j.chiabu.2014.02.010



3

defined by the researchers did not significantly affect mock jurors’ perceptions of children’s credibility. However, when mock jurors’ perceptions of victim demeanor were considered, child victims displaying calm behavior but who nevertheless were perceived as emotional were rated as more credible than child victims displaying crying, and as more credible than child victims perceived as unemotional. This result is intriguing, as it suggests that victims’ overt emotional display may be less relevant than their perceived emotionality in affecting credibility judgments. Moreover, it suggests that intense display of emotion (i.e., crying) may not always be associated with the greatest perceived credibility. Indeed, when Golding and collaborators (Golding, Fryman, Marsil, & Yozwiak, 2003) investigated the effects of victims’ crying behavior (as depicted in an alleged courtroom drawing) on mock jurors’ perceptions of child credibility, they similarly found teary children to be believed more than children displaying calm behavior or hysterical crying. The intensity of expressed emotion should thus be taken into account when addressing perceptions of child victims’ credibility. Child victims’ emotional behavior affects not only lay-people’s evaluations of children, but also legal authorities’ evaluations. In a survey of professionals, Melinder, Goodman, Eilertsen, and Magnussen (2004) found that judges, prosecutors, and police officers, among others, considered strong emotional expressions as a more reliable sign of witness credibility than calm behavior (but see Wessel, Drevland, Eilertsen, & Magnussen, 2006). Thus, empirical evidence seems to indicate that lay people and professionals alike may hold stereotypical beliefs and normative expectations regarding the proper emotional reaction that an alleged victim should display. However, no study to date has directly addressed whether child victims’ perceived emotional behaviors affect prosecutors’ evaluations of children as potential witnesses. This is an important issue, especially to the extent that it may affect prosecutors’ decisions to press charges. Because neutral or stunted emotional behavior potentially has a negative effect on jurors’ views and adversely affects child victim credibility (but see Cooper, 2012), prosecutors may feel hesitant to pursue a case when the victim does not display sufficient emotion. Indeed prosecutors often observe forensic interviews to determine if the child victim will have “jury appeal,” that is, whether a child will be a credible witness, with credibility possibly defined in part by whether or not the child displays emotional behavior. Given that research reveals that children in forensic interviews often display neutral affect when disclosing, it is important to address how their perceived emotional behavior may affect the likelihood of their cases being prosecuted. A handful of studies have investigated prosecutorial decisions in child sexual abuse cases (e.g., Finkelhor, 1983; MacMurray, 1989; Spears & Spohn, 1996; Walsh, Jones, Cross, & Lippert, 2010) and have found that case characteristics consistently predict likelihood of prosecution. Specifically, more severe alleged abuse (in terms of number of incidents, severity of the offense, and use of force) has been found to increase likelihood of prosecution (Cross, De Vos, & Whitcomb, 1994), as has the presence of additional evidence (e.g., other victims/witnesses, defendant’s confession, physical evidence; see Walsh et al., 2010). Defendant characteristics, such as being a “non-family” offender or having a prior police record, also increase the likelihood of prosecution (Finkelhor, 1983; Spears & Spohn, 1996; but see MacMurray, 1989). Victim age is a consistent predictor of prosecution, such that children older than 7 years of age are more likely to see their case prosecuted than are younger children or teenagers (e.g., Cross et al., 1994). Analyzing substantiated sexual abuse cases from the National Reporting System for Child Abuse and Neglect, Finkelhor (1983) indeed reported a bimodal distribution for child age, such that district attorneys were more likely to prosecute cases involving child victims between the ages of 7 and 12 compared to cases involving younger and older children (see also Walsh et al., 2010). However, to our knowledge, no previous studies have investigated the role of alleged victims’ perceived emotional behavior in prosecutors’ decision-making.

The Present Study Two main goals of our study were to explore children’s perceived emotional behaviors when disclosing alleged abuse in a forensic interview, and the potential predictors of perceived emotional intensity. Two main goals of our study were to explore children’s perceived emotional behaviors when disclosing alleged abuse in a forensic interview, and the potential predictors of perceived emotional intensity. Similarly to Katz et al. (2012), we separately coded the rapport-building and substantive phases of the interviews where neutral and abuse-related topics are discussed, respectively. This procedure ensured we could compare perceived emotional behaviors at disclosure phase with, in effect, a preceding baseline measure (rapport-building phase), facilitating detection of potential changes in perceived affect and allowing us to address whether the perceived changes were congruent with the topics of conversation (e.g., if the children appeared more distressed when they disclosed abuse compared to when they discussed neutral topics during rapport building). We were also interested in examining whether child and case characteristics are associated with perceived emotional behaviors. Several predictors were considered. Research indicates that the ability to understand emotion display rules and regulate one’s expressive behavior gradually develops during childhood (Cole, 1986; Saarni, 1984; Sodian, 1991). Indeed, older children were found to display more signs of distress than younger children in forensic interviews (Katz et al., 2012). Thus, we expected child age to be associated with greater perceived negative emotional behavior (i.e., distress) at disclosure. Severity and frequency of abuse, and relationship to perpetrator, were also considered as possible predictors of perceived emotional behaviors. Severe and frequent abuse, as well as a close relationship to the perpetrator, has been linked to poorer adjustment (Browne & Finkelhor, 1986; Kendall-Tackett, Williams, & Finkelhor, 1993; Mennen & Meadow, 1995). During a forensic interview, particularly during disclosure, poorer adjustment could be associated with greater feelings of distress. Please cite this article in press as: Castelli, P., & Goodman, G.S. Children’s perceived emotional behavior at disclosure and prosecutors’ evaluations. Child Abuse & Neglect (2014), http://dx.doi.org/10.1016/j.chiabu.2014.02.010



4

Thus, we explored whether these variables would be associated with greater perceived negative emotional behavior during disclosure. Forensic interviews may affect prosecutorial decisions (e.g., to charge the case) and may be presented as evidence in legal proceedings in addition to or in place of the child’s live testimony (but see Crawford v. Washington, 2004). Therefore, an additional goal of of our study was to investigate whether perceived emotional behavior influenced prosecutors’ evaluations of the alleged child victims and recommendations for prosecution. Previous research shows that lay people and professionals alike are less likely to believe abuse victims when they fail to display emotions, or display emotions incongruent with the topic disclosed (e.g., Regan & Baker, 1998). We therefore predicted that prosecutors would be more likely to provide positive evaluations of child victims as potential witnesses when children’s perceived emotional behaviors matched those generally expected from victims who disclose abuse (e.g., sadness, crying). We also explored whether perceived emotional behavior would influence the likelihood of attorneys’ recommendations of prosecution. To achieve these goals, videotapes of forensic interviews of child abuse victims were coded for perceived emotional behaviors, and prosecutors’ judgments were also measured. Method Participants Forensic interviews of 98 alleged child victims of sexual abuse (79.6% females) were included in the present study. Children’s age at the time of interview ranged from 4 to 17 years (M = 10.06; SD = 3.59). Forty-five percent of the children were Caucasian, 24% were African American, 21% were Hispanic, 3% were Asian American, and 7% were of other ethnicity. In the majority of cases (68%), the alleged abuse was perpetrated by a parent/step-parent or another relative. Of the disclosed abuse, 50% involved genital penetration (47% in the prosecutor-rated subsample). Sixty-eight percent of children were allegedly abused more than once. In 83 of the 98 cases, prosecutors made a recommendation. Specifically, 5% of these cases were not recommended for filing, 23% were conditional to finding additional corroboration (and thus not recommended for prosecution at the time of our coding), and the remaining 72% received a recommendation of the charges to file. Among these 83 cases, 69% involved some form of physical or psychological coercion. Corroboration (i.e., medical evidence, witnesses, or confession from defendant) was present only in 11% of these cases. However, in 30% of the cases, other victims were involved, and in 18% of the cases, the defendant had a prior record of some kind. Materials Videotaped interviews. The videotapes were made at a Child Advocacy Center (CAC) that serviced a large county in a west coast region of the United States. State law required all alleged child victims under age 14 to be interviewed at the CAC for forensic purposes. From among the forensic interviews conducted within the most recent past two years of the study, we selected cases that satisfied the following requirements: (a) a sufficiently good quality audio/video recording was available to ensure valid coding, (b) the child disclosed the alleged abuse, (c) case information was available (e.g., type of alleged abuse, relationship to the perpetrator), and (d) DAs’ evaluative comments (n = 95 cases) or recommendations about prosecutions (n = 83 cases) were obtainable. Behavioral Indicators of Emotion Rating Scale (BIERS). A coding system for behavioral indicators of emotion was developed for the present study. Nine emotion categories were included in the coding system: Happiness (e.g., smiling, eagerness to talk), sadness (e.g., frowning, serious look on face), shame/embarrassment (e.g., hiding face, low tone of voice), anxiety/nervousness (e.g., fidgety, body rubbing), withdrawal (e.g., maintaining distance, covering mouth while talking), avoidance (e.g., avoiding eye contact, lack of conversation), distress (fidgeting, being agitated, confrontational), tears or crying, and being relaxed (keeping eye contact, relaxed posture). The 9 variables were each coded on a 5-point scale (0 = never to 4 = extremely). Note that for happiness, a score of 0 indicated the absence of signs of happiness, which can mean neutral affect. Case information. The interview center’s case files and prosecutors’ files provided access to substantial information about each case (e.g., alleged severity of abuse, use of coercion, other evidence, recommendation for prosecution). These reports were generally gathered from child protection workers and law enforcement, as well as through SCAN (Suspected Child Abuse and Neglect) multidisciplinary team meetings where information is exchanged with district attorneys, child protective service workers, law enforcement officers, and medical staff. Several variables gleaned from case files were utilized in the present study. First, relationship to the alleged perpetrator was classified as either 1 = a family member (e.g., parent, sibling, relative) or 2 = not a family member (e.g., neighbor, friend). Previous research indicates differences in likelihood of prosecution for intra-familial versus extra-familial abuse (Finkelhor, 1983; MacMurray, 1989). Severity of abuse was classified as 0 = low to moderate (i.e., no genital penetration involved) or 1 = severe (i.e., genital penetration occurred). Coercion is the intimidation of victims through physical force (e.g., holding victim down) or psychological pressure (e.g., “If you tell, I’ll be arrested”; “If you tell, you’ll go into foster care and never see your mother again”) to induce them to participate in acts against their will. It is often considered as an additional indicator of severity of abuse (Spears & Spohn, 1996). The presence of physical or psychological coercion by the perpetrator, as reported Please cite this article in press as: Castelli, P., & Goodman, G.S. Children’s perceived emotional behavior at disclosure and prosecutors’ evaluations. Child Abuse & Neglect (2014), http://dx.doi.org/10.1016/j.chiabu.2014.02.010



5

in the case file, was coded (0 = absent, 1 = present). Frequency was coded on a 4-point scale (1 = once, 2 = more than once, 3 = more than 3 times, 4 = more than 10 times), reflecting the way in which this measure was collected and reported by the forensic interviewers. From the prosecutors’ files, four variables were derived: The first two variables represented whether or not the defendant had a prior police record of any kind (i.e., priors), and whether or not other evidence (i.e., other victims, physical evidence, defendant’s confession) supported the child allegation (i.e., evidence). A third variable (i.e., prosecutor evaluation) reflected the prosecutor’s appraisal of the child witness on a 4-point scale (0 = negative evaluation, 1 = somewhat negative, 2 = somewhat positive, 3 = positive evaluation). The last variable (i.e., prosecutor recommendation) concerned whether or not the case was recommended for prosecution (0 = not recommended, 1 = not currently recommended, but recommended if corroboration is obtained, 2 = recommended).

Procedure The study was approved by the university Institutional Review Board. Prior to the researchers’ involvement, child participants were interviewed individually as a part of the forensic investigations. All interviews were conducted at a CAC by highly trained social workers. The interviews usually started with a rapport-building phase, during which the child was asked questions about relatively neutral events (e.g., his/her last birthday) and about the difference between truth and lie in order to familiarize him/her with the interviewer and with the demands of the interview setting. Instructions (e.g., “Don’t guess”) were also provided to the children. With young children, the rapport-building phase also included a brief assessment of children’s general cognitive capacity (e.g., knowledge of colors and prepositions such as up and down). After the truth and lie test, the substantive phase of the interview began. The interviews were semi-structured according to a funnel approach. In this approach, the child is first asked open-ended questions and then more specific questions about possible abuse. Leading questions are reduced to a minimum (if employed at all), while highly leading, suggestive, and/or coercive questioning is not used. Thus, any emotional behavior detected can be attributed more to the nature of the topics discussed than to an inappropriate interview style (We did not code for interview quality. Without an objective record of the assault, we could not know with certainty if questions were correctly leading or misleading, conducive, and so forth.). All children were interviewed in a child-friendly room appropriate to their age level, while being videotaped from behind a one-way mirror. An assistant district attorney, a police officer, and at times a Child Protective Service worker observed the interview from behind a one-way mirror. Parents, defendants, and others were not permitted to observe. The majority of the interviews lasted between 20 and 40 minutes, with a few exceptions (M = 34 minutes; range: 10–93). Immediately after the interview, the attorney provided a written evaluation of the child as a witness and (in 85% of the cases) a recommendation regarding prosecution. Videotapes of the forensic interviews were coded using the BIERS. Children’s behavioral expressions of emotions were rated after each interview phase (i.e., rapport building vs. disclosure phase). Specifically, after the rapport-building phase was concluded, raters stopped the videotapes and provided a global rating of the perceived intensity of children’s emotional behavior on each of the dimensions presented on the BIERS. The same procedure was repeated at the end of the disclosure phase of the interview. Because emotion is conveyed not only via nonverbal behavior but also, of considerable importance, through the voice, the videotapes were coded with the audio on. This promoted valid coding of emotion and also served our interest in the generalizability of our findings to lay persons’ and prosecutors’ judgments of child witnesses, as such fact finders would also be able to hear the children’s voices. To establish inter-rater reliability, the first author and another researcher (blind to study hypotheses) first jointly coded several tapes (excluded from the current sample) and resolved disagreements through discussion. Then, each rater independently coded 10 videotaped interviews (10.2%). The raters were blind to each other’s rating and to the case information. Reliabilities were calculated within a 1-scale-point as an agreement. Proportion of agreement was higher than .90 on the majority of variables. Exceptions were for relaxed behavior (.80), and shame/embarrassment (.85). Once reliability was established, and due to the CAC’s imposed restrictions on access to the tapes, the first author coded the remaining interviews. Two independent coders, blind to the study hypotheses, coded the prosecutor file information. Proportion of agreement between the coders was higher than .87 on all variables. Prosecutors’ evaluations were coded in two ways. First, the overall comments were considered and were given a score of 3 when they contained all positive evaluations (e.g., “Very effective. Great memory. She’ll carry case”). A score of 2 was given to comments containing mostly positive evaluations (e.g., “Articulate, believable, but somewhat inconsistent”), a score of 1 was assigned when comments contained mostly negative statements (e.g., “All over the place. Has a hard time describing stuff. Inconsistent, and not a good historian, but I think she’s credible”), and a score of 0 was given when the comments contained only negative evaluations (e.g., “Reluctant to discuss, inconsistencies. Doubt she can carry the case”). Thus, the prosecutor evaluation variable mainly reflected prosecutors’ perception of children’s potential effectiveness on the witness stand (from least effective to most effective). Second, recurrent themes in prosecutors’ comments were also identified. Four main themes emerged: Child appearance (e.g., cute, looks mature), which was mentioned in 21% of the cases; child emotion (e.g., soft spoken, shy, reluctant, emotional), mentioned in 39% of the cases; child credibility (e.g., credible, believable), present in 52% of the cases; child report (e.g., Please cite this article in press as: Castelli, P., & Goodman, G.S. Children’s perceived emotional behavior at disclosure and prosecutors’ evaluations. Child Abuse & Neglect (2014), http://dx.doi.org/10.1016/j.chiabu.2014.02.010



6

Table 1 Mean intensity ratings of discrete emotional behaviors by interview phase (N = 98).

Emotional behavior (0 = never to 4 = extremely) Ashamed/embarrassed Anxious/nervous Withdrawn Avoidant Sad Distressed Crying Happy Relaxed

Rapport building

Disclosure

.49 (.80) .40 (.63) .22 (.63) .14 (.56) .48 (.76) .05 (.30) .00 (.00) .66 (.81) .96 (.88)

.95 (1.12) .56 (.91) .63 (1.11) .57 (1.09) 1.22 (1.15) .34 (.80) .16 (.56) .12 (.44) .68 (.90)

Note: Standard deviations in parentheses.

consistent, articulate), mentioned in 59% of the cases. Four dummy variables were thus created to reflect whether or not each category was mentioned in the prosecutors’ comments (i.e., 0 = not mentioned and 1 = mentioned). Results The analysis plan relied on analysis of variance (ANOVA) to examine behavioral indicators of emotion displayed by the children during the forensic interview. Regression analyses were then conducted to identify unique predictors of the children’s emotional displays, prosecutors’ evaluations of children, and recommendation to file charges. All significant effects are reported. Preliminary Analyses Three mean composite scores were created from the BIERS ratings and used in further analyses: Sadness/distress (sad, distress, cry; Cronbach ˛ = .66), Embarrassment/withdrawal (ashamed/embarrassed, anxious/nervous, withdrawn, avoidant; Cronbach ˛ = .77), and Positive affect (Happy, relaxed; Cronbach ˛ = .53). Preliminary analyses revealed no significant difference in perceived emotion expression between males and females; thus, gender differences are not considered further. Behavioral Indicators of Emotion during Forensic Interviews As can be gleamed from the means displayed in Table 1, children’s perceived emotionality was generally not intense. Indeed, except for sadness, which 65% of the children were perceived to display at least “a little bit,” the majority of children were not judged to display any behavioral sign of emotions. To assess whether a change in overall perceived emotional behavior occurred when children began talking about the abuse allegations, mean-composite scores for perceived emotional behaviors were entered as dependent variables in a 2 (Abuse severity: penetration vs. no penetration) × 3 (Emotion: sadness/distress vs. embarrassment/withdrawal vs. positive affect) × 2 (Interview phase: rapport-building vs. disclosure) mixed ANOVA, with abuse severity as a between-subjects factor, and expressed emotion and interview phase as within-subjects factors. Children’s age was entered as a covariate in the model. There were significant Emotion × Interview Phase and Emotion × Abuse Severity interactions, Fs (2, 95) ≥ 4.15, ps < .05, 2p > .04, which were further qualified by a significant Emotion × Interview Phase × Abuse Severity interaction, F (2, 95) = 3.70, p < .05, 2p = .04 (see Table 2). Simple effects analyses indicated that, compared to children in the rapport-building phase, children in the disclosure phase were perceived as more sad/distressed (rapport-building: M = .18, SD = .27; disclosure: M = .58, SD = .67), more embarrassed/withdrawn (rapport-building: M = .31, SD = .41; disclosure: M = .68, SD = .80), and less likely to display positive affect (rapport-building: M = .80, SD = .71; disclosure: M = .40, SD = .59), Fs (1, 97) ≥ 31.84, ps < .001, Table 2 Means for the Emotion × Interview Phase × Abuse Severity interaction (N = 98).

Sadness/distress Rapport building Disclosure Embarrassment/withdrawal Rapport building Disclosure Positive affect Rapport building Disclosure

No-penetration

Penetration

.17 (.23) .47 (.56)

.19 (.32) .70 (.76)

.25 (.37) .46 (.72)

.38 (.45) .90 (.83)

.84 (.74) .53 (.68)

.76 (.69) .27 (.43)

Note: Standard deviations in parentheses.


G Model

ARTICLE IN PRESS

CHIABU-2738; No. of Pages 12

P. Castelli, G.S. Goodman / Child Abuse & Neglect xxx (2014) xxx–xxx

7

Table 3 Correlation matrix for regression variables (perceived emotional behavior and prosecutor evaluations and recommendations). Age Age Ab-sev R-perp Ab-freq Priors Evid Coerc Sad/stress Emb/with Posaff Evals Prosrec

1 −.10 −.18 −.07 .12 .02 .01 .21* −.09 .17 .26* .04

Ab-sev

R-perp

Ab-freq

Priors

Evid

Coerc

Sad/stress

Emb/with

Posaff

Evals

Prosrec

1 .06 .20* .03 .11 .15 .16 .28** −.23* −.14 .27*

1 −.24* −.20* .02 .14 −.40** .05 .15 −.06 .11

1 .04 −.14 .11 .10 .07 −.28** −.06 −.05

1 −.08 .20* .09 .03 .06 .07 −.06

1 .13 .08 .06 .21* −.11 −.09

1 .02 .13 −.01 −.11 .30*

1 .34** −.39** .05 .06

1 −.32** −.44** .08

1 .01 −.11

1 .36**

1

Note: Ab-sev, abuse severity; R-Perp, relationship to perpetrator; Ab-freq, abuse frequency; Evid, evidence; Coerc, coercion; Sad/stress, sadness/distressed; Emb/with, embarrassment/withdrawal; Posaff, positive affect; Evals, prosecutors’ evaluation; Prosrec, prosecutor recommendation. * p < .05. ** p < .01.

2ps ≥.25. Analyses of simple effects further clarified that children’s perceived embarrassment/withdrawal was greater and their perceived positive affect was lower, when they were disclosing more severe compared to less severe alleged abuse, Fs (1, 96) ≥ 7.88, ps < .01, 2ps ≥.05. There was no significant change in perceived sadness/distress depending on the severity of the abuse disclosed. Predictors of Emotional Display at Disclosure To explore whether child and case characteristics would predict perceived emotional behavior at disclosure, mean composite scores were entered as dependent variables in a series of three separate regressions (see Table 3 for correlations among variables). In Model 1, child age was entered as an independent variable to control for developmental differences in the ability to display emotion. In Model 2, abuse characteristics (i.e., relationship to perpetrator, abuse severity and abuse frequency) were added. The first set of regressions investigated predictors of perceived sadness/distress (Table 4). In Model 1, which was significant, F (1, 96) = 4.32, p < .05, child age was a significant predictor, accounting for 4.3% of the variation in perceived sadness/distress. With age, children were rated as expressing greater sadness/distress. Adding case characteristics in Model 2, explained an additional 15% of variation in perceived sadness/distress, and this change in R2 was significant. Relationship to perpetrator was the most important predictor, uniquely explaining 11% of the variation in perceived sadness/distress at disclosure and suggesting that victims of intra-familial abuse were perceived as expressing more sadness/distress. Abuse severity approached significance (p = .07), indicating a tendency for victims of more severe abuse to be perceived as displaying more sadness/distress. Taken together, the variables in Model 2 explained 19% of the variability in perceived sadness/distress. The second set of regressions addressed predictors of perceived embarrassment/withdrawal. Neither Model 1 nor Model 2 was significant (Table 4). The third set of regressions investigated predictors of perceived positive affect (Table 4). Model 1 was not significant. Introducing case characteristics in Model 2, F (4, 96) = 3.74, p < .01, explained an additional 11% of variation in perceived positive affect, and this change in R2 was significant. Only abuse frequency was a reliable predictor, such that more frequent abuse was associated with lower perceived positive affect. Variables in Model 2 explained 14% of the variation in perceived positive affect at disclosure. Predictors of Prosecutors’ Ratings To examine whether the alleged victims’ perceived emotional behavior influenced prosecutors’ evaluations of them and the likelihood that the case would be recommended for prosecution, prosecutors’ evaluations, and prosecutors’ recommendations were entered as dependent variables in two separate series of regressions, respectively (see Table 3 for correlations among variables). The first set of regressions examined predictors of prosecutors’ evaluations of children as potential witnesses. Model 1 controlled for the effect of child age. Case characteristics previously found to affect perceived emotionality (i.e., relationship to perpetrator, abuse severity, and abuse frequency) were entered in Model 2. Model 3 tested for the additional effects of children’s perceived emotional behaviors. Model 1 was significant, F (1, 92) = 6.53, p < .05. Child age accounted for 7% of the variations in prosecutors’ evaluations, indicating that older children were more likely to be evaluated as effective witnesses than younger children (ˇ = .26, p < .05). The inclusion of abuse characteristics in Model 2 did not significantly improve the model. Perceived emotional behaviors entered in Model 3, F (7, 86) = 5.01, p < .001, explained an additional 21% of variation in prosecutors’ evaluations, and this Please cite this article in press as: Castelli, P., & Goodman, G.S. Children’s perceived emotional behavior at disclosure and prosecutors’ evaluations. Child Abuse & Neglect (2014), http://dx.doi.org/10.1016/j.chiabu.2014.02.010

G Model

ARTICLE IN PRESS



8

Table 4 Predictors of perceived emotional behavior at disclosure (N = 96). Variables Sadness/distress Model 1 Intercept Age Model 2 Intercept Age R-perp Ab-sev Ab-freq Embarrassment/withdrawal Model 1 Intercept Age Model 2 Intercept Age R-perp Ab-sev Ab-freq Positive affect Model 1 Intercept Age Model 2 Intercept Age R-perp Ab-sev Ab-freq * ** ***

B(SE)

.18 (.20) .04 (.02) .68 (.33) .42 (.02) −.48 (.14) .23 (.13) −.00 (.06)

.88 (.24) −.02 (.02) .62 (.41) −.01 (.02 −.02 (.18) .43 (.16) .01 (.07)

ˇ

R2

F for R2

.03

.04

4.32*

.15

.15

5.47**

−.00

.01

.74

.04

.07

2.41

.03

2.69

.11

4.01**

R2

.21*

.23* −.34*** .17 −.01

−.09*

−.06 −.01 .27 .01 02

.13 (.18) .03 (.02)

.17 .10

.32 (.29) .02 (.02) .15 (.13) −.20 (.12) −.10 (.05)

.14 .12 −.17 −.21*

p ≤ .05. p ≤ .01. p ≤ .001.

change in R2 was significant, F (3, 86) = 8.49, p < .001. Perceived embarrassment/withdrawal (ˇ = −.51, p < .001) was the only significant predictor, such that children perceived as less embarrassed/withdrawn received more positive evaluations from prosecutors. Taken together, variables in Model 3 explained 29% of the variation in prosecutors’ evaluation of the child. The second set of regressions explored potential predictors of recommendations for prosecution. Model 1 controlled for child age. Because children between the ages of 7 and 12 are more likely to see their cases prosecuted compared to younger and older children (see Finkelhor, 1983), a dummy variable was created for age (1 = 7- to 12-year-olds, 0 = others) and used in the current analysis. Abuse characteristics found to influence prosecution (i.e., relationship to perpetrator, abuse severity, abuse frequency, and coercion) were entered in Model 2. Other variables previously found to influence prosecutorial decisions (i.e., evidence, priors) were entered in Model 3. Lastly, Model 4 tested the contribution of perceived emotional behaviors in predicting likelihood to recommend filing charges. Model 1 was not significant. Entering abuse characteristics in Model 2, F (5, 76) = 2.96, p < .05, significantly improved the model, explaining an additional 16% of variation in prosecutorial decision. This change in R2 was significant, F (4, 76) = 3.71, p < .01. Abuse severity and coercion were the only reliable predictors (ˇ = .26 and ˇ = .28, respectively, ps < .05), indicating that cases involving penetration and use of coercion were more likely to receive a filing recommendation. Entering other case characteristics in Model 3, F (7, 74) = 2.89, p < .05, did not significantly improve the model. Similarly, none of the perceived emotional behaviors added in Model 4, F (10, 71) = 2.00, p < .05, were reliable predictors. Taken together, the variables in Model 4 explained 22% of the variation in prosecutors’ decision to recommend filing charges. Additional Analyses of Prosecutors’ Ratings A new series of regressions was conducted to explore whether the likelihood of prosecutors’ commenting on the child’s appearance, emotion, credibility, or report would predict their likelihood to recommend filing charges (see Table 5 for correlations among variables). Model 1 included child age. Model 2 added variables that were found to affect prosecution or perceived emotional behavior in previous analyses (i.e., relationship to perpetrator, abuse severity, abuse frequency, and coercion). Model 3 explored the added contribution of variables reflecting the presence of each of the recurring themes in prosecutors’ evaluations (e.g., child emotion). Model 1 and Model 2 mimicked the results reported in the previous analysis of prosecuting recommendations. Adding the variables in Model 3 reflecting whether prosecutors commented on children’s appearance, emotion, credibility, and Please cite this article in press as: Castelli, P., & Goodman, G.S. Children’s perceived emotional behavior at disclosure and prosecutors’ evaluations. Child Abuse & Neglect (2014), http://dx.doi.org/10.1016/j.chiabu.2014.02.010

G Model

ARTICLE IN PRESS



9

Table 5 Correlation matrix for regression variables (prosecutors’ comments). Age Age Ab-sev R-perp Ab-freq Coerc Evals Prosec Appear Emotion Credib Report

1 −.10 −.18 −.07 .01 .26* .04 −.19 .18 .26* −.04

Ab-sev

R-perp

1 .06 .20* .15 −.14 .27* .00 .15 −.02 .02

1 −.24* .14 −.06 .11 −.03 −.05 −.14 −.08

Ab-freq

Coerc

Evals

Prosec

Appear

Emotion

Credib

1 −.11 .30* −.20 .11 .02 −.05

1 .36** .13 −.15 .30** .03

1 .08 .19 .13 −.13

1 −.26* −.30** .01

1 −.06 −.13

1 −.10

Report

1 .12 −.06 −.05 −.11 −.15 .15 .08

1

Note: Appear, child appearance; Emotion, child emotion; Credib, child credibility; Report, child report. * p < .05. ** p < .01.

report, explained an additional 11% of variation in prosecuting decisions. The change in R2 was significant, F (4, 72) = 2.65, p < .05. Child appearance, child emotion, and child credibility were all significant predictors of prosecuting decisions (ˇ = .26, ˇ = .23, and ˇ = .25, respectively; ps < .05). Whether or not the prosecutors commented on the child’s appearance, emotion, and credibility significantly predicted likelihood of recommending the case for prosecution. Together, the nine variables in Model 3 explained 27% of the variability in recommendation for prosecution (Analyses were also conducted on a subsample of subjects for which we had corroborating evidence of the abuse, or other circumstantial evidence [i.e., other alleged victims]. The analyses mainly replicated the pattern of results reported.). Discussion Two main goals were to investigate children’s perceived emotional behavior during disclosure of alleged abuse in forensic interviews and predictors of perceived emotion intensity. Past research has shown that in formal forensic interviews, children tend to express neutral demeanor during disclosure (e.g., Katz et al., 2012; Wood et al., 1996). Our results suggest that children’s emotional behavior at disclosure, although not extremely intense, may be evident enough to be detected by observers. This is especially true when changes in behavior are considered. Indeed, as also reported by Katz et al. (2012), children’s emotional behaviors were perceived as more intense when they talked about the abuse compared to when they discussed more neutral topics. The detected change in perceived emotionality also appeared to be congruent with the severity of disclosed abuse incidents, such that disclosure of more severe abuse (i.e., involving penetration) was accompanied by an increase in perceived withdrawal/avoidance, and a decrease in perceived positive affect. This finding is important, given that lay people hold expectations about the emotional behavior abuse victims should express, and are particularly sensitive to its congruence with the topics discussed (Kaufmann et al., 2003; Nadler & Rose, 2003). Our results suggest that the children’s behavior did not violate such expectations. Consistent with previous research (Regan & Baker, 1998; Sayfan et al., 2008), child and case characteristics predicted the intensity of perceived emotional behavior during disclosure. Specifically, children’s age, relationship to perpetrator, and, to a lesser extent, abuse severity, predicted the intensity of perceived sadness/distress at disclosure. As children’s understanding of display rules develops with age (Cole, 1986), older children might be more likely to express affect congruent with the valence of the topics being discussed, increasing the chances that their emotional behavior will be detected by observers. Victims of intra-familial abuse may have more difficulty disclosing, as family members may pressure children to drop the allegation, children may feel they are betraying a loved one, and/or children may feel distressed at the possibility of sending a relative to jail and affecting family cohesion. Regarding the present study, this could explain the greater likelihood of detecting behavioral signs of sadness/distress when the alleged victims disclosed this type of abuse as opposed to abuse from someone external to the family. There was also a tendency for more severe abuse to be associated with greater perceived expressed sadness/distress. Severe abuse (i.e., involving penetration) is indeed more likely to have an emotional impact on the victim, sometimes triggering serious adverse reactions, such as PTSD or dissociation (Kendall-Tackett et al., 1993; Putnam, 1997). These emotional sequelae may influence emotional responses during disclosure in children (e.g., Sayfan et al., 2008). We found more frequent abuse to be related to less intense perceived positive affect at disclosure. It has been suggested that frequently abused children may learn to hide their emotional states from others to prevent further harm to themselves or family members or have psychic numbing from the abuse, often engaging in coping strategies to reduce their emotional awareness (Briere, 1992; Cole, Zahn-Waxler, & Smith, 1994). Indeed, victims of frequent abuse generally report feeling detached and numb (Long & Jackson, 1993) and could thus be expected to have a stunted affect at disclosure (see also Sayfan et al., 2008). Other major goals of our study were to examine whether perceived emotional behavior at disclosure would relate to prosecutors’ evaluations of alleged victims and whether the alleged victims’ perceived emotion behaviors (as coded by Please cite this article in press as: Castelli, P., & Goodman, G.S. Children’s perceived emotional behavior at disclosure and prosecutors’ evaluations. Child Abuse & Neglect (2014), http://dx.doi.org/10.1016/j.chiabu.2014.02.010


10


raters or detected by prosecutors) predicted recommendations for prosecution. Past research suggests that lay people’s credibility judgments of witnesses are influenced by witnesses’ emotional behaviors (Cooper, 2012; Kaufmann et al., 2003), and incongruent emotions may negatively affect potential jurors’ feelings toward victims/witnesses (Tsoudis & Smith-Lovin, 1998). Prosecutors may thus take displayed emotional behavior into consideration when judging alleged victims’ appeal as witnesses and when making recommendations for prosecution. Having prosecutors’ files available, we were in the favorable position to explore this issue. Prosecutors’ evaluations of alleged victims as potential witnesses were significantly associated with the children’s ages and perceived embarrassment/withdrawal at disclosure. Older children were more likely to be positively evaluated, whereas as the intensity of children’s perceived embarrassment/withdrawal increased, prosecutors were more likely to evaluate children negatively. This latter association held after statistically controlling for the effects of age. Developmental changes in memory, language, comprehension, and resistance to suggestion permit older children to better relate experienced events (Bauer & Fivush, 2014). This may explain, at least in part, why increased age was associated with a more positive overall evaluation of the children as potential witnesses. Indeed prosecutors were also more likely to mention credibility in their evaluation of older children. It is important to note that age was also associated with increased perceived sadness at disclosure. Thus, it is possible that older children’s greater appeal as witnesses may be partly attributed to their greater ability to display affect congruent with the disclosed topic. Conversely, greater embarrassment and withdrawal can be associated with behaviors that may or may not be considered effective on the witness stand, such as nervousness, low tone of voice, and gaze aversion. Although such behaviors may be considered signs of distress due to victimization, lay people may also interpret them as potential indicators of deception. Moreover, embarrassment/withdrawal may be related to greater reluctance to disclose, thus making the children less effective potential witnesses and explaining the greater likelihood of children who were perceived as embarrassed/withdrawn to be negatively evaluated. When considering findings on prosecutorial decisions, our results partly confirm previous research (e.g., Cross et al., 1994). Abuse severity and presence of coercion (itself generally considered an additional index of abuse severity) were found to predict prosecutors’ filing recommendations. It is reasonable to assume that with more severe abuse, the likelihood of having physical evidence is also greater, thus providing corroboration for the children’s testimony, and possibly increasing the chances of a positive trial outcome (although we could not directly test this possibility in our sample due to the small number of cases where information on physical evidence was provided). Similarly, prosecutors may feel that cases involving coercion are more likely to lead to a guilty verdict. Indeed the presence of coercion may be an indication of the alleged victims’ resistance of the abuse, thus reducing their potential blame and increasing jurors’ sympathy toward them. Prosecutors’ commenting on children’s credibility and appearance also predicted likelihood of prosecution. Indeed physical attractiveness is generally associated with greater perceived credibility and greater perceived trustworthiness (e.g., Goodman et al., 1998; Mazzella & Feingold, 1994). Thus, prosecutors may think that testimony from cute children has a higher chance of being believed by jurors. Credible witnesses undoubtedly have a greater chance to influence jurors’ guilty verdicts. It is thus not surprising that cases with more credible-seeming witnesses were also more likely to be prosecuted. Contrary to previous research, other characteristics such as children’s age or defendants’ prior record were not found to predict prosecutors’ recommendations in our sample (although children’s age did predict prosecutors’ evaluations). By exploring the role of perceived emotional behavior in prosecutorial decisions, we found a discrepancy in the likelihood that victims’ perceived emotionality would influence recommendations for prosecution. Specifically, when coded emotional behaviors were used as predictors, no significant effect on prosecutors’ decisions to file charges emerged. However, whether or not prosecutors mentioned the children’s emotional behaviors in the prosecutors’ evaluations was associated with prosecuting decisions, such that alleged child victims whose emotionality was noted by prosecutors were more likely to see their cases recommended for prosecution, even when other factors contributing to prosecuting decisions or associated with perceived emotional behavior were statistically controlled. This result appears to mimic those reported by Cooper (2012). In Cooper’s study, abuse victims’ credibility was not influenced by their displayed emotion. Rather, whether or not mock jurors perceived victims as emotional (regardless of actual display) affected credibility ratings. Similarly, prosecutors perceiving alleged victims as emotional when disclosing (as indicated by the prosecutors having commented on emotion behaviors) appeared to affect their decisions to prosecute. Taken together, the results of our study suggest that prosecutors may take lay people’s expectations about witnesses into consideration when evaluating the potential jury appeal of alleged victims, and when making decisions about whether or not a case should be pursued. This may be troublesome when the role of emotional behaviors is being considered. Indeed, whereas research indicates that victims displaying affect congruent with the disclosed abuse are more likely to be believed (Kaufmann et al., 2003), the effects of displayed emotion may still be somewhat idiosyncratic to the observing jurors (Cooper, 2012). Thus, what appears emotional to the observing prosecutors may not appear as such to a trial juror, and vice versa. Research on mock jurors’ perceptions of alleged adult rape victims also shows that, whereas display of congruent emotion affected jurors’ independent judgments of the witness’ credibility, jury deliberation resulted in greater credibility being attributed to the victim displaying a neutral emotional expression (Dahl et al., 2007). It is also important to note that children are likely not disclosing the abuse for the first time during the forensic interview, which may affect the perceived intensity of their emotional behavior. However, when testifying in court, factors other that disclosing the alleged abuse (e.g., having to face the defendants) may affect children’s perceived emotional behavior. Thus, alleged victims who are not perceived as Please cite this article in press as: Castelli, P., & Goodman, G.S. Children’s perceived emotional behavior at disclosure and prosecutors’ evaluations. Child Abuse & Neglect (2014), http://dx.doi.org/10.1016/j.chiabu.2014.02.010



11

intensely emotional in forensic interviews could actually appear emotional when on the witness stand. This suggests that perceived emotional display during forensic interviews may not be a successful screening criterion for prosecutable cases, and prosecutors should be alerted to the risk of relying on this criterion to determine whether to file charges or not, as it may prevent less emotional, but nonetheless effective witnesses from seeing their cases prosecuted. It is important to note that the small sizes underlying the detected effects, as well as the correlational nature of our study, preclude firm conclusions about the role of alleged victims’ emotionality on prosecutorial decision. The findings, however, do suggest that the area deserves further investigation, given the potential implication of relying on such decision-making heuristics. In summary, the results of this study further our understanding of children’s perceived emotion behaviors during disclosure of abuse, supporting previous findings of general neutral affect at disclosure and suggesting that the intensity of victims’ emotional behavior is influenced by factors such as victim age, relationship to perpetrator, abuse severity, and abuse frequency. Of importance, we shed initial light on how alleged victims’ emotional behaviors may affect prosecutors’ perceptions of children as witnesses, and likelihood to file charges. Limitations Some potential limitations in our study are worth noting. First, the videotaped interviews were coded with the audio on, and it may be contended that the content of the interviews might have influenced coders’ perceptions of expressed emotional behavior. As mentioned earlier, important emotional cues are present in one’s voice, which we would have missed with the sound muted. Moreover, given that both prosecutors and potential jurors also evaluate witnesses’ emotional behavior in relation to the verbal content of their testimony, we believe this approach is ecologically valid. Second, we were unable to determine whether the same or different prosecutors witnessed each of our coded interviews, or the reason for prosecutors’ absence or lack of comments during certain interviews. It may be possible that the cases where the prosecutor was present and left a comment in the record were already cases more likely to be prosecuted. Third, as we mentioned earlier, children in the videotaped forensic interviews were probably not disclosing the abuse for first time. It is reasonable to assume that with repeated recounting less emotional behavior will be displayed, thus influencing the intensity of children’s perceived emotion. Unfortunately, we lacked information regarding the time elapsed since the children’s first disclosure, or the number of times they were previously interviewed. Lastly, it is important to note that the effect sizes for prosecutors’ recommendations were fairly small. This may be a consequence of the small number of cases in our sample that were not recommended for prosecution, a limitation that might have also prevented other predictors (i.e., corroborating evidence) to emerge as significant. Future research should address these issues. Conclusions Recent research has begun to explore victims’ emotional expressions when recounting alleged abuse in formal interview contexts, as well as factors that may influence perceived emotional behaviors in such contexts. Findings converge in indicating that the commonly held belief that abuse disclosure is accompanied by intense negative emotion displays is not always supported. Previous research on the influence of victims’ perceived emotional behavior on prosecutors’ evaluations, however, is scarce, if not non-existent. Our results not only provide additional evidence about factors that may influence perceived emotional behavior in child victims when disclosing sexual abuse, but also begin to fill gaps in scientific knowledge of how victims’ emotional displays affect prosecutors’ decisions. Although firm conclusions are premature, our results provide initial evidence that prosecutors may rely, at least in part, on children’s emotional expressions during disclosure to evaluate whether the children will have sufficient appeal as witnesses and to determine whether the case warrants prosecution. Given the potential implications of our findings, it is important that the issues be further investigated, including in controlled experiments. Acknowledgments We thank the director and staff of the child abuse evaluation center where this study took place for their encouragement and support. We also thank Dr. Liat Sayfan, Dr. Emilie Mitchell, and Michelle Culver for their help with this project. References Associated Press. (2005). Emotional Shanley witness returns to court. Fox News. Retrieved December 12, 2010 from http://www.foxnews. com/story/0,2933,145700,00.html Baldry, A. C. (1996). Rape victims’ risk of secondary victimization by police officers. Issues in Criminological & Legal Psychology, 25, 65–68. Bauer, P. J., & Fivush, R. (Eds.). (2014). The Wiley handbook on the development of children’s memory. Malden, MA: Wiley Blackwell. Bollingmo, G. B., Wessel, E., Sandvold, Y., Eilertsen, D. E., & Magnussen, S. (2009). The effect of biased and non-biased information on judgments of witness credibility. Psychology, Crime and Law, 15, 61–71. Bothwell, R., & Jalil, M. (1992). The credibility of nervous witnesses. Journal of Social Behavior and Personality, 7, 581–586. Briere, J. N. (1992). Child abuse trauma: Theory and treatment of the lasting effects. In Interpersonal violence: The practice series. Thousand Oaks, CA: Sage Publications, Inc.



12


Browne, A., & Finkelhor, D. (1986). Impact of child sexual abuse: A review of the research. Psychological Bulletin, 99, 66–77. Cole, P. M. (1986). Children’s spontaneous control of facial expression. Child Development, 57, 1309–1321. Cole, P. M., Zahn-Waxler, C., & Smith, K. D. (1994). Expressive control during a disappointment: Variations related to preschoolers’ behavior problems. Developmental Psychology, 30, 835–846. Cooper, A. (2012). Big kids don’t cry. . .but what happens when they do? The effects of child witness demeanor, age, and gender on jurors’ perceptions in a child sexual abuse case. Order No. AAI3457257. Dissertation Abstracts International: Section B: The Sciences and Engineering, 4971. Retrieved from http://www.proquest.com (1001919431; 2012-99040-142). Crawford v. Washington. (2004). 541 U.S. 36. Cross, T. P., De Vos, E., & Whitcomb, D. (1994). Prosecution of child sexual abuse: Which cases are accepted? Child Abuse & Neglect, 18, 661–677. Dahl, J., Enemo, I., Drevland, G. C. B., Wessel, E., Eilertsen, D. E., & Magnussen, S. (2007). Displayed emotions and witness credibility: A comparison of judgments by individuals and mock juries. Applied Cognitive Psychology, 21, 1145–1155. Finkelhor, D. (1983). Removing the child – Prosecuting the offender in cases of sexual abuse: Evidence from the national reporting system for child abuse and neglect. Child Abuse & Neglect, 7, 195–205. Goodman, G. S., Myers, J. E. B., Qin, J., Quas, J. A., Castelli, P., Redlich, A. D., & Rogers, L. (2006). Hearsay versus children’s testimony: Effects of truthful and deceptive statements on jurors’ decisions. Law and Human Behavior, 30, 363–401. http://dx.doi.org/10.1007/ s10979-006-9009-0 Goodman, G. S., Taub, E. P., Jones, D. P. H., England, P., Port, L. K., Rudy, L., & Prado, L. (1992). Testifying in criminal court: Emotional effects on child sexual assault victims. Monographs of the Society for Research in Child Development, 57, v-142 Goodman, G. S., Tobey, A., Batterman-Faunce, J., Orcutt, H., Thomas, S., Shapiro, C., & Sachsenmeier, T. (1998). Face-to-face confrontation: Effects of closedcircuit technology on children’s eyewitness testimony and jurors’ decisions. Law and Human Behavior, 22, 165–203. Golding, J. M., Fryman, H. M., Marsil, D. F., & Yozwiak, J. A. (2003). Big girls don’t cry: The effect of child witness demeanor on juror decisions in a child sexual abuse trial. Child Abuse & Neglect, 27, 1311–1321. Katz, C., Hershkowitz, I., Malloy, L. C., Lamb, M. E., Atabaki, A., & Spindler, S. (2012). Non-verbal behavior of children who disclose or do not disclose child abuse in investigative interviews. Child Abuse & Neglect, 36, 12–20. Kaufmann, G., Drevland, G. C. B., Wessel, E., Overskeid, G., & Magnussen, S. (2003). The importance of being earnest: Displayed emotions and witness credibility. Applied Cognitive Psychology, 17, 21–34. Kendall-Tackett, K. A., Williams, L. M., & Finkelhor, D. (1993). Impact of sexual abuse on children: A review and synthesis of recent empirical studies. Psychological Bulletin, 113, 164–180. Long, P. J., & Jackson, J. L. (1993). Initial emotional responses to childhood sexual abuse: Emotion profiles of victims and relationship to later adjustment. Journal of Family Violence, 8, 167–181. Lyon, T. D., & Dente, J. (2012). Child witnesses and the confrontation clause. Journal of Criminal Law and Criminology, 102, 1181–1232. MacMurray, B. (1989). Criminal determination for child sexual abuse: Prosecutor case-screening judgements. Journal of Interpersonal Violence, 4, 233–244. Mazzella, R., & Feingold, A. (1994). The effects of physical attractiveness, race, socioeconomic status, and gender of defendants and victims on judgments of mock jurors: A meta-analysis. Journal of Applied Social Psychology, 24, 1315–1338. McWilliams, K., Augusti, E.-M., Dion, J., Block, S. D., Melinder, A. M., Cashmore, J., & Goodman, G. S. (2014). Children as witnesses. In G. B. Melton, A. Ben-Arieh, J. Cashmore, G. S. Goodman, & N. K. Worley (Eds.), The Sage handbook of child research (pp. 285–299). London, UK: Sage. Melinder, A., Goodman, G., Eilertsen, D. E., & Magnussen, S. (2004). Beliefs about the child witness: A survey of professionals. Psychology, Crime and Law, 10, 247–365. Mennen, F. E., & Meadow, D. (1995). The relationship of abuse characteristics to symptoms in sexually abused girls. Journal of Interpersonal Violence, 10, 259–274. Myers, J. E. B., Redlich, A. D., Goodman, G. S., Prizmich, L. P., & Imwinkelried, E. (1999). Jurors’ perceptions of hearsay in child sexual abuse cases. Psychology, Public Policy, & Law, 5, 388–419. Nadler, J., & Rose, M. R. (2003). Victim impact testimony and the psychology of punishment. Cornell Law Review, 88, 419–456. Putnam, F. W. (1997). Dissociation in children and adolescents: A developmental perspective. New York: Guilford Press. Regan, P. C., & Baker, S. J. (1998). The impact of child witness demeanor on perceived credibility and trial outcome in sexual abuse cases. Journal of Family Violence, 13, 187–195. Saarni, C. (1984). An observational study of children’s attempt to monitor their expressive behavior. Developmental Psychology, 55, 1504–1513. Sayfan, L., Mitchell, E. B., Goodman, G. S., Eisen, M. L., & Qin, J. (2008). Children’s expressed emotions when disclosing maltreatment. Child Abuse & Neglect, 32, 1026–1036. Sodian, B. (1991). The development of deception in young children. British Journal of Developmental Psychology, 9, 173–188. Spears, J. W., & Spohn, C. C. (1996). The genuine victim and prosecutors’ charging decisions in sexual assault cases. American Journal of Criminal Justice, 20, 183–205. The Boston Channel. (2005). Alleged victim testifies at ex-priest’s rape trial. The Boston Channel. Retrieved December 12, 2010, from http://www. thebostonchannel.com/r/4131834/detail.html The Washington Post. (2005). Nation in brief. The Washington Post. Retrieved December 12, 2010, from http://www.washingtonpost.com/ wp-dyn/articles/A40074-2005Jan26.html Tsoudis, O., & Smith-Lovin, L. (1998). How bad was it? The effects of victim and perpetrator emotion on responses to criminal court vignettes. Social Forces, 77, 695–722. Vrij, A., & Fischer, A. (1996). The role of displays of emotions and ethnicity in judgments of rape victims. International Review of Victimology, 4, 255–265. Walsh, W. A., Jones, L. M., Cross, T. P., & Lippert, T. (2010). Prosecuting child sexual abuse: The importance of evidence type. Crime & Delinquency, 56, 436–454. Winkel, F. W., & Koppelaar, L. (1991). Rape victims’ style of self-presentation and secondary victimization by the environment: An experiment. Journal of Interpersonal Violence, 6, 29–40. Wessel, E., Drevland, G. C. B., Eilertsen, D. E., & Magnussen, S. (2006). Credibility of the emotional witness: A study of court judges. Law and Human Behavior, 30, 221–230. Wood, B., Orsak, C., Murphy, M., & Cross, H. J. (1996). Semistructured child sexual abuse interviews: Interview and child characteristics related to the credibility of disclosure. Child Abuse & Neglect, 20, 81–92.


Emotional disclosure and victim blaming.

Preliminary study of perceived self-disclosure.

Physical attractiveness, self-disclosure, and helping behavior.

Emotional behavior and the brain.

Competitive behavior and perceived aggression.

Privacy Concerns and Disclosure Behavior in a Health Setting.

Self-Perceived Emotional Distress and Diabetes Risk Among Young Men.

Iron and mechanisms of emotional behavior.

Childrens Dentists Houston tx

Shoppers' perceived embeddedness and its impact on purchasing behavior at an organic farmers' market.

Psychosocial effects of perceived emotional synchrony in collective gatherings.

Social reactions to sexual assault disclosure and problem drinking: mediating effects of perceived control and PTSD.

Changes in distressing behavior perceived by family of persons with schizophrenia at home - 25 years later.

Prosecutors and Use of Restorative Justice in Courts: Greek Case.

Emotional correlates of type A behavior pattern.

Emotional Disclosure Through Journal Writing: Telehealth Intervention for Maternal Stress and Mother-Child Relationships.

A multidimensional look at sex and target differences in disclosure.

Insights From Twin Research: Conference at Royal Childrens Hospital, Melbourne, Australia.

Investigating the Effect of Gaze Cues and Emotional Expressions on the Affective Evaluations of Unfamiliar Faces.

Objective and Perceived Weight: Associations with Risky Adolescent Sexual Behavior.

Child Perceived Parenting Behavior: Childhood Anxiety and Related Symptoms.

Relationship between perceived discrimination and sedentary behavior in adults.

Compulsory citizenship behavior and organizational citizenship behavior: the role of organizational identification and perceived interactional justice.

Perceived parental behavior and the social desirability response set.