“NEED TO TALK”: A LATENT CLASS ANALYSIS OF SEXUAL VICTIMIZATION DISCLOSURE TO A NATIONAL SEXUAL ASSAULT ONLINE HOTLINE By Hannah Feeney A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Psychology—Doctor of Philosophy 2019 “NEED TO TALK”: A LATENT CLASS ANALYSIS OF SEXUAL VICTIMIZATION DISCLOSURE TO A NATIONAL SEXUAL ASSAULT ONLINE HOTLINE ABSTRACT By Hannah Feeney Sexual violence is a pervasive social problem in the United States that affects has long-term negative health consequences for children, adolescents, and adults. While some survivors choose to disclose their victimization to informal or formal resources, others choose to access a third, less studied source of support: online sexual assault crisis lines. Anonymous online hotlines allow survivors a confidential space to disclose sexual victimization and may be particularly beneficial for those who have not previously disclosed or are facing barriers to accessing other services. The current study utilized data from a national anonymous online hotline to answer two main research questions, guided by Liang and colleagues (2005) Model of Helpseeking and Change. First, are there latent classes of hotline sessions that differ based on victim and assault characteristics, and second, do these latent classes account for variation in disclosure behaviors among victims. Results revealed a four-class solution and relationships between class membership and disclosure behavior were identified. Findings suggest that anonymous online sexual assault hotlines are an instrumental resource that can both address survivors’ immediate needs and build bridges to sustainable, long-term support networks. Implications for practice are discussed. ACKNOWLEDGEMENTS This dissertation was made possible by the support of many, and I hope to recognize each for their contributions to this work. First and foremost, to my adviser and chair, Dr. Rebecca Campbell. I have grown as a writer, scholar, and person because of your thoughtful mentorship. You have worn many hats throughout this journey, and your patience with and commitment to my goals has shaped who I am today. I will never be able to thank you enough for believing in me and my ideas. To my committee members: Drs. Deb Bybee, Sarah Ullman, and Heather McCauley. I feel incredibly honored to have received your meaningful feedback, guidance, and support for my research. It has been a privilege to work with each of you. To Kimberly Goodman and RAINN, thank you for trusting me with this project. The foundation of this dissertation was made possible by your hard work and vision. To my colleagues, friends, and family. I thank you deeply for all of the support, love, and encouragement along the way. Katie, Val, Dani, and the "advisee group" – thank you for being my sounding board, cheerleaders, and biggest motivators. I look forward to collaborating with each of you in the future. Erin, Leana, Jess, Jamie, and the Busstacys – your friendship has made the challenging parts of this work possible. You have brought joy to every late night and early morning. To my parents and brother – thank you for letting my carve my own path and for always trusting me to do so. To my partner, Nathan. When we signed up for this expedition six years ago, I could not have anticipated all of the ways in which you would demonstrate dedication and patience to the process. Our wonderful life is possible because of you, and I am so incredibly grateful that you have chosen to pursue my dreams with me. I share this accomplishment with you. To Rowan. Every achievement, every day, every high low surprise, every-single-thing is so much brighter because of you. Thank you for the joy you bring to my world. This dissertation is dedicated to survivors of sexual victimization – both those seeking help and those still finding their voice. Your courage inspired this work and your resilience continues to motivate me every day. I am humbled by the opportunity to share part of your story. iii TABLE OF CONTENTS LIST OF TABLES ............................................................................................................................... vi LIST OF FIGURES ............................................................................................................................. vii INTRODUCTION & OVERVIEW ........................................................................................................1 CHAPTER 1: LITERATURE REVIEW .................................................................................................4 Definitions of Sexual Violence ...................................................................................................4 Prevalence and Incidence of Sexual Violence .............................................................................5 The Impact of Sexual Violence ..................................................................................................7 Psychological impact on adults ......................................................................................7 Psychological impact on adolescents and children..........................................................8 Physical impact on adults ..............................................................................................8 Physical impact on adolescents and children .................................................................9 Psychosocial impact on adults ..................................................................................... 10 Psychosocial impact on adolescents and children ........................................................ 10 Disclosure of Sexual Violence ................................................................................................. 11 Informal disclosure ...................................................................................................... 11 Formal disclosure ........................................................................................................ 13 Other forms of disclosure ............................................................................................ 16 Model of Helpseeking and Change ........................................................................................... 18 Problem recognition and appraisal ............................................................................... 19 Decision to seek help ................................................................................................... 20 Support selection ......................................................................................................... 21 The Current Study .................................................................................................................... 22 CHAPTER 2: METHOD ...................................................................................................................... 27 Sample ..................................................................................................................................... 27 Procedures ............................................................................................................................... 30 Measures ................................................................................................................................. 30 Data Analytic Plan ................................................................................................................... 33 CHAPTER 3: RESULTS ...................................................................................................................... 35 Sample Characteristics ............................................................................................................. 35 Are There Underlying Latent Classes of Hotline Sessions? ....................................................... 37 Model formulation and selection .................................................................................. 38 Model interpretation .................................................................................................... 43 Do Latent Classes Explain Variation In Disclosure Behaviors?................................................. 46 First and general disclosure findings ............................................................................ 47 Informal disclosure findings ........................................................................................ 48 Formal disclosure findings ........................................................................................... 48 Additional Exploratory Regressions ......................................................................................... 52 CHAPTER 4: DISCUSSION ................................................................................................................ 54 Key Findings and Contributions to Helpseeking Theory & Research ........................................ 56 Class 1 (Minors) .......................................................................................................... 56 Class 2 (Adults Recent) ............................................................................................... 59 iv Class 4 (Adults Past) ................................................................................................... 61 Class 3 (Not Discussed) ............................................................................................... 63 Overall model .............................................................................................................. 64 Context, Limitations, and Implications for Future Research ...................................................... 66 Implications for Practice .......................................................................................................... 69 Conclusion ............................................................................................................................... 71 APPENDIX .......................................................................................................................................... 72 REEFERENCES .................................................................................................................................. 82 v LIST OF TABLES Table 1: Victim and assault characteristic variables of interest ............................................................... 31 Table 2: Disclosure variables of interest................................................................................................. 32 Table 3: Frequencies & descriptives for all victim and assault characteristics variables .......................... 36 Table 4: Frequencies & descriptives for all disclosure variables ............................................................. 37 Table 5: Victim and assault characteristics latent class analysis fit indices.............................................. 40 Table 6: Posterior probabilities of the final 4-class solution.................................................................... 45 Table 7: Number and percentage of disclosure variables endorsed by class ............................................ 47 Table 8: Binary logistic regression results of disclosure regressed on latent class dummy variables ........ 51 Table 9: Number and percentage of exploratory dependent variables endorsed by class ......................... 53 vi LIST OF FIGURES Figure 1: Model of Helpseeking and Change (Liang et al., 2005) ........................................................... 19 Figure 2: Sampling criteria .................................................................................................................... 28 Figure 3: Model-estimated, class-specific item probability profile-plot for 4-class solution .................... 42 Figure 4: Highlighted Model of Helpseeking and Change (Liang et al., 2005) ........................................ 55 Figure 5: Amended Model of Helpseeking and Change for Class 1 (Minors) .......................................... 56 Figure 6: Amended Model of Helpseeking and Change for Class 2 (Adults Recent) ............................... 60 Figure 7: Amended Model of Helpseeking and Change for Class 4 (Adults Past) ................................... 62 Figure 8: Amended Model of Helpseeking and Change for Class 3 (Not Discussed) .............................. 63 Figure 9: Amended Model of Helpseeking and Change ......................................................................... 65 vii INTRODUCTION & OVERVIEW Sexual violence is a broad term that encompasses all acts of a sexual nature perpetrated in the absence of explicit consent, including sexual assault, rape, and sexual exploitation. National epidemiological data indicates that individuals of all ages and genders are at risk of experiencing these varying types of sexual violence; however, some groups are at higher risk for specific types of victimization than others. Females, adolescents, bisexual individuals, and transgender individuals are at particularly high risk of sexual victimization (James et al., 2016; NISVS, 2010; Planty, Langton, Krebs, Berzofsky, & Smiley-McDonald, 2013; Rennison & Rand, 2003). Such violence may cause negative physical, psychological, and psychosocial health outcomes. These impacts differ for adults, adolescents, and children, but consequences may be severe and long-lasting regardless of age group (Campbell & Townsend, 2011). Survivors may choose to seek support for their victimization through disclosure to various types of helping systems, including informal systems (e.g., family, friends, intimate partners) or formal systems (e.g., law enforcement, medical professionals, mental health professionals). While many survivors do disclose to informal systems (Jacques-Tiura, Tkatch, Abbey, & Wegner, 2010; London, Bruck, Ceci, & Shuman, 2007; Rickert, Wiemann, & Vaughan, 2005), most do not disclose to formal helpers (Casey & Nurius, 2006; Jacques-Tiura et al., 2010; Kilpatrick, Saunders, & Smith, 2003; Lonsway & Archambault, 2012). Survivors who are not ready or not able to disclose to formal helping systems are therefore also unable to receive the potential positive benefits of formal disclosure (e.g., access to healing resources). However, survivors may also consider using anonymous hotlines, which are easily accessible and offer considerable privacy and confidentiality. Sexual assault crisis phone lines have been successfully supporting survivors for decades (Macy, Giattina, Sangster, Crosby, & Montijo, 2009) and recently, with new developments in technology, this service has evolved into internet-based sexual assault crisis lines. Less is known about these online crisis hotlines, but initial research suggests that they may be particularly beneficial to those who have not previously sought support for their sexual victimization due to the potential barriers of other helping services (Finn, Garner, & Wilson, 2011; Finn & Hughes, 2008). 1 How a survivor chooses to disclose to one or more of these helping systems (e.g., friends, law enforcement, anonymous online hotlines) is a complex process. According to Liang and colleagues’ (2005) Model of Helpseeking and Change for victims of intimate partner violence (IPV), deciding whether to disclose and to whom, is likely a non-linear, dialectical process that includes three overarching stages: 1) problem recognition and appraisal, 2) the decision to seek help, and 3) support selection. Each of these stages may be thought of as distinctly influenced by individual-, interpersonal-, and sociocultural- factors to form an interactive, ongoing feedback loop that can affect whether a survivor seeks help. While the model was developed for IPV victims, it may also be applied to survivors of sexual victimization and their decisions to disclose acts of sexual violence. Victims may choose to share their assault experiences with informal, formal, or other resources at different points post-assault, but little is known about the factors that may influence these choices and there may be many components affecting the process between Stage 2 (decision to seek help) and Stage 3 (support selection) that are currently unexplored. Understanding these relationships and patterns is crucial to identifying potential gaps in post-assault support provision. To date, the literature on disclosure of and help seeking for sexual victimization has focused on Stage 1 (problem recognition and appraisal) and Stage 2 (decision to seek help). From this body of work, we know that survivors differ in their “acknowledgement” of sexual victimization and that the decision to pursue assistance for non-consensual sexual experiences varies based on victim and assault characteristics. Building on this prior work, the current study explored a potential gap in Liang and colleagues’ (2005) Model of Helpseeking and Change, one between Stage 2 of the model (decision to seek help) and Stage 3 (support selection). This study assessed the possibility of underlying groupings (latent classes) of users who choose an anonymous online hotline as a support provider, and whether such groups who have chosen to use the national anonymous online hotline service demonstrate differences in their disclosure behaviors. The current study utilized latent class analysis and binary logistic regression with national-level anonymous online hotline data to answer these questions. Studying victims’ use of online hotlines in this way further contextualizes Liang and colleagues’ (2005) framework, insofar that it 2 describes which survivors may require an additional “landing pad” between Stage 2 (decision to seek help) and Stage 3 (support selection). These research questions contribute to the disclosure and helpseeking literature by identifying types of anonymous hotline sessions and establishing how disclosure to hotlines relates to other help-seeking behaviors. To set the stage for the current study, a review of the relevant literature will be presented. First, various types of sexual victimization will be operationalized. Next, the current research on the prevalence and impact of sexual victimization will be discussed, noting findings specific to adult, adolescent, and child survivors. The literature review will then cover survivors’ post-assault help-seeking with informal systems, formal systems, and other forms of disclosure. Finally, the guiding framework for the research, Liang and colleagues (2005) Model of Helpseeking and Change, will be outlined. A description of the current study, methods, findings, and conclusions follow. 3 CHAPTER 1: LITERATURE REVIEW Definitions of Sexual Violence Sexual violence is a broad term that encompasses all acts of a sexual nature perpetrated in the absence of explicit consent. Acts may be considered nonconsensual when they occur against a victim’s will by force, threat, or coercion, such as by violence, fear of violence, isolation, or abuse of power (United Nations [UN], 2016). An individual may also be unable to give consent due to age, disability, or if they are under the influence of drugs or alcohol (National Center for Victims of Crime [NCVC], 2012; UN, 2016). Acts of sexual violence may include, but are not limited to, sexual assault, rape, and sexual exploitation (UN, 2016). While these terms are often used interchangeably and vary by state, they each have distinctly different meanings (Campbell & Townsend, 2011). Sexual assault is most commonly defined as an act of unwanted and/or nonconsensual sexual penetration, contact, or behavior, which may include sexual touching, fondling, or exposure to exhibitionism or pornography (Krug, Mercy, Dahlberg, & Zwi, 2002; NCVC, 2012). Child sexual assault, also referred to as child sexual abuse or molestation, is a specific type of sexual assault that involves sexual penetration, contact, or behavior (e.g., forced watching of masturbation) with a minor (Child Welfare Information Gateway, 2016). When the sexual contact occurs between family members (regardless of age), it is referred to as incest. Rape, another distinguishable type of sexual assault, may be defined as nonconsensual vaginal, anal, or oral penetration by a body part or object (Koss & Achilles, 2008; NCVC, 2012). Rape is distinct from sexual assault in that sexual assault does not necessarily entail penetration while rape always does (Koss & Achilles, 2008; UN, 2016). Sexual exploitation differs from sexual assault and rape in that the abuse is rooted in a victim’s vulnerability and typically results in the offender profiting sexually, monetarily, or socially from the abuse (Estes & Weiner, 2012; Sexual Harassment/Assault Resource & Education [SHARE], 2018; UN, 2016). This may come in the form of a power differential between victim and offender (e.g., an individual being forced to watch their employer masturbate; an individual being coerced to take nude photos by threat of blackmail) or the victim trusting the offender to have their best 4 interests in mind (e.g., an adolescent being prostituted by their intimate partner). It does not necessarily involve physical contact between the perpetrator and victim but rather includes interactions such as forced prostitution, trafficking, nonconsensual voyeurism, or distributing sexual material (e.g., nude photos) without consent or as blackmail (Estes & Weiner, 2012; SHARE, 2018; UN, 2016). As an example, in the case of forced prostitution the person prostituting the victim would be committing sexual exploitation as they are profiting monetarily from the exploitation of an individual, while the person engaging in sexual acts with the victim would be committing sexual assault and/or rape (if penetration occurred). Prevalence and Incidence of Sexual Violence National epidemiological data indicates that individuals of all ages and genders are at risk of experiencing these varying types of sexual violence and some groups are at higher risk for specific types of victimization than others. Research on adult survivors of sexual assault has found that as many as 43.9% of women and 23.4% of men have experienced sexual violence other than rape in their lifetime (e.g., sexual assault; Breiding, 2014). In addition, 1 in 5 (19.3%) women and 1 in 71 (1.7%) men have reported being raped in during their life (Breiding, 2014). As these statistics suggest, women have been found to be at higher risk for all types of sexual violence when compared to men with 91% of victims of sexual assault and rape being female (Rennison & Rand, 2003). Transgender individuals have been found to be at the highest risk for sexual violence, with as many as 47% of transgender individuals being sexually assaulted at some point in their lifetime (James et al., 2016). Gay men and bisexual individuals are also at higher risk of sexual victimization when compared to heterosexual individuals and lesbian women (NISVS, 2010). In one national study, 40% of gay men and 47% of bisexual men reported sexual victimization other than rape (21% of heterosexual men), while 46% of bisexual women reported being raped (17% of heterosexual women; 13% of lesbians; NISVS, 2010). Sexual assault and rape also disproportionately affect rural communities and lower income communities (44% of victims have a household income less than $25,000; Planty et al., 2013). Available research suggests that racial and ethnic groups are victimized at similar rates, aside from American Indians and Alaskan Natives who are 3.5 times more likely to be sexually assaulted when compared to 5 other races (Planty et al., 2013). However, this finding may also reflect differential rates of disclosure among diverse communities. While rates of sexual violence against adults are alarmingly high, studies have found that adolescents at even greater risk. In fact, adolescent females aged 12-17 have been found to be 2.7 times more likely to experience sexual victimization when compared to older adult women aged 35-64 (Planty et al., 2013). Adolescents are also at greater risk when compared to children; when reviewing rates of sexual victimization for those under the age of 18, 66% of victims are between the ages of 12-17 (Greenfeld, 1997). Overall, 26.6% of women, 5.1% of men, and 12% of transgender individuals have experienced some kind of sexual violence before the age of 17 (Finkelhor, Shattuck, Turner, & Hamby, 2014; Grant et al., 2011). Regarding rape more specifically, as many as 40.4% of female survivors and 21.3% of male survivors reported experiencing their first completed rape before the age of 18 (rates of rape are currently unknown for transgender youth; Breiding, 2014). Rates of child sexual victimization are more difficult to estimate than adult and adolescent sexual victimization as violence against children frequently goes unreported (Hanson, Resnick, Saunders, Kilpatrick, & Best, 1999). However, Child Protective Services agencies encounter as many as 63,000 substantiated cases of child sexual victimization annually (USDHHS, 2018). In 2000 specifically, approximately 10% of reported child maltreatment cases were determined to be child sexual abuse (Putnam, 2003). Self-report studies of adults have found that as many as 12-25% of females and 16-28% of males recall an experience of child sexual assault or rape (Black et al., 2011; Dube et al., 2005). In addition to rates of victimization, the characteristics of sexual violence differ by age group as well. All age groups are most likely to be victimized by individuals known to them, and this is especially true for children. Adults and adolescents are most likely to be assaulted by friends or acquaintances (78% of adults; 74% of adolescents) and least likely to be assaulted by strangers (22% of adults; 23.2% of adolescents; Jones, Rossman, Wynn, Dunnuck, & Schwartz, 2003; Kilpatrick et al., 2003; Morgan & Kena, 2018). Conversely, children are most likely to be assaulted by family members or acquaintances (90%); they are also least likely to be assaulted by strangers (10%; Finkelhor, 2012). The younger victims 6 are, the more likely they are to be assaulted by a family member specifically; 50% of children under the age of six who have experienced sexual assault and 23% of those aged 12-17 who have experienced sexual assault, were victimized by a family member (Snyder, 2000). The family member most likely to commit incest is a parent (80%; USDHHS, 2018). Notably, perpetrators of child sexual assault are not necessarily adults and may also include adolescent family members (e.g., siblings, cousins) or playmates (RAINN, 2018). Given that adolescents and children are more likely to be sexually victimized by family members or friends, it would be unsurprising if they were also more likely than adults to be living with their perpetrator at the time of the assault, however the current literature has yet to explore this relationship. Children have been shown to be most likely to experience ongoing, repeated sexual abuse when compared to other age groups (Filipas & Ullman, 2006; Hudson & Nelson, 1986). Relatedly, adolescents and children are less likely to be assaulted by use of weapons or physical violence when compared to adults (Snyder, 2000). This is likely due to the fact that most adolescents and children are victimization by individuals known to them and known offenders are less likely to use these tactics (Koss et al., 1988; Stermac, Del Bove, & Addison, 2004; Ullman & Siegel, 1993). The Impact of Sexual Violence Victims of sexual violence may demonstrate a wide range of negative post-assault health outcomes. These impacts are significant, well documented, and have the potential to permeate all aspects of a survivor’s life, including the psychological, physical, and behavioral realms. These outcomes have been shown to have differential impact based on when in the lifespan the victimization occurred. Psychological impact on adults. In the immediate aftermath of an assault, survivors may experience shock, confusion, fear, or social withdrawal and most victims (70%) experience moderate to severe distress after their victimization (Herman, 1992; Langton & Truman, 2014). Over time, these symptoms may evolve and many victims will ultimately meet the diagnostic criteria for depression, anxiety, and/or post-traumatic stress disorder (PTSD). Campbell and colleagues’ 2009 review of the literature found that 73-82% of victims developed fear or anxiety disorders with 12-40% meeting the 7 diagnostic criteria for generalized anxiety disorder, 13-51% developing depression, 23-44% experiencing suicidal ideation with 2-19% actually attempting suicide. A recent meta-analysis of sexual assault victimization psychopathology indicated sexual assault resulted in the highest risk for suicidality when compared to other traumatic conditions (Dworkin, Menon, Bystrynski, & Allen, 2017). Moreover, as many as 7-65% of victims develop symptoms consistent with PTSD after an act of sexual victimization. Such findings indicate that rape survivors are actually one of the largest groups of individuals in the United States living with PTSD (Campbell, Dworkin, & Fehler-Cabral, 2009; Dworkin, et al., 2017; National Center for PTSD, 2018). Some victims will also experience increased cognitive difficulties, such as slower thinking, difficulty concentrating, or memory impairment (Barrera, Calderón, & Bell, 2013; Trickett, Noll, & Putnam, 2011; Daignault & Hebert, 2009). Finally, victims of sexual violence are more likely than their non-victimized counterparts to develop bipolar conditions, obsessive-compulsive disorder (OCD), or disordered eating (Bonomi, Nichols, Kammes, & Green, 2018; Dworkin et al., 2017; Fischer, Stojek, & Hartzell, 2010; Martin, Macy, & Young, 2010). Psychological impact on adolescents and children. Immediately after an assault, adolescent and child victims may experience many of the same symptoms as adults (e.g., shock, confusion, fear) and as time passes, they may also develop altered stress responses, eating disorders, tendencies to self-harm, depression, anxiety, PTSD, and/or attempt suicide (Basile et al., 2006; Danielson & Holmes, 2004; Trickett et al., 2011). In fact, adult survivors of child of sexual assault are three times more likely to develop depression and four times more likely to develop PTSD when compared to their non-victimized counterparts (Zinzow et al., 2012). Physical impact on adults. Victims of sexual violence are also vulnerable to myriad of physical health concerns. First, the act of violence itself may result in physical trauma, such as anogenital injuries or physical injuries (i.e., injury to body areas not encompassed by the anogenital area), with 31-58% of female victims incurring some type of injury as a result of sexual assault (Planty et al., 2013; Tjaden & Thonnes, 2000). Physical injuries tend to be more common than anogenital injuries in adults; studies of women presenting to emergency departments for sexual assault have found that just over half (52%) incur 8 physical injuries and 20% suffer from anogenital injuries (41% have no detected injuries; Sugar, Fine, & Eckert, 2004). Once these initial physical wounds heal, victims may still experience physical complications. For example, adult victims of sexual violence may experience gynecological problems, issues related to sexual health, gastrointestinal issues, neurological symptoms, or cardiopulmonary problems (Campbell & Townsend, 2011). Physical impact on adolescents and children. Adolescent and child victims of sexual violence are also susceptible to the physical impacts of trauma. In fact, studies of female adolescents who sought medical care for an act of sexual violence have found that as many as 27-37% of patients had documented physical injuries and 66-85% had documented anogenital injuries (Adams, Girardin, & Faugno, 2001; Adams & Knudson, 1996; Baker & Sommers, 2008; Jones et al., 2003; Markowitz, 2012; Lynch & Duval, 2010; Sugar et al., 2004). Notably, anogenital injuries are more common than physical injuries in adolescents, a finding that is opposite from that of adults. While not explicitly explored in the literature, it is possible that this difference is due to differences in assault characteristics between the two age groups. Adolescents are less likely to experience physical violence during an assault when compared to their adult counterparts, and therefore may incur fewer physical injuries (Eckert, Sugar, & Fine, 2001; Hanson et al., 2003; Jones et al., 2003; Snyder & Sickmund, 2006; Tjaden & Thoennes, 2006). Children are less likely to have their injuries detected overall, with approximately 5-39% of victims having documented anogenital injury (rates of physical injury unknown; Campbell, Patterson, Dworkin, & Digel, 2010). It has been suggested that lower rates of injury detection in children are due, at least in part, to the delays in reporting that children often experience (Campbell et al., 2010). Over time, adolescent and child survivors may experience other health problems, including sleep difficulties, chronic pain, chronic headaches, or digestive issues (McCauley et al., 1997; Polusny & Follette, 1995). Adult survivors of adolescent or child sexual assault are 30% more likely than those who were not assaulted as adolescents or children to develop diabetes, heart problems, stroke, or hypertension (Sachs-Ericsson, Blazer, Plant, & Arnow, 2005). 9 Psychosocial impact on adults. Survivors of sexual violence have been found to have negative behavioral outcomes as well, such as withdrawal from their usual activities or increased emotional outbursts (Gutner, Rizvi, Monson, & Resick, 2006). More than a third of victims experience problems at work (38%) or with family and friends (37%) (Langton & Truman, 2014). Problem drinking and drug abuse are higher amongst individuals who have experienced sexual violence (Campbell et al., 2009; Najdowski & Ullman, 2009). In fact, sexual assault victims have been found to be 6 times more likely to use cocaine, 5.3 times more likely to use prescription drugs non-medically, and 10 times more likely to use other major drugs when compared to their non-assaulted counterparts (Kilpatrick, Edmunds, & Seymour, 1992). Community samples indicate that as many as 44% of victims use prescription drugs post-rape, with many using the drugs to “self-medicate” as a coping mechanism (Sturza & Campbell, 2005). Psychosocial impact on adolescents and children. Adolescents and children who have experienced sexual victimization are at great risk for developing even more negative psychosocial impacts. In the short term, adolescent and child victims are likely to report academic problems, such as attention issues, lower academic achievement, and fewer positive feelings about school (Daignault & Hebert, 2009; Reyome, 1994). Lower self-esteem, poor social skills, physical aggression, delinquency resulting in arrest, and running away from home are also reported in adolescent and child survivors (Lalor & McElvaney, 2010; National Research Council, 1993). Adolescent and child survivors may develop distorted views of sex and mistrust of adults, report higher rates of risky sexual behavior, and are at higher risk for teenage pregnancy than their non-assaulted peers (Lalor & McElvaney, 2010). These impacts have the potential to cause significant disruption to healthy social development and can have long-lasting impacts. For example, they are also less likely to attend four-year universities and are at higher risk for revictimization as adults (Lalor & McElvaney, 2010). Adult survivors of adolescent and child sexual assault often exhibit symptoms of problem drinking over time and are 4 times more likely than their non- victimized counterparts to develop symptoms of drug abuse (Ullman, 2016; Wadsworth & Records, 2013; Zinzow et al., 2012). 10 Disclosure of Sexual Violence Given the various negative outcomes a victim of sexual violence may experience, it is unsurprising that some may seek support after their victimization. Survivors may choose to disclose their experience of violence to others, a process that has been described as finding the strength to “break the silence and speak out” about their victimization (Ahrens, 2006). Disclosure is a complicated process impacted by many individual, interpersonal, and sociocultural factors (Chen & Ullman, 2014; Long & Ullman, 2013; Starzynski, Ullman, Townsend, Long, & Long, 2007). These disclosures may occur voluntarily or by invitation and may be to various types of helping systems after an assault, the first being informal sources (Campbell, Greeson, Fehler-Cabral, & Kennedy, 2015) Informal disclosure. Informal helping sources typically consist of individuals who are not trained to respond to disclosures of sexual violence and are not obligated to file a formal report on the disclosure. This often includes family (e.g., parents, grandparents, siblings), friends, or intimate partners (e.g., spouse, boyfriend), but may also include acquaintances (e.g., classmates, co-workers, a friend’s parent) or strangers (e.g., someone encountered immediately post-assault). Research suggests that most survivors who choose to disclose, will do so to informal sources at some point, with as many as 96% of adult survivors, 60-85% of adolescent survivors, and 34% of child survivors confiding in informal sources during their lifetime (Jacques-Tiura et al., 2010; London et al., 2007; Rickert et al., 2005). Notably, children are less likely to disclose to informal sources than other groups, with younger children being the least likely (London, Bruck, Ceci, & Shuman, 2003). In fact, the younger the survivor, the more likely they are to have their victimization detected “by accident” as opposed to through purposeful telling by the child (Alaggia, Collin-Vézina, & Lateef, 2017). Children may not disclose for fear of their perpetrator (particularly if the perpetrator is a family member or close to the family), causing trouble, upsetting their family, or for a lack of understanding or opportunity (Crisma, Bascelli, Paci, & Romito, 2004; Goodman-Brown, Edelstein, Goodman, Jones, & Gordon, 2003; London et al., 2003; Malloy, Brubacher, & Lamb, 2011; McElvaney, 2008). Indeed, Alaggia and colleagues’ (2017) literature review of child sexual abuse disclosures found that barriers outweighed any facilitators allowing children 11 to come forward with a disclosure of sexual assault. When survivors of child sexual assault do disclose, it is often more than five years after the assault (33-69%) with many disclosing well into adulthood or never disclosing at all. Only 21-33% of child victims (as compared to 40% of adolescent victims) will disclose within a month of the assault (Broman-Fulks et al., 2007; Hébert, Tourigny, Cyr, McDuff, & Joly, 2009; London et al., 2007; Schönbucher, Maier, Mohler-Kuo, Schnyder, & Landolt, 2012). One study of adult male survivors of child sexual assault found that men wait 21 years on average to disclose abuse experienced during childhood and 28 years to have an in-depth discussion about their victimization (Easton, 2012). If the perpetrator of the sexual victimization was a family member, a child is more likely to delay disclosure or never disclose at all (Goodman-Brown et al., 2003; Kogan, 2004; Lyon, Ahern, Malloy, & Quas, 2010; Ullman, 1996). When survivors do choose to disclose, adult victims, adult survivors of child sexual assault, and adolescent victims are all most likely to turn to their friends (Ahrens, Campbell, Ternier-Thames, Wasco, & Sefl, 2007; Fehler-Cabral & Campbell, 2013; Jacques-Tiura et al., 2010; Kogan, 2004; Orchowski & Gidycz, 2012; Stein & Nofzinger, 2008; Ullman, 1996, 2010; Ullman & Filipas, 2001). Young people deciding to disclose tend to be hesitant about discussing acts of sexual violence with their parents or other adults, as such a disclosure involves talking about potentially stigmatized behavior (e.g., sex, drug or alcohol use; Clements, Speck, Crane, & Faulkner, 2004; Fehler-Cabral & Campbell, 2013; Smith & Cook, 2008). However, the peers to whom adolescents disclose often encourage survivors to seek additional help from parents (Campbell et al., 2015). While fewer child survivors of sexual assault disclose during childhood, of those who do, more than half will disclose to a parent, particularly their mother, and about 40% will tell a peer (Broman-Fulks et al., 2007; Hershkowitz et al., 2007; London et al., 2003; Malloy, Brubacher, & Lamb, 2013). When survivors disclose, adults, adolescents, and children will typically do so in an attempt to receive support, advice, or assistance (Ahrens, et al., 2007; Jacques-Tiura et al., 2010). Adults cite disclosing to informal sources when they feel particularly close to an individual, need to talk about the assault, and/or “just feel like the person should know” about the victimization (Jacques-Tiura et al., 12 2010). Fehler-Cabral & Campbell found that adolescents disclose to friends or family from a desire to “vent and seek validation,” (2013, p. 77). In a separate qualitative study, interviews with adolescent and child survivors of sexual victimization revealed younger survivors will disclose to informal sources for five main reasons: 1) the abuse was noticed or asked about by others, 2) they could no longer “handle" the abuse on their own, 3) they learned about sex and healthy relationships, developing reference points and language with which they could talk about the abuse, 4) the nature of the abuse changed, or 5), they wanted to protect others, like a sibling, from experiencing the abuse (Allnock & Miller, 2013). Adult survivors of child sexual abuse will also disclose to informal sources when they remember forgotten instances of victimization (Allnock & Miller, 2013). The actual impact of the disclosure tends to depend on how the disclosure recipient responds to the survivor (Ullman, & Peter‐Hagene, 2014). Friends have been found to have the most positive, healing reactions to informal disclosures of sexual victimization (Ahrens, Cabral, & Abeling, 2009; Filipas & Ullman, 2001). Conversely, family members and partners provide reactions that are more mixed; while some are positive and helpful, others may be perceived as providing blaming, controlling, or egocentric responses (Filipas & Ullman, 2001). These more negative reactions may have a harmful effect on survivors, resulting in a worse recovery and/or damaged relationship between the survivor and their disclosure recipient (Ullman, & Peter‐Hagene, 2014; Milliken, Paul, Sasson, Porter, & Hasulube, 2016). However, informal helpseeking, when responded to well, can have great benefits for the survivor. Social support from family or friends can mitigate some of the negative psychological and health impacts of an assault (Campbell et al., 2015; Broman-Fulks et al., 2007; Ullman, & Peter‐Hagene, 2014). Adolescents and children in particular are at reduced risk for major depressive episodes and delinquency if they disclose within a month of the assault (Broman-Fulks et al., 2007; Ullman, 2007). Prompt disclosure also reduces the risk for additional victimization in cases of ongoing abuse, particularly for children (Kogan, 2005). Formal disclosure. A minority of those who disclose to informal support sources will go on to seek formal help (Stein & Nofzinger, 2008). These formal sources may include the legal system (e.g., 13 police, detectives, a Title IX office), the medical system (e.g., doctors, nurses), the mental health system (e.g., psychologists, counselors, social workers), the education system (e.g., teachers, principals), advocacy services (e.g., rape crisis centers, sexual assault advocates), or any other source that may be required to file a formal report on the disclosure. While telling informal sources about an assault may provide a survivor with social support, telling formal sources may connect a survivor with important opportunities for professional assistance by way of: opportunities to pursue prosecution (i.e., legal), care for physical injuries and other health needs (i.e., medical), and/or assistance with mental health impacts (i.e., mental health; Campbell, 2008; Ullman, 2010). Victims may also choose to disclose to advocacy services, which are confidential resources that assist survivors in navigating complex legal, medical, and/or mental health systems (Townsend & Campbell, 2018). Few victims actually pursue these means of formal care though, with less than 24% of adult survivors and 8-14% of adolescent and child survivors disclosing to formal resources (Casey & Nurius, 2006; Jacques-Tiura et al., 2010; Kilpatrick et al., 2003; Lonsway & Archambault, 2012). These low rates of formal reporting may be attributed to a number of factors, including fear of shame, blame, stigma, and/or confidentiality concerns (Finkelhor & Wolak, 2003; Fisher, Cullen, & Turner, 2000; Rickwood, Deane, Wilson, & Ciarrochi, 2005; Wilson & Deane, 2001). Victims who have non-stereotypical assaults (e.g., non-stranger rapes without weapons), survivors of color, victims of low socioeconomic status, and male survivors are less likely than others to report their assaults to formal systems (Campbell, Wasco, Ahrens, Sefl, & Barnes, 2001; Martin, 2005; Starzynski, Ullman, Filipas, & Townsend, 2005). Finally, access issues, such as being unable to access services independently or lack of knowledge surrounding available resources, may be considered barriers to formal helpseeking that survivors of sexual victimization may experience (Holland & Cortina, 2017). The small percentage of survivors who do seek formal help, do so for select reasons. Victims in need of imminent medical care, for example, will disclose to the medical system in order to receive treatment (Campbell, 2008; Ullman, 2010). Similarly, victims who hope to pursue formal prosecution of their perpetrator will disclose to the criminal justice system (Campbell, 2008; Ullman, 2010). However, 14 disclosure to formal systems is a complex process and how and to whom victims disclose may be affected by aspects of the abuse (e.g., co-occurrence of domestic violence), their sociodemographic characteristics (e.g., race/ethnicity or socioeconomic status), or even prior disclosures (Chen & Ullman, 2014; Long & Ullman, 2013; Starzynski et al., 2007). Connections to formal helping systems are often encouraged or initiated by informal sources to whom the victim had previously disclosed (Campbell et al., 2015; Stein & Nofzinger, 2008). In the case of adolescents, qualitative research has shown how peers will often give advice suggesting that the survivor seek help from an adult, who will then assist in seeking formal help (Campbell et al., 2015). While this process happens voluntarily for some survivors (i.e., they welcome the assistance from peers and adults), it occurs involuntarily for others. In these circumstances, peers will disclose to adults against the victim’s wishes and adults will make the victim seek formal help. In other scenarios, peers will seek help on the victim’s behalf if the victim was unconscious at the time of the assault (Campbell et al., 2015). Children experience similar pathways to formal help as adolescents and are most likely to make formal reports when they first disclose to a parent (Stein & Nofzinger, 2008). Similar to informal disclosure, the result of formal helpseeking are wholly dependent on the reaction of the disclosure recipient. When law enforcement, in particular, have a supportive response and express belief in the victim, victims are much more likely to have a positive experience and stay engaged with the formal process (Campbell et al., 2015; Feeney, Campbell, & Cain, 2018). Unfortunately, the majority of survivors who choose to disclose to formal helping systems do not have positive experiences. Research on disclosures of rape to law enforcement and medical professionals have found many victims encounter victim blaming behaviors that leave them feeling bad about themselves, depressed, violated, distrustful of others, distressed, frustrated, embarrassed, and, ultimately, reluctant to seek further help (Campbell, 2005; Campbell & Raja, 2005; Konradi, 2007). In fact, some survivors state that they would not have reported if they knew what the experience would have been like (Logan, Evans, Stevenson, & Jordan, 2005). Adolescents and children may be particularly susceptible to these types of secondary victimization (i.e., victim blaming attitudes, practices, and behaviors by community service providers; Campbell, 2008). Rape myths endorse the idea that adolescents may lie about sexual violence in an 15 attempt to “cover-up” for other behaviors (e.g., underage drinking, truancy; Feeney et al., 2018; Shaw, Campbell, Cain, & Feeney, 2017) and that children are “unreliable” in their disclosures (Berliner & Barbieri, 1984). Overall, most victims do not report to formal helping systems. Survivors who are not ready or are not able to disclose to these resources due to fear of judgment, confidentiality concerns, or lack of knowledge or access, are therefore also unable to receive the potential positive benefits of formal disclosure. However, survivors may also consider using anonymous hotlines, which are easily accessible and defined by the privacy and confidentiality they offer. Less is known about hotlines, but a handful of studies have explored the anonymous resource and the ways in which survivors use it. Other forms of disclosure. Anonymous phone crisis lines have been effectively used to elicit disclosure of difficult topics for decades. Allowing survivors to quickly and confidentially access trained sexual assault advocates, the intent of such crisis lines are generally to help ensure survivor safety, educate on effective coping mechanisms, and share information on local support resources (Macy et al., 2009). Phone hotlines have been shown to be flexible in the types of support they provide, such that they can respond to various types callers and crises (Ingram et al., 2008). Some survivors may reach out for immediate post-assault assistance, while others call in to process the long-term effects of victimization (Townsend & Campbell, 2018). Hotlines have been found to offer benefits to survivors of sexual violence that other forms of disclosure may be unable to provide. The resource can be particularly helpful for victims who may have accessibility concerns (e.g., those with disabilities, those living in rural communities) or fear that services may be too expensive or not applicable to their needs (e.g., male survivors; Finn et al., 2011; Young, Pruett, & Colvin, 2016). Importantly though, these hotlines can also help connect survivors to resources they may not have known about otherwise (Young et al., 2016). Wasco and colleagues (2004) asked survivors who used a sexual assault hotline how helpful the call had been to them and 84% shared that it had given them “somewhat more” or “a lot more” information about resources. Overall, survivors find 16 these interactions useful; 94% of the survivors in Wasco and colleagues’ (2004) study indicated that the call had provided “some support” or “a lot of support.” Sexual assault crisis phone lines have been providing advocacy services to survivors for decades (Macy, Giattina, Sangster, Crosby, & Montijo, 2009) and recently, this resource has evolved with technology and text and internet-based crisis lines have advanced the advocacy landscape. Limited research on these anonymous texting hotlines has suggested that survivors may prefer these methods to phone hotlines due to their increased discretion; not only is the service completely confidential, but no one can overhear a survivor detailing an account of sexual assault when written via text (Noble, 2016). Internet-based disclosures are also becoming increasingly popular and taking a variety of forms. Through the use of hashtag activism and the #MeToo movement, Twitter users have begun accessing the space for survivor disclosure and to provide social support for acts sexual victimization (Bogen, Bleiweiss, & Orchowski, 2018; Hosterman, Johnson, Stouffer, & Herring, 2018; Maas, McCauley, Bonomi, & Leija, 2018). Emerging research on disclosures to Twitter suggests that individuals may use the platform in an attempt to gain social support not otherwise available to them in-person (Schneider & Carpenter, 2019). Similarly, Andalibi and colleagues (2016) found that survivors also seek out anonymous written disclosure online though development of “throwaway” accounts on the website Reddit. Here, survivors document their victimization experiences through anonymous accounts and subsequently seek social support from other Reddit users. The authors found that those with anonymous accounts were significantly more likely to disclose sexual abuse when compared to those with non- anonymous accounts (Andalibi, Haimson, De Choudhury, & Forte, 2016). Survivors may also use anonymous online hotlines, services that take similar form to phone and texting hotlines. Victims can access the chat service through select advocacy websites and “chat” with an advocate anonymously, in real time. These types of anonymous written disclosure may be beneficial to those who have not previously sought support for their sexual victimization due to the potential barriers of other helping services, including phone hotlines (Finn & Hughes, 2008; Finn et al., 2011). For instance, online hotlines may accommodate those with social anxiety or who find phone calls a more “formal” form of 17 communication (Noble, 2016). In addition, the discretion of anonymous written communication may be particularly important for survivors who are proximally close to their perpetrators (e.g., children living with their abusers). The added level of anonymity appears to be an appealing aspect of these a non-phone hotlines (Andalibi et al., 2016; Gould, Munfakh, Lubell, Kleinman, & Parker, 2002; Noble, 2016) and research has suggested that this level of confidentiality allows survivors to engage in more support seeking when compared to those who are not anonymous in their interactions (Andalibi et al., 2016). Written disclosure (i.e., the process by which individuals write about emotional experiences), like that done via text, message boards, or online hotlines, has been found to have positive effects on a survivor’s emotional well-being, particularly because the writing is shared with others (e.g., other Reddit users or an advocate; Pennebaker, 1997). As such, it is possible that these written services may be able to serve survivors equally as well as more-established forms of anonymous crisis intervention. Model of Helpseeking and Change Up to this point, this literature review has discussed the various choices a survivor has for disclosure: informal, formal, or other, more in-between forms of disclosure, such as hotlines. These descriptions, however, do not capture the complexity of how a survivor chooses a disclosure recipient(s). Disclosure is an iterative process and survivors are faced with multiple decision points when trying to decide if and to whom they will disclose an act of sexual victimization. Next, the process by which survivors may ultimately select one or more sources as a support provider will be reviewed. For those survivors who choose to disclose their experiences of victimization, many decisions are made regarding when and to whom such a disclosure occurs. As discussed, most adult and adolescent victims choose to disclose to friends, while fewer individuals disclose to formal resources and little is known about rates of disclosure to crisis hotlines. According to Liang and colleagues’ (2005) conceptual framework of Helpseeking and Change for victims of intimate partner violence (IPV; see Figure 1), these decisions are likely a non-linear, dialectical process that includes three overarching stages: 1) problem recognition and appraisal, 2) the decision to seek help, and 3) support selection. Each of these stages may be thought of as distinctly influenced by individual-, interpersonal-, and sociocultural-factors to form an 18 interactive, ongoing feedback loop that ultimately results in a survivor’s decision of whether or not to seek help. While the model was developed for IPV victims, it may also be applied to survivors of sexual victimization and their decisions to disclose acts of sexual violence. A description of such application is described below. Figure 1: Model of Helpseeking and Change (Liang et al., 2005) Problem recognition and appraisal. The first step of Liang and colleagues’ (2005) model is that of problem recognition and appraisal. At this stage, victims interpret their situation by defining or labeling their experiences as problematic (or otherwise), and then evaluate the severity of said problem. For instance, an individual who experiences nonconsensual sexual touching may evaluate whether they find the behavior inappropriate, and if so, how inappropriate they thought it to be. This is a crucial stage in the process of disclosure as the way in which a victim defines their experience ultimately dictates the ways they do, or do not, seek assistance for the problem. Individual factors, such as a survivor’s personal knowledge of various types of sexual victimization, may influence their ability to define harmful acts. For example, younger victims (e.g., adolescents or children) may not yet understand how to label non- consensual sexual acts and may struggle to communicate their experiences. Interpersonal factors may also 19 influence problem recognition and definition, such as when others’ opinions reinforce or conflict with the appraisal made by the individual. In the circumstance of a younger victim, an abuser may reinforce the confusion felt on an individual level, by suggesting that the child will get in trouble if they disclose the non-consensual act (NCVC, 2011). Finally, sociocultural factors may influence the ways in which individuals view acts of violence on a broad scale, like how some cultures are more likely to endorse rape myths, potentially causing confusion for victims trying to define the acts committed against them as wrong (Kennedy & Gorzalka, 2002). This stage of problem recognition and definition has been well researched among survivors of sexual victimization, particularly sexual assault. Studies have shown that victims of sexual violence vary in the degree to which they “acknowledge” their experiences as wrong, but overall victims of more “traditional” assaults (e.g., penetrative stranger rape with physical violence), are most likely to label the violence as such (e.g., Anderson, 2007; Bondurant, 2001). Conversely, survivors who were intoxicated during their assaults, had prior consensual romantic involvement with their perpetrator, did not experience physical force, did not verbally or physically resist their perpetrator, or otherwise blame themselves for the victimization, are less likely to label their experiences as sexual assault (Bondurant, 2001; Cleer & Lynn, 2013; Kahn, Jackson, Kully, Badger, & Halvorsen, 2003; Koss, 1985) Decision to seek help. The next stage of Liang and colleagues’ (2005) model is the actual decision to reach out for help. This stage does not include selecting a person or provider to seek help from, but rather is an individual’s determination of whether assistance for a problem is necessary. According to the framework, two individual-level conditions are required for this decision-making process: 1) recognizing the problem as undesirable, and 2) seeing the problem as unlikely to go away without assistance. The first of these requirements is directly related to the ways in which a victim defines their circumstances (i.e., Stage 1 of problem definition and appraisal) – if they do not acknowledge their experience as wrong, they will not decide to seek help. If they do acknowledge the experience as wrong, they may then consider the next prerequisite for disclosure. This second requirement, seeing the problem as unlikely to go away without help from others, is rooted in whether a survivor feels as though they can 20 “handle” their victimization and its effects on their own. Interpersonal and sociocultural factors may influence the decision to seek help, in that victims may weigh others’ experiences of disclosure when making a determination about how they would like to move forward. An example of this would be a survivor hearing about a friend being shamed after a disclosure of sexual assault, and factoring this into their decision-making process. General awareness of people or providers from which one could seek help may also play a role in a victim’s decision. This stage in the model may be particularly complex for survivors of sexual violence. First, at the individual-level, victims may identify the “problem” for which they may or may not seek help in a variety of ways. For some, the problem may be the event of violence itself: if the violence is an ongoing occurrence (e.g., repeat sexual abuse), a victim may be more likely to believe that the sexual assaults will continue unless assistance is obtained, and will therefore be more likely to decide to seek help. However, if the violence is a one-time event, as many acts of sexual victimization are, the “problem” may be defined a bit differently. Here, the victim may not be trying to stop an explicit act of sexual violence, but rather would be making decisions regarding negative post-assault consequences, like physical injuries or PTSD symptoms. In this circumstance, the decision to seek help would be rooted in whether a survivor could identify their symptoms as problematic and if they felt they could handle the effects of the victimization on their own. Research on sexual victimization has shown that particular victim and assault characteristics do influence individuals’ disclosure decisions in this way. For example, survivors who acquire injuries as a result of their assaults are more likely to decide to seek assistance from formal services (Campbell et al., 2001), while adolescent survivors are less likely to decide to seek post-assault care for fear of not being believed or lack of confidentiality (Konradi, 2007; Rickwood et al., 2005; Wilson & Deane, 2001). Support selection. The third and final stage of Liang and colleagues’ (2005) conceptual framework is the selection of a help provider. This involves the actual identification of an informal, formal, or other source of support. At the individual level, this stage involves the victim weighing the costs and benefits of reaching out to potential support sources. Meanwhile, at the interpersonal level, 21 selecting a support provider may include appraising the perceived availability of a person or organization to which the victim may disclose: a survivor looking for emotional support may consider whether any family members or friends may be receptive of a disclosure. Finally, sociocultural factors, like cultural differences in seeking help for “private” matters, may contribute to the selection of a support provider (Kennedy & Gorzalka, 2002). Regarding sexual victimization in particular, survivors’ preference of selecting informal help providers over formal help providers has been well documented in the literature and generally, all victims are more likely to choose family or friends over or before disclosing to more formal sources (e.g., law enforcement or medical providers; Aherns et al., 2007; Kogan, 2004; Stein & Nofzinger, 2008). Beyond this, less is known about how the circumstances of victimization may affect support selection. The current literature could benefit from research on how other assault characteristics (e.g., relationship between the victim and their offender, the frequency of the abuse, whether the victim lives with their perpetrator, the time since the assault, etc.) and relationships between these characteristics are relevant to support selection. The literature also has yet to document support selection at varying points in a survivors’ disclosure history; victims may choose to share their assault experiences with varying informal, formal, or other resources at different points post-assault, but little is known about the factors that may influence these choices. Essentially, there may be many components affecting the process between Stage 2 (decision to seek help) and Stage 3 (support selection) that have yet to be explored. Knowing about such relationships and patterns may help support providers better anticipate the needs of survivors and may identify potential gaps in post-assault support provision. The Current Study To date, the literature on disclosure of and help seeking for sexual victimization has focused on Stage 1 (problem recognition and appraisal) and Stage 2 (decision to seek help). From this body of work, it is known that survivors differ in their “acknowledgement” of sexual victimization and that the decision to pursue assistance for non-consensual sexual experiences varies based on victim and assault characteristics. Yet the current study explored Liang and colleagues’ (2005) Model of Helpseeking and 22 Change to address a gap between Stage 2 of the model (decision to seek help) and Stage 3 (support selection). More specifically, for victims who decide they should seek help (Stage 2), what affects their decisions in choosing a help provider (prior to Stage 3)? Traditionally, help providers have been thought of as either informal or formal sources of support and descriptively the research has shown that individuals are more likely to turn to informal rather than formal sources. However, the lesser-studied resource of anonymous online hotlines may also help with understanding of the challenges survivors face when selecting a source of support post-assault. Anonymous hotlines are a relatively unexplored source of support that do not fit neatly into the typical categorizations of informal or formal resources described in Stage 3 of Liang and colleagues’ (2005) Model of Helpseeking And Change. Revealing an act of sexual victimization to an anonymous hotline may be conceptualized as more of a “soft disclosure,” occurring between Stage 2 (decision to seek help) and Stage 3 (support selection), as this type of disclosure is not flanked with the same obligations as revealing an act of sexual violence to a friend, family member, or formal resource. For instance, confidentiality concerns, like those frequently experienced by adolescents and children, are non-issues when disclosing to an anonymous resource. Wanting to process legal options with a knowledgeable provider may be done through an anonymous hotline without the commitment of a formal report to law enforcement. In addition, anonymous hotlines may serve varying functions for those in different places in their healing trajectories; while some victims using a hotline may be sharing their story for the first time, others may have called upon informal and formal resources in the past. Survivors who have previously disclosed may be using a hotline to process new feelings about their victimization years after it occurred. Regardless of the rationale for using an anonymous hotline, survivors may call upon this “landing pad” after deciding to seek help (Stage 2) but before the phase of support selection (Stage 3) in order to process their experiences in a safe space, without commitment. Importantly, such processing may ultimately assist in the selection of additional, more traditional support providers. The current study addressed these gaps in the literature regarding support selection by exploring this in-between space of anonymous hotlines to gain insight into what occurs between a victim’s silence and their ultimate disclosure. 23 Studies on helpseeking and disclosure typically use a variable-centered approach to explore these research topics as such an approach aggregates data across individuals allowing one to generalize to the greater population (von Eye & Bogat, 2006). Essentially, these methods focus on how variables relate to one another with the goal of understanding how independent variables predict or influence dependent variables (Muthén & Muthén, 2000). In sexual victimization research, a variable-centered study may discern whether certain types of assault characteristics predict the recipient of a victim’s disclosure. An important caveat of this approach, however, is that variable-centered approaches cannot be used to group cases based on inter- and intra-individual differences (Bogat, Levendosky, & von Eye, 2005). The current study focused on whether these inter- and intra-individual differences are an important element of survivors’ disclosure behaviors. Victim characteristics (e.g., victim age) and assault characteristics (e.g., the relationship between the victim and the offender) are likely why victims make the decisions they do regarding support selection. Alternatively, a person-centered approach can account for inter- and intra- individual factors. A person-centered approach focuses on the relationships between cases (rather than variables) with the intention of grouping cases into similar categories (Muthén & Muthén, 2000). This assumes that individuals are unique and that behavior is, in part, person-specific (Bogat et al., 2005). A person-centered approach is particularly well-suited for use when: 1) the sample may be assumed to be drawn from more than one population; 2) attempts may be made to establish external validity of groupings; and 3) the groupings are interpreted based on theory (von Eye & Bogat, 2006). Essentially, this approach can better capture the complexities of a heterogeneous sample, like that of an anonymous online hotline, representing differing types of victimization, genders, or even geographical areas. In addition, this approach allows one to ask questions about underlying groupings of individuals (e.g., latent classes) based on their heterogeneous qualities. The latent classes developed from such person-centered analyses, can be further analyzed by variable-centered approaches. For example, a variable-centered approach can be used to explore similarities and/or differences between established latent classes. While the variable-centered perspective does assume that the inter-individual differences within each latent class are negligible or random, this assumption is appropriate at this stage given that the person-centered 24 approach grouped the cases into a latent class based on their similarities. Therefore, one may draw conclusions across individuals, generalizing to the entire latent class (von Eye & Bogat, 2006). In the current study, both of these approaches were used to address the gaps in the literature regarding victims’ selection of support providers. To accomplish these research goals, the current project used national-level anonymous online hotline data collected as part of an internal evaluation of the National Sexual Assault Online Hotline (NSAOH), RAINN’s anonymous online hotline service. A survivor may find this hotline through a number of mediums as the NSAOH is widely advertised and, at the time, was the advocacy service recommended by whitehouse.gov. Once a visitor had identified the service, they could access the NSAOH through a “need to talk?” or “chat now” button on the RAINN.org homepage (survivors who did not want to use an online hotline could choose to utilize RAINN’s anonymous phone hotline with a number provided next to the chat feature). After clicking the “need to talk?” or “chat now” button, a survivor would be prompted to read a description of service and privacy policy and would be required to check a box saying they agree to the terms (see Appendix A for the Privacy Policy) before starting the anonymous chat. If they agreed to the terms, a user would then enter a chat session with a trained advocate. The advocate would start the session by reiterating the privacy policy and asking if the survivor was in immediate danger. If necessary, the advocate would encourage the user to call 911. If the survivor was not in immediate danger, the chat would continue according to the user’s needs. Typically, this would include providing the survivor with support and/or information/referrals. To collect the data, the staff and volunteer advocates working the online hotline were prompted to complete a brief evaluation form for a sample of the sessions in which they participated (i.e., after the first session of each shift). The first session of the shift was selected regardless of the content of the session and no purposive sampling strategies were in place, thus developing the equivalent of a random sample. All evaluation forms were completely anonymous and only asked volunteers to answer questions about the session they had just completed. Questions were multiple-choice or select-all- that-apply with the option to provide an open-ended response if a response option of “other” was selected. The survey was estimated to take about 5 minutes to complete. No information for the evaluation form 25 was solicited from users of the NSAOH, so only information that was revealed by the user organically could be included in the evaluation form. This dataset is appropriate for various quantitative analyses as it provides enough variation in victim and assault types (i.e., heterogeneity) to identify underlying patterns in the data. In the current study, this dataset was used to answer two main research questions. First, are there underlying groupings of users that choose an anonymous online hotline as a support provider? More specifically, are there latent classes of hotline sessions that differ based on victim and assault characteristics? While the current published literature may speak to characteristics of those victims using other, more typical forms of post- assault support (e.g., victims of more stereotypical assaults, like those committed by strangers with a weapon, are most likely to report to law enforcement; Campbell et al., 2001; Starzynski et al., 2005), little is known about the users of anonymous online hotlines. Second, for these subgroups or classes of survivors who have chosen to use the national anonymous online hotline, are there differences in their disclosure behaviors? In other words, do the revealed latent classes explain or account for variation in disclosure behaviors among victims such that victims choose informal sources, formal sources, neither, or both? For this second research question, revealed latent classes from the first research question are the independent variable, and disclosure behaviors are the dependent variable. Learning more about victims’ use of online hotlines helps further contextualize Liang and colleagues’ (2005) model, by providing additional information about which survivors may require a “landing pad” between Stage 2 (decision to seek help) and Stage 3 (support selection). 26 Sample CHAPTER 2: METHOD The current study sought to understand the groups of users who access online hotlines for experiences of sexual victimization, so the study’s target population was survivors of sexual victimization who utilized an anonymous online hotline. The sampling frame for this target population was drawn from archival data collected from RAINN’s NSAOH. Due to the anonymous nature of the hotline, the unit of analysis for the current study was hotline sessions, as opposed to individual users. This was necessary for two reasons. First, as an anonymous resource, the hotline does not allow determination of whether single individuals have used the hotline multiple times, meaning it cannot be understood whether multiple sessions represent the same event. This raises the possibility that observations may not be independent and that multi-session users may be overrepresented, exerting more influence on the results than is warranted. While the random sampling procedures potentially minimize the impact of this by reducing the probability that multiple sessions from the same individual were sampled, it is important to avoid overestimating individual users and their experiences and therefore the focus is on sessions. In addition, because volunteers were not able to solicit information from the users for the NSAOH evaluation form, there is often incomplete information at the individual level. For example, an individual accessing the hotline to seek support for a recent rape may also have an undisclosed history of child sexual assault. The archival data collected from RAINN’s NSAOH is only able to capture the user’s disclosure of the recent rape, and is unable to document any undisclosed experiences. By defining the level of analysis be a session, the current study focused on how individuals do use a national sexual assault online hotline and not how users do not use the hotline. As such, the current study examined archival data collected on 11,500 NSAOH victim1 sessions that took place between 6/1/2015 and 1/2/2017. Only cases that 1) took place during the study’s data collection period, 2) had a primary event indicated to be a circumstance of 1 Sessions for which the individual did not identify as the victim were removed from the dataset prior to the primary investigator receiving the data. 27 sexual violence (e.g., rape, sexual assault, or sexual exploitation), 3) for which the age category of the victim was known, and 4) was 10 minutes or longer, were included in the final sample (see Figure 2). Figure 2: Sampling criteria While the NSAOH is intended to serve survivors of sexual violence, users may also access the hotline to receive help or support for other types of violence or trauma. Because the focus of the current 28 study is sexual violence and other types of victimization may be considered characteristically different from this focus, sessions for which the primary event was not indicated to be rape, sexual assault, or sexual exploitation were removed from the sample. Therefore, sessions for which the primary event was indicated to be sexual harassment, domestic violence, stalking, cyberstalking, physical assault/abuse, cult violence, or “none of the above/other” were removed from the sample (n=1,080 cases removed). Sessions where the victim did not remember any part of the assault were also removed from the sample as it could not be determined whether the session pertained to sexual violence, or some other type of victimization (n= 31 cases removed). Given that hotline users were not solicited for any information from program staff to complete the evaluation form, many sessions have substantial missing data. The assumptions of many statistical analyses require complete cases and consequently, incomplete data can produce misleading results or complicate the interpretability of findings. As such, a Missing Value Analysis (MVA) was completed in SPSS to identify the variables for which missing data could potentially present a problem. The MVA indicated that the variable “age category” (adult vs. minor) had a pattern of missing data that could become problematic for future analyses such that a large number of cases had missing data on this variable (43% of cases missing age category). Given that the relevant literature suggests age of the victim may play a critical role in support selection (e.g., children are less likely to disclose to informal sources when compared to other groups; London et al., 2003), it was crucial to examine this variable for all cases analyzed. As such, listwise deletion was used to remove all cases that had missing data on age category (n=4,465 cases removed). It is possible that the listwise of deletion of so many cases could have been avoided by reasoning that those with missing data on age category were likely adults (i.e., that if they were minors, the topic would have presented itself during the session). However, this justification has little support from the literature and had the potential to be highly consequential to the results, so imputation was not pursued in this circumstance. Listwise deletion was used to remove any cases where the sessions lasted less than 10 minutes as these sessions also tended to have substantial missing data (n= 6 cases removed). By using these inclusion criteria, there were 5,918 sessions in the final sample. 29 Procedures As is common with most archival data, some data cleaning was necessary to determine project feasibility (e.g., determine sample size). The evaluation form included skip patterns, multiple-choice questions (select-one and select-all-that-apply), closed-ended questions, and open-ended questions. Responses that had previously been left blank as the result of skip-patterns (or inaccurately grouped into the “no/unknown”) response option were re-coded as “N/A” to further contextualize the reason for missing data. Some multiple-choice questions were cleaned to use consistent variable names, labels, and response options. For example, all variables that had dichotomous no/yes response options were re-coded so that 0=no and 1=yes consistently, across each variable. Many multiple-choice questions on the evaluation form included an “other” option; volunteers were instructed to select this option if they were unsure of how to categorize the information provided by the user or none of the available response options accurately captured the information. If “other” was selected, they were then able to type the information in an open-ended response. For the current study, qualitative data produced from this “other” option were reviewed and coded by the primary investigator. For the variable Primary Event, for example, qualitative data stating “un-consensual sexual intercourse with a long-term partner” was coded as “rape,” while “nonconsensual voyeurism” was coded as “sexual exploitation.” Measures For the 5,918 sessions that met the sampling criteria, 26 variables were extracted from the dataset for the current study. Categories of variables included 1) victim and assault characteristics and 2) disclosure and reporting of the primary event (e.g., if and to whom the victim disclosed). For the first research question, only victim and assault characteristics were used to explore whether there are underlying classes of hotline sessions. Revealed latent classes served as the independent variable for the second research question. All disclosure behavior variables served as the dependent variables. Table 1 further describes each of the variables. 30 Table 1: Victim and assault characteristic variables of interest RQ Victim and Assault Characteristics Variable Name Variable Information Victim Age Victim Gender Primary Event Time Since Incident Frequency of Incident Victim-Perpetrator Relationship Living with the perpetrator at time of incident Living with the perpetrator at time of session Incident involving a weapon Incident involving drugs/alcohol Incident related injuries t l u a s s a d n a m i t c i v n o d e s a b r e f f i d t a h t s n o i s s e s e n i l t o h f o s e s s a l c t n e t a l e r e h t e r A : 1 n o i t s e u Q h c r a e s e R ? s c i t s i r e t c a r a h c A categorical variable indicating whether the victim was an adult or minor at the time of the NSAOH session A categorical variable indicating whether the victim identified themselves as female, male, or trans/gender-neutral. A categorical variable indicating the type of victimization for which the user came to the NSAOH to discuss. Only instances of rape, attempted rape, sexual assault (including incest and child sexual assault), multiple perpetrator sexual assault, and sexual exploitation were included in the current study. A categorical variable indicating the between the primary event and the NSAOH session. Options included the same day, 1-7 days ago, within the last month, within the last 6 months, within the last year, 1- 5 years ago, or greater than five years ago. A categorical variable indicating whether the primary event occurred one time, was a repeated event, but no longer occurring, or was a repeated event that was still ongoing. A categorical variable indicating the relationship between the victim and the perpetrator of the primary event. Options included: stranger, known by sight or nickname, friend or acquaintance (e.g., a coworker), family members (e.g., a parent, cousin), past or present intimate partners (e.g., spouse, ex-boyfriend), authority figures (e.g., employer, teacher, police officer), foster family members (e.g., foster parent or foster sibling), a parent’s partner (e.g., mom’s boyfriend), or relationship unknown by the victim. A dichotomous variable indicating whether or not the victim was living with their perpetrator at the time of the primary event. A dichotomous variable indicating whether or not the victim was living with their perpetrator at the time of the NSAOH session A dichotomous variable indicating whether the perpetrator used a weapon during the primary event. A dichotomous variable indicating whether the victim and/or perpetrator were using drugs and/or alcohol at the time of the primary incident. If marked as present, this variable is unable to distinguish which party was using drugs and/or alcohol. A dichotomous variable indicating whether the victim incurred physical injuries as a result of the primary event. 31 Table 2: Disclosure variables of interest RQ Disclosure Variable Name Variable Information First disclosure Disclosure General Informal Disclosure (ID) ID: Family ID: Friend ID: Intimate Partner ID: Other A dichotomous variable indicating whether the victim had ever disclosed the primary event vs. the NSAOH session being the first ever disclosure of the primary event A dichotomous variable (no/yes) indicating whether the victim discussed disclosing to informal or formal systems. A dichotomous variable (no/yes) indicating whether the victim discussed disclosing to informal sources. A dichotomous variable (no/yes) indicating whether the victim discussed disclosing to family (e.g., a parent, sister) A dichotomous variable (no/yes) indicating whether the victim discussed disclosing to a friend or acquaintance (e.g., best friend, classmate, coworker) A dichotomous variable (no/yes) indicating whether the victim discussed disclosing to a past/present intimate partner or spouse (e.g., ex-boyfriend, husband). A dichotomous variable (no/yes) indicating whether the victim discussed disclosing to a different or unspecified informal source (e.g., ‘a witness who saw the assault’). FD: Medical FD: Mental Health FD: Law Enforcement Formal Disclosure (FD) A dichotomous variable (no/yes) indicating whether the victim discussed disclosing to formal sources. A dichotomous variable (no/yes) indicating whether the victim discussed disclosing to law enforcement (e.g., police, 911, campus security). A dichotomous variable (no/yes) indicating whether the victim discussed disclosing to a mental health professional (e.g., counselor, ‘psych ward,’ support group) A dichotomous variable (no/yes) indicating whether the victim discussed disclosing to medical personnel (e.g., emergency room, doctor, nurse) A dichotomous variable (no/yes) indicating whether the victim discussed disclosing to a victim advocate (e.g., RVA, the NSAOH, RCC) A dichotomous variable (no/yes) indicating whether the victim discussed disclosing to a religious figure (e.g., a pastor) A dichotomous variable (no/yes) indicating whether the victim discussed disclosing to a teacher/principal/mentor/coach (e.g., teacher, soccer coach) A dichotomous variable (no/yes) indicating whether the victim discussed disclosing to a different or unspecified formal source (e.g., a lawyer, Title IX Office, Child Protective Services, HR/Corporate). FD: Advocate FD: Religious FD: Teacher FD: Other g n o m a s r o i v a h e b e r u s o l c s i d n i n o i t a i r a v r o f t n u o c c a r o n i a l p x e s e s s a l c t n e t a l d e l a e v e r e h t o D ? h t o b r o , r e h t i e n , s e c r u o s l a m r o f , s e c r u o s l a m r o f n i e s o o h c s m i t c i v : 2 n o i t s e u Q h c r a e s e R t a h t h c u s s m i t c i v 32 Data Analytic Plan To answer the current study’s first research question, are there underlying groupings of users that choose an anonymous online hotline as a support provider, a latent class analysis (LCA) was conducted based on categorical victim and assault characteristic variables. Latent class analysis was used to describe how the probabilities of observed categorical variables vary across unobserved groups (Muthén & Muthén, 2000). The unobserved groups revealed by the LCA are referred to as latent classes and the overall purpose of an LCA is to identify the smallest number of latent classes that adequately describe the data. While the smallest number of classes is the goal, both statistical and substantive considerations are required to determine the final number. The analysis begins with the smallest number of classes possible, adding classes stepwise until good model fit is achieved (Muthén & Muthén, 2000). The resulting model provides posterior probabilities, or estimates of class membership for each case, allowing for interpretation of each class. Absolute fit statistics, relative fit statistics (e.g., Bayesian information criterion statistics), model entropy, and substantive information from the current literature were utilized to select the final number of classes that best describe the data. It should be noted that latent class analysis is particularly suitable for person-centered variables, as the method focuses on the relationships between heterogeneous cases and categorizes them into groups with other, similar cases. This is particularly important for categorical variables, as the analysis is able to distinguish between cases that may be homogenous within a single category of variables, but heterogeneous across categories of variables (Muthén & Muthén, 2000). As such, an LCA was used to identify class membership among hotline sessions, based on the demographics of the victim and the characteristics of the primary event. Upon selection of the most parsimonious latent class model, binary logistic regressions were conducted to answer the current study’s second research question: for sessions representing user who chose to use the national anonymous online hotline, are there differences in their disclosure behaviors? First, preliminary analyses were conducted to assure adequate variance among dependent variables, assess collinearity, and determine which disclosure variables occur frequently enough in the data to be utilized in 33 analyses with the latent classes. Binary logistic regression was then used to examine how the independent variable of class membership related to the dependent categorical disclosure behavior variables (e.g., first disclosure, informal disclosure, formal disclosure). Binary logistic regressions are appropriate when the dependent variables are comprised of two mutually exclusive categories (e.g., dichotomous variables; Petrucci, 2009). To limit the inflation of error, binary logistic regressions were conducted in multiple stages. At stage 1, higher-level disclosure variables were entered into the logistic equation, including First Disclosure, Disclosure General, Informal Disclosure, and Formal Disclosure. At stage 2, all informal disclosure (ID) variables were entered. At stage 3, all formal disclosure (FD) were entered. For all stages, latent classes served as the independent variable via a single categorical variable called class membership, and the disclosure behaviors served as the dependent variables. Results guided additional exploratory analyses, specifically combinations of disclosure variables potentially explained by latent classes. Mplus was used to fit the latent class models and SPSS was used to conduct all binary logistic regressions. 34 CHAPTER 3: RESULTS This study examined whether there are underlying groupings (latent classes) of sessions with users who choose a national anonymous online hotline as a support provider, and whether these groups differ in their disclosure behaviors. To begin this presentation of findings, the characteristics of the sample will be described. Next, the results of the latent class analyses will be presented. Finally, results of binary logistic regressions will be presented, which examined whether latent class membership significantly accounted for variance in disclosure behaviors. Sample Characteristics As shown in Table 3, the final sample consisted of 5,918 anonymous online hotline sessions that were comprised largely of adult (73.7%) and female (56.6%) session users. Most survivors disclosed an experience of rape (65.7%) and 36.1% were assaulted more than a year prior, whereas 10.6% were victimized within 24 hours of the hotline session. Session users were most likely to have been victimized by a family member (30.5%) or someone known to them like a friend, co-worker, or acquaintance (21.6%). The majority were assaulted only one time (41.0%) and were living with their perpetrator at neither the time of the victimization (44.3%) nor at the time of the anonymous online hotline session (67.7%). Less than one-fifth (17.4%) of survivors discussed assault related injuries. A minority of survivors discussed incidents involving a weapon (43.8%) or drugs/alcohol (34.9%). As presented in Table 4, for just over one quarter (27.8%) of the session users, the anonymous online hotline chat was the first time they had disclosed their experience of sexual victimization. The majority (61.7%) of victims had disclosed to someone prior to the hotline session and/or during the chat itself (e.g., contacted law enforcement while using the hotline). Among those who had at least one prior disclosure (n=3,653), the majority (71.8%) had disclosed to at least one informal source of support such as family (35.7%), friends (37.5%), or intimate partners (19.5%). Fifty-nine percent had disclosed to at least one formal source, such as law enforcement (12.0%), mental health professionals (21.6%), or medical professionals (8.3%). In sum, 1,172 session users disclosed to both formal and informal sources 35 of support (46.8% of users who had disclosed to at least one person besides the anonymous online hotline advocate). Table 3: Frequencies & descriptives for all victim and assault characteristics variables n (%) (N= 5,918) 4,359 (73.7) 1,559 (26.3) 3,344 (56.5) 456 (7.7) 41 (0.7) 2,077 (35.1) Variable Victim-Perpetrator Relationship Stranger Known by sight/nickname Friend/coworker/acquaintance Family member Past/present intimate partner Authority figure/Other adult None of the above/Other Living With Perp at Time of Incident 3,890 (65.7) 1,951 (33.0) 77 (1.3) Was not living with perp Was living with perp Not discussed Living With Perp at Time of Session Was not living with perp Was living with perp Not discussed 626 (10.6) 689 (11.6) 458 (7.7) 645 (10.9) Incident Involving a Weapon 589 (10.0) 1,546 (26.1) 1,365 (23.1) No weapons present Weapons present Not discussed Incident Involving Drugs/Alcohol No drugs/alcohol present Drugs/alcohol present Not discussed 2,426 (41.0) 1,784 (30.1) 994 (16.8) 714 (12.1) Incident Related Injuries No injuries present Injuries present Not discussed n (%) (N= 5,918) 252 (4.3) 240 (4.1) 1,277 (21.6) 1,803 (30.5) 713 (12.0) 343 (5.8) 1,290 (21.8) 2,621 (44.3) 1,749 (29.6) 1,548 (26.2) 4,006 (67.7) 905 (15.3) 1,007 (17.0) 2,593 (43.8) 189 (3.2) 3,136 (53.0) 2,068 (34.9) 871 (14.7) 2,979 (50.3) 1,431 (24.2) 1,029 (17.4) 3,458 (58.4) Variable Victim age Adult Minor Victim gender Female Male Trans/gender-neutral Not discussed Primary Event Rape Sexual assault Sexual exploitation Time Since Incident Within 24 hours Within one week Within one month Within one year One to five years Greater than five years Not discussed Frequency of Incident One time Repeated, no longer occurring Repeated, still occurring Not discussed Note. All n’s and percentages refer to number of sessions. 36 Table 4: Frequencies & descriptives for all disclosure variables Variable n (%) (N= 5,918) First Disclosure No, visitor had previously disclosed Yes, the chat was the first disclosure Not discussed General Disclosure to informal and/or formal support providers No Yes 3,551 (60.0) 1,646 (27.8) 721 (12.2) 2,265 (38.3) 3,653 (61.7) Number (proportion) of sessions in which disclosure to helpseeking sources took place prior to or during the hotline session (n= 3,653) Informal Disclosure (ID) ID: Family ID: Friend ID: Past/Present Intimate Partner ID: Other Formal Disclosure (FD) FD: Law Enforcement FD: Mental Health FD: Medical FD: Advocate FD: Religious FD: Teacher/Principal/Mentor/Coach FD: Other Informal AND Formal Disclosure Note. All n’s and percentages refer to number of sessions. Are There Underlying Latent Classes of Hotline Sessions? 2,624 (71.8) 1,303 (35.7) 1,371 (37.5) 714 (19.5) 28 (0.8) 2,154 (59.0) 711 (19.5) 1,278 (35.0) 489 (13.4) 218 (6.0) 56 (1.5) 125 (3.4) 143 (3.9) 1,172 (46.8) Latent class analyses were conducted to answer the study’s first research question, are there underlying groupings of hotline sessions that differ based on victim and assault characteristics? Latent class analysis requires categorical or ordinal data, performs better with a large sample size (at least n=100), and assumes variables are uncorrelated within each class (i.e., conditional, or local, independence). Bivariate residuals were reviewed to assess local independence as residuals greater than 3.84 suggest a failure to meet this assumption (Vermunt & Magidson, 2005). Bivariate residuals were indeed greater than 3.84 on 10.1% of possible variable response option combinations, indicating a relatively small problem. Therefore, the assessment indicated that the assumption of local independence was only partially met. This is still appropriate as observed variables are often correlated for reasons 37 unrelated to the latent class structure and the assumption of conditional independence is often partially relaxed (Masyn, 2013). The method makes no assumptions regarding linearity, distribution, or homogeneity. All analyses were conducted using Mplus (Version 8.2, Muthén & Muthén, 2018) and missing data were handled through the maximum likelihood estimation algorithms integrated into Mplus software. Three variables were identified as having >50% missing data (incident involving a weapon, incident involving drugs/alcohol, incident resulting in injury), an anticipated limitation given that only information revealed organically during the anonymous online hotline session could be recorded via the data collection form. I included these three variables in the initial model, with the intention of exploring how their high rates of missingness affected results. Multiple random sets of starting values (0 0 300 20) were used to assure satisfactory replication of the best maximum log likelihood value. Model formulation and selection. The class enumeration process began by attempting to fit latent class models with all 11 victim and assault characteristic variables. However, when all 11 victim and assault characteristic variables were included, the software could not compute the likelihood ratio (LR) chi-squares because the chi-squares were too large. To combat this issue, I identified and removed three variables that had >50% missing data (incident involving a weapon, incident involving drugs/alcohol, incident resulting in injury). Removing these three variables allowed the latent class analyses to converge and as such, I decided to reduce the number of variables included in the LCA to eight. Continuing the analysis with eight victim and assault characteristic variables, I fit classes, increasing K by one, until the model ceased to be well-identified (Masyn, 2013). Ultimately, I fit six latent class models with K = 1 to K = 6 classes Three main categories of model fit were used to assess and select a final model. First was absolute fit, which compares the model’s representation of the data to the actual data (Masyn, 2013). To assess absolute fit, I examined the LR chi-square goodness-of-fit test; all solutions met this criterion with p=1, suggesting perfect model fit. Next, I reviewed relative fit, which compares the parsimony of two models with equal log likelihoods to assess efficiency of parameters and discourage over-fitting of a model (Schwarz, 1978). 38 The Bayesian Information Criterion (BIC) is a well-established method of determining relative fit, whereby the model with the smallest BIC (e.g., that with the fewest parameters) is identified as the ideal solution (Nylund, Asparouhov, & Muthén, 2007). In the current study, the BIC decreased with each additional class added to the model seemingly regardless of true fit, a problem typical of this criterion. This is due to the fact that the BIC includes an adjustment for sample size that causes standards for adding parameters to the model to increase when the sample is large (Weakliem, 1999). In these circumstances, the BIC may simply suggest that a greater number of classes fit the data better than fewer classes, and should be evaluated in light of other fit indices (Masyn, 2013; Weakliem, 1999). Third, relative entropy assesses the posterior probabilities of revealed latent classes to determine how clearly and precisely classes are delineated across the entire sample (Celeux & Soromenho, 1996). Relative entropy results range from 0-1, where 0 implies the solution is no better than random guessing and 1 suggests perfect classification. Results above 0.80 are considered acceptable (Masyn, 2013). Relative entropy reached its highest point at the 3- and 4-class solutions, suggesting that these models had the most well-defined, delineated classes. Relative entropy began to decrease again at the 5-class solution, indicating that the addition of a fifth class caused the classes to become less well-defined. This criterion therefore suggests pursuing the 3- or 4-class solution, but it should be noted that entropy was well above the acceptable 0.80 rate for all modeled solutions. Finally, the results of the Lo-Mendell-Rubin Adjusted LR test (LMR-LRT) and parametric bootstrapped LR test (BLRT) were assessed. Both provide direction on the number of latent classes to extract by testing the null hypothesis that data are drawn from k-1 latent classes instead of k latent classes. The first model that sees no improvement on LMR-LRT or BLRT p-values with the addition of another class is determined to be the “best” fitting model (Masyn, 2013; Nylund, Asparouhov, & Muthén, 2007). These tests have been rigorously explored in simulation studies (Nylund, Asparouhov, & Muthén, 2007) and significant p-values suggest model fit that is significantly better than a model with one less class. The LMR-LRT produced significant p-values for solutions with 2-, 3-, and 4-classes, suggesting that the 4- class solution as the best fitting model. All BLRT p-values were significant. 39 All fit criteria were considered in the selection of a final solution. It was determined that after K = 5 classes the models ceased to be well identified (see Table 5), however both K = 4 classes and K = 5 classes demonstrated well-defined models. While both classes met the required criteria for the LR chi- square p-value, BIC, and BLRT p-value, relative entropy and the LMR-LRT p-value suggested a 4-class solution. In addition, the 4-class solution represented the more parsimonious model (e.g., fewer free parameters and classes). To assure appropriate model selection, both solutions were interpreted to understand if an additional fifth class contributed meaningful results. Ultimately, the fifth class was not well-defined when compared to other classes in the model; it appeared to extract cases from two different well-defined classes in the 4-class solution, rather than break apart one large class or provide clarity on a vaguely defined class. No clear interpretive value was added by inclusion of this fifth class, so the more- parsimonious, better-fitting 4-class solution was selected as the final model. Table 5: Victim and assault characteristics latent class analysis fit indices Free parameters LR Chi- square p-value BIC Entropy LMR- LRT p-value BLRT p-value 8 variable model* 1 class solution 2 class solution 3 class solution 4 class solution 5 class solution 6 class solution 7 variable model* 1 class solution 2 class solution 3 class solution 4 class solution 5 class solution 6 class solution *The 8-variable model included: 1) victim age, 2) victim gender, 3) primary event, 4) time since incident, 5) frequency of incident, 6) victim-perpetrator relationship, 7) living with perpetrator at time of incident, and 8) living with perpetrator at time of session. The seven-variable model excludes victim gender. 1 106252.267 98438.846 1 92391.402 1 89360.520 1 1 88609.587 88205.333 1 95313.535 0 87565.099 0 81586.637 1 78588.716 1 1 77882.635 77464.651 1 -- 0 0 0 0.788 0.761 -- 0 0 0 0.756 0.967 -- 0.914 0.927 0.926 0.903 0.909 -- 0.916 0.930 0.932 0.906 0.914 25 51 77 103 129 155 22 45 68 91 114 137 -- 0 0 0 0 0 -- 0 0 0 0 0 All variables were assessed for appropriateness of inclusion in the final model. Collins and Lanza (2010) suggest two criteria for determining whether a particular item reliably measures the latent class variable: 1) latent class homogeneity and 2) latent class separation. Latent class homogeneity asserts that 40 a particular response characterizes membership in at least one class, demonstrated by having response probabilities that are <0.30 or >0.70 (see Figure 3). In other words, if a latent class has high homogeneity in regards to a particular item, then that item has a response option that epitomizes the class. To illustrate, Figure 3 depicts latent class homogeneity on the variable “victim-perpetrator relationship (RLTNSHIP)”; each class has a response probability of below 0.30 indicating a particular response option on this variable was meaningful for the definition of these classes. Had the response probabilities been between 0.31-0.69, the item would not be purposefully contributing to the definition of the classes. Conversely, latent class separation posits that a particular item can be used to distinguish members across classes, demonstrated by odds ratios that are <0.20 or >5.0. In other words, if a latent class has high separation in regards to a particular item, that item can differentiate the class from others in the model. For example, the variable “age” in Figure 3 illustrates latent class separation; while one class has a low rate of endorsement (Class 1), the other three classes have a high rate of endorsement (Classes 2, 3, & 4) demonstrating how a particular response option can meaningfully differentiate classes from one another. Only one item, victim gender, failed to meet both criteria; all classes had response probabilities been between 0.31-0.69 and they were not significantly separated (see Figure 3). I ran all analyses both with and without gender and found only marginal improvements in model fit (see Table 5). No changes in interpretation of the findings occurred by removing the variable. Given the importance of gender to the sexual victimization literature and the minimal impact inclusion of the variable had on fit indices, I ultimately decided to keep gender in the model. 41 Figure 3: Model-estimated, class-specific item probability profile-plot for 4-class solution 42 Model interpretation. Results suggested a 4-class solution. Table 6 presents the posterior probabilities of the selected solution, with the most likely response option for each class highlighted in bold. Figure 3 presents a class-specific item probability profile-plot. Items constituting the final model are listed along the x-axis, while the probability of a endorsing an item for sessions within a latent class are shown on the y-axis. Probabilities are reported by class. For example, looking at the item “age,” the profile-plot shows how Class 1 has a low rate of endorsement (which in this circumstance indicates that the class is constituted primarily of minors), and Classes 2, 3, and 4 have high rate of endorsement (indicating the Classes are constituted primarily of adults). When reviewed alongside Table 6, Figure 3 can serve as a visual representation of differences and similarities in Class membership. Moving forward, I will describe the classes in a different order than extracted to aid in interpretability. Class 1 (Minors), with an estimated proportion of 14.8% of all sessions, is primarily characterized by sessions where the victims were minors repeatedly raped by family members within the week (including within 24 hours). In Class 1 (Minors), the rapes were still occurring at the time of the session and victims were living with their perpetrators (both at the time of the assault and at the time of hotline session). Class 2 (Adults Recent; 43.5% of sessions) may be characterized by sessions with adult victims who were assaulted one time by someone known to them like a friend, co-worker, or acquaintance. The victims in Class 2 (Adults Recent) never lived with their perpetrator. Class 4 (Adults Past; 19.6% of sessions), may be characterized by adults who were raped repeatedly by family members five or more years before the anonymous online hotline session. The rapes were no longer occurring at the time of the session (as compared to Class 1, where the repeat offenses were ongoing) and while victims were living with their perpetrators at the time of the assault, they were no longer at the time of the hotline session. Given that many of the assault characteristics of this class were similar to that of Class 1 (Minors), with age of the victim and timing of the assault being critical differences, it is possible that this class represents individuals assaulted as minors, disclosing as adults. Finally, Class 3 (Not Discussed), with an estimated proportion of 22.0% of all sessions, may be characterized by what session users did not discuss during their anonymous online hotline session. These victims were adults who disclosed 43 experiences of rape or sexual assault. While some victims shared that they were victimized five or more years prior to the chat, more victims in this class did not discuss how much time had lapsed between the incident and the session. They discussed few additional assault characteristics. It is instead possible that these victims used their anonymous online hotline session to discuss topics outside the scope of the current study (e.g., barriers to helpseeking, obtaining resource referrals). 44 Table 6: Posterior probabilities of the final 4-class solution Variable Size of class (%) Victim age Adult Minor Victim gender Female Male Trans/gender-neutral Not discussed Primary Event Rape Sexual assault Sexual exploitation Time Since Incident Within 24 hours Within one week Within one month Within one year One to five years Greater than five years Not discussed Frequency of Incident One time Repeated, no longer occurring Repeated, still occurring Not discussed Victim-Perpetrator Relationship Stranger Known by sight/nickname Friend/co-worker/acquaintance Family member Past/present intimate partner Authority figure/Other adult None of the above/Other Living With Perp at Time of Incident Was not living with perp Was living with perp Not discussed Living With Perp at Time of Session Was not living with perp Was living with perp Not discussed Class 1: Minors 878 (14.8) 0.284 0.716 0.608 0.025 0.002 0.365 0.827 0.146 0.027 0.339 0.220 0.098 0.053 0.015 0.016 0.260 0.133 0.028 0.796 0.042 0.007 0.001 0.032 0.682 0.172 0.048 0.058 0.003 0.995 0.002 0.025 0.967 0.008 Class 2: Adults Recent 2,567 (43.5) 0.768 0.232 0.634 0.072 0.008 0.286 0.556 0.437 0.007 0.110 0.173 0.124 0.171 0.117 0.150 0.155 0.719 0.158 0.084 0.039 0.092 0.091 0.437 0.045 0.136 0.090 0.108 0.962 0.026 0.011 0.997 0.001 0.002 Class 3: Not Discussed 1,301 (22.0) 0.843 0.157 0.400 0.093 0.006 0.511 0.511 0.480 0.009 0.035 0.040 0.025 0.082 0.110 0.305 0.404 0.292 0.279 0.046 0.384 0.006 0.004 0.075 0.106 0.073 0.026 0.710 0.009 0.013 0.979 0.332 0.000 0.668 Class 4: Adults Past 1,172 (19.6) 0.890 0.110 0.566 0.108 0.009 0.317 0.914 0.625 0.020 0.000 0.000 0.018 0.047 0.113 0.640 0.181 0.072 0.847 0.016 0.065 0.001 0.000 0.025 0.810 0.101 0.031 0.032 0.116 0.675 0.208 0.847 0.046 0.106 45 Do Latent Classes Explain Variation In Disclosure Behaviors? Binary logistic regressions were conducted to answer the study’s second research question, do the revealed latent classes explain or account for variation in disclosure behaviors among victims? Table 7 shows the number and proportion of sessions within each class for which each dependent variable was endorsed. For example, the number and proportion of sessions in Class 1 (Minors) that involved a first disclosure was n = 444, 58.6%. As an additional example, for the sessions in Class 1 (Minors) in which disclosure to helpseeking sources took place prior to or during the hotline session (n = 3,653), n = 140, 16.0% had disclosed to family. To build the regression models, membership in the identified latent classes was dummy coded and used as independent variables. Dummy coding recodes categorical variables into dichotomous variables so they can be included in a regression model. To illustrate, to create Dummy Variable 1, every session assigned to Class 1 is coded as 1, and all other groups are recoded as 0. For Dummy Variable 2, every session assigned to Class 2 is coded as 1, and all other groups will be recoded as 0 and so on. To perform analyses with dummy variables, k – 1 dummy variables are entered into the regression; the dummy variable left out of the regression model serves as the reference group (e.g., the group to which all other groups are compared). For the current study, hierarchical binary logistic regressions were performed. Initially the study planned for binary logistic regressions, however in an attempt to incorporate the three assault characteristic variables that were removed from the latent class analysis (incident involving a weapon, incident involving drugs/alcohol, incident resulting in injury) into the current study, these variables were integrated into the regression analyses. As such, in Block 1 of the model, binary disclosure variables were regressed on the latent class dummy variables. Incident involving a weapon, incident involving drugs/alcohol, incident resulting in injury were added to Block 2 of the regression. This approach allowed me to understand if these variables explain or account for any variance above and beyond that identified by the latent classes. Each class was tested as the reference group to all possible comparisons between classes could be explored, and Class 1 (Minors) was determined to have the most interpretable findings (see Table 8 for results with Class 1 as the reference group; in-text descriptions include Class 1 as the 46 reference group unless otherwise specified). Significant findings from additional classes as the reference groups are shared in text. All analyses were conducted using SPSS. Table 7: Number and percentage of disclosure variables endorsed by class Class 1: Minors 878 (14.8%) Class 2: Adults Recent 2,567 (43.5%) Class 3: Not Discussed 1,301 (22.0%) Class 4: Adults Past 1,172 (19.6%) n = 3,653 (%) N = 5,918 (%) 444 (58.6) 340 (38.9) 224 (21.6) 832 (64.0) 793 (33.4) 1,627 (63.0) 223 (65.6) 140 (16.0) 82 (24.1) 16 (4.7) 3 (0.9) 157 (42.6) 69 (20.3) 1,288 (79.2) 543 (21.0) 793 (48.7) 338 (20.8) 13 (0.8) 904 (55.6) 378 (23.2) First Disclosure General Disclosure Informal Disclosure ID: Family ID: Friend ID: Intimate Partner ID: Other Formal Disclosure FD: Law Enforcement FD: Mental Health FD: Medical FD: Advocate FD: Religious FD: Teacher FD: Other Informal AND Formal Note. All n’s and percentages refer to the number of sessions within each class. For example, 58.6% (n=444) of sessions in Class 1 (Minors) had a user share that they were disclosing for the first time. 406 (48.8) 95 (11.4) 44 (5.3) 10 (1.2) 24 (2.9) 25 (3.0) 262 (20.1) 456 (28.0) 274 (16.8) 104 (6.4) 18 (1.1) 58 (3.6) 82 (5.0) 578 (22.4) 497 (59.7) 231 (17.8) 249 (29.9) 167 (20.1) 9 (1.1) 580 (69.7) 139 (16.7) 33 (9.7) 40 (11.8) 31 (9.1) 3 (0.9) 21 (6.2) 9 (2.6) 47 (5.4) 185 (18.0) 854 (73.7) 616 (72.1) 389 (33.6) 247 (28.9) 193 (22.6) 3 (0.4) 513 (60.1) 125 (14.6) 383 (44.8) 80 (9.4) 39 (4.6) 25 (2.9) 22 (2.6) 27 (3.2) 285 (24.6) First and general disclosure findings. Results indicated that Class 1 (Minors) was most likely to have disclosed for the first time while on the hotline, when compared to other classes (see Table 8.A). The odds that a session involved an initial disclosure were 64% lower for adults assaulted recently (Class 2), 81% lower for sessions where few assault characteristics were discussed (Class 3), and 85% lower for adults assaulted five or more years ago (Class 4) when compared to sessions with minors (Class 1). Conversely, Class 1 (Minors) was least likely to have disclosed to anyone other than the hotline (see Table 8.B, 8.5C). Class 4 (Adults Past) was the least likely to have disclosed for the first time on the hotline and the most likely to have disclosed to at least one informal or formal source of support when compared to other groups. 47 Informal disclosure findings. Sessions involving minors (Class 1; see Table 8.E) were more likely to include disclosures to family when compared to adults who had been assaulted recently (Class 2) or those who did not discuss assault characteristics during the session (Class 3). Focusing on Class 4 (Adults Past) revealed a similar pattern: adults who had been assaulted in the past (Class 4) were 1.67 (95% CI: .505-.709) times more likely to have disclosed to family when compared to adults who had been assaulted recently (Class 2). They were also 2.17 (95% CI: .375-.563) times more likely to have disclosed to family when compared to those who shared fewer detail about their victimization on the hotline (Class 3). Regarding disclosures to friends, those who discussed few details on the chat were significantly more likely to have disclosed to friends when compared to Class 1 (Minors; see Table 8.F). In addition, sessions involving adults assaulted recently (Class 2) were most likely to have disclosed to friends when compared to other groups. In fact, adults assaulted recently were three times (95% CI: .256-.436) more likely to have disclosed to friends when compared to Class 1 (Minors). They were also more than two times more likely to disclose to friends when compared to those who did not discuss assault characteristics during the session (Class 3; 95% CI: .376-.536) and adults who had been assaulted five or more years ago (Class 4; 95% CI: .359-.511). Sessions involving minors (Class 1) were least likely to have disclosed to a past or present intimate partner when compared to other groups (see Table 8.G). No other groups were more or less likely to have disclosed to intimate partners, regardless of reference group. Formal disclosure findings. Sessions involving minors (Class 1) were more likely to have disclosed to law enforcement only when compared to adults who had been assaulted five or more years ago (Class 4; see Table 8.J). However, when Block 2 was added to the model, Class 1 (Minors) also became significantly more likely to disclose to law enforcement when compared to Class 3 (Not Discussed). Both weapon use (OR= 1.19, 95% CI: 1.050-1.342) and injury (OR=1.13, 95% CI: 1.010- 1.270) were significantly related to this comparison, indicating that only when these variables were accounted for did the additional comparison between Class 1 (Minors) and Class 3 (Not Discussed) 48 become significant. Findings also indicated that sessions involving adults assaulted recently (Class 2) were 1.51 (95% CI: .534-.822) times more likely to have disclosed to law enforcement when compared to sessions where victim and assault characteristics were not discussed (Class 3). In addition, adults assaulted recently were 1.74 (95% CI: .454-.707) times more likely to have disclosed to law enforcement when compared to sessions where and sessions where adults had been assaulted five or more years ago (Class 4). Continuing to focus on Class 2, these sessions involving adults assaulted recently were most likely to have disclosed to medical professionals, when compared to other groups. In fact, adults assaulted recently were about 50% more likely to disclose to medical professionals when compared to adults assaulted five or more years ago (Class 4; 95% CI: .392-.665) and two-thirds more likely to disclose to medical professionals when compared to Class 1 (Minors; 95% CI: .462-.939) or those who didn’t discuss assault details (Class 3; 95% CI: .496-.817). Sessions involving minors (Class 1) were least likely to have disclosed to mental health professionals when compared to other groups (see Table 8.L). Adults assaulted recently (Class 2) were less likely to have disclosed to mental health professionals when compared to sessions where assault characteristics were not discussed (Class 3; OR= 2.447, 95% CI: 2.057-2.912) and adults who had been assaulted five or more years ago (Class 4; OR= 2.088, 95% CI: 1.757-2.482). Conversely, Class 1 (Minors) most likely to have a disclosure to 1) an advocate or 2) a teacher, mentor, or coach, when compared to other groups (see Table 8.M & 7.O). Finally, sessions involving adults who had been assaulted five or more years ago (Class 4) were more than 2.5 times more likely to have disclosed to a religious figure when compared to other groups (Class 1, OR= .295, 95% CI: .089- .984; Class 2, OR= .371, 95% CI: .201-.684; Class 3, OR= .403, 95% CI: .193-.845). Looking specifically at findings related to Block 2 of the model, after accounting for the effect of class membership, results indicated sessions were less likely to include first disclosures if drugs/alcohol (OR=0.89, 95% CI: .820-.985) or injury (OR=0.85, 95% CI: .784-.919) were described. Victims were 12% less likely to have disclosed to family if assault related drugs/alcohol were described during the 49 session (95% CI: .788-.974). Conversely, victims were 1.22 times more likely to disclose to friends if assault related drugs/alcohol were described during the session (95% CI: 1.097-1.362). Victims were 24% less likely to have discussed disclosing to a friend if the described victimization included a weapon (95% CI: .687-.845). Sessions were more likely to include disclosures to formal sources if weapon use, drug/alcohol use, or injury were described. More specifically, the session was more likely to include a description of a disclosure to law enforcement if the assault involved a weapon (OR= 1.19, 95% CI: 1.050-1.342) or resulted in an injury (OR=1.13, 95% CI: 1.010-1.270). Similarly, the session was more likely to include a description of a disclosure to a mental health professional if the assault involved a weapon (OR= 1.17, 95% CI: 1.048-1.301) or resulted in an injury (OR=1.23, 95% CI: 1.118-1.361). Victims were more likely to discuss disclosure to a medical professional if they also described weapon use during their victimization (OR=1.29, 95% CI: 1.121-1.492). Interestingly, the session was 18.9% less likely to describe a disclosure to a medical professional if assault-related injuries were discussed (95% CI: .713- .922). 50 D. Informal Disclosure B 0.690 -0.251 0.306 p .000 .062 .026 OR (CI) 1.99 (1.55-2.57) 0.78 (0.60-1.01) 1.36 (1.04-1.78) B -0.100 0.206 -0.956 B 1.287 2.182 2.024 B 0.669 0.130 0.183 H. Other Informal p .876 .759 .258 OR (CI) 0.90 (0.26-3.19) 1.23 (0.33-4.56) 0.39 (0.80-1.97) L. Mental Health OR (CI) p 3.62 (2.50-5.27) .000 .000 8.87 (6.04-13.01) .000 7.56 (5.15-11.10) P. Other Formal p .061 .741 .639 OR (CI) 1.95 (0.97-3.92) 1.14 (0.53-2.47) 1.20 (0.56-2.58) Table 8: Binary logistic regression results of disclosure regressed on latent class dummy variables Class 2 Class 3 Class 4 Class 2 Class 3 Class 4 Class 2 Class 3 Class 4 Class 2 Class 3 Class 4 B -1.036 -1.634 -1.864 B -0.355 -0.600 0.178 B 0.377 0.987 0.562 B -0.385 -0.586 -0.740 A. First Disclosure p OR (CI) 0.35 (0.30-0.42) 0.19 (0.16-0.24) 0.15 (0.12-0.19) .000 .000 .000 B. General Disclosure (or) B 0.982 1.025 1.481 OR (CI) p .000 2.67 (2.28-3.13) .000 2.79 (2.33-3.32) .000 4.39 (3.64-5.31) C. Informal & Formal (both) OR (CI) p .000 5.07 (3.73-6.90) .000 4.44 (3.21-6.13) .000 5.74 (4.15-7.92) B 1.623 1.490 1.747 p E. Family OR (CI) 0.72 (0.56-0.91) 0.55 (0.42-0.71) 1.19 (0.93-1.54) .006 .000 .170 B 1.096 0.296 0.247 F. Friend OR (CI) p .000 2.99 (2.29-3.91) .045 1.34 (1.00-1.79) .094 1.28 (0.96-1.71) B 1.670 1.626 1.777 G. Intimate Partner OR (CI) p 5.31 (3.17-8.89) .000 5.08 (2.99-8.64) .000 .000 5.91 (3.49-10.01) I. Formal Disclosure p OR (CI) 1.46 (1.15-1.84) 2.68 (2.07-3.48) 1.75 (1.36-2.26) .002 .000 .000 J. Law Enforcement OR (CI) p .240 1.19 (0.89-1.59) .145 0.79 (0.57-1.09) .017 0.67 (0.49-0.93) B 0.173 -0.239 -0.395 B 0.418 -0.034 -0.255 K. Medical OR (CI) p .021 1.52 (1.06-2.16) .866 0.97 (0.65-1.43) .215 0.77 (0.52-1.16) O. Teacher/Mentor/Coach B -0.577 -0.796 -0.912 OR (CI) p .028 0.56 (0.34-0.94) .009 0.45 (0.25-0.82) .003 0.40 (0.22-0.74) p M. Advocate OR (CI) 0.68 (0.45-1.03) 0.56 (0.34-0.89) 0.48 (0.29-0.78) .072 .016 .003 B 0.228 0.312 1.220 N. Religious OR (CI) p .715 1.26 (0.37-4.29) .637 1.37 (0.37-4.99) 3.39 (1.01-11.29) .047 51 Additional Exploratory Regressions Two additional exploratory dependent variables were created based on the identified classes and related literature. Results indicated that sessions involving Class 1 (Minors) and Class 4 (Adults Past) were more likely than other classes to have disclosed to family, while Class 2 (Adults Recent) was more likely than other classes to have disclosed to friends. Disclosures to these informal sources of support (family or friends) sometimes result in additional disclosures to formal resources. More specifically, research has shown how disclosures to law enforcement may be encouraged by family and/or friends and particularly so for minors (Campbell et al., 2015; Stein & Nofzinger, 2008). Given that the identified latent classes were characterized, in part, by age, I created two variables to understand if any identified classes explain or account for variation in disclosure to these informal sources and law enforcement. I created one dichotomous variable where 1 = disclosure to both family AND law enforcement and 0 = disclosure to EITHER family, law enforcement, or neither. I created a second dichotomous variable where 1 = disclosure to both friends AND law enforcement and 0 = disclosure to EITHER friends, law enforcement, or neither. Table 9 shows the number and proportion within each class for whom each of these exploratory dependent variable is endorsed (e.g., for family and law enforcement, the number and proportion of Class 1 who had disclosed to both a family member and law enforcement). Few sessions, however, endorsed each of these variables so results should be reviewed with caution. Sessions involving minors (Class 1) were the least likely to have disclosed to family and law enforcement when compared to all other groups (Class 2, OR= 3.309, 95% CI: 2.049-5.345; Class 3, OR= 1.724, 95% CI: 1.006-2.953; Class 4, OR= 2.937, 95% CI: 1.756-4.910). Class 1 (Minors) were also the least likely to have disclosed to a friend and law enforcement when compared to all other groups (Class 2, OR= 12.742, 95% CI: 4.703-34.518; Class 3, OR= 4.088, 95% CI: 1.413-11.822; Class 4, OR= 6.573, 95% CI: 2.324-18.594). Sessions involving adults who were recently assaulted (Class 2) were 2-3 times more likely to have disclosed to friends and law enforcement when compared to all other groups (Class 3, OR= .321, 95% CI: .207-.497; Class 4, OR= .516, 95% CI: .353-.755). 52 Class 3: Not Discussed 1,301 (22.0%) Class 4: Adults Past 1,172 (19.6%) Table 9: Number and percentage of exploratory dependent variables endorsed by class Class 1: Minors 878 (14.8%) Class 2: Adults Recent 2,567 (43.5%) 19 (2.2) Family AND Law Enforcement Friend AND Law Enforcement Note. All n’s and percentages refer to the number of sessions within each class. For example, 2.2% (n=19) of sessions in Class 1 (Minors) had a user discuss disclosure to family and law enforcement. 143 (5.5) 177 (6.8) 24 (1.8) 48 (3.7) 4 (0.5) 71 (6.1) 24 (2.9) 53 CHAPTER 4: DISCUSSION Anonymous hotlines for survivors of sexual victimization have not been extensively studied, but emerging data suggest that this service is a valuable resource for survivors’ disclosure and helpseeking, and particularly so for survivors who have not previously sought support (Finn & Hughes, 2008; Finn et al., 2011). Hotlines may also be helpful to those who have previously disclosed part of their on-going recovery efforts but, the varied and many ways survivors use hotlines has not been thoroughly explored. Understanding who uses anonymous hotlines, what types of abuse those users have experienced, and how hotlines might help survivors transition to formal support or provide formal support after seeking formal assistance, are key issues still to be explored in this developing literature. The purpose of the current study was to explore these topics within the context of Liang and colleagues’ (2005) Model of Helpseeking and Change. Liang and colleagues (2005) describe helpseeking as an iterative three-stage process, whereby a victim first needs to recognize a situation as problematic (Stage 1), decide external help is necessary for said problem (Stage 2), and select a support provider (Stage 3; see Figure 4). This model focuses on “the individual help-seeker’s internal, cognitive processes,” (Liang et al., 2005, pg. 73). However, if the ultimate goal of the process is to select an appropriate support provider with the intention of receiving help, there are limitations to this model as many individuals do not know about the variety of available support providers or how to gain access. The current study sought to address this limitation by exploring how anonymous online hotlines are a part of the helpseeking process as described by Liang and colleagues’ (2005). While initially one may think of the hotline as fitting neatly into the support selection phase (Stage 3) of the framework, it may be more complicated than this. It is possible that anonymous online hotlines function between Stages 1 and 2 (see Figure 4), as some survivors may utilize the hotline to confirm their experience as problematic or wrong. However, the current study cannot speak to this part of the process, as the data did not include information on survivors’ rape acknowledgement. Results based on disclosure behavior data instead suggest amending the model to reflect that some survivors call upon services, in this case 54 anonymous hotlines, between Stages 2 and 3 (see Figure 4), as they provide an interim space for survivors to process and decide which support provider may be best-suited for their needs. This interim space may also help them decide if additional sources of support are even necessary. How the current study’s findings support this argument will now be explored as it relates to each of the four identified latent classes. Importantly, latent classes are developed probabilistically therefore, not all categorizations perfectly reflect all sessions grouped into a class (e.g., Class 1, Minors contains some sessions where the victims are adults). Consequently, the following conclusions offer suggestions of how these aggregated groups may call-upon an anonymous online hotline as a support provider. Figure 4: Highlighted Model of Helpseeking and Change (Liang et al., 2005) 55 Key Findings and Contributions to Helpseeking Theory & Research Class 1 (Minors). The first group identified by the current study’s latent class analysis consisted of sessions where children and adolescents discussed being repeatedly raped by family members. For these children and adolescents, the repeat rapes were still occurring when they accessed the hotline and they were typically living with their perpetrators (at both the time of the assault and the time of the hotline session). The findings of this study suggest three possible variations on survivors’ pathways to helpseeking as outlined in Liang et al.’s Model of Helpseeking and change. Figure 5 highlights how the sessions in this class illustrate three distinct uses of hotlines for child/adolescent survivors. Figure 5: Amended Model of Helpseeking and Change for Class 1 (Minors) Figure 5.A depicts a first way in which those represented by sessions in Class 1 (Minors) may utilize an anonymous online hotline in their helpseeking journey: individuals decide they need help for their victimization (Stage 2) but before proceeding to selection of a help provider (Stage 3), they access the anonymous online hotline. Results suggest this pathway may be used to circumvent traditional barriers to disclosure and helpseeking that minors experience. Hierarchical binary logistic regressions revealed that Class 1 (Minors) was the class least likely to have disclosed to informal or formal sources of 56 support, meaning they had yet to proceed to Stage 3 (selection of a support provider) of the model when they accessed the hotline. They were instead the class most likely to be disclosing for the first time while on the anonymous online hotline session. These children and adolescents had never discussed the ongoing, repeat rapes before – the first time they were discussing their victimization was with the advocate during the anonymous online hotline. These findings are consistent with prior literature suggesting minors are less likely to disclose to informal or formal networks when compared to adults (Jacques-Tiura et al., 2010; Kilpatrick et al., 2003; London et al., 2007; Lonsway & Archambault, 2012; Rickert et al., 2005). These low rates of disclosure may be attributed to access issues that differentially affect minors, including lack of knowledge surrounding available resources or the inability to utilize them independently (Holland & Cortina, 2017; Crisma, Bascelli, Paci, & Romito, 2004; Goodman-Brown, Edelstein, Goodman, Jones, & Gordon, 2003; London et al., 2003; Malloy, Brubacher, & Lamb, 2011; McElvaney, 2008). Young survivors may not know how to disclose their abuse safely and finding opportunities to do so may be overwhelming. However, the anonymous online hotline studied in the current project is publicized in spaces congruent with this age group’s internet usage (e.g., Twitter, Tumblr), allowing minors to learn about the resource organically. While the current study cannot confirm that this is how the minors represented in the sessions learned about the hotline, research has shown that youth typically receive their sexual health information through the internet, rather than formal sources (Ralph, Berglas, Schwartz, & Brindis, 2011). Additionally, an anonymous online hotline provides an opportunity for minors to discuss victimization privately; other sources of support cannot offer this type of confidentiality and even other types of anonymous resources present risks (e.g., young victims living with their perpetrators fear being overheard on phone hotlines). Anonymous online hotlines allow children and adolescents access to a helper with whom they may begin to process next steps in a safe, confidential way. Research has also established that minors are hesitant to disclose in fear of causing trouble, upsetting their family or risk more harm from their perpetrator, particularly if their perpetrator is a family member (Crisma, Bascelli, 57 Paci, & Romito, 2004; Goodman-Brown, Edelstein, Goodman, Jones, & Gordon, 2003; London et al., 2003; Malloy, Brubacher, & Lamb, 2011; McElvaney, 2008). Given that Class 1 (Minors) was characterized by sessions where the victim was experiencing repeat, on-going rapes by a family member at the time of the session, these fears may be quite salient for victims in this group. In this study, the individuals represented by Class 1 (Minors) may have had these fears and utilized the hotline to reaching out to traditional sources of support (see Figure 5.A) as a way to process these dynamics in a judgement- free space and without the heavy commitment of sharing their victimization with those close to them before they feel confident doing so. Here, they can test the waters of disclosure with a safe and confidential resource, and consider what type of response their helpseeking would receive before selecting a source of support (Stage 3; see Figure 5.A). Figure 5.B illustrates a different way Class 1 (Minors) may integrate the anonymous online hotline into their helpseeking: first, they decide they need help for the abuse (Stage 2), then select a support provider (Stage 3), but then turn to the hotline for additional support. Hierarchical binary logistic regressions examining the relationship between class membership and disclosure indicated that when those represented by Class 1 (Minors) did discuss disclosure (reached Step 3), they were likely to have discussed disclosing to family members or a teacher, mentor, principal, or coach. Minors’ disclosure to family is a finding well represented in the prior literature (particularly their mother; Broman-Fulks et al., 2007; Hershkowitz et al., 2007; London et al., 2003; Malloy, Brubacher, & Lamb, 2013). While disclosure to these latter sources is not validated by past work in the same way, minors likely have the greatest access to teachers, principals, or coaches. Therefore, these findings highlight that some child and adolescent survivors turn to the anonymous online hotline after their first attempts at disclosure to these sources support did not fully meet their needs. This class was characterized by abuse that was ongoing at the time of the session, showing that for some, prior attempts at helpseeking did not result in an end to the victimization. While the current study cannot confirm the motivation for accessing the anonymous online hotline after previous disclosure to informal or formal sources, it is possible that these minors sought out 58 the resource when their first attempts at disclosure were unsuccessful. Figure 5.B highlights how, ideally, utilization of the hotline would then evolve into the selection of an additional helpers for the minor after the hotline session was over. The third panel (C) of Figure 5 integrates a new facet into the model; a double-headed arrow to shows how this group may utilize the anonymous online hotline multiple times before feeling comfortable reaching out to informal or formal sources of support. In addition to disclosure to family and teachers, principals, or coaches, regression results indicated that when minors do disclose, they were more likely than other groups to have disclosed to an advocate. This result may seem contrary to well-documented findings asserting that minors infrequently disclose to formal sources of support (e.g., advocates), but in the current study, advocates include those staffing the current or different anonymous crisis hotlines. Therefore, this finding suggests that those with sessions in Class 1 (Minors) may be calling upon the resource repeatedly to process their options or prepare for other, additional disclosures. Overall, the anonymous online hotline may present a relatively-barrier free space for children and adolescents to seek confidential support. Class 2 (Adults Recent). The second latent class (Adults Recent) consisted of sessions where adult victims disclosed being sexually assaulted or raped by someone known to them, such as a friend, co- worker, or acquaintance. These victims were assaulted one time by the offender and never lived with their perpetrator. Figure 6 highlights how adults assaulted recently (Class 2) may have decided they needed support (Stage 2), selected a friend or formal network as a support provider (Stage 3), but then determined they required additional help (Stage 2, again). In this scenario, these survivors then turned to the anonymous online hotline for further assistance in processing their victimization (see Figure 6). 59 Figure 6: Amended Model of Helpseeking and Change for Class 2 (Adults Recent) This figure reflects the results of hierarchical binary logistic regressions, which revealed that the hotline is typically not Class 2’s first disclosure, but instead an interaction that occurs after seeking support from a number of other more-traditional resources. In fact, Class 2 (Adults Recent) were the group most likely to have disclosed to a friend, which is in accordance with past research showing that almost all (96%) adult survivors disclose to an informal helping source, with most turning to friends (Ahrens, Campbell, Ternier-Thames, Wasco, & Sefl, 2007; Fehler-Cabral & Campbell, 2013; Ullman, 1996, 2010; Ullman & Filipas, 2001). They were also likely to have discussed disclosure to medical professionals and law enforcement, which is consistent with literature suggesting that individuals assaulted as adults are more likely to disclose to these formal sources when compared to those assaulted as minors (Casey & Nurius, 2006; Jacques-Tiura et al., 2010; Kilpatrick et al., 2003; Lonsway & Archambault, 2012). The results of this study do not speak to why this group chose to call upon the anonymous online hotline after previously disclosing to these other sources of support, however prior literature on survivors’ experiences with these informal and formal helpers may provide some insight. Numerous studies have shown that survivors of sexual victimization experience victim-blaming behaviors after an assault. While friends have been found to provide the most supportive and healing reactions to informal disclosure (Ahrens, Cabral, & Abeling, 2009; Filipas & Ullman, 2001), this is not uniformly true (Ahrens & Aldana, 2012; DePrince et al., 2017). Additionally, the majority of survivors who disclose to medical professionals and law enforcement do not have positive experiences; instead, these interactions leave survivors feeling worse than they had prior to the disclosure (Campbell, 2005; Campbell & Raja, 60 2005; Konradi, 2007). For many victims of sexual assault and rape, these reactions are traumatizing and confusing. Survivors are not only left to deal with the impact of the victimization itself, but also the secondary victimization imparted by their disclosure recipients. For Class 2 (Adults Recent), this may be where the anonymous online hotline becomes important; it is possible that these survivors seek out anonymous trained advocates for a second chance at helpseeking when their first attempts were harmful or otherwise unsatisfactory (see Figure 6). Survivors who have recently disclosed to multiple traditional forms of support may be less like to utilize the hotline to become connected with additional traditional resources. Overall, the current study suggests that for these survivors, the anonymous online hotline serves as an opportunity for additional processing of their victimization in a safe, supportive space. Class 4 (Adults Past). The next latent class identified, Class 4 (Adults Past), consisted of sessions where an adult disclosed being raped five or more years prior to the session. Similarities between this class and that of Class 1 (Minors), with critical differences being age of the victim and timing of the assault, suggest that this class represents adults processing childhood victimization. It would be unsurprising that such a class emerged, as research has indicated that when survivors of child sexual assault do disclose, it is often more than five years after the assault with many disclosing well into adulthood (Alaggia et al., 2017; Easton, 2012). As portrayed in Figure 7, survivors in these sessions initially decide to seek help (Stage 2) and select traditional support providers (Stage 3). After this selection of traditional helpers, however, those represented by sessions in Class 4 (Adults Past) seem to decide they need to seek further help (Stage 2, again) and call upon the anonymous online hotline. In ideal circumstances, hotline providers would then be able to connect these survivors with sustainable, long- term helping resources. 61 Figure 7: Amended Model of Helpseeking and Change for Class 4 (Adults Past) Hierarchical binary logistic regressions examining the relationship between Class 4 (Adults Past) membership and disclosure indicated that sessions involving adults who were assaulted five or more years prior were the group most likely to have disclosed to at least one informal or formal source of support, particularly mental health professionals, and/or religious figures. For those sessions represented by Class 4 (Adults Past), utilization of the hotline after prior disclosures to other support providers (Stage 3) may represent the desire to process a victimization over time. The limited research on anonymous hotlines has indicated that the resource may be used for similar reasons and that some survivors will utilize the services to process the long-term effects of victimization (Townsend & Campbell, 2018). While survivors in Class 4 (Adults Past) have disclosed to a variety of other sources, healing from trauma is not a linear process and the adults represented by sessions in this group may feel as though they require additional support across the lifespan. While the current study cannot assert survivors’ rationale, prior literature on adult disclosure of childhood or adolescent victimization would suggest that these survivors may have also experienced secondary victimization preventing them from feeling safe disclosing to new informal or formal support networks (Berliner & Barbieri, 1984; Campbell, 2008; Feeney et al., 2018; Shaw, Campbell, Cain, & Feeney, 2017). Instead of returning to their previous support providers, these victims called upon the anonymous online hotline to discuss their past rapes in a safe, non-judgmental environment. For these survivors, disclosure to the hotline may or may not result in the selection of additional support providers (Stage 3). 62 Class 3 (Not Discussed). The final latent class represents sessions where users disclosed few victim or assault characteristics (Class 3). These victims were primarily adults who were sexually assaulted or raped, but few additional details were disclosed during the session. Hierarchical binary logistic regressions indicated that his class did not have any particularly salient disclosure experiences relative to the other classes. Instead, this class is defined by the absence of discussion regarding topics and experiences raised in other class’ sessions. As such, it is unclear exactly what these sessions suggest for the processes articulated in the Model of Helpseeking and Change, but Figure 8 depicts three processes suggested by the literature that may explain the experiences of this class. Figure 8: Amended Model of Helpseeking and Change for Class 3 (Not Discussed) In the first of these, Figure 8.A, survivors may decide they need help (Stage 2), but conclude that an anonymous hotline may be more in line with their helpseeking goals than traditional forms of support (Stage 3). Given that these survivors chose to share few details about their assault experiences with the anonymous hotline, more-traditional forms of helpseeking may have felt too intrusive. Instead, these survivors may have benefited from the safe, anonymous, external space (the hotline) to discuss other 63 topics with a trained advocate. The victims represented by sessions in this class may also have decided that they need help (Stage 2), but do not know how to access traditional support providers (Stage 3) and call on the hotline in an attempt help bridge this gap (see Figure 8.B). Survivors requesting specific information about traditional sources of support would not necessarily need to disclose details specific to their victimization experiences to gain the information they were seeking. Prior literature has shown that victims utilize hotline resources not only to process their victimization, but also to learn about coping mechanisms or local support resources (Macy et al., 2009). It is possible that the individuals represented by sessions in this class utilized the online hotline for informational purposes, which do not necessarily require disclosure of specific victim or assault characteristics. Finally, these survivors may have had experiences similar to that of other classes, where they sought traditional forms for help, but felt unsatisfied with the support received, and therefore reached out to the hotline (see Figure 8.C). While the characteristics of their assaults may have been similar to those represented by other classes, survivors in Class 4 (Not Discussed) may not have felt the need to share these details to receive the support they were seeking, or may not have felt comfortable doing so. Hotlines offer all survivors a space free of judgement to talk about whatever they need to discuss. This also means that these survivors are not “required” to describe specifics of an assault or prior forms of helpseeking if they are not comfortable or do not feel the need to do so, in order to utilize the service. Overall model. Overall, anonymous online hotlines are a relatively unexplored source of support that do not fit neatly into the typical categorizations of informal or formal resources described by Liang and colleagues’ (2005) Model of Helpseeking and Change. The findings of the current study support the conceptualization of disclosure to an anonymous hotline as a “soft disclosure” occurring between Stage 2 (decision to seek help) and Stage 3 (support selection) highlight the utility of an amended Model of Helpseeking and Change for survivors of sexual victimization (see Figure 9). Incorporating anonymous online hotlines as an external resource further contextualizes Liang and colleagues’ (2005) framework, insofar that it contributes information about which survivors and which circumstances may require an 64 additional “landing pad” between Stage 2 (decision to seek help) and Stage 3 (support selection). These online hotlines may serve varying functions for those in different places in their healing trajectories; while some victims using a hotline do so to share their story for the first time (e.g., Class 1, Minors), others may use the hotline in addition to traditional forms of support or to process new feelings about their victimization years after it occurred (e.g., Class 2, Adults Recent, Class 4, Adults Past, and Class 3, Not Discussed). Figure 9: Amended Model of Helpseeking and Change 65 Context, Limitations, and Implications for Future Research It is important to acknowledge both the context in which this study was conducted, as well as several methodological limitations that may affect the strength of the conclusions that can be drawn from this work. These limitations also suggest important lines of inquiry for future research and each will be discussed. To begin, a critical challenge of working with anonymous resources is the extent of missingness in the data. Missing data are often an issue because, as was the case in the current study, only information revealed organically by the session user can be captured by advocates and analyzed. While this allows for an understanding of what survivors choose to share and how they utilize anonymous online hotlines, the approach results in incomplete information regarding each victim’s experience. For instance, the absence of discussion of injury does not necessarily mean that the victim was injury-free, only that they did not discuss injuries during their anonymous online hotline session. As such, the current study, and others using this type of data, cannot develop estimates of types of victim or assault experiences (e.g., cannot estimate the number of assaults resulting in injury), the number of unique individuals who access this resource (e.g., repeat hotline use cannot be detected), or rates of use among varying populations. The current study, for example, was unable to explore further, differential use of the hotline by victim gender given the extensive missing data on this variable; while most session users disclosed being female (56.5%), few disclosed being male (7.7%) or gender non-binary (0.7%) and most did not discuss their gender (35.1%). Results indicated that the gender variable did not meaningfully contribute to the latent class analysis, however this may have been an effect of those who did not disclose their gender. More complete data would have allowed this analysis to run separately for each group, providing better insight into how gender impacts each of latent classes. The type of missing data that accompanies an anonymous resource also creates limitations for understanding how anonymous online hotlines support diverse communities because victim characteristics like race/ethnicity, disability status, and socioeconomic status cannot be known unless explicitly disclosed by the user. Disclosure is a complex process that is impacted by each of these identities, as well as other interpersonal and sociocultural factors that the current study 66 could not take into account due to the ways in which the data were collected (Chen & Ullman, 2014; Long & Ullman, 2013; Starzynski et al., 2007). Researchers may choose to address the challenge of extensive missing data in one of three ways. First, those studying hotlines may continue to work within the limitations of the data by letting all survivors guide all conversations without interjecting questions specific to data collection. The current study worked with data under these conditions and context, as it was the practice of RAINN’s hotline not to ask specific questions of users for the purposes of research/evaluation. Therefore, researchers then must address the problems of missing data through quantitative analysis strategies compatible with missingness (e.g., missing variable analysis, listwise deletion, bootstrapping) and/or by re-conceptualizing their research questions to account for lack of data. To work within the context of the missingness problem in the current study, the unit of analyses was the hotline session (as opposed to session user) because this allowed the research to focus how individuals do use a national sexual assault online hotline, rather than centering how users do not use the hotline. Future research may consider similar methods to addressing large amounts of missing data with anonymous online hotlines. Some researchers, however, may wish to take a different approach to the problem of missing data by systematically asking survivors questions for research purposes during the anonymous session. Doing so would minimize missing data challenges for select variables. Prior evaluations of anonymous phone hotlines have used this strategy, and in one such project, Wasco and colleagues (2004) had advocates staffing a crisis line obtain oral consent before asking a series of evaluation questions at the very end of a crisis call. Wasco and colleagues (2004) noted that the approach was effective in that it allowed for more complete record keeping and their questions were considered minimally invasive, particularly because they were asked at the end of the session. Those hoping to take a similar approach should take great care to continue to protect the confidentiality of session users and put the needs of the victim before the goals of data collection (Gondolf et al., 1997; Riger et al., 2002). 67 Not all approaches to addressing missing data must occur via the hotline though. Those hoping to study this resource could also contemplate a third approach via gathering retrospective data about victims’ use of anonymous online hotlines. In interview-based studies with survivors regarding their helpseeking and disclosure experiences, researchers could incorporate questions specifically regarding hotline usage, which would help us understand which survivors call-upon anonymous online hotlines and what types of information they choose to share with the resource. This approach it not without its own limitations though, as those interviewed would represent a sub-sample of anonymous hotline users who eventually disclose to additional resources and ultimately decide to participate in a research project. Taken together, these three strategies for addressing the problem of missing data can help improve future research on hotline utilization. Even when missing data is appropriately accounted for, there are additional limitations to using anonymous online hotline data. The results of the current study do not generalize to other types of advocacy services as the characteristics that draw individuals to an anonymous online hotline are likely specific to this resource. The way victims can access the hotline without barriers and with complete confidentiality (e.g., without being overheard) is likely a reason why survivors choose an anonymous online hotline. Further inquiry is necessary to determine what motivates survivors to use anonymous online hotlines instead of or in conjunction with other more traditional forms of support. Exploration into whether other types of anonymous disclosure networks (e.g., anonymous online chatrooms, anonymous phone hotlines) hold a similar space in the amended Model of Helpseeking and Change would benefit modern interpretations of the framework. Finally, the current study is admittedly exploratory and multiple hierarchical binary logistic regressions were run because of its exploratory nature. This strategy was intentionally used to be able to provide practical conclusions regarding possible relationships between the identified classes and all disclosure variables (e.g., minors disclosing to teachers, adults disclosing to law enforcement, or adults assaulted as minors disclosing to mental health personnel). However, because some dependent variables 68 are likely related to one another (e.g., victims may have disclosed to both medical personnel and law enforcement), the findings are not completely independent. This is problematic because the failure to meet assumptions of complete independence mean that some observations could influence or affect the value of others, overemphasizing the relationships between some variables as well as inflating Type I error. To avoid issues of independence in future work on disclosure in relationship to hotline use, research should pursue scaling disclosure data to create composite variables of informal disclosure and formal disclosure. Future studies should also consider extending research on anonymous online hotlines to include the advocate perspective. Exploring how advocates experience the anonymous online hotline sessions and how effective they perceive the sessions to be by latent class, could inform areas for growth moving forward. Identifying classes of sessions advocates feel less prepared to support would ultimately allow for more effectively targeted trainings. Implications for Practice Despite these limitations, the current study offers practical guidance for RAINN and other similar large-scale anonymous online hotline providers (e.g., State of MI Sexual Assault Voices 4 Hotline). First, at the organizational level, findings emphasize of the importance of these large-scale advocacy organizations making strong connections to local resources. Results identified a substantial group of sessions where young victims were disclosing experiences of ongoing sexual victimization for the first time. These survivors were in acute crisis situations during their anonymous online hotline sessions, which means that advocates had an opportunity to affect the trajectory of the victims’ helpseeking. As shown by the current study, anonymous online hotlines may serve as a “landing pad” for many survivors, bridging the gap between silence and disclosure to more tangible sources of support. As such, is important that large, national-scale organizations, such as RAINN, or statewide hotline providers have strong ties to a variety of local resource agencies. Findings also highlighted how session users not currently experiencing ongoing abuse may still need warm hand-offs to local helping sources. Prior 69 literature details the ways in which the reactions of a disclosure recipient can differentially impact a survivor’s healing process (Ullman, & Peter‐Hagene, 2014), and large-scale anonymous hotline services need partnerships with local resources that can facilitate long-term healing. Once organizational structures are in place to link survivors to local resources, hotline staff and volunteers must be trained about on the most effective ways create these connections. Results of the current study suggest that victim and assault characteristics described during the session should inform the helpseeking recommendations advocates provide (e.g., those in acute crisis should receive different suggestions than those processing long-term trauma). While most anonymous hotline trainings already attend to the differences between imminent threat to safety and past sexual abuse, it is crucial that these nuances are carried through when helping victims select informal and formal support providers. This may be particularly important for minors as results of the current study indicate that some minors are experiencing ongoing victimization at the time of their hotline session. Minors disclosing ongoing sexual abuse for the first time may require special care: both the concerns and barriers to reporting this age group may face, as well as the challenge of discussing sensitive topics with children and adolescents, must be tended to. As such, advocates staffing the anonymous online hotline must be trained in developmentally appropriate crisis intervention to effectively guide young individuals to safety. One way hotlines could train their staff and volunteers in these areas is through case study vignettes. Hotline trainings often use vignettes to help volunteers practice their advocacy skills (CCASA, 2014). The latent classes documented by this study provide sample scenarios that could be used to inspire this portion of trainings. Each class represents real, aggregated session content that advocates may encounter during their hotline shifts and advocates can practice appropriate responses and understand common dynamics of each group. Advocates should review these practice sessions in light of other helping sources these groups may have utilized in conjunction with the hotline. The results of this study also invite reflection on how advocates should balance the role of supportive listener versus an active agent in helpseeking. For example, it is important for trainings to explore the differing approaches needed 70 to support individuals who have previously disclosed to multiple informal and formal networks over their healing trajectory (e.g., adults assaulted as minors) versus those who are disclosing for the first time during the session (e.g., minors). Beyond the hotline itself, the current study’s findings also provide suggestions for how large- scale advocacy organizations could be targeting their outreach and education efforts. The current study documented rates of disclosure to specific informal and formal sources of support, which can help guide outreach and educational programming. For example, 6.2% of Class 1 (Minors) disclosed to a teacher, principal, mentor, or coach, 23% of Class 2 (Adults Recent) disclosed to law enforcement, and almost 45% of Class 4 (Adults Past) disclosed to mental health professionals. These findings point to how RAINN and other community partners can maximize their outreach efforts by training these recipient groups on how to respond to disclosures and to prepare them for their role in victims’ helpseeking. RAINN offers the country’s only national-level online sexual assault hotline, and the results of this project, particularly those regarding first-time disclosures, underscore the importance this kind of outreach educational programming. Conclusion Given high rates of sexual victimization among children, adolescents, and adults, it is critical that research and practice work to identify ways to bridge the gap between a victim’s silence and their disclosure. For diverse sets of abuse victims (e.g., varying ages, contexts), anonymous online sexual assault hotlines are an instrumental resource that can tend to survivors’ immediate needs and build bridges to sustainable, long-term support networks. 71 APPENDIX 72 RAINN’s Online Privacy Policy & Information on “What We’ve Done to Make it Safe” This Online Privacy Policy (the “Policy”) is intended to describe how the Rape, Abuse & Incest National Network ("RAINN", “we” or “us” ) collects, uses and discloses information gathered at www.rainn.org ("RAINN.org") and the National Sexual Assault Online Hotline (the "Online Hotline" and, together with RAINN.org, the “Sites” or each a “Site”), which can be reached through RAINN.org, https://online.rainn.org, or other related sites of RAINN or its partners. By visiting the Sites, you agree to follow the terms and conditions of this Policy. RAINN may change these terms and conditions from time to time without prior notification, and such changes will be effective upon posting, so please consult this Policy each time you visit a Site. If you do not agree with it, you should not use the Sites. Your privacy and security are very important to us. This Policy will help you understand what information we collect via the Sites, and how we use and disclose that information once we collect it. This Policy will tell you if we disclose that information to anyone, and what choices you have regarding how we use that information. Please note that this Policy applies only to the Sites, and not to the websites of any other organizations to which we may link or that may link to us. If you have any questions or suggestions about this Policy, please contact RAINN at 202.544.1034 or info@rainn.org. Alternatively, you can mail RAINN at: Privacy Coordinator RAINN 1220 L St. NW Suite 505 Washington, DC 20005 1. Information You Provide to Us "Personally Identifiable Information" is information that could reasonably be used to identify you, such as your name or address. General: In general, visitors may browse RAINN.org without providing any Personally Identifiable Information. However, there are certain pages on RAINN.org, such as donation pages and the Prevention Navigator, where we may request Personally Identifiable Information about you, including your name, address, phone number, e-mail address, or credit card number. RAINN may seek such information in order to register you as a supporter of RAINN, to collect your rating of a third-party service, to enable your making of a donation or purchasing merchandise, or to use certain parts of the RAINN.org. Finally, we may use your e-mail address to send you updates or newsletters apprising you of recent events and developments relating to RAINN and its mission, or receipts for your charitable donations. Other than as set forth in this Policy, RAINN will not sell, trade or share a donor's Personally Identifiable Information with anyone else, nor will we send donor mailings on behalf of another organization, unless you have given us explicit permission to do so; this Policy applies to all donations made by mail, via our Site, via employee withholding, or any other means. 73 Registration: In order to use some of the features of RAINN.org, you will be asked to register with RAINN. The information you submit as part of the registration process is used to update and maintain our files. We collect and save such information so that you are not required to submit it every time you use a feature. When you register with us, you will be asked to provide your e-mail address and select a password. You will also be asked to provide certain other Personally Identifiable Information so that we may readily identify you. Our Sites are designed to process and maintain accurately the information that you share with us. You are invited to access our Sites 24 hours a day, 7 days a week, to change your profile information, and are encouraged to provide us with feedback. Online Hotline: We operate the Online Hotline using a combination of trained staff, volunteers and contractors (collectively, “Trained Staff”). Our Trained Staff includes both individuals who provide you with one-to-one help in sessions, and hotline supervisors. At times, the Rape Crisis Center for Children & Adults ("RCCCA") also provides hotline supervisors to help monitor online sessions. Any information that you share during an Online Hotline session will be viewed by the individual assisting you, and may also be viewed by one or more supervisors for quality control purposes. Online Hotline sessions may also be viewed by other RAINN staff. THE ONLINE HOTLINE IS INTENDED TO BE AN ANONYMOUS SERVICE AND YOU WILL NOT BE ASKED TO PROVIDE YOUR NAME, ADDRESS, TELEPHONE NUMBER, E-MAIL ADDRESS, OR ANY OTHER PERSONALLY IDENTIFIABLE INFORMATION AT ANY TIME DURING YOUR USE OF THE ONLINE HOTLINE. In fact, RAINN specifically requests, for your own safety and anonymity, that you do not provide any such Personally Identifiable Information when you use the Online Hotline. In addition, the Online Hotline does not capture any Personally Identifiable Information about you or your computer, before or during your use of the services. The information conveyed by you to our Trained Staff during your Online Hotline session is intended to be part of a private conversation between you and the applicable Trained Staff. This conversation will NOT be recorded, stored, or saved. During this conversation you are not required to provide any information that you do not feel comfortable with. We use non- identifying information to improve our services and training based on user needs. We may share aggregate information (e.g., patterns and trends of topics discussed) with academic researchers and the public, to help them understand the needs of sexual assault survivors. Further, in order to determine usage levels and average session length, and for other recordkeeping purposes, an entry will be made in a database containing the start and stop time of each session. The RCCCA may also collect information for its own recordkeeping purposes. Such information will not identify you personally. Please note, however, that some states require us by law to record and report information regarding the abuse of someone less than 18 years of age, or vulnerable or disabled adults, including persons over 60 years of age, or persons who are in danger of committing suicide. If you convey such circumstances, and provide information that could lead us to identify your name and location, we may have to record such information and provide it to the appropriate authorities. 74 If you provide feedback through an anonymous feedback survey following your session, that feedback may be shared with the Trained Staff assisting you, and with other personnel at RAINN. However, the feedback is voluntary and anonymous. We will not ask for your name or any Personally Identifiable Information, or attach any feedback to information that you provide during your Online Hotline session. 2. Information We Collect Through Technology In addition to the above, RAINN.org may also collect information, such as your computer's IP address, through technical means. (An IP address is often associated with the portal through which you enter the Internet.) We automatically gather, or aggregate, certain usage information, such as the number and frequency of visitors to our Sites, length of stay, and other anonymous data, the topics discussed and areas of importance to users. This data helps us determine usage levels on different parts of our Sites, and other useful statistics. We may use such information to analyze site usage for research, internal recordkeeping and service improvement purposes. In addition, we may aggregate such data and disclose it in a non-personally identifiable manner to donors, sponsors, supporters, and other third parties, to generate support for RAINN, its mission and its services, or as required by law. In these situations, we do not disclose any information that could be used to identify you personally. Cookies and How We Use Them In general, a "cookie" is a small piece of text data stored on a user's computer by a website, in order to give that computer a unique identity to the website while performing certain processes, such as filling an online shopping cart and checking out, or calculating the number of unique visitors at a given time. Cookies may contain personal information, like email addresses, usernames, or passwords, or completely anonymous information, like a randomly generated number. They may be deleted once a user leaves the site, leaves a section of the site, or closes the browser; or they may remain, so that the website "remembers" the user when he or she next visits. The Online Hotline does NOT place any permanent or "persistent" cookies on your computer. Although RAINN.org does use session cookies, as explained below), the cookies placed by RAINN.org cannot be used to identify you as a user of the Online Hotline. RAINN.org uses two session ID cookies. The first is used to measure the amount of overall traffic on our website; we use this data only in the aggregate, meaning no individual user can be identified from the data. This cookie is in no way linked to Personally Identifiable Information, and is deleted from the user's computer when the user closes his or her browser. The second session ID cookie is only used when a user is making a donation, purchasing merchandise, completing a form, or posting a review. This cookie carries a user's form entries from one page to another, and is deleted once the user leaves the relevant section of RAINN.org (unless the user explicitly chooses to remember it for future visits). We do not use permanent cookies, which retain information past a user's current visit. 3. How We Use Personally Identifiable Information We do not publish, sell, share, or rent information about you except as set forth in this Policy. We, our Trained Staff and the RCCCA, may be required to disclose information to the 75 government or third parties under certain circumstances. Such circumstances include (and you hereby authorize us to) disclosure of any information about you that we possess to law enforcement or other government officials as we, in our sole discretion, believe is necessary or appropriate, in connection with the investigation of a crime, mandatory reporting laws, fraud, intellectual property infringement, or other activity that is illegal, or may expose us, or you, to liability. Also, if we, or a member of our Trained Staff or RCCCA, fear that you are in danger of committing suicide, or are under 13 years old or over 60 years old, we, or the Trained Staff or RCCCA member, may be required to disclose whatever information you have provided to the appropriate authorities. Trained Staff are located across the United States, and may not be licensed in your state or trained in your state's law. The laws of the Trained Staff’s state may apply to your communications with that Trained Staff. The Trained Staff’s state laws may not protect your communications to the same degree, or in the same way, as the laws of your state. For instance, your state might recognize a privilege to communications with the Online Hotline — meaning that the state may allow you to keep such communications private — but the staff’s state might not recognize such a privilege. It is possible that a court of law will not keep private, and may require us to disclose, information that you provide during your Online Hotline session. If you are concerned about this possibility, please use the telephone hotline instead, at 1.800.656.HOPE (4673); the Trained Staff or partner agency who answers your phone call will be able to provide you with further information regarding the privileged communication laws in your state. Other Uses: We may use third party service providers to facilitate our services, including companies and individuals who perform functions on our behalf, such as website hosting, data analysis, contacting members of Congress, and providing customer service. We may provide your Personally Identifiable Information to such parties, however they will only have access to such Personally Identifiable Information needed to perform their functions; such parties may not use it for any other purpose and are obligated to protect the confidentiality of such Personally Identifiable Information. We use Google Analytics, a web analytics service provided by Google, Inc. ("Google"), which uses cookies and tracking code to help analyze how visitors use our Site. The information collected (including what browser and operating system you are using, your IP address, the pages on our Site that you visit, your number of page views, and the time you spend on our Site) is transmitted to and stored by Google. Google uses this information for the purpose of evaluating visitors' use of our Site in an aggregate manner, compiling reports on website activity for website operators and providing other services relating to website activity and Internet usage. Google may transfer this information to third parties where required to do so by law, or to third parties that process information on Google's behalf. You can read Google's privacy policy at http://www.google.com/intl/en/privacy.html 4. How We Protect Your Security RAINN stores all information on a secure server that is only accessible by RAINN employees or contractors, pursuant to the terms described above. Again, we do not store transcripts of communications over the Online Hotline. 76 RAINN employs a variety of security measures and techniques to ensure that all personal information is protected from unauthorized access, both online and offline, including the use of Secure Sockets Layer (SSL) encryption. SSL technology establishes an encrypted communication pipeline between your device and our servers to ensure the privacy, integrity, and security of information being exchanged. All pages on our website are made available through the use of SSL certificates. For technical details, click the lock icon in the address bar of your browser. You should be aware, however, that "perfect security" does not exist anywhere on the Internet. ALTHOUGH WE BELIEVE WE TAKE APPROPRIATE MEASURES TO SAFEGUARD AGAINST UNAUTHORIZED DISCLOSURES OF INFORMATION, WE CANNOT ASSURE YOU THAT YOUR PERSONALLY IDENTIFIABLE INFORMATION OR COMMUNICATIONS WITH THE ONLINE HOTLINE WILL NEVER BE DISCLOSED IN A MANNER INCONSISTENT WITH THIS POLICY, AND MAKE NO REPRESENTATIONS OR WARRANTIES REGARDING THE SUFFICIENCY OF OUR SECURITY MEASURES. Always be careful and responsible regarding your personal information. We are not responsible for, and cannot control, the use by others of any information which you provide or divulge to others, purposefully or inadvertently, including your e-mail address and password for RAINN.org or the Online Hotline. For more information on how to keep your computer safe, please read our User Safety Procedures. While we may list and/or link to third-party websites, including non-profit and for-profit prevention programs and other services on our Sites, we do not endorse, support, represent or guarantee the completeness, truthfulness, accuracy, or reliability of any content or communications posted on our Sites or endorse any opinions expressed via reviews on our Sites. Instead, any program data, descriptions of prevention courses, video excerpts, assertions about its effectiveness, and client lists or ratings are provided for you for evaluation purposes only when you make your own decisions about the identity and suitability of others whom you contact or interact with. We do not monitor or have any control over, and make no claim or representation regarding, third-party websites. You understand that by using our Sites, you may be exposed to content that might be offensive, harmful, inaccurate or otherwise inappropriate, or in some cases, postings that have been mislabeled or are otherwise deceptive. You acknowledge sole responsibility for and assume all risk arising from your use of any such websites or resources. Other Information Collectors: Except as otherwise expressly included in this Policy, this document addresses the collection, use and disclosure of information. If you disclose your information to others, different rules may apply to their use or disclosure of such information. We do not control the privacy policies of others. We encourage you to ask questions before you disclose your personal information to others. 5. Limitations of Liability IN NO EVENT SHALL RAINN OR ITS AFFILIATES AND THEIR RESPECTIVE PARTNERS, EMPLOYEES, VOLUNTEERS, DONORS, AFFILIATES AND AGENTS (COLLECTIVELY “RELATED PERSONS”) BE LIABLE FOR ANY CLAIMS, LIABILITIES, LOSSES, COSTS, DAMAGES OR INJURIES, INCLUDING, BUT NOT LIMITED TO, ANY 77 DIRECT, INDIRECT, PUNITIVE, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL (“DAMAGES”) ARISING OUT OF OR IN CONNECTION WITH ANY ACCESS, USE (OR INABILITY TO USE) OR DISTRIBUTION OF OUR SITES AND THEIR CONTENTS, OR A SITE LINKED TO OUR SITES, INCLUDING ANY DAMAGES OR INJURY CAUSED BY ANY FAILURE OF PERFORMANCE, ERROR, OMISSION, INTERRUPTION, DELETION, DEFECT, DELAY IN OPERATION OR TRANSMISSION, COMPUTER/DEVICE VIRUS, COMMUNICATION LINE FAILURE, OTHER COMPUTER/DEVICE MALFUNCTION, THEFT, DESTRUCTION OR UNAUTHORIZED ACCESS TO, ALTERATION OF OR USE OF ANY ASSET, WHETHER FOR BREACH OF CONTRACT, TORTIOUS BEHAVIOR, NEGLIGENCE OR UNDER ANY OTHER CAUSE OF ACTION, OR ANY CONTENT OBTAINED THROUGH USE OF THE SITE. THIS IS TRUE EVEN IF RAINN OR RELATED PERSONS HAVE BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES OR LOSSES. Disclaimer: The material on our Sites may include technical inaccuracies or typographical errors. We may shut down our Sites and end these Terms of Use at any time. However, in the event of such a termination, the Limitations of Liability, Disclaimer, and Governing Law sections would continue to apply. OUR SITES ARE PROVIDED BY RAINN ON AN "AS IS" AND "AS AVAILABLE" BASIS. RAINN MAKES NO REPRESENTATIONS OR WARRANTIES OF ANY KIND, EXPRESS OR IMPLIED, AS TO THE OPERATION OF OUR SITES OR THE INFORMATION INCLUDED ON OUR SITES INCLUDING ANY REPRESENTATIONS OR WARRANTIES WITH RESPECT TO MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, TITLE, NON-INFRINGEMENT, AVAILABILITY, SECURITY, ACCURACY, FREEDOM FROM VIRUSES OR MALWARE, COMPLETENESS, TIMELINESS, FUNCTIONALITY, RELIABILITY, SEQUENCING OR SPEED OF DELIVERY. YOU EXPRESSLY AGREE THAT YOUR USE OF OUR SITES IS AT YOUR SOLE RISK. INFORMATION AND ADVICE PROVIDED THROUGH OUR SITES IS NOT INTENDED AND SHALL NOT CONSTITUTE PROFESSIONAL COUNSELING. ONLINE SESSIONS MAY BE CONDUCTED BY STAFFERS (INCLUDING VOLUNTEERS) WHO ARE NOT PROFESSIONALS AND MAY NOT BE AUTHORIZED TO PROVIDE COUNSELING SERVICES IN CERTAIN STATES. ONLINE SESSIONS ARE NOT A SUBSTITUTE FOR PROFESSIONAL COUNSELING SERVICES. THE INFORMATION CONTAINED ON OUR SITES IS NOT INTENDED OR IMPLIED TO CONSTITUTE MEDICAL ADVICE, DIAGNOSIS OR TREATMENT. IN ADDITION, NOTHING ON OUR SITES IS INTENDED TO CONSTITUTE LEGAL ADVICE. If you need professional services or counseling, please call the National Sexual Assault Hotline at 800.656.4673 to discuss how to locate a legally-certified professional. 6. How to Review, Update or Delete Your Information and Opt-Out Procedures When you register at RAINN.org, we may send you e-mails or newsletters, to update you on events and developments pertinent to RAINN and our charitable goals. If you do not want to receive such e-mails, please email our Privacy Coordinator at info@rainn.org or at the address listed above and we will remove you from our lists. Note that donation and/or volunteer 78 information will remain in RAINN's databases. Please contact our Privacy Coordinator to obtain, change or update our records of your personal information. 7. Our Policy Towards Children We are concerned about the safety and privacy of children who use the Internet. Consistent with the Children's Online Privacy Protection Act of 1998, we will never knowingly request Personally Identifiable Information from anyone under the age of 13 without prior verifiable parental consent. If we become aware that a child under 13 has provided us with Personally Identifiable Information through the Sites without verifiable parental consent, we will use our best efforts to remove such information from our files. If a parent or guardian becomes aware that his or her child has provided us with Personally Identifiable Information through the Sites without verifiable parental consent, he or she should contact our Privacy Coordinator at info@rainn.orgor at the address listed above. 8. Do Not Track RAINN does not track its visitors over time and across third party websites to provide targeted advertising and therefore does not respond to Do Not Track (“DNT”) signals. However, some third party sites do keep track of your browsing activities when they serve you content, which enables them to tailor what they present to you. If you are visiting such sites, your browser allows you to set the DNT signal so that third parties (particularly advertisers) know you do not want to be tracked. You should consult the help pages of your browser to learn how to set your preferences so that websites do not track you. 9. International Use RAINN makes no claims that materials or services on our Sites are appropriate or may be downloaded for use in locations outside the United States. Access to the Sites from countries or territories where such access is illegal is prohibited. Furthermore, [our databases are located in the United States]. If you access this website from outside the United States, you do so at your own risk. By sending us your data, you consent to its transfer to and storage within the United States. Those who access our Sites from outside the United States do so on their own initiative and are responsible for compliance with local laws, rules and regulations. 10. Governing Law You agree that your use of our Sites, this Policy and any disputes relating thereto shall be governed in all respects by the laws of the District of Columbia. Any dispute relating to this Agreement shall be resolved solely in the courts located in the District of Columbia. You agree to waive trial by jury in any such action. 79 80 81 REFERENCES 82 REFERENCES abuse. Archives of Pediatrics & Adolescent Medicine, 150(8), 850–857. American Journal of Community Psychology, 38(3-4), 263-274. disclosure on survivors' relationships with friends, family, and partners. Journal of Trauma & Dissociation, 13(2), 226-243. interpretations of social reactions from support providers. Psychology of Women Quarterly, 33(1), 81-94. Adams, J. A., Girardin, B., & Faugno, D. (2001). Adolescent sexual assault: Documentation of acute injuries using photo-colposcopy. Journal of Pediatric and Adolescent Gynecology, 14(4), 175 -180. Adams, J. A., & Knudson, S. (1996). Genital findings in adolescent girls referred for suspected sexual Ahrens, C. E. (2006). Being silenced: The impact of negative social reactions on the disclosure of rape. Ahrens, C. E., & Aldana, E. (2012). The ties that bind: understanding the impact of sexual assault Ahrens, C. E., Cabral, G., & Abeling, S. (2009). Healing or hurtful: Sexual assault survivors' Ahrens, C. E., Campbell, R., Ternier-Thames, N. K., Wasco, S. M., & Sefl, T. (2007). Deciding whom to Alaggia, R., Collin-Vézina, D., & Lateef, R. (2017). Facilitators and barriers to child sexual abuse (CSA) Allnock, D., & Miller, P. (2013). No one noticed, no one heard: A study of disclosures of childhood Andalibi, N., Haimson, O. L., De Choudhury, M., & Forte, A. (2016, May). Understanding social media disclosures of sexual abuse through the lenses of support seeking and anonymity. In Proceedings of the 2016 CHI Conference on Human Factors in Computing Systems (pp. 3906-3918). San Jose, CA. Anderson, I. (2007). What is a typical rape? Effects of victim and participant gender in female and male Baker, R. B., & Sommers, M. S. (2008). Relationship of genital injuries and age in adolescent and young Barrera, M., Calderón, L., & Bell, V. (2013). The cognitive impact of sexual abuse and PTSD in children: Basile, K. C., Black, M. C., Simon, T. R., Arias, I., Brener, N. D., & Saltzman, L. E. (2006). The association between self-reported lifetime history of forced sexual intercourse and recent health risk behaviors: Findings from the 2003 National Youth Risk Behavior Survey. Journal of Adolescent Health, 39(5), 752-e1. adult rape survivors. Journal of Obstetric, Gynecologic, & Neonatal Nursing, 37(3), 282–289. tell: Expectations and outcomes of rape survivors' first disclosures. Psychology of Women Quarterly, 31(1), 38-49. disclosures: A research update (2000–2016). Trauma, Violence, & Abuse. Advance online publication. abuse. London, England: National Society for the Prevention of Cruelty to Children. a neuropsychological study. Journal of Child Sexual Abuse, 22(6), 625-638. rape perception. British Journal of Social Psychology, 46, 225-245. 83 Social Issues, 40(2), 125-137. factors. Violence Against Women, 7(3), 294-314. in college women with a mental health and/or behavior disability. Journal of Women's Health, 27(3), 359-368. violence: Person‐oriented and variable‐oriented perspectives. American Journal of Community Psychology, 36(1-2), 49-70. to disclosures of sexual victimization on twitter. Psychology of Violence, 9(1), 127-137. Stevens, M.R. (2011). The National Intimate Partner and Sexual Violence Survey (NISVS): 2010 Summary Report. Atlanta, GA: National Center for Injury Prevention and Control, Centers for Disease Control and Prevention. Berliner, L., & Barbieri, M. K. (1984). The testimony of the child victim of sexual assault. Journal of Black, M.C., Basile, K.C., Breiding, M.J., Smith, S.G., Walters, M.L., Merrick, M.T., Chen, J., & Bogat, G. A., Levendosky, A. A., & Eye, A. V. (2005). The future of research on intimate partner Bogen, K. W., Bleiweiss, K., & Orchowski, L. M. (2018). Sexual violence is# NotOkay: Social reactions Bondurant, B. (2001). University women's acknowledgment of rape: Individual, situational, and social Bonomi, A., Nichols, E., Kammes, R., & Green, T. (2018). Sexual violence and intimate partner violence Breiding, M. J. (2014). Prevalence and characteristics of sexual violence, stalking, and intimate partner violence victimization—National Intimate Partner and Sexual Violence Survey, United States, 2011. Morbidity and mortality weekly report. Surveillance summaries (Washington, DC: 2002), 63(8). Broman-Fulks, J. J., Ruggiero, K. J., Hanson, R. F., Smith, D. W., Resnick, H. S., Kilpatrick, D. G., & Campbell, R. (2005). What really happened? A validation study of rape survivors’ help‐seeking Campbell, R. (2008). The psychological impact of rape victims’ experiences with the legal, medical, and Campbell, R., Dworkin, E., & Cabral, G. (2009). An ecological model of the impact of sexual assault on Campbell, R., Greeson, M. R., Fehler-Cabral, G., & Kennedy, A. C. (2015). Pathways to help: Adolescent Campbell, R., Patterson, D., Dworkin, E., & Digel, R. (2010). Anogenital injuries in childhood sexual Saunders, B. E. (2007). Sexual assault disclosure in relation to adolescent mental health: Results from the National Survey of Adolescents. Journal of Clinical Child and Adolescent Psychology, 36(2), 260-266. mental health systems. American Psychologist, 68, 702‐717. women's mental health. Trauma, Violence, & Abuse, 10(3), 225-246. sexual assault victims’ disclosure and help-seeking experiences. Violence Against Women, 21, 824-847. abuse victims treated in a pediatric Forensic Nurse Examiner (FNE) program. Journal of Forensic Nursing, 6, 188-195. experiences with the legal and medical systems. Violence & Victims, 20, 55‐68. 84 Helpseeking experiences in military and civilian social systems. Psychology of Women Quarterly, 29, 97‐106. Renzetti, C. M., Edleson, J. L., & Kennedy Bergen, R. (Eds.) Sourcebook on Violence Against Women, Second Edition (pp. 95-109). Sage Publications. rape”: rape survivors' experiences with community service providers. Journal of Interpersonal Violence, 16(12), 1239-1259. Campbell, R., & Raja, S. (2005). The sexual assault and secondary victimization of female veterans: Campbell, R., & Townsend, S. M. (2011). Defining the scope of sexual violence against women. In Campbell, R., Wasco, S. M., Ahrens, C. E., Sefl, T., & Barnes, H. E. (2001). Preventing the “Second Casey, E. A., & Nurius, P. S. (2006). Trends in the prevalence and characteristics of sexual violence: A Celeux, G., & Soromenho, G. (1996). An entropy criterion for assessing the number of clusters in a Chen, Y., & Ullman, S.E. (2014). Women’s reporting of physical assaults to police: A brief report. Child Welfare Information Gateway. (2016). Definitions of Child Abuse and Neglect. Retrieved from: Cleere, C., & Lynn, S. J. (2013). Acknowledged versus unacknowledged sexual assault among college https://www.childwelfare.gov/topics/systemwide/laws-policies/statutes/define/ Journal of Aggression, Maltreatment and Trauma, 23, 854-868. mixture model. Journal of Classification, 13(2), 195-212. cohort analysis. Violence and Victims, 21(5), 629-644. women. Journal of Interpersonal Violence, 28(12), 2593-2611. training guide. Denver, CO: CCASA. social, behavioral, and health sciences (Vol. 718). John Wiley & Sons. Clements, P. T., Speck, P. M., Crane, P. A., & Faulkner, M. J. (2004). Issues and dynamics of sexually assaulted adolescents and their families. International Journal of Mental Health Nursing, 13(4), 267-274. Collins, L. M., & Lanza, S. T. (2010). Latent class and latent transition analysis: With applications in the Colorado Coalition Against Sexual Assault (CCASA) (2014). Sexual assault advocacy & crisis line Crisma, M., Bascelli, E., Paci, D., & Romito, P. (2004). Adolescents who experienced sexual abuse: Daignault, I.V. & Hebert, M. (2009). Profiles of school adaptation: Social, behavioral, and academic Danielson, C. K., and Holmes, M. M. (2004). Adolescent sexual assault: an update of the literature. DePrince, J.D., Dmitrieva, J., Gagnon, K.L., Labus, J., Srinivas, T., & Wright, N. (2017). Expanding use of the social reactions questionnaire among diverse women: Summary overview. Washington, DC: National Institute of Justice. Retrieved from https://www.ncjrs.gov/pdffiles1/nij/grants/251459.pdf fears, needs and impediments to disclosure. Child Abuse & Neglect, 28(10), 1035-1048. functioning in sexually abused girls. Child Abuse & Neglect, 33, 102-115. Current Opinion in Obstetrics and Gynecology, 16(5), 383-388. 85 psychopathology: A review and meta-analysis. Clinical Psychology Review, 56, 65-81. sources. Violence and Victims, 16(6), 673. and Mexico: An Executive Summary of the U.S. National Study. Retrieved from http://www.gemsgirls.org/Estes%20Wiener%202001.pdf Untested Sexual Assault Kits and victim notification of rape survivors assaulted as adolescents. Victims & Offenders, 13(5), 651-674. Journal, 41(4), 344-355. patients. Obstetrics & Gynecology, 97(4), S53. Long-term consequences of childhood sexual abuse by gender of victim. American Journal of Preventive Medicine, 28(5), 430-438. Dube, S. R., Anda, R. F., Whitfield, C. L., Brown, D. W., Felitti, V. J., Dong, M., & Giles, W. H. (2005). Dworkin, E. R., Menon, S. V., Bystrynski, J., & Allen, N. E. (2017). Sexual assault victimization and Easton, S. D. (2013). Disclosure of child sexual abuse among adult male survivors. Clinical Social Work Eckert, L., Sugar, N., & Fine, D. (2001). Correlates of physical injury in sexual assault Estes, R, & Weiner, N. A. (2012). The Commercial Sexual Exploitation of Children in the U. S., Canada Feeney, H., Campbell, R., & Cain, D. (2018). Do you wish to prosecute the person who assaulted you?: Fehler-Cabral, G., & Campbell, R. (2013). Adolescent sexual assault disclosure: The impact of peers, Finkelhor, D. & Shattuck, A. (2012). Characteristics of Crimes Against Juveniles. Durham, NH: Crimes Finkelhor, D., Shattuck, A., Turner, H. A., & Hamby, S. L. (2014). The lifetime prevalence of child Filipas, H. H., & Ullman, S. E. (2001). Social reactions to sexual assault victims from various support Filipas, H. H., & Ullman, S. E. (2006). Child sexual abuse, coping responses, self-blame, posttraumatic stress disorder, and adult sexual revictimization. Journal of Interpersonal Violence, 21(5), 652 -672. Finkelhor, D., & Wolak, J. (2003). Reporting assaults against juveniles to the police: Barriers and Finn, J., Garner, M. D., & Wilson, J. (2011). Volunteer and user evaluation of the national sexual assault Finn, J., & Hughes, P. (2008). Evaluation of the RAINN national sexual assault online hotline. Journal of Fischer, S., Stojek, M., & Hartzell, E. (2010). Effects of multiple forms of childhood abuse and adult sexual abuse and sexual assault assessed in late adolescence. Journal of Adolescent Health, 55(3), 329-333. catalysts. Journal of Interpersonal Violence, 18(2), 103-128. online hotline. Evaluation and Program Planning, 34(3), 266-272. Technology in Human Services, 26(2-4), 203-222. families, and schools. American Journal of Community Psychology, 52(1-2), 73-83. against Children Research Center. sexual assault on current eating disorder symptoms. Eating Behaviors, 11(3), 190-192. 86 model of children’s disclosure of sexual abuse. Child Abuse & Neglect, 27(5), 525–40. U.S. Department of Justice, Office of Justice Programs, Bureau of Justice Statistics. internet during adolescence. Journal of the American Academy of Child & Adolescent Psychiatry, 41(10), 1182-1189. relationship to the perpetrator, and posttraumatic distress in female crime victims. Journal of Traumatic Stress, 19(6), 813-823. Washington, DC: U.S. Department of Justice, Office of Justice Programs, National Institute of Justice. Fisher, B. A., Cullen, F. T., & Turner, M. G. (2000). The Sexual Victimization of College Women. Goodman-Brown, T., Edelstein, R., Goodman, G., Jones, D. and Gordon, D. (2003). Why children tell: A Gould, M. S., Munfakh, J. L. H., Lubell, K., Kleinman, M., & Parker, S. (2002). Seeking help from the Grant, J. M., Mottet, L. A., Tanis, J., Harrison, J., Herman, J. L., & Keisling, M. (2011). Injustice at Every Turn: A Report of the National Transgender Discrimination Survey. Washington, DC: National Center for Transgender Equality and National Gay and Lesbian Task Force. Retrieved from: http://www.thetaskforce.org/static_html/downloads/reports/reports/ntds_full.pdf. Greenfeld, L.A. (1997). Sex Offenses and Offenders: An Analysis of Data on Rape and Sexual Assault. Gutner, C. A., Rizvi, S. L., Monson, C. M., & Resick, P. A. (2006). Changes in coping strategies, Hanson, R. F., Kievit, L. W., Saunders, B. E., Smith, D. W., Kilpatrick, D. G., Resnick, H. S., & Hanson, R. F., Resnick, H. S., Saunders, B. E., Kilpatrick, D. G., & Best, C. (1999). Factors related to the Hébert, M., Tourigny, M., Cyr, M., McDuff, P., & Joly, J. (2009). Prevalence of childhood sexual abuse Herman, J L. (1992). Complex PTSD: A syndrome in survivors of prolonged and repeated trauma. Hershkowitz, I., Lanes, O., & Lamb, M. E. (2007). Exploring the disclosure of child sexual abuse with Holland, K. J., & Cortina, L. M. (2017). “It happens to girls all the time”: Examining sexual assault Hosterman, A. R., Johnson, N. R., Stouffer, R., & Herring, S. (2018). Twitter, social support messages, Hudson, J., & Nelson, K. (1986). Repeated encounters of a similar kind: Effects of familiarity on and timing of disclosure in a representative sample of adults from Quebec. The Canadian Journal of Psychiatry, 54(9), 631-636. alleged victims and their parents. Child Abuse & Neglect, 31(2), 111-123. survivors’ reasons for not using campus supports. American journal of community psychology, 59(1-2), 50-64. and the #Metoo movement. The Journal of Social Media in Society, 7(2), 69-91. children’s autobiographic memory. Cognitive Development, 1(3), 253-271. Ruggiero, K. J. (2003). Correlates of adolescent reports of sexual assault: Findings from the National Survey of Adolescents. Child Maltreatment, 8(4), 261-272. reporting of childhood rape. Child Abuse & Neglect, 23(6), 559-569. Journal of Traumatic Stress, 5, 377‐391. 87 2015 U.S. Transgender Survey. Washington, DC: National Center for Transgender Equality. experiences of women who do or do not label their sexual assault as rape. Psychology of Women Quarterly, 27, 233-242. with crisis across the lifespan: The role of a telephone hotline. Journal of Child and Family Studies, 17(5), 663-674. adult versus adolescent sexual assault: Epidemiology and patterns of anogenital injury. Academic Emergency Medicine, 10(8), 872-877. Characteristics and implications for posttraumatic stress symptoms among African American and Caucasian survivors. Journal of Trauma & Dissociation, 11(2), 174-192. Ingram, S., Ringle, J. L., Hallstrom, K., Schill, D. E., Gohr, V. M., & Thompson, R. W. (2008). Coping Jacques-Tiura, A. J., Tkatch, R., Abbey, A., & Wegner, R. (2010). Disclosure of sexual assault: James, S. E., Herman, J. L., Rankin, S., Keisling, M., Mottet, L., & Anafi, M. (2016). The Report of the Jones, J. S., Rossman, L., Wynn, B. N., Dunnuck, C., & Schwartz, N. (2003). Comparative analysis of Kahn, A. S., Jackson, J., Kully, C., Badger, K., & Halvorsen, J. (2003). Calling it rape: Differences in Kennedy, M. A., & Gorzalka, B. B. (2002). Asian and non-Asian attitudes toward rape, sexual harassment Kilpatrick, D. G., Edmunds, C. N., Seymour, A. K. (1992). Rape in America: A Report to the Nation. Arlington, VA: National Victim Center and Crime Victims Research and Treatment Center. Kilpatrick, D. G., Saunders, B. E., & Smith, D. W. (2003). Youth Victimization: Prevalence and Kogan, S. M. (2004). Disclosing unwanted sexual experiences: Results from a national sample of Kogan, S. M. (2005). The role of disclosing child sexual abuse on adolescent adjustment and Konradi, A. (2007). Taking the stand: Rape survivors and the prosecution of rapists. Westport, CT: Koss, M. P. (1985). The hidden rape victim: Personality, attitudinal, and situational characteristics. Implications, Research In Brief. Washington, DC: U.S. Department of Justice, Office of Justice Programs, National Institute of Justice. and sexuality. Sex Roles, 46(7-8), 227-238. adolescent women. Child Abuse & Neglect, 28(2), 147–165. revictimization. Journal of Child Sexual Abuse, 14(2), 25-47. Greenwood Publishing Group. Psychology of Women Quarterly, 9(2), 193-212. Online Resource Center on Violence against Women. Koss, M., & Achilles, M. (2008). Restorative justice responses to sexual assault. VAWnet: The National Koss, M.P., Dinero, E., Seibel, C.A., & Cox, S.L. (1988). Stranger and acquaintance rape: Are there Krug, E. G., Mercy, J. A., Dahlberg, L. L., & Zwi, A. B. (2002). The world report on violence and health. differences in the victim's experience? Psychology of Women Quarterly, 12, 1-24. The Lancet, 360(9339), 1083-1088. 88 behavior, and prevention/treatment programs. Trauma, Violence, & Abuse, 11(4), 159-177. Justice, Office of Justice Programs, Bureau of Justice Statistics. survivors of rape. Journal of Interpersonal Violence, 20(5), 591-616. understanding help‐seeking processes among survivors of intimate partner violence. American Journal of Community Psychology, 36(1-2), 71-84. of the contemporary empirical literature. In M.E. Pipe, M. E. Lamb, Y. Orbach, & A.C. Cederborg (Eds.), Child Sexual Abuse: Disclosure, Delay, & Denial (pp. 11-39). Mahwah, NJ, US: Lawrence Erlbaum Associates Publishers. Lalor, K., & McElvaney, R. (2010). Child sexual abuse, links to later sexual exploitation/high-risk sexual Langton, L., & Truman, J. L. (2014). Socio-emotional Impact of Violent Crime. US Department of Liang, B., Goodman, L., Tummala-Narra, P., & Weintraub, S. (2005). A theoretical framework for Logan, T. K., Evans, L., Stevenson, E., & Jordan, C. E. (2005). Barriers to services for rural and urban London, K., Bruck, M., Ceci, S., & Shuman, D. (2003) Disclosure of child sexual abuse: What does the research tell us about the ways that children tell? Psychology, Public Policy, & Law, 11(1), 194 226. London, K., Bruck, M., Ceci, S. J., & Shuman, D. W. (2007). Disclosure of child sexual abuse: A review Long, L., & Ullman, S.E. (2013). The impact of multiple traumatic victimization on disclosure and Lonsway, K. A., and Archambault, J. (2012). The “justice gap” for sexual assault cases: Future directions Lynch, V. A., & Duval, J. B. (2010). Forensic Nursing Science. Maryland Heights, MO: Mosby Lyon, T. D., Ahern, E. C., Malloy, L. C., & Quas, J. A. (2010). Childrens’ reasoning about disclosing Maas, M. K., McCauley, H. L., Bonomi, A. E., & Leija, S. G. (2018). “I was grabbed by my pussy and its Macy, R. J., Giattina, M., Sangster, T. H., Crosby, C., & Montijo, N. J. (2009). Domestic violence and Malloy, L. C., Brubacher, S. P., & Lamb, M. E. (2011). Expected consequences of disclosure revealed in Malloy, L. C., Brubacher, S. P., & Lamb, M. E. (2013). “Because she’s one who listens”: Children discuss disclosure recipients in forensic interviews. Child Maltreatment, 18(4), 245-251. adult transgressions: Effects of maltreatment, child age, and adult identity. Child Development, 81(6), 1714-1728. coping mechanisms for Black women. Feminist Criminology, 8, 295-319. of research and reform. Violence Against Women, 18, 145-168. #NotOkay”: A Twitter backlash against Donald Trump’s degrading commentary. Violence Against Women, 24(14), 1739-1750. sexual assault services: Inside the black box. Aggression and Violent Behavior, 14(5), 359-373. investigative interviews with suspected victims of child sexual abuse. Applied Developmental Science, 15(1), 8-19. Incorporated. 89 AEquitas; Strategies in Brief, 13, 1–3. New York, NY: Routledge. doctoral dissertation). Trinity College, Dublin. recipients’ experiences: emotional distress and changes in the relationship with the victim. Violence & Victims, 31(3), 457. (BJS), US Department of Justice, Office of Justice Programs, & United States of America. Handbook of Quantitative Methods, Volume 2: Statistical Analysis (pp. 551-611). New York, NY: Oxford University Press. (1997). Clinical characteristics of women with a history of childhood abuse: unhealed wounds. JAMA, 277(17), 1362-1368. Markowitz, J. (2012).Absence of anogenital injury in the adolescent/adult female sexual assault patient. Martin, S. L., Macy, R. J., & Young, S. K. (2010). The impact of sexual violence against women: Health and economic consequences. In J. W. White, M. P. Koss, & A. Kazdin (Eds.), Violence Against Women and Children, Volume 1. Washington, DC: American Psychological Association. Martin, P.Y. (2005). Rape work: Victims, gender, and emotions in organization and community context. Masyn, K. E. (2013). Latent class analysis and finite mixture modeling. In T.D. Little (Ed.), The Oxford McCauley, J., Kern, D. E., Kolodner, K., Dill, L., Schroeder, A. F., DeChant, H. K., ... & Bass, E. B. McElvaney, R. (2008). How Children Tell: Containing The Secret of Child Sexual Abuse (Unpublished Milliken, J., Paul, L. A., Sasson, S., Porter, A., & Hasulube, J. (2016). Sexual assault disclosure Morgan, R. E. & Kena, G., (2018). Criminal Victimization, 2016: Revised. Bureau of Justice Statistics Muthén, B.O., & Muthén, L. K. (2000). Integrating person‐centered and variable‐centered analyses: Muthén, L.K. and Muthén, B.O. (1998-2018). Mplus User’s Guide. Version 8.2. Los Angeles, CA: Najdowski, C. J., and Ullman, S. E. (2009). Prospective effects of sexual victimization on PTSD and National Center for PTSD (2018). Violence and abuse. Retrieved from: https://www.ptsd.va.gov/understand/types/violence_abuse.asp National Center for Victims of Crime (NCVC). (2012). About sexual assault. Retrieved from: National Center for Victims of Crime (NCVC). (2011). Grooming dynamics. Retrieved National Research Council. (1993). Understanding child abuse and neglect. Washington, DC: National Growth mixture modeling with latent trajectory classes. Alcoholism: Clinical and Experimental Research, 24(6), 882-891. Muthén & Muthén problem drinking. Addictive Behaviors, 34(11), 965-968. http://victimsofcrime.org/our-programs/dna-resource-center/untested-sexual-assault kits/about-sexual-assault from:http://victimsofcrime.org/media/reporting-on-child-sexual-abuse/grooming- dynamic-of-csa Academy Press. 90 thesis). State University of New York College, Plattsburgh. Science, 8(3), 162-166. regression. Journal of Social Service Research, 35(2), 193-205. sexual violence, 1994-2010. Special Report (No. NCJ 240655). Washington, DC: Bureau of Justice Statistics. US Department of Justice. analysis and growth mixture modeling: A Monte Carlo simulation study. Structural Equation Modeling: A Multidisciplinary Journal, 14(4), 535-569. National Center for Injury Prevention and Control. Division of Violence Prevention. Center for Disease Control. NISVS (2010). An overview of 2010 findings on victimization by sexual orientation. Atlanta, GA: Noble, A. (2016). Utilization of a Text-Based Sexual Assault Crisis Hotline (Unpublished master’s Nylund, K. L., Asparouhov, T., & Muthén, B. O. (2007). Deciding on the number of classes in latent class Orchowski, L. M., & Gidycz, C. A. (2012). To whom do college women confide following sexual assault? A prospective study of predictors of sexual assault disclosure and social reactions. Violence Against Women, 18(3), 264-288. Pennebaker, J. W. (1997). Writing about emotional experiences as a therapeutic process. Psychological Petrucci, C. J. (2009). A primer for social worker researchers on how to conduct a multinomial logistic Planty, M., Langton, L., Krebs, C., Berzofsky, M., & Smiley-McDonald, H. (2013). Female victims of Polusny, M. A., & Follette, V. M. (1995). Long-term correlates of child sexual abuse: theory and review Putnam, F. W. (2003). Ten-year research update review: Child sexual abuse. Journal of the American Ralph, L. J., Berglas, N. F., Schwartz, S. L., & Brindis, C. D. (2011). Finding teens in TheirSpace: using social networking sites to connect youth to sexual health services. Sexuality Research and Social Policy, 8(1), 38-49. RAINN (2018). Sexual assault [Definitions]. Retrieved from: https://www.rainn.org/articles/sexual Rennison, C. M., & Rand, M. R. (2003). Criminal victimization, 2002. Washington, DC: US Department Reyome, N.D. (1994). Teacher ratings of the academic achievement related classroom behaviors of Rickert, V. I., Wiemann, C. M., & Vaughan, R. D. (2005). Disclosure of date/acquaintance rape: Who Rickwood, D., Deane, F. P., Wilson, C. J., & Ciarrochi, J. (2005). Young people's help-seeking for mental maltreated and nonmaltreated children. Psychology in the Schools, 31, 253-260 reports and when. Journal of Pediatric and Adolescent Gynecology, 18(1), 17-24. of the empirical literature. Applied and Preventive Psychology, 4, 143-166. Academy of Child & Adolescent Psychiatry, 42(3), 269-278. -assault. of Justice, Bureau of Justice Statistics. health problems. Australian e-Journal for the Advancement of Mental Health, 4(3), 218–251. 91 social reactions. Equality, Diversity and Inclusion: An International Journal (online first). sexual abuse by adolescents: A qualitative in-depth study. Journal of Interpersonal Violence, 27(17), 3486-3513. abuse. Retrieved from: https://reliefweb.int/sites/reliefweb.int/files/resources/un_glossary_on_sea.pdf Swarthmore College. Retrieved from https://www.swarthmore.edu/share/what-sexual exploitation manifest in sexual assault police records. Psychology of Violence, 7(4), 602. messages about sex. Violence Against Women, 14(11), 1326-1348. and the 1-year prevalence of medical problems in the National Comorbidity Survey. Health Psychology, 24(1), 32. Sachs-Ericsson, N., Blazer, D., Plant, E. A., & Arnow, B. (2005). Childhood sexual and physical abuse Schneider, K. T., & Carpenter, N. J. (2019). Sharing# MeToo on Twitter: incidents, coping responses, and Schönbucher, V., Maier, T., Mohler-Kuo, M., Schnyder, U., & Landolt, M. A. (2012). Disclosure of child SEA Task Team of the United Nations. (2016). United Nations glossary on sexual exploitation and Sexual Harassment/Assault Resource & Education (SHARE). (2018). What is sexual exploitation. Shaw, J., Campbell, R., Cain, D., & Feeney, H. (2017). Beyond surveys and scales: How rape myths Smith, S. G., & Cook, S. L. (2008). Disclosing sexual assault to parents: The influence of parental Snyder, H. N. (2000). Sexual assault of young children as reported to law enforcement: Victim, incident, Snyder, H. N., & Sickmund, M. (2006). Juvenile offenders and victims: 2006 national report. Starzynski, L. L., Ullman, S. E., Filipas, H. H., & Townsend, S. M. (2005). Correlates of women's sexual Starzynski, L. L., Ullman, S. E., Townsend, S. M., Long, L. M., & Long, S. M. (2007). What factors Stein, R. E., & Nofzinger, S. D. (2008). Adolescent sexual victimization choice of confidant and the Stermac, L., Del Bove, G., & Addison, M. (2004). Stranger and acquaintance sexual assault of adult Sturza, M.L., & Campbell, R. (2005). An exploratory study of rape survivors’ prescription drug use as a and offender characteristics. Washington, DC: U.S. Department of Justice, Office of Justice Programs, Bureau of Justice Statistics. Retrieved from: http://www.ojp.usdoj.gov/bjs/pub/pdf/saycrle.pdf failure of authorities. Youth Violence and Juvenile Justice, 6(2), 158-177. males. Journal of Interpersonal Violence, 19(8), 901-915. predict women's disclosure of sexual assault to mental health professionals? Journal of Community Psychology, 35(5), 619-638. Washington, DC: National Center for Juvenile Justice, U.S. Department of Justice. assault disclosure to informal and formal support sources. Violence and Victims, 20(4), 417. means of coping with sexual assault. Psychology of Women Quarterly, 29, 353-363. 92 case series. American Journal of Obstetrics and Gynecology, 190(1), 71–76. comprehensive services. In Renzetti, C. M., Edleson, J. L., & Kennedy Bergen, R. (Eds.) Sourcebook on Violence Against Women, Third Edition. Sage Publications. Washington, DC: U.S. Department of Justice, Office of Justice Programs, National Institute of Justice. development: Lessons from a multigenerational, longitudinal research study. Development and Psychopathology, 23(2), 453-476. from the National Violence Against Women Survey. Washington, DC: U.S. Department of Justice, Office of Justice Programs, National Institute of Justice. Sugar, N. F., Fine, D. N., & Eckert, L.O. (2004). Physical injury after sexual assault: Findings of a large Tjaden, P., & Thoennes, N. (2000). Findings from the national violence against women survey. Tjaden, P., & Thoennes, N. (2006). Extent, nature, and consequences of rape victimization: Findings Townsend, S., & Campbell, R. (2018) Services for survivors of sexual violence: Moving from core to Trickett, P. K., Noll, J. G., and Putnam, F. W. (2011). The impact of sexual abuse on female Ullman, S. E. (1996). Do social reactions to sexual assault victims vary by support provider? Violence & Ullman, S. E. (2007). Comparing gang and individual rapes in a community sample of urban women. Ullman, S. E. (2010). Talking about sexual assault: Society’s response to survivors. Washington, DC: Ullman, S. E. (2016). Sexual revictimization, PTSD, and problem drinking in sexual assault survivors. Ullman, S. E., & Filipas, H. H. (2001). Correlates of formal and informal support seeking in sexual Ullman, S. E., & Peter‐Hagene, L. (2014). Social reactions to sexual assault disclosure, coping, perceived control, and PTSD symptoms in sexual assault victims. Journal of Community Psychology, 42(4), 495-508. Ullman, S. E., & Siegel, J. M. (1993). Victim-offender relationship and sexual assault. Violence & U.S. Department of Health & Human Services, Administration for Children and Families, Administration on Children, Youth and Families, Children’s Bureau. (2018). Child maltreatment 2016. Retrieved from: https://www.acf.hhs.gov/cb/research-data-technology/statistics-research/child maltreatment. Vermunt, J.K. & Magidson, J. (2005). Chapter 6: Model and model summary output. In Latent GOLD 4.0 User's Guide (pp. 103-162). Belmont, Massachusetts: Statistical Innovations Inc. Victims, 11, 143–156. Violence & Victims, 22, 43‐51. American Psychological Association. Addictive Behaviors, 53, 7-10. Victims, 8(2), 121. assault victims. Journal of Interpersonal Violence, 16, 1028-1047. 93 Sociological Methods & Research, 27(3), 359-397. and development. Merrill-Palmer Quarterly (1982-), 390-420. von Eye, A., & Bogat, G. A. (2006). Person-oriented and variable-oriented research: Concepts, results, Wasco, S. M., Campbell, R., Howard, A., Mason, G. E., Staggs, S. L., Schewe, P. A., & Riger, S. (2004). A statewide evaluation of services provided to rape survivors. Journal of Interpersonal Violence, 19(2), 252-263. Wadsworth, P., & Records, K. (2013). A review of the health effects of sexual assault on African American women and adolescents. Journal of Obstetric, Gynecologic, & Neonatal Nursing, 42(3), 249-273. increasing appropriate help engagement. Journal of Educational and Psychological Consultation, 12(4), 345–364. Weakliem, D. L. (1999). A critique of the Bayesian information criterion for model selection. Wilson, C. J., & Deane, F. P. (2001). Adolescent opinions about reducing help-seeking barriers and Young, S. M., Pruett, J. A., & Colvin, M. L. (2018). Comparing help-seeking behavior of male and Zinzow, H. M., Resnick, H. S., McCauley, J. L., Amstadter, A. B., Ruggiero, K. J., & Kilpatrick, D. G. (2012). Prevalence and risk of psychiatric disorders as a function of variant rape histories: Results from a national survey of women. Social Psychiatry and Psychiatric Epidemiology, 47(6), 893 -902. female survivors of sexual assault: A content analysis of a hotline. Sexual Abuse, 30(4), 454-474. 94