The Open Collections website will be unavailable July 27 from 2100-2200 PST ahead of planned usability and performance enhancements on July 28. More information here.

Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Caregiver support in child sexual abuse : an investigation into factors that impact delays of disclosure Wallis, Cassidy R. D. 2019

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata


24-ubc_2019_september_wallis_cassidy.pdf [ 860.56kB ]
JSON: 24-1.0379793.json
JSON-LD: 24-1.0379793-ld.json
RDF/XML (Pretty): 24-1.0379793-rdf.xml
RDF/JSON: 24-1.0379793-rdf.json
Turtle: 24-1.0379793-turtle.txt
N-Triples: 24-1.0379793-rdf-ntriples.txt
Original Record: 24-1.0379793-source.json
Full Text

Full Text

  Caregiver Support in Child Sexual Abuse: An Investigation into Factors that Impact Delays of Disclosure  By Cassidy R. D. Wallis B.A., MacEwan University 2016  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF  THE REQUIREMENTS FOR THE DEGREE OF  MASTER OF ARTS in THE COLLEGE OF GRADUATE STUDIES (Psychology) THE UNIVERSITY OF BRITISH COLUMBIA (Okanagan)  Supervised by Dr. Michael Woodworth  July 2019  © Cassidy R. D. Wallis, 2019 ii  The undersigned certify that they have read, and recommend to the College of Graduate Studies for acceptance, a thesis entitled: Caregiver Support in Child Sexual Abuse: An Investigation into Factors that Impact Delays of Disclosure   Submitted by Cassidy Wallis in partial fulfilment of the requirements of the degree of Master of Arts.  Dr. Michael Woodworth Supervisor, Professor Dr. Paul Davies Supervisory Committee Member, Professor Dr. Zach Walsh Supervisory Committee Member, Professor Dr. Harry Miller University Examiner, Professor Dr. Grant Charles External Examiner, Professor   iii  Abstract Child Sexual Abuse (CSA) is a pervasive and devastating offence estimated to impact 12% of children globally (Stoltenborgh, van IJzendoorn, Euser, & Bakermans-Kranenburg, 2011). For perpetrator conviction or disruption of contact, victim testimony is often required in leu of physical evidence (Tashjian, Goldfarb, Goodman, Quas, & Edelstein, 2016). However, disclosure of CSA can be difficult as children face multiple barriers to reporting abuse. Victim characteristics such as age and gender impact disclosure (Lippert, Cross, Jones, & Walsh, 2009; Leach, Powell, Sharman, & Anglim, 2017), as well as abuse-specific factors (e.g., relation to perpetrator, severity and frequency of abuse; Hershkowitz, Lanes, & Lamb, 2007). Non-offending caregiver support has been shown to mitigate the adverse psychological outcomes of abuse (e.g., Bolen, & Lamb, 2008) and may also increase disclosure in investigative interviews. Although these factors have been studied in relation to rates of disclosure, relatively little is known about their impact on delays in reporting CSA. Research was conducted through special access to classified RCMP case files on CSA. All predictors were found to significantly impact delays of disclosure. Contrary to previous studies (Lippert et al., 2009) as age increased or if victims were female, delays of disclosure decreased. Alternatively, if abuse was more severe, occurred at a higher frequency, or if perpetrators were more closely related, delays of disclosure increased. Unique to the current study severity and frequency of abuse were measured and considered separately. Finally, non-offending caregiver support was found to be a protective factor and as levels of support increased so did the likelihood of disclosure. The current study provides potential for tailored approaches according to individual needs and vulnerabilities in CSA investigations and treatments. It also demonstrates the need for supportive programs created for caregivers whose families have been impacted by CSA.  iv  Preface This research was the result of partnerships between the University of British Columbia, Okanagan and both the RCMP and the Child Advocacy Centre of Kelowna. For the present thesis, I was responsible for coding cases of child sexual abuse, the generation of the SPSS dataset, analysis, and the written portion herein. This research has been approved by the UBC Okanagan Behavioural Research Ethics Board under Ethics Certificate number H18-00273. No findings of this study have been published.     v  Table of Contents Abstract .…………………………………………………………………………………….…..iii Preface …………………………………………………………………………………………...iv Table of Contents ……………………………………………………………………………….v  List of Tables …………………………………………………………………………………...viii List of Figures ……………………………………………………………………………….…..ix Acknowledgements ………………………………………………………………………...…….x Dedication ……………………………………………………………………………………….xi CHAPTER 1 Introduction ……………………………………………………………………...1 1.1 Pattern of Disclosure ………………………………………………………………..2 1.2 Barriers to Disclosure ……………………………………………………………….3 1.2.1 Victim Characteristics ……………………………………………….….3 Victim Age …………………..………………………….………...3 Victim Gender ………………...………………………………….6 1.2.2 Abuse-Specific Factors …………………………………………………..7 Severity and frequency of Abuse …………………………….....7 Intra versus Extrafamilial Abuse …..…………………………..8 1.2.3 Non-offending Caregiver Support ……………………………………...9 1.3 Present Study ……………………………………………………………………….11 CHAPTER 2 Methods ……………………………………………………………….…………14  2.1 Participants …………………………………………………………………………14  2.2 Measures ………………………………………………………………………….…14   2.2.1 Victim Characteristics …………………………………………….……...14 vi    2.2.2 Abuse-Specific Factors ……………………………………………….…..15   2.2.3 Non-offending Caregiver Support ………………………………………16   2.2.4 Investigative Outcomes …………………………………………………..17  2.3 Procedure …………………………………………………………………………...17  2.4 Analysis ………………………………………………………………………….…..18 CHAPTER 3 Results ……………………………………………………………………….…..20  3.1 Descriptive Findings …………………………………………………………….….20   3.1.1 Victim and Perpetrator Characteristics ………………………………...20   3.1.2 Abuse Specific Factors …………………………………………………...20   3.1.3 Non-offending Caregiver Support ………………………………………22 3.1.4 Investigative Outcomes …………………………………………………..22  3.2 Item Response Theory Analysis …………………………………….……………..23  3.3 Multivariate Cox Regression Analysis ……………………………………………25 3.4 Binomial Logistic Regression ………………………………………………….…..27 CHAPTER 4 Discussion ……………………………………………………………………….28  4.1 Patterns of Disclosure ……………………………………………………………...29  4.2 Victim Characteristics ……………………………………………………………..30 4.2.1 Victim Age ……………………………….……………………………….30 Ages 1 – 6 ……………………………………………………….30 Ages 10 – 11 …………………………………………………….33 Ages 13 – 18 …………………………………………………….33 Perpetrator Age ………………………………………………..34   4.2.2 Gender ……………………………………………………………………35 vii   4.3 Abuse-Specific Factors …………………………………………………………….38   4.3.1 Severity and Frequency of Abuse ………………………….……………38 Severity of Abuse ……………………………………………….38 Frequency of Abuse ……………………………………………39 The interrelation between Severity and Frequency of Abuse ……………………………………………………………………………40   4.3.2 Intra versus Extrafamilial Abusers ……………………………….…….41  4.4 Non-offending Caregiver Support ………………………...………………………43  4.5 Investigative Outcomes …………………………………………………………….46  4.6 Implications ………………………………………………………………………...47  4.7 Limitations and Future Directions ………………….…………………………….50  4.8 Conclusion ………………………………………………………………………….52 Tables …………………………………………………………………………………………...55 Figures …………………………………………………………………………………………..64 Appendices .……………………………………………………………………………………..73 References ………………………………………………………………………………………75          viii  List of Tables Table 1      Mean’s for all Variables under Consideration Across Males and Females .………..55 Table 2      Frequency of Abuse Specific Factors ………………………………………………56             Table 3      Percentage of Most Severe Sexually Abusive Behaviour according to Age of Child at First Incident of Abuse ……………………………………………………………………….....57 Table 4      Frequency of Occurrence of Severe Incidents of Abuse according to Total Incidents of Abuse …………………………………………………………………………………………58 Table 5      Frequency and Percentages of Degree of Relation between Victims and Perpetrators and Percentages of occurrence of Frequency of Abuse according to the Category of Degree of Relation …………………………………………………………………………………………59 Table 6      Frequency of Response Options for each Supportive Item …………………………60 Table 7      Item Parameter Estimates for 2PL on the Non-offending Caregiver Support Items ……………………………………………………………………………………………………61 Table 8      Statistics for Multivariate Cox Regression Model …..………………………………62 Table 9      Percentage of Role of the Abuser in the Child’s Life according to Age of Child at First Incident of Abuse …………………………………………………………………………..63     ix  List of Figures Figure 1      Item Characteristic Curves for Non-offending Caregiver Support Items …….....…64 Figure 2      Item Information Curves for Non-offending Caregiver Support ………….....…….65 Figure 3      Total Item Characteristics Curves, item information Curves, and Test response .....66 Figure 4      Multivariate Cox regression analysis survival functions and hazard functions …….67 Figure 5      Age at first incident of sexual abuse and the resulting mean delay in disclosure measured in days ………………………………………………………………………………..68 Figure 6      Mean delay in disclosure in days according to severity of most abusive behaviour ……………………………………………………………………………………………………69 Figure 7      Mean delays in disclosure according to number of total incidents of abuse ……….70 Figure 8      Average delays in disclosure in days according to degree of relation between victim and perpetrator …………………………………………………………………………………..71 Figure 9      Mean delay in disclosure according to level of non-offending caregiver support ....72        x  Acknowledgments First, I would like to acknowledge the Child Advocacy Centre of Kelowna and the Kelowna RCMP, without whose support this research would not be possible. It is because of their partnership and willingness to conduct research that I have been given this amazing opportunity to study childhood maltreatment and potentially help this vulnerable population. I commend the frontline workers at the RCMP and Child Advocacy Centre for their outstanding service to community.   To my supervisor, Dr. Michael Woodworth, I would like to say thank you for your continued support and patience. I am so grateful for all your help throughout this process. Whenever I entered a meeting feeling less than confident, I always left feeling positive and reinvigorated. Thank you for being the fantastic role model and teacher that I needed.   Thank you to my committee members – Dr. Woodworth, Dr. Walsh and Dr. Davies – for devoting your time to this project and for all your help and guidance throughout.   Finally, I would also like to acknowledge my lab mates and peers for their much-needed support. Thank you for study breaks, late night edits, advice and the fun you have brought to this process. Thank you to my family and friends for your encouragement and care, and to my daughter who has been my inspiration.    xi  Dedication For George who was there through it all.       1  CHAPTER 1 Introduction Intervention in child sexual abuse (CSA) cases is heavily reliant on victim testimony, as there is rarely substantial physical evidence (Tashjian, Goldfarb, Goodman, Quas & Edelstein, 2016). Even in the presence of evidence (e.g., video recordings, medical examinations, or presence of STDs), children may deny allegations, recant statements, or simply fail to disclose during investigative interviews (Elliot & Briere, 1994; Sjoberg & Lindblad, 2002). This is in great part due to the barriers that youth face in reporting abuse, including fear, lack of support or understanding of abuse, and embarrassment (Lemaigre, Taylor, & Gittoes, 2017). In fact, 1 in 5 victims of abuse do not disclose until adulthood, if at all (Hébert, Tourigny, Cyr, McDuff, & Joly, 2009; McElvaney, 2013; Smith et al., 2000; Sorenson & Snow, 1991). Nondisclosure has been linked to three fundamental aversive outcomes. First, nondisclosure may result in poor mental health and posttraumatic stress over time (Nemeroff, 2016). In cases of prolonged abuse, victims risk symptoms of complex trauma in the wake of no psychological support (O’Leary, Coohey, & Easton, 2010; Ullman, 2007). Relatedly, an uninterrupted pattern of abuse prevents victims, and their families, from receiving professional support services (e.g., counselling). Secondly, nondisclosure delays a formal investigation, which may risk loss of valuable evidence (McElvaney, Greene, & Hogan, 2014; Thackeray, Hornor, Benzinger, & Scribano, 2011). Third, perpetrators are unlikely to be held responsible without disclosure of abuse from the child. An arguably even greater consequence of the lack of disclosure is that youth are continually put in dangers way (Olafson & Lederman, 2006). In order to maximize the ease of the experience, accuracy and efficiency of victim disclosure, it is imperative to better understand factors that both aggravate and mitigate delays of disclosure and recantation in cases of CSA.   2  1.1 Pattern of Disclosure While disclosure can initiate an investigation into CSA, only a small number (10%-18%) of abuse cases reach the investigation stage (London, Bruck, Ceci, & Shuman, 2005). Disclosures may be made in full to individuals within (i.e. parent, sibling) or outside (i.e. teacher, counsellor) of the family. In the past, it would appear that disclosure has primarily been viewed as dichotomous (i.e. disclosure, versus nondisclosure). However, this is a narrow view of how a child communicates his or her experiences. For example, children may disclose tentatively or partially (London et al., 2005). That is more information may be given during each subsequent interview or victims may minimize details of the abuse out of embarrassment. Children who tentatively disclose, or neither confirm or deny the incident, can be viewed as non-credible and unreliable (Anderson, 2016). This is exacerbated in investigative circumstances when disclosure is withheld or revised over time (Paine & Hansen, 2002). At both the investigative and court level, there are serious implications for incomplete or missing details of the CSA incident(s), such as delays or acquittals. Research often finds disclosure to be a process that occurs on a spectrum and may take multiple interviews in order to elucidate full details of the offence (London et al., 2005; Paine & Hansen, 2002). This is further complicated when children recant previous statements. Often children will recant when consequences to the family are illuminated once an investigation is underway (Wolfteich & Loggins, 2007). Once recantation occurs children may redisclose and in many cases more details to the abuse will be revealed (Gonzalez, Waterman, & Kelly, 1993). However, relatively little is known about patterns of disclosure and the factors that impact them.    3  1.2 Barriers to Disclosure Victims of CSA often carefully consider the consequences of disclosure. For example, they often express feelings of shame and self-blame, and concern for not being believed, all of which may be furthered by a perpetrator’s refusal to acknowledge any incidence of abuse (Hershkowitz et al., 2007; McElvaney et al., 2014). The child may also fear for themselves or others due to direct or indirect perpetrator threats (McElvaney et al., 2014). Responses to reporting can be as difficult as the disclosure itself; some parents have described their reaction to disclosure of abuse as one of disbelief and shock, even in instances where they claim to believe their child. Disclosure is also dependent on numerous internal and external factors such as the child’s memory for the incident(s), understanding of the incident(s), emotional response to the incident(s), family support, insight into consequences of disclosure, and effectiveness of CSA investigation (McElvaney et al., 2014). These factors are moderated by both characteristics of the victim (e.g., age, gender), abuse-specific factors (e.g., intra versus extrafamilial sexual abuse, severity and frequency of abuse), and non-offending caregiver support (Alaggia, 2005; Anderson, 2016; Hershkowitz et al., 2007; Lippert, Cross, Jones, & Walsh, 2009), which will be explored in more detail below.  1.2.1 Victim Characteristics Victim Age Age of victim at time of forensic interview is a significant determinant of both memory, understanding, and ability to describe an incident of sexual abuse (Brilleslijper-Kater, Friedrich, & Corwin, 2004; Lippert et al., 2009). Consequently, younger children (e.g., ages 1 through 6) are likely to have difficulties disclosing abuse. An inability to fully understand abuse is further exasperated by confusing statements made by the perpetrator such as “it’s a game” (Brilleslijper-4  Kater et al., 2004). It is understandable then, that younger victims are more likely to disclose accidently rather than purposefully (Paine & Hansen, 2002). In such cases, multiple interviews are often required before younger interviewees can provide an account. It then follows that older children and adolescents, due to their greater ability to understand the goal of forensic interviews and typically provide a more cohesive narrative of the abuse, should be better able to disclose CSA (Orbach & Lamb, 2007). However, barriers to disclosure affect all children, regardless of age, and previous research on likelihood to disclose at various ages has been somewhat mixed. Hershkowitz and colleagues’ (2007) found that older children (10–12 years) were less likely to disclose in fewer interviews than a lower age group (7–9 years). Later research however, offered opposite findings with children who had disclosed being on average three years older than those who had not yet made a disclosure (Lippert et al., 2009). More recently, Leach, Powell, Sharman and Anglim, (2016) found that disclosure rates peak at age eleven, after which there is a sharp decline. While there has been a lack of clarity in previous studies, it is important to obtain a more refined understanding of the influence of age-related differences in disclosure. For example, lower disclosure in older children and adolescents may be due to the awareness of social norms and taboos, as well as escalated feelings of shame or embarrassment. Older children and teenagers (e.g., ages 10 and on) are often perceived to be more at fault for CSA as they are seen as more able to resist advances through physical or verbal means (Theimer & Hansen, 2017). Additionally, those aged 10 and over often state that they feel responsible for the abuse and are fearful of the negative repercussions of disclosure (Goodman-Brown, Edelstein, Goodman, Jones, & Gordon, 2003). Negative repercussions can vary from personal costs such as emotional turmoil, to familial costs, such as financial issues or removal of the child from the home. 5  Age at time of disclosure (forensic interview) may also be affected by the age at which the abuse began. If the gap between age at onset and age of forensic interview is large, disclosure may be impacted as memory limitations occur over time and younger victim’s memory is more malleable (La Rooy, Pipe, & Murray, 2007). Likelihood of full disclosure is decreased by a lengthier interval between onset of abuse and disclosure (Salmon & Pipe, 2000) and has been found to peak when abuse begins at an intermediate age. For example, Kogan (2004) found that 35% of teenagers who experienced CSA between the ages of 14 to 17 years had never disclosed compared to 14% for those aged 7 to 10 years. When age at abuse onset was under 6 years this resulted in lowest levels of disclosure across age groups. It is important to discern the exact relationship between age and disclosure as steps can be taken to account for barriers to disclosure, such as inability to recount abuse in a fluid narrative or correcting inconsistencies in children’s understanding of the incident(s). In addition, if particular age groups are found to be more vulnerable to certain barriers to disclosure, agencies, which play crucial roles in supporting and initiating investigations into childhood maltreatment, such as Child Advocacy Centres or forensic investigators can focus additional efforts in supporting such individuals. Younger individuals can hardly be faulted for failing to provide relevant details of the offense if it is not developmentally appropriate (e.g., preschool children have difficulties recalling temporal attributes of events) and so other retrieval aids should be employed (e.g., free recall; see Middleton, 2017 for a detailed review). However, as mentioned earlier,  findings are not consistent across studies (e.g. Bottoms, Rudnicki, & Epstein, 2007; Goodman-Brown et al., 2003; Hershkowitz et al., 2007; Leach et al., 2017) and the exact impact victim age has on disclosure remains somewhat unclear as outlined above. These contradictory findings may be due to the moderating effects of other victim characteristics (such as gender) and abuse-specific 6  factors, indicating a need to study these factors in tandem. As such gender of victim and perpetrators are important variables to consider when investigating factors that alter willingness to disclose and patterns of disclosure. Victim Gender  In a 2011 meta-analysis on worldwide prevalence of CSA, it was found that 18% of girls experience CSA compared to 7.6% of boys (Stoltenborgh, Vanijzendoorn, Euser & Bakermans-Kranenburg, 2011). Despite considerable prevalence of CSA in males, females are the primary focus of CSA research, indicating rates of CSA may be underestimated for males. It is important to consider gender differences as they can moderate a child’s willingness to disclose, with male victims continually being found to disclose less often (Gagnier & Collin-vézina, 2016; Hershkowitz, Horowitz, & Lamb, 2005; Lippert et al., 2009; Priebe & Svedin, 2008). Research has highlighted three themes exclusive to the inhibition of male victim disclosure: (a) fear of being labelled a homosexual; (b) fear of being stigmatized or isolated due to their victimized status; and (c) fear of continuing the cycle of abuse themselves (Priebe & Svedin, 2008; Sivagurunathan, Orchard, MacDermid, & Evans, 2018). Overwhelmingly, males are more often offended against by other older males. From the male victims’ perception this often brings into question both their masculinity and sexuality (Alaggia, 2005; Gagnier & Collin-vézina, 2016; Priebe & Svedin, 2008) and may contribute to mounting feelings of shame. Even in the instances in which the perpetrator is female, this form of sexual exploitation may be mistakenly viewed as desirable by outsiders (Alaggia, 2005), leading the victim to feel confused in regards to their negative perception of the event. Reaching a better understanding of gender differences in both patterns and barriers to disclosure is crucial as it may shape how interviewers obtain victim testimonies and the emphasis they place on specific barriers. Resources may also be geared 7  towards different gender beliefs and norms when encouraging victims to come forward (e.g., the myth that males who are offended against by other males will be labelled homosexual). This is particularly important, as is highlighted above, males may be less likely to disclose or will disclose more tentatively than females because of the differential stigmatisation of CSA across genders. These feelings may be exacerbated for both sexes when the abuse is more severe and frequent.   1.2.2 Abuse-Specific Factors Severity and Frequency of Abuse Abuse-specific factors have the potential to impact disclosure rates (Bottoms et al., 2016). Serial incidents and those of greater severity have been found to often lead to greater difficulty in obtaining a disclosure (Hershkowitz et al., 2007; Kogan, 2004). The delay in disclosure can even double in time in some circumstances when physical aggression is also experienced (Paine & Hansen, 2002). In a large-scale survey conducted by Priebe and Svedin (2008), it was found that girls who had experienced non-contact sexual abuse were more likely to disclose than those who had experienced contact sexual abuse. However, contact sexual abuse included both penetrative and non-penetrative sexual acts. Kogan (2004) found that girls who experienced penetrative sexual abuse were, in fact, more likely to disclose. The finding that penetrative acts increase rates of disclosure has since been replicated (Leclerc & Wortley, 2015). Disclosure in this case may have been inevitable due to the overt signs of the abuse such as bleeding, presence of STI’s, pregnancy or other psychological symptoms. Indeed, severity and frequency of abuse have the potential to be integral variables that influence how disclosure occurs, if at all. To our knowledge few studies have examined these variables independently (e.g., one incident of abuse can be extremely severe without reoccurring, just as reoccurring 8  abuse may not necessarily be severe). Due to the impact these variables have on rates and delays of disclosure the present study will examine these factors separately in order to parse apart their effects. There is also a deficit of research examining the progression of severity of sexual abuse when the abuse is reoccurring and if abusers may work up to more severe acts. It is important to discern the unique impact severe or frequent abuse each have on the development and severity of psychiatric and substance use disorders (Bulik, Prescott, & Kendler, 2001; Zanarini et al., 2002). As aforementioned, victims often weigh the consequences of disclosure and non-disclosure, and more severe forms of abuse may lead to greater consequences for the victim and a need for protection that outweigh costs to the family (Kogan, 2004; McElvaney et al., 2014). The number of consequences to the family unit depend upon whether the abuse is intra or extra familial, with most serious and severe abuse usually conducted by perpetrators who are family members (Ullman, 2007). Intra versus Extrafamilial Abuse  Greater frequency and longer duration of CSA indicates that the perpetrator may have easier access to the child in question, as is the case in abusers who are related (Bergh, 2017). In fact, there are alarmingly fewer cases of CSA perpetrated by strangers (40%) than familiars (60%; Hershkowitz et al., 2007). Intrafamilial perpetrators may include anyone related to the victim (e.g., parents, step-parents, siblings) whereas extrafamilial perpetrators includes those outside the home (e.g. strangers, teachers, friends; Bergh, 2017). In nearly one third of CSA cases, the perpetrator is a family member (Stoltenborgh, et al., 2011) with the most common relative perpetrators being either fathers or step-fathers (Rice & Harris, 2002). In 77% of intrafamilial CSA cases there are multiple incidents, 73% of which endure abuse for more than one year (Fischer & McDonald, 1998). Further, intrafamilial abuse often consists of what are 9  considered more severe sexual acts at a higher frequency. This may include penile rubbing (10% versus 7%), oral sex (14% versus 12%), and most severe, penetrative sex (54% versus 32%; Bergh, 2017).  Research has found that evidence provided in intra versus extrafamilial cases varies greatly. For example, confessions of intrafamilial abuse are less likely to be obtained than extrafamilial (35% versus 49% respectively; Bergh, 2017). Intrafamilial abuse is also less likely to involve pornographic photo documentation as evidence (i.e., 15% versus 48%; Bergh, 2017). In these cases, victim disclosure and testimony is paramount in order to secure a conviction. Unfortunately, research has found that when perpetrators are familiar, victims will often delay disclosure and only disclose following prompts (Hershkowitz et al., 2007; Kogan, 2004; Sjöberg & Lindblad, 2002). Nondisclosure and delays in disclosure, further increase when the perpetrator lives within the home or is a parent of the victim (Goodman-Brown et al., 2003; Tashjian et al., 2016). However, research rarely investigates the degree of relation between the victim and perpetrator. Instead, the majority of research focuses on intra versus extrafamilial abuse as categorical. The current study attempts to examine the exact relationship between time until disclosure and degree of relation. Many explanations for decreased disclosure rates are centered around consequences for the victims and their families (e.g., financial hardship, removal of children from home; Bolen, & Lamb, 2008). These consequences may be mitigated by the presence of non-offending caregiver support.  1.2.3 Non-offending Caregiver Support How other’s respond to the disclosure of CSA is important to improve the long-term outcomes for victims. If positive and supportive non-offending caregiver responses are present it can help to restore the abused child’s sense of security and mitigate serious psychological 10  consequences (Bick, Zajac, Ralston, & Smith, 2014; Cyr et al., 2014). Likewise, lack of caregiver support may inhibit or delay disclosure, even when abuse is extrafamilial (Tashjian et al., 2016).  Expectedly, parents of a child who reports abuse face tremendous difficulties. Even when a parent believes the report or when explicit evidence is available they sometimes still describe their initial reactions as dubious (McElvaney et al., 2014). As Summit (1993) states “protective denial surrounding sexual abuse can be seen as a natural consequence (of) …the need of almost all adults to insulate themselves from the painful realities of childhood victimization” (p. 179). In other words, disbelief can be a defence mechanism against the consequences of CSA. Particularly in intrafamilial cases, the consequences for the family may be difficult to overcome and could directly influence support given to the child (Hershkowitz et al., 2007). In some instances, children are directed by their parents not to discuss the abuse with others. For example, Bolen and Lamb (2008) found that children who experience these unsupportive reactions describe their families as disbelieving. This can result in emotional, verbal and physical abuse, lack of protection, and potentially be told to remain silent. Likewise, if victims are aware of the consequences to the family unit it will also directly impact their decision to disclose. McElvaney (2015) found that some victims expressed concern that disclosure would upset their parents, while others voiced concern about the consequences to their family.  Positive non-offending caregiver support and positive perceptions of parental bonding have been found to increase disclosure rates (Bolen et al., 2008; Hunter, 2015; McCarthy, Cyr, Fernet, & Hebert, 2019). In Hershkowitz and colleagues (2007) study, 71% of children whose parents reported responding anxiously disclosed only after being prompted, while supportive reactions required no prompting. Non-offending caregiver support has four major dimensions: 11  believing the child, protecting the child, emotionally supporting the child, and obtaining resources for the child (Cyr et al., 2014; McCarthy et al., 2019; Priebe & Svedin, 2008). Further, caregiver support may be demonstrated directly through actions (e.g., reporting abuse and restricting contact between the child and perpetrator; Lippert, Cross, Jones, & Walsh, 2009). The present study will further examine the possible protective effect that non-offending caregiver support can have on disclosure in forensic investigations of CSA. The protective effect of this variable may be mitigated by the degree of relation between the victim and perpetrator as the number of consequences of disclosure increases. Further, as outlined in the literature above, age and ability to understand these consequences could also impact our findings.  The complexity of factors related to CSA make it exceedingly difficult for children to disclose experienced abuse (Bottoms et al., 2016). Especially when specific combinations of victim characteristics and abuse-specific factors have the potential to aggravate cases of non-disclosure. For instance, when working with 16-year-old male victims who are capable of understanding societal norms, they may fear being labelled homosexual for disclosing abuse perpetrated by another male. Conversely, working with a four-year-old would present different challenges such as articulating a fluid narrative of reality. Due to court emphasis on testimony for conviction, it is crucial to investigate factors that influence disclosure rates. Understanding the circumstances in which children do or do not disclose is invaluable and has relevance in legal, investigative and treatment circumstances. 1.3 Present Study Disclosure can initiate an investigation into CSA, however only a small number (10%-18%) of abuse cases reach the investigation stage. Indicating a need to study factors that could increase the speed with which forensic investigative disclosures are made. Once potentially 12  highly dependent barriers to disclosure have been more comprehensively and thoroughly investigated, interventions and forensic interviews may be better able to focus in on specific barriers for each individual. In response to the issues highlighted herein, the current study will examine a multitude of factors impacting disclosure in cases of CSA and how they might relate to each other. Specifically, age, gender, relatedness of perpetrator (intra vs. extra familial and degree of relation), severity and frequency of abuse, and caregiver support will be included. This study is unique in five aspects. (1) It will attempt to define victim-abuse profiles which will help to identify children who may take longer to disclose. This is important to understand as it may better inform investigative practice and the need for specific investigative and support measures to increase comfort of victims and conviction rates of offenders. (2) Additionally, whereas other studies have merely examined disclosure versus non-disclosure this study will establish a timeline of disclosure (e.g., time between initial abusive event and disclosure in forensic interview) that will be moderated by the other variables under investigation. (3) While a vast amount of research has found that non-offending caregiver support leads to more positive prognoses in treatment, there is a decided lack of research on non-offending caregiver support and its impact on delays of disclosure in forensic interviews. Furthermore, previous research has relied on perceptions of parental support or self report scales. (3a) The current study will rely on behavioural indicators of non-offending caregiver support rather than emotional reports from third parties. Behaviours of interest are those involving initiation and cooperation with an investigation into CSA. (4) This study will attempt to disentangle two important variables, severity and frequency of abuse. These variables are often studied in tandem but, as discussed earlier, abuse that is severe is not necessarily frequent and one incident of abuse can be very 13  severe without being frequent. It is important to parse apart these two variables as they may impact disclosure differently. (5) Finally, the current study will examine how victim characteristics and abuse-specific factors impact patterns of disclosure and investigative outcomes.  Four primary hypotheses will be examined. First, it is hypothesized that time between the abuse and disclosure will increase if the abuse is severe, except when that abuse includes penetrative acts as the need to be protected against further abuse will outweigh the potential consequences of disclosing. Likewise, more frequent abuse will result in longer delays of disclosure as more time is required to commit additional offences. Additionally, it is predicted that those who experienced abuse more often would also on average experience a greater severity of abuse. Second, intrafamilial abuse will result in longer periods between the incident and disclosure as a function of degree of closeness. Third, age and gender will impact delays in reporting CSA. With delay in disclosure decreasing until age 11 after which it increases (unless abuse is severe). Gender will moderate time until disclosure negatively when victims are male. It is also proposed that levels of caregiver support will be instrumental in decreasing time until disclosure. Non-offending caregiver support will function as a protective factor. It is also predicted that increased levels of support will lead to higher rates of convictions as the parent acts as an advocate for the child during the investigation. Finally, the current study will explore the relation between victim-abuse specific factors and their impact on disclosure patterns, specifically rates of recantation and nondisclosure. While conducting these analyses it is expected that a specific victim-abuse profile will emerge, in that particular combinations of variables may lead to a lower or higher likelihood of and time until disclosure.    14  CHAPTER 2 Methods 2.1 Participants A total of 278 archival cases of child sexual abuse were extracted from the RCMP database system for the greater Okanagan area in British Columbia, Canada. Cases were collected based on the criteria that they must have included a youth victim of a sexual offence and that the case must be concluded. Cases were coded by the primary researcher from the years 2008 – 2017. Each case was coded on victim-abuse specific factors and investigative outcomes. Of the 278 cases only 212 contained information regarding non-offending caregiver supportive or unsupportive behaviours. Cases were excluded when the level of support was unknown, and it was not possible to infer if there had or hadn’t been caregiver support. The current sample was further broken down into child-on-child (48 cases) and adult-on-child (164 cases) sexual assault, which in initial analyses were found to be fundamentally different samples with significantly different means on abuse-specific factors (e.g., relation between victim and abuse, p < .0005, and age at first incident p < .0005) and investigation outcomes (e.g., investigation length, p < .0005, arrest, p < .0005, and conviction, p < .005). In total any cases which did not include information on non-offending caregiver support or in which abuse was child-on-child were removed from current analysis but will be utilized in follow-up studies. This resulted in a final 164 cases which were retained for analysis.  2.2 Measures 2.2.1 Victim Characteristics RCMP case files were coded and analyzed for specific independent variables and disclosure types using a researcher created coding form. The independent victim characteristics to be measured included: age of victim at initial incident of abuse, as a continuous variable in 15  order to discern age differences in disclosure. Age was measured at abuse onset, age at last occurrence of abuse and age at first forensic interview. This helped to establish a timeline between initial incident and disclosure. Gender of victim was the only categorical variable in the cox regression analysis and was recorded as either male or female.   2.2.2 Abuse-Specific Factors Abuse-specific factors under investigation included severity and frequency of abuse, separately and relation to perpetrator. In this case coding forms were adapted from the Abuse Dimensions Inventory (Chaffin, Wherry, Newlin, Crutchfield, & Dykman, 1997). The specific subscales of sexual abuse behaviours, number of total incidents of abuse, and role of abuser(s) in child’s life were utilized. All remaining subscales (e.g., coercion to gain secrecy, reaction of abuser, and use of force or coercing) were not used for two reasons. Either they were not pertinent to the current study (as the current study strictly measured variables theoretically related to disclosure), or they were not possible to measure with the information available in the provided case files. This inventory was chosen because it provided the most comprehensive list of abusive behaviours as well as abuser information such as coercion tactics used and reactions of abuser. In addition, it rates the relationship between the abuser and victim along a continuous scale (intra versus extrafamilial). Severity was rated on a 1 – 11 scale, with 1 being sexually suggestive talk, hugs or kissing, and 11 being paraphilic sex or exploitation (with scores of 9 or greater indicating a penetrative act has occurred). Frequency of abuse was measured according to total incidents of abuse, from 1 incident to 10 incidents. Once incidents surpassed 10 they were measured at intervals of 5 (e.g., 15, 20, 25). This was done as once the total number of incidents surpassed 10, it became more difficult for victims to remember the exact number of incidents and so they often estimated. Role of abuser is child’s life was also measured along a continuum of 16  degree of closeness to the victim, from 1 being not related at all and 7 being primary caregiver. For example: in-home primary caretaker was a 7, noncustodial parent…, teacher…, stranger was a 1 (Chaffin et al., 1997). 2.2.3 Non-offending Caregiver Support Caregiver support was also measured along a continuum of supportive versus non-supportive. The current study was the first of its kind to utilize non-offending caregiver supportive behaviours. Knowledge of abuse was the indicator that non-offending caregiver support could be measured in each casefile. If caregiver had no knowledge of the abuse, then support could not be coded for that file. Caregiver support was coded based on behaviours which foster investigation into CSA. All behaviours were coded dichotomously as either supportive, that is they engaged in that behaviour, or unsupportive, they did not engage in that behaviour. Behaviours that were coded were: believing the child, reporting abuse immediately (e.g., within a few days if initial disclosure occurred), discontinuing contact between the perpetrator and child, and cooperation with the investigation (e.g., through attending appointments and returning follow up calls). Further, victim blaming was reverse coded and was measured by presence of comments in interviews regarding how the child was dressed or suggestions that the child had acted in a manner that initiated, or justified abuse (such as comments like, if the child had listened this would not have happened). These variables were measured in a two-step process (See Appendix A). The first step was to look for any mention of that behaviour. The second step was to note if that behaviour was supportive or not. For example, with reporting abuse to authorities the primary researcher would first look to see if there is mention of who reported the abuse and would next code this variable as the caregiver reporting or not. These items were created specifically for the current study and so analyses determining their accuracy will be run. 17  As it could not be determined if behaviours not present in case files were supportive or not, scores on behaviours were prorated. That is, final scores on non-offending caregiver support were measured in reference to how many behaviours were measurable per case file.  2.2.4 Investigative Outcomes Investigative outcomes included case timelines from open to close and number of forensic interviews until disclosure. The dependent variables used were disclosure (non-disclosure versus disclosure), conviction (conviction versus no conviction), recanting (recantation versus no recanting) and delays of disclosure (time lapse between initial incident of abuse and disclosure).  2.3 Procedure  The primary researcher was given special permission to access classified RCMP casefiles from their database for the greater Okanagan area. Records were viewed at the Kelowna City Police Services building to assure the security of the sensitive data being gathered by the researcher. Once case files were selected (by the RCMP employee prior to coding sessions) the researcher coded for the variables described above onto coding sheets held on an encrypted, password protected laptop. The same RCMP employee who selected the case files was on hand if needed. Having an RCMP employee available to the primary researcher was ideal as they have extensive knowledge and familiarity with the RCMP case files. See Appendix A and B for specific coding schemes. Coding occurred for no more than three hours at a time due to the sensitive and emotionally taxing content under investigation. Records were coded and analyzed for specific independent variables and disclosure patterns. All variables were coded by the primary researcher following a researcher created coding sheet and the adaptation of the Abuse Dimensions Inventory (see Apendix A and B; Chaffin et al., 1997). It is also important to note 18  that while inter-rater reliability is always preferable when reviewing files, in this case it was not possible. The process that the main researcher had to go through was not feasible for another individual as it entailed a strict security clearance which was required by the RCMP to view their case files. Having two researchers access these case files was impossible due to the RCMP’s concern around the amount of time and resources that they would have to contribute to the project, as well as the sensitive and confidential nature of the files (that are almost never made available to anyone outside of the RCMP). After initial coding was complete all coding forms were converted into purely quantitative information to be statistically analyzed. 2.4 Analysis First descriptive statistics were used to examine the characteristics of the data. The current study looked at both frequencies of occurrence and variations among the data. With a focus on rate of occurrence for victim characteristics (i.e., age and gender) and abuse specific factors (i.e., intra versus extrafamilial, severity and frequency).   Item Response Theory Analysis was run on all non-offending caregiver support items in order to confirm that they all measured different levels of the latent trait non-offending caregiver support. This was done as these variables were not obtained from a previously validated questionnaire but were specific to the current study (see Appendix B). First, models are fit to the data and then ANOVA’s may be used to determine which model has a better fit. IRT analysis results in two main parameters of interest. The first parameter of interest is difficulty and second is discrimination. Item difficulty describes how each item behaves along the ability scale. Item characteristic curves that are easy to endorse (e.g., even those at the low end of support may engage in these behaviours) are shifted to the left of the scale. Item discrimination indicates the rate at which the probability of endorsing an item changes depending on ability level. In total, if 19  the items accurately capture the continuum of caregiver support, then the item characteristics curves should be spread out and the item information curves should have relatively high peaks and also be spread out.   A multivariate Cox regression analysis was used in order to determine how the independent variables related to time until disclosure. As with other survival analyses this type of analysis allows us to examine amount of time until an event, based on variables impacting an individual (Wright, 2000). This type of procedure also allows researchers to take into account censoring. Censoring occurs when the event does not occur as is sometimes the case for disclosure in forensic interviews. The multivariate Cox regression can handle multiple continuous or categorical variables. The result of this type of statistic involves a survival function which demonstrates the effects of each predictor while adjusting for the effects of all other predictors in the model (i.e., the adjusted survival functions). This will allow us to determine how much each variable affects time until the event (in this case disclosure). In addition, a odds ratio is given which represents the probability that an event will occur divided by the probability of that event not occurring.  Three binary logistic regressions were run on the three dichotomous outcome variables, disclosure versus nondisclosure, recantation present versus not, and conviction versus not. They were used to determine whether the independent variables significantly predicted the categorical outcomes variables. The null hypothesis being that our independent variables did not predict the outcome any better than chance.     20  CHAPTER 3 Results 3.1 Descriptive Findings 3.1.1 Victim and Perpetrator Characteristics Our final sample consisted of 148 cases of female victims and 16 cases of male victims. Perpetrator gender consisted of 163 males and 2 females. On average males experienced their first incident of abuse at age 7.75, whereas females experienced the first incident of abuse at age 10.76 (see Table 1). Males experienced more incidents of abuse, 9.86 versus 5.99 for females. Females also reported receiving higher levels of support (.72) from non-offending caregivers, than males (.66).  Average age for victims when the first incident of abuse occurred was 10.48 years (SD = 4.49), with average age of abusers at first incident being 37.47 years (SD = 13.77). The average age difference was 27.02 years (SD = 15.41) between perpetrator and victim. Ethnicity of victims consisted of 65.4% Caucasian, 16.4% Unknown, 12.1% Aboriginal, 2.4% South Asian, 1.2 % Black, 1.2% Hispanic, 0.6 % Asian, and 0.6% other. Most visible minorities are underrepresented in this population as they make up 8.5 % of the total Kelowna population (Statistics Canada, 2016). However, this total does not account for areas surrounding Kelowna which would have been amalgamated into the RCMP database. Ethnicity of perpetrators consisted of 75.7% Caucasian, 9.7% Unknown, 7.3% Aboriginal, 2.4% Asian, 2.4% South Asian, 1.2% Black, 0.6% East Indian, and 0.6% Hispanic.  3.1.2 Abuse Specific Factors The average duration of abuse was 503.29 days (SD = 973.74) with the average days between initial incident and disclosure being 1227.63 days (SD = 2325.41). As expected, the standard deviations for these are quite large as cases can include single incidents of abuse to 21  years of abuse. Likewise, children may disclose the day after the initial incident or may wait until well into adulthood to disclose. For frequency of occurrence for each level of all variables see Table 2.  Severity of abuse ranged from sexually suggestive talk to Paraphilic sex or exploitation (See Table 2). These acts can be more broadly divided into penetrative with 59 victims (36%) and the less severe non-penetrative acts with 105 victims (64%). The impact that severity of abuse has on length of delays of disclosure can help to be explained by examining average severity of abuse across age ranges. Age ranges 1 – 3, 13 – 15, and 16 – 18 experienced a larger percentage of penetrative acts compared to less severe acts, 42%, 46%, and 40% respectively (see Table 3). However, in the older ages there was some variance in average level of severity.  The average number of incidents of abuse was 6.33 (SD = 7.91). It is important to note that once abusive incidents reach 10, there tends to be a lack of specificity in number of incidents. Once incidents totalled 10 or more this number was coded as the smaller grouping number in order to assure number of incidents were not exaggerated and that only incidents that had occurred were coded. However, it is important to note that 70 victims (40.3%) only experienced one incident of abuse (see Table 2). In these cases, in which one incident of abuse was experienced, 39% reported a penetrative act. When there was more than one incident of abuse 34% reported at least one penetrative act (see Table 4). Rates of penetrative acts were similar whether victims experienced one or more acts of abuse.  Role of abuser ranged from stranger to primary caregiver. Relation between victim and perpetrator can also be more broadly divided into perpetrators who live within the home, intrafamilial, 78 (47.9%) to perpetrators from outside the home, extrafamilial, 85 (52.1%). In 18.4% of cases the perpetrator was a primary caregiver (see Table 2). However, in only 13 22  (8.0%) of cases was the perpetrator a stranger, indicating that in 151 cases (92%) the perpetrator had a previous relationship with the victim. Further 80% of primary caregiver(s) committed more than one incident of abuse against their child and 60% committed 10 or more incidents of abuse (see Table 5). 3.1.3 Non-offending Caregiver Support Over half the current sample, 89 cases (54.3%) displayed only supportive behaviours in casefiles, with 15 cases (9%) displaying unsupportive behaviours, the remaining 60 cases (36.7%) contained a mixture of both supportive and unsupportive behaviours (see Table 2). The high number of caregivers that were supportive may have resulted from the current studies inability to differentiate those at the high end of the latent trait non-offending caregiver support, which will be discussed in the section on IRT. ANOVA analysis indicated that on average caregivers were less supportive when abuse was intrafamilial .61 (SD = .38), over extrafamilial .81 (SD = .32), F(1, 161) = 13.09, p < .0005.  3.1.4 Investigative Outcomes Average investigation length from the time the case was reported to case conclusion was 628.95 days (SD = 499.43). The timelines reported here include time from first report of abuse to final jury decisions, if the case proceeds to court. A full disclosure was made in 128 cases (78%), in 4 cases (2.5%) it was unclear, and in 32 cases (19.5%) an incomplete disclosure was not made, due to either errors in memory or refusal. Of those 32 cases in which a full disclosure was not made, 14 did not disclose at all. From all cases 7 children (4.3%) denied the abuse occurred and in 3 cases (1.8%) children recanted their previous statements. In 15 cases (9%) children received two or more interviews. In 7 cases (4%) received no interviews, either because they refused to provide a statement, or they were too young to provide one. Finally, 142 cases (87%) received just one forensic interview. Of the 164 cases which were reported to police, arrests occurred in 23  100 cases (61%). Of the 100 arrests made only 48 cases (29% of all cases included in analysis) resulted in a conviction. 3.2 Item Response Theory Analysis As mentioned above, the items included in the current study to assess non-offending caregiver support have not been utilized in previous studies. An IRT analysis was carried out in R, to assess the psychometric properties of the items of this scale created for the current study. As our scale involved dichotomous variables the 2PL model was employed. This was done to assure all items measured the same latent trait and to assess whether our scale was more precise at certain locations along the latent trait continuum (Toland, 2014). The presumed latent trait being measured is non-offending caregiver support. Our scale had only 5 items (α = .637). If there was no clear endorsement of an item, it was considered null and was not measured. As endorsement indicates how difficult an item is and how likely someone is to respond to it (i.e., high difficulty indicates participants are less likely to respond “correctly” to an item). For proportions of each level of response see Table 6. Item characteristic curves and item information curves see Figure 1 and Figure 2, respectively and the estimated parameters for our model are presented in Table 7. The a parameters (i.e., item discrimination) ranged from 0.71 to 24.47 and the b parameters (i.e., item difficulty) varied from -4.87 to -.05. When examining the item characteristic curves (Figure 1), it is evident that this scale is most reliable on the below average level of the latent trait. Two items in particular have high discriminative properties, inhibiting contact (a = 24.47) and reporting abuse to authorities (a = 18.02). These items have the ability to differentiate between subjects who provide caregiver support and those who do not provide support. However, these two items were relatively easy to endorse, indicating they are easier items to respond to. While blaming the child was highly difficult to endorse, it helped to 24  differentiate between those at the extremely low end of non-offending caregiver support. To view the amalgamated item characteristic curves, item information curves and test response function see Figure 3. For these figures we would want to see all curves spread out along the latent trait continuum, which shows that all levels of the presumed latent trait are measured. However, it is evident that the spread of the current items is low on the latent trait continuum.  One main inference from this statistical analysis is that while the current non-offending caregiver support items are not good at differentiating between subjects on the high end of supportive behaviours, they do a good job of differentiating between subjects on the low end of the ability continuum. That is, the current items are better at differentiating between those demonstrating low supportive behaviours and those demonstrating extremely low supportive behaviours, but not between high and extremely high. This is acceptable for the current study as IRT has demonstrated that our items assess different levels non-offending caregiver support. All items will be retained for further analysis, as results of the IRT analysis indicate each item accounts for a different location along the latent trait continuum and are of differing difficulty. One important finding to take away from this analysis is that the current items assessing non-offending caregiver support were sufficient in determining differences between those who give low levels of support. This may result in findings that more caregivers demonstrate highest levels of support, but these caregivers will already be high in supportive behaviours, it will not inaccurately classify caregivers who are low in supportive behaviours as high. For the purposes of the current study this is acceptable as all findings will be discussed in terms of if little to no, or if high levels of caregiver support are present, which our items can accurately gauge. The current items can accurately differentiate between what are high levels of caregiver support compared to low, but are not particularly good at gauging within high levels. Although more information at 25  the high end of caregiver support would be ideal, for the current study this was not possible as the amount of information given in case files did not allow for more items to be added to the current scale.  3.3 Multivariate Cox Regression Analysis  Prior to the multivariate Cox regression analysis, all theoretically relevant items were examined using a One-way ANOVA to test whether items had a significant impact on the outcome variable delays of disclosure measured in days. All items were measured on a continuous scale, except victim gender which was dichotomous. The delay in disclosure was statistically significantly different for different levels of the victim characteristics, gender, F(1, 148) = 10.448, p = .002, and age, F(15, 131) = 3.166, p = .000. As for abuse specific factors, all 4 theorized factors, severity of abuse, F(10, 139) = 1.925, p = .047, frequency of abuse, F(11, 136) = 3.656, p = .000, relation of abuser and victim, F(6, 142) = 2.374, p = .032, and caregiver support, F(6, 143) = 2.452, p = .028), had a significant impact on delay of disclosure. As a result of these findings all variables were included in the multivariate Cox regression analysis.  It is also important to note that there were concerns over the youngest ages of 1 and 2 skewing findings, as it would be impossible for these ages to disclose sexual abuse. Children aged 1 and 2 have both a cognitive and developmental inability to communicate and so disclosure for them is impossible. In order to determine whether ages 1 and 2 skewed the data a multivariate Cox regression analysis was conducted with them removed. Analysis revealed that age was still a significant factor even when those aged 1 and 2 were removed, b = .053, p = .035, OR = 1.055 (95% CI: 1.004, 1.108). Despite an inability to communicate this group still experiences CSA at a similar rate to those in older age groups and not including them in the 26  current study would have meant removing a crucial part of the data that provided relevant information in regard to victim characteristics and delays of disclosure. Cases available for the multivariate Cox regression analysis were 147 (89.6%), with 17 cases (10.4%) displayed as missing data, in which no disclosure was ever made. No cases were right censored as our timeline was quite inclusive (i.e., included time from incident until disclosure in a forensic interview). Of the final 147 cases included in the survival analysis 135 cases (91.8%) were female and the remaining 12 cases (8.2%) were male. The model including all 5 variables was found to be statistically significant, x2(6) = 65.558, p = .000, meaning that the overall model was able to predict the length of time until disclosure As is displayed in Table 8 all variables were found to significantly impact delays in disclosure. Exposure to the predictors age of victims at first incident, b = .064, p = .005, OR = 1.066 (95% CI: 1.019, 1.115), gender of victims, b = .832, p = .011, OR = 2.299 (95% CI: 1.214, 4.354), and level of non-offending caregiver support, b = .570, p = .020, OR = 1.767 (95% CI: 1.094, 2.856), increases the risk of disclosure occurring. Figure 4 displays the survival functions and hazards functions for all findings. As two of these variables are measured continuously, age and level of support, for every one unit increase in these variables (e.g., 3 to 4 years old, and ¼ to 2/4 supportive behaviours) will result in the greater likelihood of a disclosure occurring. For gender, this indicates that being a female victim will result in greater likelihood of a disclosure than male victims. Exposure to the predictors, severity of abuse, b = -.072, p = .016, OR = .930 (95% CI: .877, .986), total incidents of abuse, b = -.037, p = .002, OR = .964 (95% CI: .942, .987), and role of the abuser in child’s life, b = -.169, p = .000, OR = .845 (95% CI: .769, .928), decreased the chance of a disclosure occurring. That is for every one unit increase in these variables (e.g., feeling over top of clothes to feeling under clothes, 3 to 4 incidents of abuse, and 27  stranger to acquaintance perpetrators) there will be a decreased likelihood of a disclosure occurring.  3.4 Binomial Logistic Regressions It was intended that binomial logistic regressions would be used to test if the variables in our model would be able to accurately predict whether participants would either, disclose versus not disclose, and whether recantation was present versus not. However, the outcome variables of non-disclosure and recantation occurred in relatively few cases, 32 and 3, respectively. While conviction rates occurred at a higher rate than the other two dichotomous outcome variables at 49 cases this number was still insufficient for analysis considering there are 6 main variables of interest. These numbers do not meet the minimum requirements of 15 cases per independent variable. As binomial logistic regressions rely on maximum likelihood estimations, reliability declines significantly when there are few cases available (Psutka & Psutka, 2015). As such, this statistical procedure was not utilized in the current study.      28  CHAPTER 4 Discussion The current study examined how victim characteristics (i.e., age and gender) and abuse specific factors (i.e., severity and frequency of abuse, relation between perpetrator and victim, and non-offending caregiver support) impact delays in reporting CSA. All of the main variables considered were found to be significantly related to delays of disclosure. Additionally, most hypotheses were supported by the current findings. Factors which significantly and positively impacted (i.e., shortened) the speed of disclosure were the age and gender of the child, and high levels of caregiver support. Specifically, the older the victim at the time of the first incident of abuse, the faster a disclosure was made. Female victims reported CSA after a shorter delay than males. However, it is important to note that there was a substantive lack of males in the current sample. When non-offending caregivers are observed to provide more support, this also facilitated a more rapid disclosure. In contrast, abuse-specific factors, specifically the severity and frequency of abuse, and the victim’s relation to perpetrator, led to an increase in delays of disclosure. When abuse was more severe or frequent, children delayed disclosure longer. The closer the relation between victim and perpetrator, the longer a disclosure took to make. In summary, child disclosures of sexual abuse will occur faster in cases when the child is older, female, and receiving support from their non-offending caregiver. However, in cases where the victim is male, and experiencing particularly severe or frequent abuse from a close relative, there can be an expected delay in disclosure. The findings from this study are important to better understand how to assist and prepare for the particularly challenging (and in some cases, more unlikely) process of disclosure.    29  4.1 Patterns of Disclosure   The ways that victims disclose sexual abuse are numerous. Victims can disclose fully, partially, or in some cases not at all. In the current study, only 9% of children received a follow-up forensic interview. As research has found disclosure to be a process that may take multiple interviews, this number is somewhat worrying (Anderson, 2016; Paine & Hansen, 2002). Typically, further interviews were conducted if the first interview took place in a non-forensic setting and/or the investigators needed to gather more information. Multiple interviews can be both beneficial (e.g., providing further detail) or detrimental (e.g., testimony may be revised over time or recanted) to investigations and the decision to do so is made carefully by investigators (Anderson, 2016; Paine & Hansen, 2002). However, it is important to note that investigators walk a thin line between protecting victims against further harm, which may accompany recounting traumatic events, and obtaining enough testimonial evidence to advance cases through the judicial system. Additionally, when more detail is added in subsequent interviews this can make the interviewee seem less credible. It is unrealistic to view CSA disclosure as a dichotomous process that only takes one interview to either obtain or forfeit a disclosure. Further, ability to accurately recount abuse can be exacerbated when more time goes by between initial incident of abuse resulting in memory limitations (La Rooy, Pipe, & Murray, 2007). It is imperative that investigations and court proceedings take into account the spectrum of disclosure and the limitations that victims face when recounting events of CSA.  Counter to providing more evidence, in some instances, victims will not disclose at all or fully, despite how many interviews were conducted. In the current study 32 victims (19.5%) did not give a full disclosure and 14 (8.5%) made no disclosure at all. As an inability to communicate applies to the ages of 1 - 3 this can only account for at most 9 of those 30  nondisclosures. Due to the relative rarity of nondisclosures within the current sample researchers were not able to determine if victim-abuse specific factors had any impact on whether a victim chooses to report or not. Even though the current sample displayed low rates of nondisclosures this does not mean they do not occur often in the general population (Elliot & Briere, 1994; Sjoberg & Lindblad, 2002). It is important for future studies to sample differently and examine how victim-abuse specific factors impact nondisclosures. Reasons behind nondisclosure in cases of CSA may be similar to those for recantation.  Recantation and denial of abuse are most concerning to investigations. In the current study 7 children (4%) denied the abuse had occurred and in 3 cases (2%) children recanted their previous statements. Unfortunately, the current study could not determine what impact the current variables have on recantation or non-disclosure, due to the low rate of occurrence for these types of disclosure. Denial and recantation can occur for multiple reasons, most often when consequences to the family are evident, or become evident once an investigation has started, statements will be withdrawn or not given at all (Wolfteich & Loggins, 2007). Non-offending caregiver support may provide a positive avenue for decreasing denials and recantation as it would counteract the negative consequences experienced by the family for the abuse and subsequent disclosure. Factors that impact disclosures will be discussed in more detail below, beginning with the impact age has on disclosure.  4.2 Victim Characteristics 4.2.1 Victim Age Ages 1 - 6 As hypothesized, victim age was found to significantly impact delays in reporting CSA such that the older a child, the shorter delays in disclosure were (see Figure 5). Age at time of 31  forensic interview has been found to be a substantial determinant of both memory, understanding, and ability to describe an incident of sexual abuse (Brilleslijper-Kater, Friedrich, & Corwin, 2004; Lippert et al., 2009). It makes sense then that younger children (aged 1 – 6) may have difficulties disclosing, with one consequence being prolonging the time before a police report. As the gap between age at onset of abuse and age at forensic interview grows larger, testimonies may be negatively impacted. While memory distortions or loss occur over time for everyone, younger victims’ memories are already more malleable (La Rooy, Pipe, & Murray, 2007). Further, at very young ages (1 – 3) it may be difficult or impossible to communicate that abuse has occurred due to verbal and cognitive limitations. The inability children have to communicate at ages 1 – 3 is particularly concerning in light of current findings that 42% experienced penetrative acts of sexual abuse. Despite verbal and cognitive deficits, children as young as 3 have been found to evidence abuse through behavioural cues, such as early sexualized behaviours or aggression (Friedrich, Urquiza, & Beilke, 1984). Further, children from older age categories may experience developmental delays or deficits (e.g., autism) that make it extremely challenging for them to disclose CSA. While removing participants because of age or developmental deficits would make the sample more homogeneous, it would not enable a comprehensive representation of the full spectrum of children that experience CSA. Once most children develop the ability to communicate around ages 3 – 4 they may be able to report abuse to an adult. The finding that younger children (ages 3 – 6) disclose after longer delays may be explained by a lack of understanding of societal sexual taboos, and so, they may not recognize that what is happening to them is wrong or inappropriate (Goodman-Brown et al., 2003). As a result, children aged 3 – 6, sometimes make disclosures accidentally, that is they do not intent to disclose. Accidental disclosures are typically made through inadvertent cues to 32  the abuse, as they often do not understand the severity of the situation (Paine & Hansen, 2002). Relatedly, young children ages 3 - 6 may not experience feelings of embarrassment or shame which frequently accompany disclosures of this kind (Goodman et al., 2003; Saywitz, Goodman, Nicholas, & Moan, 1991). However, in the current study this potential lack of awareness did not decrease delays of disclosure for ages 3 - 6, but rather it may have prevented children from communicating the abuse sooner. It should also be noted that when abuse is committed by close relatives this can cause further confusion about the deviancy of the abuse.  The relationship between age and delay of disclosure was further examined in order to determine if any of the other main variables, such as relation to perpetrator and severity of abuse, may help to explain these findings. Younger victims, ages 1 – 3 and 4 – 6, were found to be perpetrated against more often by intrafamilial perpetrators, and most often the primary caretaker(s), 42% and 28%, respectively (see Table 9). At these ages, adults outside of the family have less access to younger children (Bergh, 2017). In contrast, none of the cases of CSA in the age range 16 – 18, were offended against by primary caretakers. Taken together, the finding that younger children (e.g., ages 1 – 6)  take longer to disclose and they are also offended against more often by intrafamilial offenders, support previous research which has found that disclosure will be delayed longer (or not occur at all) when victims are younger and perpetrators are intrafamilial (Goodman-Brown et al., 2003; Hershkowitz et al., 2007; Kogan, 2004; Leclerc & Wortley, 2015; Sjöberg & Lindblad, 2002; Tashjian et al., 2016). Intrafamilial offenders also tend to perpetrate abuse against victims on multiple occasions. Often these individuals will have  continued access to the child and these incidents have the potential to become increasingly more severe (Fischer & McDonald, 1998).    33 Ages 10 - 11        The current hypothesis regarding age was that delays in disclosure would decrease until age 11, after which there would be an increase in delays. Counter to this, results indicated that there was a brief increase in length of delays around age 10 to 11 that lasted only a couple years, length of delays declined afterward (see Figure 5). This brief increase in length of delays may be due to the new realization of social norms and increased feelings of shame that occur around the ages 9 and up (Leach, et al., 2016; Theimer & Hansen, 2017). Social norms considered here are those surrounding appropriate sexual behaviour and gender differences. The development of knowledge surrounding sex and sexual behaviour coincides with the presence of sexual education in Canadian schools. The drop in length of delays after age 11 could be due to the consistency of perpetrator populations, in that once children reach ages 12 and up, abuse can be committed more often by distantly related and stranger perpetrators. Ages 13 - 18 Unlike previous research in which older age groups were less likely to disclose (e.g., those aged 13 – 18; Hershkowitz et al., 2007), in the current study they evidenced the shortest lengths of delay overall. As discussed earlier, decreased delays in disclosure at older ages may be for several reasons. First, this decrease in length of delays when abuse is severe may be in response to the need to be protected against further incidents (Kogan, 2004; McElvaney et al., 2014). Findings in the current sample demonstrated that older victims experience a greater variation of sexually abusive behaviours, from fondling over and under children’s clothing (age 13 – 15, 37%, ages 16 – 18, 50%), to penetrative acts (age 13 – 15, 46%, ages 16 – 18, 40%; see Table 3). These results indicate that while abuse is severe for some, it is not for all and this may result in the shorter delays of disclosure. Second, this finding of shortest lengths of delays of 34  disclosure at older ages may be a result of the relation between the victim and perpetrator. While age significantly impacted delays of disclosure, degree of relation may help to explain this impact. Victims aged 13 – 18 at time of first incident of abuse are offended against more often by strangers and extrafamilial perpetrators such as friends, neighbors and other acquaintances. In cases such as these there are often fewer consequences to the family unit for disclosure, resulting in shorter delays (Hershkowitz et al., 2007). In the current study none of the victims within the age range 13 – 18 experienced sexual abuse from a primary caretaker, but the sample was derived from reported cases of CSA. It is likely then that those aged 13 – 18 may have not yet disclosed abuse that was perpetrated by caregivers. More comprehensive samples would help to rectify this issue and provide more evidence on the relationship between relation and age in cases of CSA. Previous studies have consistently found extrafamilial abuse, as well as less severe abuse, to result in shorter delays of disclosure (Goodman-Brown et al., 2003; Hershkowitz et al., 2007; Kogan, 2004; Tashjian et al., 2016). When the perpetrator is a child, abuse is more often extrafamilial and can occur outside of the home (e.g., school, daycare, clubs). Perpetrator Age Age of perpetrators can have a substantial impact on sexual abuse experienced and the resulting disclosure. Although not considered in the current study, initial analyses revealed that victims aged 1 - 12 had higher frequencies of adult-on-child sexual abuse compared to the older ages 13 – 18 where there is a drop in delay of disclosure and the majority of offences are committed by other children. While the initial analysis of the data included child-to-child abuse and adult-to child-abuse, there were enough differences between these two groups that it was decided for the purposes of this study to focus solely on adult-to-child abuse. The child-to-child abuse will be subsequently analyzed as a separate study. 35  It is apparent from the present study and other research that age seems to be a substantial influence in the time it takes to disclose sexual abuse (Brilleslijper-Kater, Friedrich, & Corwin, 2004; Leach et al., 2017; Lippert et al., 2009; Orbach & Lamb, 2007; Paine & Hansen, 2002). This information is beneficial as it points to the need for age-appropriate expectations in forensic interviews. Further there is a need to further develop and research strategies that can be employed in such interviews to elicit the greatest amount of pertinent detail for the CSA incident (Sheen, Andrews, Stoltenberg, & Lyon, 2015; Brown, Lewis, Stephens, & Lamb, 2017). For example, the use of open-ended interviews and employing wh- questioning techniques (i.e., wh- questioning refers to the use of words like what or when in interviews; Ahern, Andrews, Stolzenberg, & Lyon, 2018; Brown et al., 2017).  4.2.2 Gender An individual’s understanding of their gender, similar to social norms, develops over time (Martin & Little, 1990; Martin & Ruble, 2004). As such, younger children may not be as impacted by gender barriers as mid to older aged children. In the present study, there were several gender differences across the variables under consideration (see Table 1 for a direct comparison of males and females). On average¸ males experienced their first incident of sexual abuse 3 years younger than females, beginning at age 7 for males and age 10 for females. While males and females were found to experience similar severities of abuse, males endured more total incidents on average (10 versus 6). These differences are notable as previous studies have found that different variables predict non-disclosure across genders. For example, Priebe & Svedin, (2008) found severity and frequency of sexual abuse, and relation to the abuser impacted rates of nondisclosure to a larger degree in females. For males, being in an educational program, and family structure are better predictors of nondisclosure. For both genders, levels of caregiver 36  support were instrumental in obtaining a disclosure (Priebe & Svedin, 2008). One reason male disclosure may be delayed, or not occur at all, could be the disproportionate levels of non-offending caregiver support. Males received less support than their female counterparts. As caregivers provide the first line of defence against the negative impact CSA can have on children this is worrying. However, this finding should be considered in light of the current scale’s inaccuracy at higher levels of the non-offending caregiver support continuum. Differences across gender are important to consider for targeted investigations which seek to obtain a disclosure. In addition, understanding differences between males and females may provide clues as to why males disclose at a lower rate and after longer delays than females.  Despite the small number of male victims who reported CSA, gender was still found to significantly impact delays of disclosure. When victims were male, disclosure was delayed two times longer than their female counterparts. Current results confirm previous studies that have found that females report sexual abuse more often and after shorter delays than males (Gagnier & Collin-vézina, 2016; Hershkowitz, Horowitz, & Lamb, 2005; Lippert et al., 2009; Priebe & Svedin, 2008). Especially based on previous research that found while males report experiencing CSA less often than females, the number who experience it is still somewhat comparable (Stoltenborgh et al., 2011). A ratio of approximately 1 male to every 3 female victims can be extrapolated from previous findings, which is well below those reported in the current study. This low rate of male reported CSA suggests that male victims may still be under reporting sexual abuse.  While current findings are intriguing, without a comparison sample of males who have not reported, it is impossible to tell why these victims may have disclosed over others. Results from the current study suggest a need for further investigation into what specifically prevents 37  male victims of CSA from coming forward and how to mitigate those barriers. Three myths have been highlighted in the past as the main barriers to male specific nondisclosure. These are: fear of, being labelled a homosexual, stigmatization due to victimization, and victimizing others (Easton, Saltzman, & Willis, 2014; Priebe & Svedin, 2008). Greater endorsement of these myths has been associated with increased feelings of shame and embarrassment, which often accompany CSA (Easton, Saltzman, & Willis, 2014). Future studies could examine the impact that educational programs, which demystify male CSA and address the myths associated with this group, have on willingness to disclose sexual abuse. It would also be pertinent to examine factors such as educational programs and family structures, as these have been linked to male disclosures (Priebe & Svedin, 2008). Considering current and previous findings, public health initiatives and preventative educational programs should address different barriers to disclosure for males and females. In the case of female CSA, nondisclosures and delays of disclosure have been directly connected to severity and frequency of abuse in past literature (Priebe & Svedin, 2008). 4.3 Abuse-specific Factors 4.3.1 Severity and Frequency of Abuse Severity of Abuse  In line with previous research (Hershkowitz et al., 2007; Kogan, 2004; Priebe & Svedin, 2008), severity of abuse was found to impact delays of disclosure. When abuse increased in severity, disclosures took longer to obtain. The most severe form of abuse, paraphilic sex or exploitation, resulted in the longest average delays of disclosure, with lengths of delays surpassing 15 years (see Figure 6). It was expected that because of the serious health complications that can arise with more severe forms of abuse, this would result in more 38  expedient disclosures, as victims would find these consequences aversive and they would be more noticeable by doctors or teachers (Kogan, 2004). However, it is apparent from the current study that despite the extremely negative consequences of severe abuse, this did not result in a decrease in delays of disclosure, even when abuse was penetrative, as was predicted. This can be explained by the link between severity of abuse to the use of avoidant coping strategies (Fortier et al., 2009). Further, more severe abuse has been linked to increased feelings of shame due to more sustained, physical and emotional harm, and degradation by the perpetrator (Firing & Tasks, 2005). This coupled with negative perceptions from others regarding victims’ culpability for the abuse, may result in the current finding of delayed disclosure in more severe cases of CSA (Theimer & Hansen, 2017). Feelings of shame have been found to be highly predictive of the development of psychopathology and can impede the emotional processing required to recover from abusive events. Moreover, penetrative acts of abuse may involve some form of force to gain secrecy which may worsen delays of disclosure. When force is used or when CSA is coupled with other forms of abuse (e.g., physical or neglect), disclosures become even more difficult to obtain (Paine & Hansen, 2002). The use of manipulation to gain secrecy may be more common in abuse which is frequent. Frequency of Abuse  As with severity, frequency of abuse was found to significantly impact the delay until disclosure. Delays of disclosure were longer when there were more incidents of abuse (see Figure 7). This is not surprising as more time is required to commit more offences. However, in this case more frequent abuse may be a symptom, rather than a cause, of longer delays of disclosure. Since the abuse is not reported immediately, if the perpetrator has continued contact with the child, they will still have the opportunity to commit further abuse.  39  One reason for the increase in delay could be the perpetrators use of manipulation to gain secrecy. The practise of grooming typically proceeds frequent acts of abuse (Young, 1997). Grooming involves the systematic desensitization of the child to the abuse through a careful process of trust and rapport building to create more opportunities for contact (Wolf & Pruitt, 2018). Establishing a relationship through grooming can often better ensure that victims are compliant and keep the abuse secret. Over time, interactions with the child will progress to sexual contact, as acts building up to this (e.g., touching) are normalized through the perpetrator-formed connections and daily routines, or in the guise of playing a game (Leclerc & Felson, 2016; Young, 1997). In some cases, abusers may use aggression, fear, or threats to obtain secrecy (Lang & Frenzel, 1988). For example, in the current study 2 files indicated that the perpetrator threatened to assault siblings if the victims disclosed. In previous studies, cases in which threatening or violent grooming tactics were used the most severe trauma symptoms surfaced (Wolf & Pruitt, 2018) and increased delays of disclosure were present (Paine & Hansen, 2002).  In summary, disclosures may take longer to obtain in cases of frequent sexual abuse because grooming techniques may have been employed. The interrelation between Severity and Frequency of Abuse In line with previous studies (e.g., Bottoms et a., 2016; Hershkowitz et al., 2007; Kogan, 2004; Priebe & Svedin, 2008), both severity and frequency of abuse were found to considerably impact delays of disclosure. Further, while severity and frequency of abuse were related and impacted disclosure in a similar manner, these factors also accounted for their own distinct impact on delays of disclosure. For instance, longer delays of disclosure can result from more frequent abuse that is not severe. Likewise, longer delays may result from more severe abuse that only occurs once. This is notable as previous research has typically reported severity and 40  frequency of abuse together (Kogan, 2004; O’Leary et al., 2010; Zanarini et al., 2002). However, these two variables may also have different underlying contributing factors which impact delays of disclosure differently. For example, the experience of multiple incidents of abuse may result from a closer relation between victim and perpetrator, whereas single incidents may be committed more often by strangers and acquaintances.   Despite differences between severity and frequency of abuse (e.g., grooming, relation to perpetrator), these two variables were in some cases connected. The impact that severity has on delays of disclosure could be related to the development of more severe abuse over time and across incidents. Counter to the prediction of severity increasing over time, those who experienced a single incident of abuse reported a comparable level of severity to those who reported multiple incidents. When there was only a single incident of abuse, 39% of this group reported a penetrative act, similar to those who reported more than one incident of abuse, 34% (see Table 4). While in many instances even initial abusive acts are very severe, in a substantial amount of other cases, more time is required before more severe abusive acts are committed.  This finding very clearly denotes the need for helping the child to feel comfortable and willing to disclose as expediently as possible. 4.3.2 Intra versus Extrafamilial Abusers Frequency of abuse is inextricably tied to the role the abuser plays in the child’s life, as more frequent abuse almost always requires consistent contact with the child, which is unlikely to happen with a stranger (Bergh, 2017). The closer the relation between victim and perpetrator the longer disclosures took to obtain. When a child is offended against by a trusted adult this results in longest delays of disclosure (see Figure 8). This is especially true when the adult is a primary caregiver(s). In almost half of all cases the perpetrator was intrafamilial and in 18% of 41  all cases the perpetrator was a primary caregiver (see Table 5). These findings help illuminate the challenges law enforcement and support agencies face in obtaining disclosures considering the majority of cases will be with a family member, primary caregiver or a known individual in a position of trust. Indeed, consistent with previous statistics, strangers committed sexual abuse against victims considerably less often at 8% than other relations (see Table 5; Hershkowitz et al., 2007). In these cases, victims are most often older, when the probability of a stranger having contact with the child is heightened. As a result of longer delay between initial incident of abuse and disclosure, victims who are offended against by an intrafamilial adult have been found to experience more frequent abuse, with previous findings demonstrating most frequent abuse committed by primary caregivers (Fischer & McDonald, 1998). In line with previous research, 80% of those offences in the present study committed by primary caregivers involved more than one incident of abuse and 60% involved a troubling 10 or more incidents (see Table 5; Fischer & McDonald, 1998). Longer delays of disclosure may very well be a result of the trust built between the offending caregiver and victim. However, previous research has found that poor relationships with related children can also be a precursor to CSA (Daly & Wilson, 1998; Seto et al., 2015). Fathers who sexually offend against their children are reported to be dominant, aggressive, and are typically not very involved in parenting (Herman, 2012; Seto et al., 2015). When relationships with caregivers are detached or hostile it may indicate that trust between the perpetrator and victim is weak and so may not impact the decision to disclose. If breach of trust does not entirely explain longer delays in disclosure, some other variable(s) may. For example, if the perpetrator is a primary caregiver it may not be trust that results in longer delays, but rather the use of grooming, 42  specifically fear and intimidation (such as consequences of disclosure for the whole family) that lead to the delay of disclosure (Lang & Frenzel, 1988; Leclerc & Felson, 2016).  As previously considered, when abuse is intrafamilial, consequences to the family can involve loss of primary bread maker or removal of the child and siblings from the home (Bolen &Lamb, 2008). Research has found that for older victims the decision to disclose is made carefully and the benefits and costs of their disclosure is considered (Hershkowitz et al., 2007; McElvaney et al., 2014). Likewise, victims often state their concern over how their disclosure will impact their family, concerns which could be increased through grooming practises (McElvaney, 2015; Wolf & Pruitt, 2018; Young, 1997). It is important to take into account the relation between the alleged perpetrator and victim, as outing a family member as an offender can have serious consequences and may be difficult to believe. Even if disclosing abuse would not have serious consequences, the child can be led to believe that it will (Lang & Frenzel, 1988; Leclerc & Felson, 2016; Wolf & Pruitt, 2018; Young, 1997). If the primary caregiver is the offender this obviously leaves the child with a compromised support system. However, when non-offending caregiver relationships are strong, they can act as a protective factor, mitigating delays of disclosure (Bolen et al., 2008; Hunter, 2015; McCarthy et al., 2019).   4.4 Non-offending Caregiver Support Non-offending caregiver support was found to be instrumental in decreasing delays of disclosure (see Figure 9). When victims received more support from non-offending caregivers delays in disclosure became shorter. While other variables may indicate that a victim may be at risk to delay or not disclose sexual abuse, non-offending caregiver support shows promise as a protective factor against this. Encouragingly, over half of the current sample (54%) displayed only supportive behaviours, which is a far greater number than one other previously mentioned 43  study that found only 37% of sampled caregiver(s) were supportive (Hershkowitz et al., 2007). The remaining 33% of caregivers displayed behaviours that were moderately supportive but still lacked behaviours conducive to the supportive group. Current findings of a greater percentage of those displaying only supportive behaviours, is most likely due to two reasons. First, the inability of our scale to differentiate between those who displayed a high level of supportive behaviours and those who displayed the highest level of supportive behaviours. This would result in more individuals falling in the high end of supportive behaviours with little differentiation. However, this would only effect those on the high end of supportive behaviours and as demonstrated through IRT those demonstrating low levels of supportive behaviours were easily differentiated from each other. Alternatively, 13% of caregivers displayed more unsupportive behaviours than supportive ones. Second, it should be noted that the high rates of non-offending caregiver support may be a result of the current sample. This sample is one in which investigations into CSA have been initiated and so either a previous disclosure has occurred, or other evidence has been presented confirming allegations. Logistically this means that cases in which unsupportive behaviours may have prevented children from making it to the investigative stage of CSA would not be included in analysis.  Several variables have been found to impact the amount of support non-offending caregivers give to their victimized children (Hershkowitz et al., 2007). In the present study, on average non-offending caregivers were less supportive when the abuser was intrafamilial. When abuse is perpetrated by someone close to the family, it can be difficult for the caregiver to reconcile these two views of the perpetrator: a loved one and an abuser. These feelings may then potentially be dealt with through cognitive processes such as denial and disbelief (McElvaney et al., 2014; Summit, 1993). In the current study 23 (14%) caregivers demonstrated disbelief in 44  response to their child’s disclosure. This can directly impact the amount of support (e.g., believing the child, reporting the abuse) given to the child. In extreme cases children may be told to remain silent about the abuse. Even just the act of not discussing the abuse can lead to greater feelings of shame in the child, and these feelings can persist and sometimes develop into psychopathology such as depression and anxiety (Feiring & taska, 2005). Factors that impact delays in reporting are crucial to study, as avoidant coping strategies have been found to lead to the greater presentation of trauma symptoms and likelihood of being revictimized as an adult (Fortier et al., 2009). Previous studies have indicated that more severe abuse that is frequent may lead to more unsupportive caregiver behaviours (Hershkowitz et al., 2007). However, this was not found to significantly impact amount of support given by non-offending caregivers in the present study. Previous findings by Hershkowitz and colleagues (2007) found that children are able to accurately predict parental reactions to disclosures of sexual abuse. This finding indicates there may be other familial or behavioural variables that occur pre abuse, to consider when determining amount of support given. For example, considering previous rebellious behaviour committed by the child, or strength of relationship pre-abuse between caregiver and perpetrator. Just as supportive behaviours made by non-offending caregivers can help to mitigate negative psychological consequences of abuse, unsupportive behaviours can aggravate them (Feiring & Taska, 2005). Future research should attempt to assess the strength of relationships between caregivers and children pre-CSA disclosure, as this could aid in determining if this impacts support given post disclosure.  Non-offending caregiver support has four major dimensions: believing the child, protecting the child, emotionally supporting the child, and obtaining resources for the child (Cyr 45  et al., 2014; McCarthy et al., 2019; Priebe & Svedin, 2008). The current study was able to assess three out of four of these supportive dimensions, believing the child, protecting the child and providing resources for the child. These are accomplished through behaviours parents take to initiate and cooperate with an investigation into CSA. Although the use of behaviours to gauge caregiver support is novel, it can be hypothesized that such behaviours show the victim that not only is what has happened to them wrong, but that there are people on their side. This show of support, even from non-caregivers, can result in increased disclosures and more positive psychological outcomes. A recent example of this is in the #MeToo movement and literature (Rotenberg & Cotter, 2018).   Coinciding with the introduction of the #MeToo movement on social media in 2017, police-reported sexual assault peaked in Canada, with the greatest number of reported incidents of sexual assault occurring in the 2 and 3 months following (Rotenberg & Cotter, 2018). This included an increase in reports of sexual assault which occurred over a decade ago, or in which the perpetrator was related to the victim. Further, children were more likely to report incidents of sexual abuse that occurred on school property by child perpetrators. The impact this movement has had on increasing rates of disclosure demonstrate that even support obtained from outside of the home, such as public statements and sentiments, can help to increase the likelihood of a disclosure occurring. That is, even at the societal level, changes that denote understanding or support can make a big difference for victims of all forms of sexual assault. The finding that different forms of support as well as support from various sources can impact disclosure, provide potential application of current findings. Community agencies, such as Child Advocacy Centres, can have a positive impact on victims of CSA, as they can not only be a support in the absence of one, but also an additional one. This can be especially important in cases of intrafamilial abuse, 46  as support may be harder to obtain from caregivers (Bolen & Lamb, 2008; Hershkowitz et al., 2007). Further, one of the main objectives of Child Advocacy Centres is to educate the community surrounding childhood maltreatment, helping to encourage disclosures at the community level.  4.5 Investigative Outcomes The current study was not able to determine what impact non-offending caregiver support had on convictions. Despite having someone to advocate for these victims, this did not appear to impact investigative outcomes, at least in a manner that could be reflected in number of convictions. More specifically, while arrests occurred in 61% of cases, there were convictions in only 29% of cases. Although specific caregiver behaviours were the reason an investigation may have been initiated (e.g., reported abuse), it is more likely that another factor such as the presence of more evidence may have more impact on the outcome of CSA cases. Neither investigative outcome, arrest or conviction could have occurred without a disclosure. Further, although the primary variables do not impact investigative outcomes directly, the presence of disclosures do. The relatively low rate of convictions that occur in CSA cases may also be the result of delays of disclosure. For example, the more time that passes from incident of abuse to disclosure, the more valuable evidence is potentially lost, whether it be physical or corroborative evidence (McElvaney et al., 2014; Thackeray et al., 2011). The current research helps to explain those delays of disclosure in hopes that more targeted techniques can be employed for victims which help to decrease delays in reporting and increase amount of evidence available.  4.6 Implications Even in the presence of substantial evidence of sexual abuse, investigations are still heavily reliant on victim testimony for conviction (Elliott & Briere, 1994; Sjoberg & Lindblad, 47  2002; Tashjian et al., 2016). Although disclosure is not guaranteed, enhancing the conditions that can foster a more expedient disclosure can have a positive impact on both the safety of the child and on the perpetrator receiving justice. The present research provides evidence to support that speed of disclosure is influenced by multiple factors such as victim characteristics and abuse-specific factors. Talking about sexual health and behaviour with children can be difficult for some caregivers. Even though discussing past sexual behaviour can be challenging, caregivers clearly play an instrumental role in disclosures. So much so, that even the child’s perception of a strong relationship with their caregiver(s) can result in higher rates of disclosure (Bolen et al., 2008; Hunter, 2015; McCarthy et al., 2019). Even when parent’s responses to CSA disclosures are seen as anxious by their child, once prompted by caregivers, children are more likely to disclose (Hershkowitz et al., 2007). The finding that children still disclose, even to seemingly anxious caregivers, demonstrates that frank and open communication between children and their parents surrounding healthy and appropriate sexual behavior can help to improve disclosure rates. However, parents in the current study stated they were embarrassed or just didn’t know how to broach the subject with their children. Education and support for parents may be incredibly useful. Education on factors that may make children more vulnerable to nondisclosure and ways to approach the subject of CSA with children may be highly beneficial. These messages could be communicated either through public health announcements or through agencies such as Child Advocacy Centres. Further, Child Advocacy Centres can aid in supporting parents both pre and post CSA disclosures with information and resources. These resources may also help to increase the believability of disclosures of CSA. 48  Caregivers are not the only ones who may require further education and support. Frontline workers such as teachers or social workers are sometimes the first to notice symptoms or behaviors which indicate CSA may have occurred. In situations when CSA is suspected but has not yet been disclosed, it would be important for other professionals involved in the child’s life to know risk factors for longer delays of disclosure. Unfortunately, there is usually limited resources available to frontline workers and Child Advocacy Centres, so it is crucial to know how and when extra resources should be spent (i.e., on cases when disclosure is more difficult to obtain). Utilizing current findings agencies such as Child Advocacy Centres or the police can identify when a victim may need additional supports or specific strategies that can be aimed at that vulnerability in order to obtain a disclosure and ensure mental well being post abuse. For example, strategies aimed at dealing with the myths behind male CSA would be most beneficial for older male victims. Conversely, strategies aimed at obtaining a complete narrative with no leading questions would be ideal for young children (ages 3 – 6). When combined with findings regarding the impact of the #MeToo movement on disclosure (Rotenberg & Cotter, 2018), current findings demonstrate that frontline workers may be an excellent source of support for victims of CSA.  Once caregivers or frontline workers have been alerted to cases of CSA an investigation may be initiated. Investigators are often pushed to obtain an expedient disclosure as evidence would often be more readily available and recounts by victims are more easily corroborated. However, this may not be ideal, especially when dealing with children who have delicate memories and are easily manipulated (La Rooy et al., 2007), because some children may never disclose even when evidence is available. The current study has identified factors that make children vulnerable to delaying disclosure such as severity and frequency of abuse, that when 49  present, may override the need for a speedier disclosure as the victim’s well-being becomes a priority. Additionally, perpetrators often offend against multiple victims (Elliott, Browne, & Kilcoyne, 1995). If disclosures can be obtained sooner it is possible that not only could abuse against other potential victims be prevented but that additional victims could be available to testify against a single perpetrator, increasing the likelihood of proceeding to court as well as obtaining a conviction. Furthermore, corroboration may be difficult to obtain from caregivers if they do not believe the abuse has occurred, as is the case in 14% of our current sample. If parents who respond to disclosure with disbelief are asked to come in for questioning, they may give inaccurate details of the case. Delaying the initiation of an investigation, through delayed disclosures, means the loss of potentially strong evidence. Proper support and open communication with caregivers and their children may mean CSA is reported more quickly and evidence is more readily available. Support and education around CSA could be beneficial in assuring caregivers they are not at fault for their child’s victimization and for obtaining more accurate corroborative evidence.  Those cases where factors suggest a high risk to avoid disclosure, or to have a prolonged period of time go by until a disclosure should be considered in a court of law. If more time has elapsed between abuse and court hearings, memory limitations in testimony should be anticipated. Additionally, findings of the present study indicate that the risk of nondisclosure is higher if the victim is younger, male, experiences severe or frequent abuse, is abused by a close relative, and if they receive little support. As these factors are found to be more predictive of nondisclosure, expert witnesses may identify at risk children and utilize this research in court. In short, the implications of current findings are numerous and extend to the home, frontline workers, investigations, and court proceedings.  50  4.7 Limitations and Future Directions  Findings of the current study must be considered in the wake of several limitations. First, it is important to note that the current sample is based upon police reported cases of child sexual abuse. This means that in most cases either a disclosure was made, or evidence was brought forward indicating that CSA had occurred. As CSA goes largely unreported, we were unable to examine cases in which no disclosure has ever been made and evidence has not come forward to initiate an investigation. Factors that significantly predict a non-disclosure (versus a delayed disclosure) may be fundamentally different from those included herein (Connolly, et al., 2015). Further, victims who make disclosures in forensic interviews may be different from those who make disclosures in counselling settings, or those who do not disclose at all. Previous studies have collected their samples from telephone surveys, meaning that they may recruit participants who have not disclosed until the time of the study (i.e., when prompted). These studies have also been found to have a higher number of male participants, which is the opposite to the present study. Studies which sample from the general population, rather than an investigative sample, would be potentially more accurate at capturing hard to reach individuals for study. Future studies should also consider attempting to collect data from the general population in order to capture cases in which no disclosure has ever been made or disclosures were not reported to the police. A more inclusive sample would allow for a higher prevalence of males and a more thorough investigation into disclosure patterns. It would also be beneficial to compare investigations in which evidence was uncovered but no disclosure was made, and investigations in which a disclosure initiated an investigation.  Analyses were also restricted by the amount of emotional information and detail that the case files provided. Specifically, emotional properties (e.g., victim emotions during disclosure) 51  which can be useful to consider, were absent from most files. Due to a lack of documented emotion from the caregivers as well, caregiver support could only be measured behaviourally through actions recorded in the RCMP files. While this approach rectified an existing lack of investigation into behavioural indicators of support (Lippert et al., 2009), the lack of information on emotional support is unfortunate as disclosure has been found to be impacted by it (Bolen & Lamb, 2008; Hunter, 2015; Lippert et al., 2009; McCarthy et al., 2019). Traditionally emotional support is measured through interviews of non-offending caregivers or through RCMP investigator reports (Lippert et al., 2009). However, these measures have their own limitations. Non-offending caregivers may give biased answers as there is much at stake for appearing to have low support for the child (e.g., child is taken into protective custody). Caregiver’s perceived low levels of support can directly impact investigative opinions of parents. If they are seen as unsupportive to the investigator the same types of consequences (such as their child being taken away) could apply. The current study examined caregiver behaviour taken to initiate and cooperate with an investigation into CSA. However, results from the IRT analysis indicate that these items are better at measuring those low on the caregiver support continuum. Future studies should take a more holistic approach to examining non-offending caregiver support, measuring both the behavioural and emotional aspects through multiple avenues of data collection (i.e., children, observation). Perhaps emotional support would help to better assess those at the high end of caregiver support.  According to the World Health Organization, “Child sexual abuse is the involvement of a child in sexual activity that he or she does not fully comprehend, is unable to give informed consent to, or for which the child is not developmentally prepared and cannot give consent, or that violates the laws or social taboos of society” (7. Child sexual Abuse, pg. 75). This definition 52  is very inclusive and could involve acts that are committed by other children. The current study noted several significant differences between the child-on-child and adult-on-child perpetrator groups. For example, age significantly impacts delays of disclosure but there is also an increase in child-on-child abuse, as children age. These differences demonstrated that while this child-on-child sexual abuse may fall under a more inclusive definition of CSA, a resulting disclosure could be impacted by vastly different factors. For example, consequences to the family unit for disclosure may not be as substantial. The primary researchers will examine differences between child-on-child and adult-on-child sexual abuse in future studies, however they were not the focus of the current study.  Finally, as the role of non-offending caregiver support is integral to not only obtaining a disclosure but also with post abuse emotional care, future projects could expand on the current findings by examining ways in which caregivers can feel more supported. Supports for caregivers could involve more education centered around how to speak with children about healthy and appropriate sexual behaviour or the development of support groups for parents dealing with CSA in their families. This type of project can be accomplished through the partnership with community agencies already in place to support victims and their families, such as Child Advocacy Centre’s. 4.8 Conclusion The present study aimed to identify and understand factors that contribute to the speed at which a formal disclosure of child sexual abuse occurs. As a result, it contributes additional support for a more refined consideration of developmental, cognitive and socio-emotional factors when attempting to obtain a disclosure of CSA from victims (Goodman-brown et al., 2003). The current study outlines several important variables to consider when obtaining (or attempting to 53  obtain) a CSA disclosure including, age at first incident, gender, severity and frequency of abuse, role of abuser in child’s life and most importantly non-offending caregiver support. The findings of the current project suggest that there is no “one size fits all” approach, and that multiple and complex factors may play a role in when a child will provide a statement (and reasons why he or she may not until a later point in time). Specifically, this study is one of the first to validate a profile of victim and abuse that can inform both families and the legal field about risk and helping factors in a timely disclosure. For example, child-specific factors, such as age (e.g., 10-11-years) and gender (male), were found to be influential in delaying disclosure of CSA. Perpetrator- and abuse-related factors were also uncovered, such that children who were abused by a close adult and who experienced severe and more frequent abuse, were slower to disclose. One of the most encouraging and important findings is that non-offending caregiver support served a positive role in expediting the speed of formal disclosure. This is a meaningful and applicable finding for caregivers who may suspect, or who may become aware, of a child’s sexual assault. Preparing caregivers with information on how to support and respond to CSA disclosures is a positive investment in the safety of children and may improve the complex (and potentially frightening) legal process associated with reporting CSA. We consider these findings in wake of how a child’s self-disclosure may lead to the identification of a dangerous offender, interruption of abuse, and prevention of abuse against future victims (Paine & Hansen, 2002). Understanding factors that slow or expedite disclosure can inform education within communities to help end the cycle of abuse. For example, 44% of victims who did not disclose to another after the initial incident have been found to be subsequently perpetrated against repeatedly by the same individual (Sas, Cunningham, Hurley, Dick, & Farsnworth, 1995). The current study indicates several ways in which agencies such as Child Advocacy Centre’s can help to support 54  victims and their families. Agencies can help through preventative education, identifying victims who may be vulnerable to nondisclosure and delays of disclosure, targeting those specific vulnerabilities, and supporting caregivers before and after CSA has been uncovered. While the present study contributes to extant knowledge, continued efforts are needed to understand how to best support child victims and their families to enable resources such as child advocacy centers to make safe our young citizens.   55  Tables Table 1  Mean’s for all Main Variables under Consideration Across Males and Females  Males (16) Females (148)  Mean (SD) Mean (SD) Delay in Disclosure 2892.33 (4423.76) 1067.20 (1952.79) Age at first Incident 7.75 (3.84) 10.76 (4.45) Severity of Abuse 6.69 (2.33) 6.36 (3.03) Total Incidents 9.86 (7.87) 5.99 (7.86) Relation to Abuser 4.13 (1.78) 4.03 (2.08) Level of Support .66 (.37) .72 (.36)                       56  Table 2   Frequency of Abuse Specific Factors   Frequency Percent Cumulative Percent Severity of sexual abuse behaviours    1 7 4.3 4.3 2 3 1.8 6.1 3 24 14.6 20.7 4 33 20.1 40.9 5 2 1.2 42.1 6 18 11.0 53.0 7 2 1.2 42.1 8 16 9.8 64.0 9 19 11.6 75.6 10 39 23.8 99.4 11 1 0.6 100.0 Number of total incidents    1 70 43.8 43.8 2 12 7.5 51.3 3 21 13.1 64.4 4 3 1.9 66.3 5 5 3.1 69.4 6 2 1.3 70.6 7 2 1.3 71.9 10 15 9.4 81.3 13 1 0.6 81.9 15 4 2.5 84.4 20 11 6.9 91.3 25 14 8.8 100.0 Relation between victim and perpetrator 0 1 0.6 .6 1 12 7.4 8.0 2 38 23.3 31.3 3 34 20.9 52.1 4 3 1.8 54.0 5 25 15.3 69.3 6 20 12.3 81.6 7 30 18.4 100.0 Level of Support    .00 15 9.1 9.1 .25 2 1.2 10.4 .33 5 3.0 13.4 .50 24 14.6 28.0 .67 9 5.5 33.5 .75 20 12.2 45.7 1.00 89 54.3 100.0 Note. Severity of abuse was coded as the most severe abusive behaviour present and does not encompass all abusive behaviours experienced. Values of 0 indicate unknown, except in level of support in which it means non-offending caregivers were unsupportive.   57  Table 3    Percentage of Most Severe Sexually Abusive Behaviour according to Age of Child at First Incident of Abuse.  Severity of Sexual Abusive Behaviour  1 2 3 4 5 (Penetrative Acts)  Total Ages 1 – 3 - 26% 16% 16% 42% 19 Ages 4 – 6 3% 31% 28% 10% 28% 29 Ages 7 – 9 11% 50% 14% 7% 18% 28 Ages 10 – 12 11% 25% 8% 17% 39% 36 Ages 13 – 15 5% 37% 2% 10% 46% 41 Ages 16 – 18 - 50% 10% - 40% 10 Note. It is important to note that this does not mean that the initial incident of abuse was the most severe, but that severity may have developed over the course of the abuse if it was frequent.          58  Table 4   Frequency of Occurrence of Severe Incidents of Abuse according to Total Incidents of Abuse.    Severity of Abuse  Total Incidents of Abuse 1 2 3 4 5 (Penetrative Acts) Total Cases per incident category 1 Incident of Abuse  7% 40% 6% 8% 39% 70 2 – 4 3 13 7 2 11 36 5 – 9 - 3 - 2 4 9 10 – 14 1 7 1 2 5 16 15 - 19 - - 1 - 3 4 20 – 24 1 2 3 1 4 11 25 + - 3 3 4 4 14 More than 1 Incident 6% 31% 17% 12% 34% 90 Note. Categories are totaled by number of incidents of abuse.       59  Table 5  Frequency and Percentages of Degree of Relation between Victims and Perpetrators and Percentages of occurrence of Frequency of Abuse according to the Category of Degree of Relation.  Degree of Relation between Victim and Perpetrator  1 2 3 4 5 6 7 Frequency 13 38 34 3 25 20 30 Percentages 8% 23% 21% 2% 15% 12% 19% Frequency of Abuse        Unknown - - - - 4% 10% 3% 1 92% 62% 41% 67% 20% 35% 17% 2 – 4 8% 24% 20% - 24% 20% 30% 5 – 9 - 3% 3% - 4% 5% 17% 10 – 14 - 5% 15% - 12% 15% 10% 15 - 19 - - 12% - - - - 20 – 24 - 3% 9% - 20% 10% - 25 + - 3% - 33% 16% 5% 23% Note. The first row of frequency indicates how often cases fell into one of the 7 degree of relation categories. Row two indicates what percent of individuals fell into each degree of relation category. The frequency of abuse columns are totalled according to that category of degree of relation.   60   Table 6         Frequency of Response Options for each Supportive Item.    0 1 miss Contact interrupted 0.377 0.623 0.6303 Reported Abuse to Authorities 0.4615 0.5385 0.1333 Believed Child 0.162 0.838 0.1393 Cooperation with Investigation 0.1259 0.8741 0.1333 Blames Child 0.0339 0.9661 0.2848                      61   Tables 7       Item Parameter Estimates for 2PL on the Non-offending Caregiver Support Items.    a b Contact interrupted 24.47 -0.07 Reported Abuse to Authorities 18.02 -0.05 Believed Child 1.58 -1.45 Cooperation with Investigation 2.63 -1.3 Blames Child 0.71 -4.87 Note. a = item discrimination parameter, b = item difficulty parameter.          62  Table 8   Statistics for Multivariate Cox Regression Model      95% CI (eb) Variable b SE Z eb Lower Upper Age at first incident .064 .023 .005 1.066 1.019 1.115 Gender .832 .326 .011 2.299 1.214 4.354 Severity of Abuse -.072 .030 .016 .930 .877 .987 Total Incidents -.037 .012 .002 .964 .942 .987 Role of Abuser -.169 .048 .000 .845 .769 .928 Level of Support .570 .245 .020 1.767 1.094 2.856           63  Table 9  Percentage of Role of the Abuser in the Child’s Life according to Age of Child at First Incident of Abuse.  Role of Abuser in Child’s Life  Stranger Acquaintance, babysitter, neighbor, friend Teacher, minister, daycare worker, relative, etc. In-home sibling or step-sibling In-home adult, except parent/stepparent Noncustodial In-home primary caretaker Total Ages 1- 3 - 11% 5% - 5% 26% 42% 19 Ages 4 – 6 - 14% 24% 3% 17% 14% 28% 29 Ages 7 – 9 4% 22% 33% - 19% 11% 11% 27 Ages 10 – 12 6% 28% 22% - 19% 8% 17% 36 Ages 13 – 15 15% 27% 22% 5% 10% 10% 12% 41 Ages 16 - 18 40% 40% - - 10% 10% - 10 Note. Totals are across age categories.       64  Figures  Figure 1. Item Characteristic Curves for Non-offending Caregiver Support Items.  a = item characteristics curve for cooperation with investigation b = item characteristics curve for inhibiting contact between perpetrator and victim c = item characteristics curve for reporting abuse to proper authorities immediately (within 1 day of disclosure) d = item characteristics curve for believing the child’s disclosure e = item characteristics curve for blaming the child, the only reverse coded variable  65              Figure 2. Item Information Curves for Non-offending Caregiver Support. a = Item information curve for cooperation with investigation b = Item information curve for inhibiting contact between perpetrator and victim c = Item information curve for reporting abuse to proper authorities immediately (within 1 day of disclosure) d = Item information curve for believing child e = Item information curve for blaming the child, the only reverse coded variable a. b. c. d. e. 66          Figure 3. Total Item Characteristics Curves, item information Curves, and Test response Function for all non-offending caregiver support items. a = Amalgamated item characteristic curves b = Amalgamated item information curves c = Test response function    Figure 3. Total Item Characteristics Curves, item information Curves, and Test response Function for all nonoffending caregiver support items a. b. c. 67  a. b.   c. d.   Figure 4. Multivariate Cox regression analysis survival functions and hazard functions. a. Survival function for all variables and their cumulative survival in days until disclosure. b. Hazard function for all variables and their cumulative hazard in days until disclosure. c. Survival function for males and females separately, on all variables and their cumulative survival in days until disclosure. d. Hazard function for males and females separately, on all variables and their cumulative hazard in days until disclosure.        68    Figure 5. Age at first incident of sexual abuse and the resulting mean delay in disclosure measured in days.     69   Figure 6. Mean delay in disclosure in days according to severity of most abusive behaviour.     70   Figure 7. Mean delays in disclosure according to number of total incidents of abuse.    71    Figure 8. Average delays in disclosure in days according to degree of relation between victim and perpetrator.   72   Figure 9. Mean delay in disclosure according to level of non-offending caregiver support.                73  Appendices Appendix A    Disclosure PatternsAfter initial disclosure: Next InterviewMore detail added: Y/N Conflicting details: Y/NMore perpetrators revealed: Y/NLength of interviews: hours and minutes Denial of Abuse: Y/NHistory of drug use:Gender:    Offended Against: Y/NHow long:# of Siblings: Ages:Famiy History: divorce/deathLiving Situation: Family, Foster home...Setting of interview: detachment, school, etc.Investigation CharacteristicsQualitative information from interviews: caregiver feelings, beliefs about the futureSibling Interviews: Y/NForensic Medical Exam: Y/N Who conducted it:  Did caregiver report abuse: Y/NWas contact interupted: Y/N Did caregiver believe child: Y/NFamily Support - non-offending caregiverKnowledge of abuse: Y/NNumber of Interviews:Multidsiciplinary Team: Y/NBehavioural Concerns: Behavioural Concerns:Perpetrator InformationEthnicity:Previous Criminal Convictions:Age:Family ArchetypeAge, first offence: Age, most recent:Perpetrator Members of team:How long into interview:Full Disclosure: Y/N Recantation present: Y/NAge:2. Child at RiskAge:General Occurance InformationMain Offence:Occured between: Case closed:                  /year            /month            /dayReported on:                 /year                      /month          /dayCCJS InformationLocation Type: of crime Family Violence:Victim Information1. Child at Risk Initial disclosure to whom:Drugs/alcohol involved:Gender:      Ethnicity: Gender:    Ethnicity:Psychological/Medical: Psychological/Medical:Weapons used:Relation to the accused:Charges laid:Reasons for no charges:Gender:    Findings:MDT case review: Y/N Who was present:Children forcefully removed from home: Y/NInvestigative outcomesArrest: Y/N Conviction:Other Evidence available:Number of Interviews till disclosure:74  Appendix B  Adapted Abuse Dimension Inventory Coding Sheet     x12345789101112Number of TOTAL Incidents (all types)XX124567Ritual and satanic abuse or sexualized tortureTotalParaphilic sex (e.g., urine, feces, bondage) or exploitation (e.g. prostitution)Simulated intercourse over clothes6 Child masturbating abuser or involved in abuser's masturbation or simulated intercourse under clothes (penis between child's legs, no penetration)Vaginal or anal intercourse (including unsuccessful attempts)Approximate number of incidentsUnknownFondling over child's clothesFondling under child's clothesUnknownVisual exposing of genitals (abuser or child), voyeurism or viewing pornographyIn-home sibling or step-siblingIn-home adult, except parent/stepparentNoncustodialIn-home primary caretaker (e.g., custodial parent, stepparent or guardian)Victim Perpetrator RelationshipRole of Abuser (s) in Child's LifeUnknownStrangerAcquaintance, babysitter, neighbor3 Teacher, minister, daycare worker, relative, or other involved adult living out of the homeOral contact - abuser to child's genitals Oral contact - child to abuser's genitalsDigital or object penetrationSexual Abuse BehaviorsSexually suggestive talk, hugs, or kissing75  References Ahern, E. C., Andrews, S. J., Stolzenberg, S. N., & Lyon, T. D. (2018). The productivity of wh- prompts in child forensic interviews. Journal of interpersonal violence, 33(13), 2007-2015. doi: 10.1177/0886260515621084 Alaggia, R. (2005). Disclosing the trauma of child sexual abuse: A gender analysis. Journal of Loss and Trauma, 10(5), 453–470. Anderson, G. D. (2016). The Continuum of Disclosure : Exploring Factors Predicting Tentative Disclosure of Child Sexual Abuse Allegations During Forensic Interviews and the Implications for Practice , Policy , and Future Research The Continuum of Disclosure : Exploring Factors P. Journal of Child Sexual Abuse, 25(4), 382–402. Bergh, C. (2017). Intra-Familial and Extra-Familial Child Sexual Abuse : Differences in Swedish Court Cases. Bolen, R. M., Lamb, J. L., Bolen, R. M., & Lamb, J. L. (2008). Parental Support and Outcome in Sexually Abused Children Parental Support and Outcome in Sexually Abused Children, 8712. Bottoms, B. L., Peter-Hagene, L. C., Epstein, M. A., Wiley, T. R. A., Reynolds, C. E., & Rudnicki, A. G. (2016). Abuse Characteristics and Individual Differences Related to Disclosing Childhood Sexual, Physical, and Emotional Abuse and Witnessed Domestic Violence. Journal of Interpersonal Violence, 31(7), 1308–1339. Brilleslijper-Kater, S. N., Friedrich, W. N., & Corwin, D. L. (2004). Sexual knowledge and emotional reaction as indicators of sexual abuse in young children: Theory and research 76  challenges. Child Abuse and Neglect, 28(10), 1007–1017. Brown, D., Lewis, C., Stephens, E., & Lamb, M. (2017). Interviewers’ approaches to questioning  vulnerable child witnesses: The influences of developmental level versus intellectual disability status. Legal and Criminological Psychology, 22(2), 332-349. doi:10.1111/lcrp.12104 Bulik, C. M., Prescott, C. A., & Kendler, K. S. (2001). Features of childhood sexual abuse and  the development of psychiatric and substance use disorders. The British Journal of Psychiatry, 179(5), 444-449. Cantwell, H. B. (1988). Child sexual abuse: Very young perpetrators. Child abuse &  neglect, 12(4), 579-582.  Carolina, S. (2015). Associations with Youth Psychosocial Adjustment, 38(3), 479–487. Chaffin, M., Wherry, J. N., Newlin, C., Crutchfield, A., & Dykman, R. (1997). The Abuse Dimensions Inventory: Initial data on a research measure of abuse severity. Journal of Interpersonal Violence. Cyr, M., Hébert, M., Frappier, J.-Y., Tourigny, M., McDuff, P., & Turcotte, M.-E. (2014). Parental Support Provided by Nonoffending Caregivers to Sexually Abused Children: A Comparison Between Mothers and Fathers. Journal of Child Custody, 11(3), 216–236. Daly, M., & Wilson, M. (1998). The truth about Cinderella: A Darwinian view of parental love.  New Haven, CT: Yale University Press. Easton, S. D., Saltzman, L. Y., & Willis, D. G. (2014). “Would you tell under circumstances like  77  that?”: Barriers to disclosure of child sexual abuse for men. Psychology of Men & Masculinity, 15(4), 460-469. Elliott, D. M., & Briere, J. (1994). Forensic sexual abuse evaluations of older children: Disclosures and symptomatology. Behavioural Sciences & the Law, 12(3), 261–277. Elliott, M., Browne, K., & Kilcoyne, J. (1995). Child sexual abuse prevention: What offenders tell us. Child abuse & neglect, 19(5), 579-594. Feiring, C., & Taska, L. S. (2005). The persistence of shame following sexual abuse: A  longitudinal look at risk and recovery. Child Maltreatment, 10(4), 337-349. doi: 10.1177/1077559505276686 Fischer, D. G., & McDonald, W. L. (1998). Characteristics of intrafamilial and extrafamilial child sexual abuse. Child Abuse and Neglect, 22(9), 915–929. Gagnier, C., & Collin-vézina, D. (2016). The Disclosure Experiences of Male Child Sexual Abuse Survivors The Disclosure Experiences of Male Child Sexual Abuse. Journal of Child Sexual Abuse, 25(2), 221–241. Gonzalez, L. S., Waterman, J., Kelly, R. J., McCord, J., & Oliveri, M. K. (1993). Children's  patterns of disclosures and recantations of sexual and ritualistic abuse allegations in psychotherapy. Child Abuse & Neglect, 17(2), 281-289. Goodman-Brown, T. B., Edelstein, R. S., Goodman, G. S., Jones, D. P. H., & Gordon, D. S. (2003). Why children tell: A model of children’s disclosure of sexual abuse. Child Abuse 78  and Neglect, 27(5), 525–540. Hébert, M., Tourigny, M., Cyr, M., McDuff, P., & Joly, J. (2009). Prevalence of Childhood Sexual Abuse and Timing of Disclosure in a Representative Sample of Adults from Quebec. The Canadian Journal of Psychiatry, 54(9), 631–636. Herman, J.L. (2012). Father–daughter incest. Cambridge, MA: Harvard University Press Hershkowitz, I., Horowitz, D., & Lamb, M. E. (2005). Trends in children’s disclosure of abuse in Israel: A national study. Child Abuse and Neglect, 29(11), 1203–1214. Hershkowitz, I., Lanes, O., & Lamb, M. E. (2007). Exploring the disclosure of child sexual abuse with alleged victims and their parents. Child Abuse and Neglect, 31(2), 111–123. Hunter, S. V. (2015). Perceptions of the role of mothers in the disclosure and nondisclosure of child sexual abuse: A qualitative study. Journal of Child Sexual Abuse, 24(8), 887–907. Johnson, T. C. (1988). Child perpetrators—Children who molest other children: Preliminary  findings. Child abuse & neglect, 12(2), 219-229. Kaseweter, K., Woodworth, M., Logan, M., & Freimuth, T. (2016). High-risk sexual offenders: Towards a new typology. Journal of Criminal Justice, 47, 123–132. Kogan, S. M. (2004). Disclosing unwanted sexual experiences: Results from a national sample of adolescent women. Retrieved from 79  Lang, R.A. & Frenzel, R.R. (1988). How sex offenders lure children. Annals of Sex Research, 1,  303- 317.  La Rooy, D., Pipe, M. E., & Murray, J. E. (2007). Enhancing children’s event recall after long delays. Applied Cognitive Psychology, 21(1), 1–17. Leach, C., Powell, M. B., Sharman, S. J., & Anglim, J. (2017). The Relationship Between Children’s Age and Disclosures of Sexual Abuse During Forensic Interviews. Child Maltreatment, 22(1), 79–88. Leclerc, B., & Felson, M. (2016). Routine activities preceding adolescent sexual abuse of  younger children. Sexual Abuse, 28(2), 116-131. Leclerc, B., & Wortley, R. (2015). Predictors of victim disclosure in child sexual abuse:  Additional evidence from a sample of incarcerated adult sex offenders. Child Abuse & Neglect, 43, 104-111. Lippert, T., Cross, T. P., Jones, L., & Walsh, W. (2009). Telling interviewers about sexual abuse: Predictors of child disclosure at forensic interviews. Child Mal, 14(1), 100–113. London, K., Bruck, M., Ceci, S. J., & Shuman, D. W. (2005). Disclosure of Child Sexual Abuse: What Does the Research Tell Us About the Ways That Children Tell? Psychology, Public Policy, and Law, 11(1), 194–226. Martin, C. L., & Little, J. K. (1990). The relation of gender understanding to children's sex‐typed  preferences and gender stereotypes. Child Development, 61(5), 1427-1439. Martin, C. L., & Ruble, D. (2004). Children's search for gender cues: Cognitive perspectives on  80  gender development. Current Directions in Psychological Science, 13(2), 67-70. McCarthy, A., Cyr, M., Fernet, M., & Hébert, M. (2019). Maternal Emotional Support following the Disclosure of Child Sexual Abuse: A Qualitative Study. Journal of child sexual abuse, 28(3), 259-279. McElvaney, R., Greene, S., & Hogan, D. (2014). To Tell or Not to Tell? Factors Influencing Young People’s Informal Disclosures of Child Sexual Abuse. Journal of Interpersonal Violence, 29(5), 928–947. Middleton, J. (2017). Memory Development and Trauma in Preschool Children : Implications for Forensic Interviewing Professionals - A Review of the Literature, 4(1), 1–6. Nemeroff, C. B. (2016). Paradise Lost: The Neurobiological and Clinical Consequences of Child Abuse and Neglect. Neuron, 89(5), 892–909. O’Leary, P., Coohey, C., & Easton, S. D. (2010). The Effect of severe child sexual abuse and disclosure on mental health during adulthood. Journal of Child Sexual Abuse, 19(3), 275–289. Orbach, Y., & Lamb, M. E. (2007). Young children’s references to temporal attributes of allegedly experienced events in the course of forensic interviews. Child Development, 78(4), 1100–1120. Paine, M. L., & Hansen, D. J. (2002). Factors influencing children to self-disclose sexual abuse. Clinical Psychology Review, 22(2), 271–295. 81  Paolucci, E. O., Genuis, M. L., & Violato, C. (2001). A meta-analysis of the published research  on the effects of child sexual abuse. The Journal of Psychology, 135(1), 17-36. Priebe, G., & Göran, C. (2008). Child Abuse & Neglect Child sexual abuse is largely hidden from the adult society An epidemiological study of adolescents ’ disclosures ଝ, 32, 1095–1108. Psutka, J. V., & Psutka, J. (2015, September). Sample size for maximum likelihood estimates of Gaussian model. In International Conference on Computer Analysis of Images and Patterns (pp. 462-469). Springer, Cham. Rotenberg, C., & Cotter, A. (2018). Police-reported sexual assaults in Canada before and after #MeToo, 2016 and 2017. Statistics Canada. Retrieved from: Salmon, K., & Pipe, M. E. (2000). Recalling an Event One Year Later: The Impact of Props, Drawing and a Prior Interview. Applied Cognitive Psychology.<99::AID-ACP639>3.0.CO;2-5 Saywitz, K. J., Goodman, G. S., Nicholas, E., & Moan, S. F. (1991). Children’s memories of a  physical examination involving genital touch: Implications for reports of child sexual abuse. Journal of Consulting and Clinical Psychology, 59, 682–691. Seto, M. C., Babchishin, K. M., Pullman, L. E., & McPhail, I. V. (2015). The puzzle of  intrafamilial child sexual abuse: A meta-analysis comparing intrafamilial and extrafamilial offenders with child victims. Clinical Psychology Review, 39, 42-57. 82  Sivagurunathan, M., Orchard, T., MacDermid, J. C., & Evans, M. (2019). Barriers and  facilitators affecting self-disclosure among male survivors of child sexual abuse: The service providers’ perspective. Child abuse & neglect, 88, 455-465. Sjöberg, R. L., & Lindblad, F. (2002). Limited disclosure of sexual abuse in children whose experiences were documented by videotape. American Journal of Psychiatry, 159(2), 312–314. Sjoberg, R., & Lindblad, F. (2002). Limited disclosure of sexual abuse in children who’s experiences were documented by videotape. American Journal of Psychiatry, 159(February), 312–314. Smith, D. W., Letourneau, E. J., Saunders, B. E., Kilpatrick, D. G., Resnick, H. S., & Best, C. L. (2000). Delay in disclosure of childhood rape: Results from a national survey. Child Abuse & Neglect, 24(2), 273-287. Sorenson, T., & Snow, B. (1991). How children tell: The process of disclosure in child sexual abuse. Child Welfare, 70, 3-15. Stoltenborgh, M., van IJzendoorn, M. H., Euser, E. M., & Bakermans-Kranenburg, M. J. (2011). A global perspective on child sexual abuse: Meta-analysis of prevalence around the world. Child Maltreatment, 16(2), 79–101. Summit, R. C. (1993). Abuse of the child sexual abuse accommodation syndrome. Journal of  Child Sexual Abuse, 1(4), 153-164.  Tashjian, S. M., Goldfarb, D., Goodman, G. S., Quas, J. A., & Edelstein, R. (2016). Delay in  disclosure of non-parental child sexual abuse in the context of emotional and physical maltreatment: A pilot study. Child Abuse and Neglect, 58, 149–159. 83 Thackeray, J. D., Hornor, G., Benzinger, E. A., & Scribano, P. V. (2011). Forensic Evidence Collection and DNA Identification in Acute Child Sexual Assault. Pediatrics, 128(2), 227–232. Theimer, K., & Hansen, D. J. (2017). Attributions of Blame in a Hypothetical Child Sexual Abuse Case : Roles of Behaviour Problems and Frequency of Abuse. Toland, M. D. (2014). Practical guide to conducting an item response theory analysis. The  Journal of Early Adolescence, 34(1), 120-151. doi: 10.1177/0272431613511332 Ullman, S. E. (2007). Relationship to perpetrator, disclosure, social reactions, and PTSD  symptoms in child sexual abuse survivors. Journal of child sexual abuse, 16(1), 19-36. Wolf, M. R., & Pruitt, D. K. (2019). Grooming hurts too: the effects of types of perpetrator  grooming on trauma symptoms in adult survivors of child sexual abuse. Journal of child sexual abuse, 1-15. Wolfteich, P., & Loggins, B. (2007). Evaluation of the children’s advocacy center model:  Efficiency, legal and revictimization outcomes. Child and Adolescent Social Work Journal, 24(4), 333-352. doi: 10.1007/s10560-007-0087-8 Wright, R. E. (2000). Survival analysis. In L. G. Grimm & P. R. Yarnold (Eds.), Reading and  understanding MORE multivariate statistics (pp. 363-407). Washington, DC, US: American Psychological Association. World Health Organization. (2003). Guidelines for medico-legal care of victims of sexual  violence.  84  Young, S. (1997). The use of normalization as a strategy in the sexual exploitation of children by  adult offenders. Canadian Journal of Human Sexuality, 6(4). Retrieved from  Zanarini, M. C., Yong, L., Frankenburg, F. R., Hennen, J., Reich, D. B., Marino, M. F., &  Vujanovic, A. A. (2002). Severity of reported childhood sexual abuse and its relationship to severity of borderline psychopathology and psychosocial impairment among borderline inpatients. The Journal of Nervous and Mental Disease, 190(6), 381-387. doi: 10.1097/01.NMD.0000018963.57744.7E  


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items