Sexual Abuse Within Employment Settings: A Comparison of Work-Related,Intra- and Extra-Familial Child Molesters

Abstract

Child sexual abuse (CSA) is a serious problem not only in the community but also in institutional settings such as youth-serving organizations, churches, and schools. Although research has started to examine the problem of abuse in institutional settings, there remains a dearth of information about the nature and context of CSA in different employment settings, including those that do not specifically cater to children. In addition, research on the similarities and differences between perpetrators who work with children and other sex offenders is scarce. As such, the present study compared offenders on variables relating to financial/employment lifestyle stability, risk/dangerousness level, abuse opportunity, and victim selection. Data revealed that child abusers who worked with children tended to be better educated, were less likely to be married, had fewer nonsexual convictions, and were more likely to abuse male post-pubescent children compared with intra- and extra-familial offenders who did not work with children. Implications for future research, prevention of CSA, and clinical practice are discussed.

Keywords

sex offenders who work with children child sexual abuse institutional abuse intra-familial professional perpetrators extra-familial

Introduction

Child sexual abuse (CSA) is a serious problem, often producing long-term negative consequences for the child experiencing the abuse, and for society at large (Currie & Widom, 2010). Internationally, 7% to 36% of women and 3% to 29% of men report a history of CSA (Finkelhor, 1994; Pereda, Guilera, Forns, & Gómez-Benito, 2009). Although the estimated number of CSA cases in the United States has declined (from 150,000 estimated cases in 1992 to 89,500 cases in 2000; a 40% decrease), CSA remains a significant global problem (see Finkelhor, 1994; Jones, Finkelhor, & Kopiec, 2001; Pereda et al., 2009). The majority of research aimed at addressing this issue focuses on the victim, perpetrator, and offense characteristics of community sex offenders. However, recent research identifies a specific type of CSA perpetrated by sex offenders who groomed victims through employment positions (i.e., individuals working with children; CSA-W), highlighting the need for research examining the nature and context of CSA committed in work places (Moulden, Firestone, Kingston, & Wexler, 2010; Wurtele, 2012).

Characteristics of CSA Within Employment Settings

Gallagher (2000) first defined the abuse perpetrated by an individual who works in a setting with access to children (i.e., institutionalized abuse) as the sexual abuse of a child by an adult who has a working relationship with the victim. This nature of the work can be paid or voluntary and can occur in public or private sectors, including residential and nonresidential settings. This form of CSA is particularly dangerous and less detectable. Perpetrators of CSA-W place themselves in a position of trust and authority. The attribution of authority and trust frequently results in unsupervised access to children, and often deters children from disclosure of CSA (Moulden, Firestone, & Wexler, 2007).

Whereas the overall prevalence of CSA within work settings is unknown, news and media reports highlight that institutional CSA is a serious and widespread, global issue (Gallagher, 2000; Wurtele, 2012). Within the scope of CSA-W cases, the recent Catholic Church crisis illustrates that CSA perpetrated by nonbiologically related abusers, specifically males working with children, is an ongoing and enduring issue (John Jay College Research Team, 2004). Solely within the U.S. Catholic Church, some 10,667 individuals made allegations of CSA perpetrated by priests between the years 1950 and 2002. Similar scandals have been reported in Canada, the United Kingdom, the Netherlands, Belgium and Australia, among others (Barnardos et al., 2010; Death, 2015; Department of Justice and Equality, 2009). As a result of the high number of clergy CSA cases, a large body of research exclusively focuses on cleric sexual abuse characteristics and prevention strategies (Calkins, Fargo, Jeglic, & Terry, 2015; John Jay College Research Team, 2004; Mercado, Tallon, & Terry, 2008; Perillo, Mercado, & Terry, 2008; Terry, Mercado, & Perillo, 2008).

Cases of CSA have also been reported in nonclergy work settings such as sporting organizations and the Boy Scouts of America (BSA; n.d.). In addition, 6% to 10% of U.S. school children have reported CSA perpetrated by teachers or other individuals working within school settings (Shakeshaft, 2004). Despite these reports, research on nonclergy CSA-W offenders is scarce. In light of recent highly publicized cases of CSA involving teachers and athletic coaches (e.g., Jerry Sandusky), it is imperative to get a better understanding of CSA perpetrated in work settings that involve access to children (Boyle, 2014).

Subtyping CSA Offenders

By definition, CSA-W offenders are a subtype of extra-familial sex offenders (CSA-E), as CSA-E offenders are defined as individuals who sexually victimize a child with whom no familial offender–victim relationship exists. There are several commonalities between CSA-W and CSA-E offenders. Both groups are significantly more likely to sexually abuse male—or male and female—children compared with intra-familial offenders (CSA-I; offenses against immediate family members; incest), and both groups share similar levels of emotional congruency with children (Sullivan, Beech, Craig, & Gannon, 2011) as well as sexual preoccupation (i.e., sex drive; Thornton, 2002). Consequently CSA-W offenders have typically been studied within CSA-E offender research (Magalhães et al., 2009).

Recent findings have challenged the long-standing belief that CSA-W offenders belong in the group of CSA-E sex offenders. One argument suggests that CSA-W offenders may exhibit characteristics, including risk factors and criminogenic needs, that more closely resemble those of CSA-I offenders rather than CSA-E offenders (Sullivan et al., 2011). Researchers (Leclerc, Proulx, & McKibben, 2005; Sullivan & Beech, 2004; Sullivan et al., 2011; Turner et al., 2014) note that particular work environments and authoritative positions of trust may create an atmosphere of loco parentis, placing the abuser into a parent-like role. Furthermore, CSA-W offenders often resemble CSA-I offenders in justification strategies and grooming techniques. Both groups often make use of noncoercive grooming strategies, involving affection, love, and gifts and seem to be less reliant on cognitive distortions that justify and validate the sexually abusive behavior (Sullivan et al., 2011).

Recent research has identified various characteristics that distinguish CSA-W offenders from CSA-E and CSA-I offenders, suggesting that CSA-W offenders may constitute a separate group of offenders altogether (Sullivan & Beech, 2004; Sullivan et al., 2011; Turner, Rettenberger, Lohmann, Eher, & Briken, 2014). For example, CSA-W offenders are better educated, and less likely to be married and to have children than CSA-E and CSA-I offenders. Furthermore, CSA-W offenders exhibit higher levels of pedophilic sexual interest and abuse a higher frequency of male victims as compared with CSA-E or CSA-I offenders. In addition, CSA-W offenders tend to have fewer prior sexual and nonsexual convictions compared with CSA-E and CSA-I offenders, and different recidivism patterns (Turner et al., 2014). The link between criminal history, specifically a high number of previous convictions, and antisocial and psychopathic personality traits (Falkenbach, 2004; Hanson & Morton-Bourgon, 2005), led Turner and colleagues (2014) to conclude that CSA-W offenders display fewer of those personality tendencies compared with CSA-E and CSA-I offenders, adding to the argument that CSA-W offenders constitute a unique subpopulation. In light of these considerations, a detailed characterization of a variety of CSA-W offender categories, along with a comparison of CSA-W, CSA-E, and CSA-I offenders is necessary to understand the unique characteristics of sex offending in employment settings. This comparative analysis may ultimately aid in the refinement of risk, treatment, and prevention strategies, so that we can better target the specific needs of this offender subpopulation.

The Present Study

To our knowledge, only four studies have compared multifaceted samples in CSA-W, CSA-I and CSA-E populations, including various categories of CSA-W employment settings (Sullivan & Beech, 2004; Sullivan et al., 2011; Turner, Rettenberger, Lohmann, Eher, & Briken, 2014; Turner et al., 2014). However, in contrast to those previous studies, the present study examines a variety of professions and voluntary activities in which sexual abuse occurs, including school settings, ecclesiastic settings, professional child caretaking services, athletic facilities, child-centric work places, and non-child-centric work places. The present research pursues several objectives. First, this CSA-W sample comprises a heterogeneous list of employment settings to explore those that are not typically the focus in CSA research. Next, to expand our knowledge of within-group differences of CSA-W offenders, the present research provides a description of the types of employment settings within which sexual abuse occurs and the different variables associated with the various types of employment. Furthermore, we seek to examine between-group differences (CSA-I, CSA-E, CSA-W) on common variable categories associated with CSA: financial/employment lifestyle stability, risk level/dangerousness, victim selection, and abuse opportunity. Finally, for a more nuanced look at group differences, the specific influences of predictor variables were considered by analyzing which variables predicted membership into the CSA-I, CSA-E, and CSA-W groups. Given the scarcity of research in this area, the present research is descriptive in nature. However, it was anticipated that the CSA-W sample would present with distinct characteristics from those observed in CSA-E and CSA-I samples.

Method

Participants

Archival data for this study was gathered as part of a larger study examining sex offender placement within a state criminal justice system (see Mercado, Jeglic, & Markus, 2011). All offenders (N = 3,194) were incarcerated and served time in either a prison-based sex offender treatment facility or the general population of the state prison, and were released from custody between 1996 and 2007. Consistent with the State legislation (Department of Law and Public Safety & Farmer, 2001), the present research included only those sex offenders who had committed the index offense against at least one child below the age of 17. Only those offenders (18 years and older) who, as part of their index offense, had committed sexual offenses exclusively against (a) a biological child (CSA-I), (b) an acquaintance (i.e., knew their victim more than 24 hr; CSA-E), or (c) a child who they accessed through their position of work (CSA-W), were included in this study. For a case to be included in the CSA-W group, the offender had to have established contact with their victim(s) through the institution or setting in which they worked in a paid or voluntary capacity (i.e., if not by means of the work position, the abuse would have not occurred). In addition, the CSA-W category did not require an authoritative child-caring offender–victim relationship, allowing for the inclusion a broader swath of employment contexts (e.g., candy store, city bus, postal work). After following these inclusion criteria, the final sample comprised 1,161 individuals who had committed an offense against a child.

The final sample consisted of offenders who, at the time of index offense ranged in age from 18 to 87 years old, with a mean age of 33.26 (SD = 11.67). Racial backgrounds included 43.2% White (n = 502), 34.3% African American (n = 398), 20.8% Latino (n = 241), 0.6% Asian/Pacific Islanders (n = 7), and 1% Unknown/Other ethnicity (n = 11). Of the final sample, 231 (20%) individuals exclusively sexually abused a biological child in a nonemployment setting (CSA-I), 876 (75%) molested an acquaintance in a nonemployment setting (CSA-E), and 54 (5%) sexually victimized a minor with whom they had contact through their employment setting (CSA-W). For a racial and ethnic breakdown of the CSA groups, see Table 3.

Measures

Employment settings

For the CSA-W offender sample, the identified employment settings were classified into six mutually exclusive categories: school setting, ecclesiastic setting, professional child caretaking services, athletic facilities, child-centric work places, and non-child-centric work places. The school setting category was defined as including any person who was working at a school organization, regardless of type of position. Ecclesiastic settings included any person who was working or volunteering at an ecclesiastic institution, regardless of trust/authority position. Professional child caretaking employment services included any person who performed a professional role aimed at providing services to children, including those aimed at improving their psychological or medical well-being. The athletic facility work setting was defined as including any athletic coaches, regardless of institution affiliation. Child-centric work places were defined as any employment setting that involved predominantly child-centric contacts, though, in contrast to professional caretaking services, not in a caretaking capacity such as a candy store owner or bus driver for children. Finally, non-child-centric work places included individuals who performed a job that is not predominantly child-centric, such as convenience store owners or taxi drivers. In this setting, the working individual was assumed to interact with both children and adults in similar frequency.

Variables of interest

Financial/employment lifestyle stability

The lifestyle stability category comprised educational level, marital status, and income prior to incarceration. Level of education was grouped into three categories: less than high school degree (less than 8th grade, some high school, General Education Diploma), high school degree/some college (high school degree, trade school degree/certificate, some college), and beyond high school/college (2-year college degree, 4-year college degree, graduate degree, and professional degree). Marital status at the time of the index offense was measured as being single (never married), married (married, lived with partner), or being separated (divorced, widowed). Finally, annual income prior to incarceration was categorized as less than US$20,000; US$20,000-US$40,000; or greater than US$40,000.

Risk/dangerousness level

Risk was assessed by age at index offense, history of psychiatric problems (yes/no), prior sexual and nonsexual convictions (yes/no), and Static-99 total score. The Static-99 (Hanson & Thornton, 2000) is an actuarial risk assessment tool developed to predict violent and sexual recidivism in adult male sex offenders. The measure uses 10 static variables, yielding a total score (ranges from 0 to 12) that transformed into risk categories (low, moderate-low, moderate-high, high). For the purpose of our study, only the overall risk categories were considered.

Abuse opportunity

Opportunity was assessed by biological children or step-children (yes/no), living with anyone at the time of the index offense (yes/no), the number of children that the offender had, and the number of children the offender lived with.

Victim selection

Victim selection variables included victim gender, victim age, and number of victims. Victim gender was measured as having only male, male and female, or only female victims. The majority of offenders had one victim; therefore, to provide uniformity across groups, the age of the first victim was recorded.

Procedure

Data were archival in nature and extracted from police reports, psychiatric evaluations, criminal history records, sentencing information, prison records, and therapeutic intake and termination reports. Data were coded by trained graduate level research assistants. Using an emergent coding strategy, a team of five graduate level students independently examined the data and organized the CSA-W offender job types into employment type categories. Cohen’s kappa statistic was calculated to determine interrater reliability of coding the specific jobs into these broad categories of employment. There was moderate agreement between the two raters, κ =.61. All discrepancies were discussed and coded to consensus.

Results

Within CSA-W Offender Sample

Types of CSA-W employment settings

As shown in Table 1, the employment setting categorization revealed that school setting (n = 16, 29%) accounted for the largest number of CSA-W cases, although non-child-centric work places (n = 12, 22%) and professional caretaking services (n = 10, 18%) comprised a sizable proportion of cases. A smaller number of cases constituted the ecclesiastic settings (8, 15%), child-centric settings (5, 9%), and athletic facilities (4, 7%) categories.

Table 1.

Professions and Voluntary Activities of CSA Offenders Working With Children.

Professions and voluntary activities (N = 55)	n	% of sample	(% of category)
School setting	16	29
Teacher	12	21.81	(75.00)
Bus driver	2	3.63	(12.50)
Maintenance worker	1	1.81	(6.25)
Volunteer	1	1.81	(6.25)
Ecclesiastic setting	8	15
Clergy	4	7.27	(57.10)
Youth group leader	2	3.63	(28.60)
Bus driver	1	1.81	(14.30)
Caretaking services	10	18
Counselor	6	10.90	(60.00)
Daycare	2	3.63	(20.00)
Pediatrician	1	1.81	(10.00)
Health aid	1	1.81	(10.00)
Athletic facilities	4	7
Basketball coach	2	3.63	(50.00)
Gymnastic coach	1	1.81	(25.00)
Unknown	1	1.81	(25.00)
Child-centric setting	5	9	-
Juvenile detention guard	2	3.63	(40.00)
Candy store owner	1	1.81	(20.00)
Bus driver for kids	1	1.81	(20.00)
Carnival ride operator	1	1.81	(20.00)
Non-child-centric setting	12	22
Security guard	2	3.63	(16.70)
Citizen Band Radio worker	2	1.81	(16.70)
Bus driver	2	1.81	(16.70)
Postal worker	1	1.81	(8.30)
Convenience store owner	1	1.81	(8.30)
Taxi driver	1	1.81	(8.30)
Apartment supervisor	1	1.81	(8.30)
Ice arena owner	1	1.81	(8.30)
Bakery worker	1	1.81	(8.30)

Note. CSA = Child sexual abuse.

CSA-W occupations

Table 1 provides an overview of the specific professions and volunteer activities within the CSA-W sample. Individuals who worked as teachers (n = 12, 21.81%), professional caretaking counselors (n = 6, 10.91%), and clergy (n = 4, 7.27%) were most frequently represented within the CSA-W sample. There were a number of professions that were represented by a single offender, such as candy store owner and postal worker.

CSA-W offender characteristics

Tables 2 and 3 present the descriptive statistics for the CSA-W sample. Due to the small sample size, no statistical comparisons were conducted. However, those individuals who worked in a child-centric employment setting tended to be older at the time of index offense (M = 45.60; SD = 11.29), whereas those who worked in a professional caretaking capacity tended to be younger (M = 30.10; SD = 11.03). Individuals who worked in child-centric work places had the highest number of victims (M = 2.00; SD = 1.22), whereas those offenders who committed CSA within their ecclesiastic-related profession tended to have the lowest number of victims (M = 1.14; SD = .38). The non-child-centric group had the greatest proportion of individuals convicted for a nonsexual offense prior to index offense (n = 8, 66.67%), and the child-centric group had the fewest (n = 1, 20%). Results indicated that those individuals who worked in an ecclesiastic setting were most often convicted for a sexual offense prior to index offense (n = 3, 42.86%), whereas the professional caretaking individuals were among those least likely to have a prior sexual offense conviction (n = 1, 10%).

Table 2.

Descriptive Statistics for the Continuous Variables Within CSA-W Sample.

Variable	Study group
	School setting		Ecclesiastic setting		Caretaking services		Athletic facilities		Child-centric setting		Non-child-centric Setting
	M	SD	M	SD	M	SD	M	SD	M	SD	M	SD
Age index offense	40.38	(16.88)	36.71	(7.57)	30.10	(11.03)	30.75	(3.59)	45.60	(11.22)	43.83	(13.60)
Number of total victims	1.31	(0.60)	1.14	(0.38)	1.40	(0.70)	1.75	(1.50)	2.00	(1.22)	1.58	(0.90)
Age first victim	12.64	(3.46)	12.14	(2.54)	12.90	(3.21)	14.25	(2.21)	9.75	5.85	12.67	(3.17)
Number of children	2.54	(2.30)	2.17	(1.94)	0.85	(1.07)	0.50	(0.71)	3.25	(0.96)	2.67	(1.58)
Number of children living with	0.40	(0.91)	1.14	(1.67)	0.00	(0.00)	0.00	(0.00)	0.63	(1.29)	1.33	(1.15)

Note. Due to small sample size no statistical analyses were conducted. CSA-W = child sexual abusers working with children.

Table 3.

Descriptive Statistics of the Categorical Variables Within CSA-W Sample.

Variable	Study group
	School setting		Ecclesiastic setting		Caretaking services		Athletic facilities		Child-centric setting		Non-child-centric setting
	n	(%)	n	(%)	N	(%)	n	(%)	n	(%)	n	(%)
Ethnicity
White	9	(56.25)	2	(28.57)	6	(60.00)	3	(75.00)	1	(20.00)	7	(63.60)
African/American	4	(25.00)	5	(71.40)	4	(40.00)	1	(25.00)	3	(60.00)	1	(9.10)
Latino	3	(18.75)	0	(0.00)	0	(0.00)	0	(0.00)	1	(20.00)	2	(18.20)
Other	0	(0.00)	0	(.00)	0	(0.00)	0	(0.00)	0	(0.00)	1	(9.10)
Marital status
Single	7	(43.75)	2	(28.57)	6	(60.00)	3	(75.00)	1	(20.00)	5	(41.67)
Married	7	(43.75)	4	(57.14)	3	(30.00)	0	(0.00)	3	(60.00)	4	(33.33)
Divorced	2	(12.50)	1	(14.29)	1	(10.00)	1	(25.00)	1	(20.00)	3	(25.00)
Education
Less than high school, GED, 8th grade	4	(25.00)	2	(28.57)	1	(10.00)	1	(25.00)	1	(20.00)	5	(41.67)
High school degree, some college	2	(12.50)	3	(42.86)	5	(50.00)	3	(75.00)	3	(60.00)	5	(41.67)
2-year/4-year college, graduate, professional	10	(62.50)	2	(28.57)	4	(40.00)	0	(0.00)	1	(20.00)	2	(16.67)
Income (US$)
Less than 20.000	0	(0.00)	2	(50.00)	1	(16.67)	1	(100.00)	1	(50.00)	4	(66.67)
20.000-40.000	3	(75.00)	1	(25.00)	3	(50.00)	0	(0.00)	0	(0.00)	2	(33.33)
Beyond 40.000	1	(25.00)	1	(25.00)	2	(33.33)	0	(0.00)	1	(50.00)	0	(0.00)
Static-99 risk
Low	7	(43.75)	1	(14.29)	0	(0.00)	1	(25.00)	2	(40.00)	4	(33.33)
Medium-low	5	(31.25)	3	(42.86)	8	(80.00)	2	(50.00)	1	(20.00)	4	(33.33)
Medium-high	3	(18.75)	3	(42.86)	1	(10.00)	0	(0.00)	2	(40.00)	3	(25.00)
High	1	(6.25)	0	(0.00)	1	(10.00)	1	(25.00)	0	(0.00)	1	(8.33)
Psychiatric history
Yes	3	(18.75)	3	(42.86)	2	(20.00)	1	(25.00)	0	(0.00)	5	(41.67)
Victim gender
Male only	5	(31.25)	3	(42.86)	5	(50.00)	2	(50.00)	0	(0.00)	4	(33.33)
Female only	10	(62.50)	4	(57.14)	5	(50.00)	2	(50.00)	4	(80.00)	8	(66.67)
Male and female	1	(6.25)	0	(0.00)	0	(0.00)	0	(0.00)	1	(20.00)	0	(0.00)
Prior nonsexual conviction
Yes	5	(31.25)	4	(57.14)	4	(40.00)	1	(25.00)	1	(20.00)	8	(66.67)
Prior sexual conviction
Yes	2	(12.50)	3	(42.86)	1	(10.00)	1	(25.00)	1	(20.00)	4	(40.00)

Note. Due to small sample size no statistical analyses were conducted. CSA-W = child sexual abusers working with children.

Between-Group Comparisons

Financial/employment lifestyle stability

Chi-square tests of independence were utilized to consider group differences on financial/employment lifestyle stability variables (see Table 4). Results indicated a significant association between groups and achieved educational level prior to incarceration, χ²(4, n = 1,150) = 80.85, p < .001, Cramer’s V = .187. Post hoc analyses revealed that CSA-W offenders were significantly better educated than both CSA-E and CSA-I offenders. Significant differences were revealed in marital status between groups, χ²(4, n = 1,148) = 136.14, p < .001, Cramer’s V = .267. Post hoc analyses revealed that CSA-I offenders were more likely to be married, compared with CSA-E and CSA-W offenders. Of the small available number of offenders for which data on income prior to incarceration was available, CSA-W offenders had a significantly higher income than CSA-I and CSA-E offenders, χ²(4, n = 468) = 26.76, p < .01, Cramer’s V = .17.

Table 4.

Between-Group Comparison on Categorical Variables.

Variable	Study group						χ²/F	V
	CSA-I		CSA-E		CSA-W
	n	(%)	n	(%)	n	(%)
Financial/Employment Lifestyle Stability
Education
Less than high school, GED, 8th grade	124	(53.91)	489	(56.47)	14	(25.93)	80.85**	.19
High school degree, some college	89	(37.70)	337	(38.91)	21	(38.89)
2-year/4-year college, graduate, professional	17	(7.49)	40	(4.62)	19	(35.19)
Marital status
Single	39	(16.96)	530	(61.34)	24	(44.44)	163.14**	.27
Married	137	(59.57)	186	(21.53)	21	(38.89)
Divorced/widowed/separated	54	(23.48)	148	(17.13)	9	(16.67)
Income prior to incarceration
Less than US$20.000	50	(55.56)	267	(75.21)	9	(39.13)	26.76**	.17
US$20.000-US$40.000	31	(34.44)	70	(19.72)	9	(39.13)
more than US$40.000	9	(10.00)	18	(5.07)	5	(21.74)
Risk/Dangerousness Level
History of psychiatric problems
Yes	74	(33.18)	270	(31.58)	13	(24.07)	1.56	.03
No	149	(66.82)	585	(68.42)	41	(75.93)
Prior nonsexual convictions
Yes	135	(58.95)	610	(70.52)	23	(42.59)	26.06**	.15
No	94	(41.05)	255	(29.48)	31	(57.41)
Prior sexual convictions
Yes	48	(22.02)	228	(27.54)	12	(23.08)	3.00	.05
No	170	(77.98)	600	(72.46)	40	(76.92)
Static-99 risk
Low	89	(67.42)	133	(19.30)	15	(27.78)	131.37**	.27
Medium-low	27	(20.45)	356	(51.67)	23	(42.59)
Medium-high	10	(7.58)	156	(22.64)	12	(22.22)
High	6	(4.55)	44	(6.39)	4	(7.41)
Abuse Opportunity
Children/Step-children
Yes	240	(100.00)	513	(58.90)	29	(53.70)	105.12**	.30
No	0	(0.00)	358	(41.10)	25	(46.30)
Living with anyone at time of index offense
Yes	217	(95.59)	702	(85.19)	44	(86.27)	17.53**	.13
No	10	(4.41)	122	(14.81)	7	(13.73)
Victim Selection
Gender
Male only	19	(8.30)	159	(18.30)	19	(35.19)	29.94**	.11
Female only	198	(86.46)	689	(79.29)	33	(61.11)
Male and female	12	(5.24)	21	(2.42)	2	(3.70)

Note. CSA-I = intra-familial child sexual abusers; CSA-E = extra-familial child sexual abusers; CSA-W = child sexual abusers working with children.

p < .05. **p < .001.

Risk/dangerousness level

As indicated in Table 5, a one-way between-group ANOVA, using the Welch’s adjusted F ratio due to unequal variances, was conducted to explore the impact of CSA groups on age at the time of the index offense. There was a statistically significant difference in age at index offense for the three groups, Welch’s F(2, 132.43) = 14.50, p < .01, η² = .02. Post hoc analyses, using the Games–Howell post hoc analyses due to unequal variances, revealed that CSA-W (M = 38.54, SD = 11.67) and CSA-I offenders (M = 35.80, SD = 9.94) were significantly older at the time of index offense than CSA-E offenders (M = 32.27, SD = 11.79).

Table 5.

Between-Group Comparison on the Continuous Variables.

Variables	Study group						F	p	Post hoc	Eta²	OP
	CSA-I		CSA-E		CSA-W
	M	(SD)	M	(SD)	M	(SD)
Age index offense (years)	35.80	(9.94)	32.27	(11.79)	38.54	(13.67)	14.50	<.01	2 < 1,3	0.02	.99
Age of first victim (years)	9.11	(3.77)	11.39	(3.41)	12.53	(3.38)	39.32	.01	1 < 2,3	0.07	1.0
Number of children	3.18	(1.55)	2.12	(1.54)	2.18	(1.83)	36.72	<.01	1 > 2,3	0.08	1.0
Number of children living with	2.28	(1.42)	0.84	(1.15)	0.57	(1.3)	93.78	<.01	1 > 2,3	0.20	1.0
Total number of victims	1.26	(0.53)	1.28	(0.71)	1.46	(0.82)	1.50	.22	ns	0.03	.42

Note. CSA-I = intra-familial child sexual abusers; CSA-E = extra-familial child sexual abusers; CSA-W = child sexual abusers working with children; OP = observed power.

Chi-square tests were used to evaluate group differences on the remaining risk/dangerousness variables (see Table 4). Results revealed no significant differences between groups on the history of psychiatric problems, though, CSA-I and CSA-E offenders tended to display a slightly higher frequency of psychiatric problems than CSA-W offenders. Of the cases for which data regarding prior nonsexual convictions was available, CSA-E offenders were significantly more likely to have been convicted of a nonsexual offense prior to index offense than CSA-I and CSA-W offenders, χ²(2, n = 1,148) = 26.06, p < .001, Cramer’s V = .151. There were no significant differences in prior sexual convictions. A significant association was found between groups and Static-99 total risk scores, χ²(6, n = 875) = 131.37, p < .001, Cramer’s V = .274. Compared with CSA-E and CSA-W offenders, CSA-I offenders were significantly more often categorized as low risk (n = 89, 67.4%).

Abuse opportunity

As shown in Table 4, between-group comparisons showed significant differences with regard to having children or step-children, χ²(2, n = 1,155) = 105.12, p < .001, Cramer’s V = .30. Post hoc analyses revealed that CSA-I offenders were significantly more likely to have children/step-children compared with CSA-E and CSA-W offenders. Significant differences were also found between groups on those who were living with someone χ²(2, n = 1,102) = 17.53, p < .001, Cramer’s V = .13, with those in the CSA-I group more likely to be living with someone compared with those in the CSA-E and CSA-W groups. Of those offenders who did have their own biological/step-children, a one-way between-group ANOVA (see Table 5) revealed a significant difference in the average number of children/step-children between groups, F(2, 833) = 36.72, p < .01, η²= .08. Post hoc analyses, using the Scheffé post hoc criterion for significance revealed that CSA-I offenders (M = 3.18, SD = 1.55) had significantly more children/step-children than CSA-E (M = 2.12, SD = 1.54) and CSA-W (M = 2.18, SD = 1.83) offenders. A one-way between-group ANOVA, using a Welch’s adjusted F ratio due to unequal variances, explored the association between the CSA groups and the number of children with whom offenders were living at the time of the offense. There was a statistically significant difference between offender group and the number of children in the home, Welch’s F(2, 109.83) = 93.78, p < .01, η²= .2. Post hoc analyses, using the Games–Howell post hoc analyses due to unequal variances, revealed that CSA-I (M = 2.28, SD = 1.42) offenders lived with significantly more children than CSA-E (M = 0.84; SD = 1.15) and CSA-W (M = .57, SD = 1.13) offenders.

Victim selection

Table 4 indicates CSA-W offenders were significantly more likely χ²(4, n = 1,152) = 29.94, p < .05 to have only male victims compared with CSA-E and CSA-I offenders. When analyzing the age of the first victim, a one-way between-group ANOVA revealed that CSA-I offenders (M = 9.14, SD = 3.80) had significantly younger victims than CSA-E and CSA-W offenders, F(2, 1155) = 39.24, p < .05 (see Table 5). Post hoc analyses using the Scheffé post hoc criterion for significance revealed that CSA-W (M = 12.53, SD = 3.38) and CSA-E (M = 11.45, SD = 3.52) offenders did not differ significantly from each other with regard to mean age of first victim, p = .09. A one-way between-group ANOVA revealed that the total number of victims did not vary by offender–victim relationship, p = .22 (see Table 5).

Group membership

A multinomial logistic regression was used to analyze which variables predicted membership into the CSA-I, CSA-E, and CSA-W groups. The reference category for the outcome variables was CSA-W; each of the other two groups was compared with this reference group. The variables that were found to differentiate between the groups were included in the model and presented in Table 6. The Pearson Goodness-of-fit statistic was nonsignificant (χ² = 1557.00, df = 1502, p = .16), suggesting a good fit for the data. The model predicted 42.1% of those in the CSA-I group correctly, 95.4% of those in the CSA-E group correctly and 21.3% of those in the CSA-W group correctly.

Table 6.

Multinominal Logistic Regression.

Variable	CSA-I		CSA-E
Variable	OR [95% CI]	SE	OR [95% CI]	SE
Victim Age	0.75 [.66, .84]**	0.06	0.90 [.81, 1.00]	0.05
Age at Offense	0.95 [.91, .98]**	0.02	0.97 [.94, 1.00]	0.16
Children/Step-children
Yes	0.10 [.03, .34]**	0.63	0.53 [.22, 1.32]	0.46
No	1		1
Victim Gender
Male	0.92 [.07, 11.84]	1.30	1.21 [.12, 12.60]	1.19
Female	1.39 [.12, 16.6]	1.27	3.01 [.30, 30.61]	1.18
Male and Female	1		1
Prior Nonsexual History
Yes	0.41 [.17, 1.00]	0.46	0.45 [.21, .69]*	0.39
No	1		1
Static-99 risk
Low	3.71 [.43, 32.09]	1.10	1.26 [.19, 8.30]	0.96
Medium-low	0.48 [.07, 3.33]	0.99	0.91 [.18, 4.70]	0.84
Medium-high	0.29 [.04, 2.21]	1.05	0.72 [.13, 3.89]	0.86
High	1		1
Education
Less than high school	18.56 [4.82, 71.46]**	0.69	10.65 [4.35, 26.09]**	0.46
High school degree	14.44 [3.77, 55.28]**	0.69	6.48 [2.72, 15.41]**	0.44
College/Graduate	1		1
Marital Status
Single	0.84 [.22, 3.17]	0.68	1.60 [.50, 5.13]	0.60
Married	1.00 [.32, 3.08]	0.58	0.54 [.19, 1.49]	0.52
Divorced/Widowed	1		1
Income
Less than US$20,000	1.26 [.46, 3.49]	0.52	1.44 [.60, 3.43]	0.44
US$20,000-US$40,000	1.01 [.34, 3.02]	0.56	0.82 [.32, 2.08]	0.48
More than US$40,000	1		1
Living with anyone
Yes	1.27 [.29, 5.59]	0.78	2.14 [.69, 6.64]	0.58
No	1		1

Note. Reference Group: CSA-I = intra-familial child sexual abusers; CSA-E = extra-familial child sexual abusers; OR = Odds Ratio; 95% CI = confidence interval.

p < .05. **p < .001.

The first column in Table 6 has the outcome of CSA-I compared with CSA-W offenders. The results suggest that if the victims are younger (OR = .90), the offender is younger (OR = .95), has children (OR = .10), and a lower level of education (OR = 18.56), this significantly increases the probability that the offenders will be classified as a CSA-I offender. Victim gender, prior nonsexual history, marital status, and pre-incarceration income were not significant predictors of CSA-I classification. The second column in Table 6 has the outcome of CSA-E offenders compared with CSA-W offenders. The results suggest that if the offender has a prior nonsexual history (OR = .45) and a lower level of education (OR = 10.65), they were significantly more likely to be classified as CSA-E offenders.

Discussion

This study provided a more comprehensive description of the types of employment settings within which sexual abuse occurs, as well as the different variables associated with the various types of employment. In addition, a more heterogeneous group of CSA-W offenders was compared with CSA-I and CSA-E offenders on factors such as financial/employment lifestyle stability, risk/dangerousness level, abuse opportunity, and victim selection. Finally, key variables were analyzed to evaluate what variables predicted membership into the CSA-I, CSA-E, and CSA-W groups. Overall, our results support previous research (Sullivan & Beech, 2004; Sullivan et al., 2011; Turner, Rettenberger, Lohmann, Eher, & Briken, 2014) suggesting that CSA-W offenders tend to be better educated, have a higher pre-incarceration income, have fewer nonsexual convictions prior to incarceration, and are more likely to abuse male post-pubescent children compared with CSA-I and CSA-E offenders.

This study, which was the first to examine CSA-W within a wide variety of professions and voluntary activities, demonstrates that CSA does in fact occur in a wide range of work environments. Notably, individuals who worked in an ecclesiastic setting did not constitute our largest group of CSA-W offenders. Instead, individuals who worked as teachers or were employed by an educational organization formed the largest CSA-W group, closely followed by those individuals who worked in a setting that does not necessarily feature contact with children. This finding sheds a different light on the issue of CSA within work places. Despite the fact that some work environments do not by their nature explicitly include contact with children, we found offenders in our sample to have committed offenses while at work in more traditional (i.e., noninstitutional) settings. In institutional settings, our findings showed that those in a professional caretaking role—as well as those not necessarily in a position of trust—perpetrated abuse. For example, in school settings, not only teachers but also other school-related personnel, such as maintenance workers and school bus drivers were shown to abuse their employment position to facilitate an offense with a child. The fact that about half of the ecclesiastic sample was formed of volunteers or bus drivers instead of priests, further exemplifies that attributions of trust and authority, which are most often associated with professions such as teachers and clergy (Wurtele, 2012), are not necessary for work-related CSA to occur. Furthermore, findings of CSA-W in non-child-centric settings, such as security guard, bus driver and store workers highlight the heterogeneity of environments in which offenders may use their workplace to target children. In fact, this group demonstrated the second highest number of offenders, exceeding traditional settings such as ecclesiastic and child-care services, as well as child-centric settings. While the within-group descriptions are informative, the small sample size prevented within-group analyses, necessitating the need for further research in this area.

To assess whether CSA-W offenders constitute a unique group, they were compared with CSA-I and CSA-E offenders on relevant variables. Overall, CSA-W offenders had more lifestyle stability than CSA-I and CSA-E offenders in that they were better educated and had a higher income prior to incarceration. This is likely because these offenders abused children in the context of their employment, which meant that they were employed and earning a steady income when they committed the abuse. Furthermore, many jobs that involve work with children require a minimum of high school education, and some jobs such as teachers and pediatricians require advanced degrees.

Taking a closer look at the overall risk/dangerousness level, the present study revealed that CSA-W offenders have fewer prior nonsexual convictions compared with CSA-E (significant difference) and CSA-I (trend) offenders. Contrary to previous studies (Sullivan & Beech, 2004; Sullivan et al., 2011; Turner, Rettenberger, Lohmann, Eher, & Briken, 2014), there were no significant group differences on prior sexual convictions. These findings may be a result of specialized offending patterns that keep CSA-W offenders from versatile criminal backgrounds (i.e., high number of prior nonsexual convictions). However, consistent with their professional roles and responsibilities, it is often challenging to obtain employment with prior criminal history.

The present research revealed that CSA-W and CSA-I offenders were significantly older at the time at the index offense compared with CSA-E offenders. Past research on age and CSA has been mixed. Turner and Colleagues (2014) found no differences in age between CSA-W and non-CSA-W offenders. However, Turner and Briken (2015) found that their CSA-W group was older than the non-CSA-W group. These authors attributed this age difference to possible delays in disclosure due to use of grooming strategies and the influence of their position. They also highlighted that this CSA-W sample (similar to the present study) consisted of professional positions for which advanced degrees and training were often required, thus delaying access to potential victims. These mixed findings suggest that more research on the impact of offender age and its relationship to offending in CSA-W offending samples is needed.

When these risk factors are considered altogether, the lack of prior nonsexual convictions and older age indicate the possibility of lower risk for the CSA-W group. However, there were limited indicators of group differences on level of risk. In fact, there were no group differences on the proportion of high and medium risk offenders on the Static-99; however, the CSA-I group had a higher proportion of low-risk offenders than the CSA-W and CSA-E groups.

When abuse opportunity was considered, it was not surprising that CSA-I offenders were significantly more likely to have children/step-children, have a higher number of children, and to live with more children than the other offender groups. Interestingly, more than 50% of CSA-E and CSA-W offenders also had children/step-children of their own, and there were no group differences in living with anyone, leaving questions about why these offenders did not take advantage of the opportunity to sexually abuse their own children. Instead, CSA-W and CSA-E offenders sought out abuse opportunities within their employment or other external environments, suggesting that CSA-W and CSA-E offenders follow preferential, rather than opportunistic offending patterns.

Finally, victim selection analyses indicated that all groups abused females more than males. However, consistent with findings from previous research conducted on clergy sex offenders (Langevin, Curnoe, & Bain, 2000; Turner, Rettenberger, Lohmann, Eher, & Briken, 2014), the CSA-W offenders in the present sample were more likely than CSA-I and CSA-E offenders to abuse male victims. In addition, CSA-W offenders tended to choose victims above 12 years of age. These results are consistent with findings of Sullivan and colleagues (2011) that CSA-W offenders were more likely to offend against post-pubescent children. This trend may be the result of abuse opportunity within certain employment settings, rather than choice. For example, some of the employment contexts, such as athletic coach, taxi driver or candy shop owner would mean CSA-W offenders would come in contact with older children. Additional research regarding sexual preference using self-report of plethysmography could help to distinguish whether victim selection was based on opportunity or preference.

Taken together, the findings of this study suggest that CSA-W offenders cannot be exclusively characterized as CSA-I offenders, nor as CSA-E offenders. While there are similarities between groups there are also key differences. It should be noted the variables in our study were only able to classify 21% of offenders in the CSA-W category. Thus, it is possible that in extending the breadth of the CSA-W category, we failed to take into consideration that there may be subgroupings within the larger CSA-W group as suggested by Turner and Briken (2015). The research on CSA-W offenders in general is still in its relative naissance, and the majority of the work done to date has focused on clergy offenders; thus, the findings of this highlight that we still need to gain a better understanding of CSA-W offenders using larger or combined data sets so that we can better assess both within and between-group differences.

There are several limitations of the present study. First, archival data for this study was not primarily collected for the purpose of comparing CSA-W, CSA-I, and CSA-E offenders. Despite the resulting small sample size of CSA-W offenders, the present sample size of CSA-W offenders was comparable to sample sizes used in previous between-group studies (Sullivan & Beech, 2004; Sullivan et al., 2011; Turner, Rettenberger, Lohmann, Eher, & Briken, 2014). In addition, despite following a consensus approach in which independent raters grouped the CSA-W occupations into groups based on the nature of institution rather than on the nature of occupation, this grouping process was still imperfect. We were also unable to compare groups on previously impactful variables, such as grooming strategies, psychopathic tendencies (Howe, Falkenbach, & Massey, 2014) deviant sexual interest and IQ, as those variables were not measured as part of the archival data set used for this study. Specifically, IQ is one of the strongest discriminators of CSA-W offenders from other CSA populations (Haywood, Kravitz, Grossman, Wasyliw, & Hardy, 1996; Langevin et al., 2000; Sullivan & Beech, 2004). Educational level was higher among CSA-W offenders in the present sample, and as rough proxy for intelligence, this further corroborates previous research.

Considering the findings and limitations together, the present study has several implications for future sex offender research, prevention programs, and clinical practices. Future research should attempt to expand the focus of CSA-W offenders, acknowledging the fact that CSA does occur in less expected and less institutionalized employment settings. The CSA cases within school or ecclesiastic settings (in which some form of authoritative and trustful attributions are given) and employment settings of convenience, such as store owners and postal workers, should be compared with regard to underlying abuse motivations.

In many ways, the CSA-W offenders fail to fit the stereotypical conceptions of sexual offenders. The CSA-W individuals often present as educated and socially stable, which may in fact cause individuals to falsely perceive CSA-W offenders as safe and trustworthy. The findings of this study suggest that at least some of the individuals that offended within the context of employment settings had a previous history of sexual offending. While some of these cases occurred decades ago when employment screening practices were less stringent—it is unclear how present day employment suitability screenings may or may not identify such individuals. For example, today anyone who works with children must undergo a criminal background check to prevent those with previous sexual abuse convictions from working with children. However, such screenings are likely not necessary for an individual to work at a convenience store which children may frequent. Thus, the findings of this study suggest that it would likely be prudent to conduct criminal background checks with any employee that may incidentally come into contact with children through the nature of their work.

Following the widely publicized reports of CSA within the church, the Catholic Church issued the Charter for the protection of children and young people, in which Article 12 demands that all dioceses and eparchies introduce a training program aimed at educating children, youth, parents, ministers, educators, volunteers, and others about CSA (United States Conference of Catholic Bishops Safe Environment Work Group, 2007); however, such trainings are not mandated in many other employment settings in which employees may have incidental contact with children. The numerous variety of non-child-centric professions within our CSA-W sample as well as within school settings and caretaking services therefore suggests that those kind of training programs should also be implemented in other institutions and organizations, even in those where CSA has received limited awareness.

Finally, with regard to clinical practices, our study provides support for the issue of applying commonly used risk assessment tools within the CSA-W offender population (Perillo, Laake, & Calkins, 2015; Turner, Rettenberger, Lohmann, Eher, & Briken, 2014). In fact, as our study concludes that CSA-W offender samples can neither be exclusively characterized as CSA-I offenders, nor as CSA-E offenders, clinicians should interpret results of risk assessment tools cautiously. As only one study specifically investigated the applicability of commonly used risk assessment tools in CSA-W offenders (Turner, Rettenberger, Lohmann, Eher, & Briken, 2014), more research as well as clinical practices need to examine this specific issue.

Footnotes

Acknowledgements

The authors thank the New Jersey Department of Corrections and New Jersey Department of Human Services for their support in providing access to files.

Authors’ Note

The opinions, findings, conclusions, and recommendations expressed are those of the authors and do not necessarily reflect those of the institutions that supported this research.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported in part by Grant NIJ 2007-IJ-CX-0037 from the National Institute of Justice, Office of Justice Programs, U.S. Department of Justice.

References

Bernados, Cari, children’s RIGHTS ALLIANCE,IAYPIC, One In Four, ISPCC, . . . RCC. (2010). Saving childhood Ryan: Fighting for children together. Retrieved from http://www.barnardos.ie/assets/files/Advocacy/SAVING%20CHILDHOOD%20RYAN%20final%20report%20_2_.pdf

Boy Scouts of America. (n.d.). Youth protection. Retrieved from http://www.scouting.org/Training/YouthProtection.aspx

Boyle

(2014). How youth-serving organizations enable acquaintance molesters. Journal of Interpersonal Violence, 29, 2839-2848.

Calkins

Fargo

Jeglic

Terry

(2015). Blessed be the children: A case-control study of sexual abuse in the Catholic Church. Behavioral Sciences & the Law, 33, 580-594.

Currie

Widom

C. S.

(2010). Long-term consequences of child abuse and neglect on adult economic well-being. Child Maltreatment, 15, 111-120.

Death

(2015). Bad apples, bad barrel: Exploring institutional responses to child sexual abuse by catholic clergy in Australia. International Journal for Crime, Justice and Social Democracy, 4, 94-110.

Department of Justice and Equality. (2009). Commission of Investigation Report into the Catholic archdiocese of Dublin. Dublin, Republic of Ireland: Author. Retrieved from http://www.justice.ie/en/JELR/Pages/PB09000504

Department of Law and Public Safety, & Farmer, J. (2001). The criminal law digest—State of New Jersey. Retrieved from http://www.state.nj.us/oag/dcj/pdfs/criminal_law_digest-orginal.pdf

Falkenbach

D. M.

(2004). The subtypes of psychopathy and their relationship to hostile and instrumental aggression (Doctoral dissertation). University of South Florida.

10.

Finkelhor

(1994). The international epidemiology of child sexual abuse. Child Abuse & Neglect, 18, 409-417.

11.

Gallagher

(2000). The extent and nature of known cases of institutional child sexual abuse. British Journal of Social Work, 30, 795-817.

12.

Hanson

R. K.

Morton-Bourgon

K. E.

(2005). The characteristics of persistent sexual offenders: A meta-analysis of recidivism studies. Journal of Consulting and Clinical Psychology, 73, 1154-1163.

13.

Hanson

R. K.

Thornton

(2000). Improving risk assessments for sex offenders: A comparison of three actuarial scales. Law and Human Behavior, 24, 119-136.

14.

Haywood

T. W.

Kravitz

H. M.

Grossman

L. S.

Wasyliw

O. E.

Hardy

D. W.

(1996). Psychological aspects of sexual functioning among cleric and non-cleric alleged sex offenders. Child Abuse & Neglect, 20, 527-536.

15.

Howe

Falkenbach

Massey

(2014). The relationship among psychopathy, emotional intelligence, and professional success in finance. International Journal of Forensic Mental Health, 13(4), 337-347.

16.

John Jay College Research Team. (2004). The nature and scope of sexual abuse of minors by Catholic priests and deacons in the United States, 1950–2000. Washington, DC: United States Conference of Catholic Bishops.

17.

Jones

L. M.

Finkelhor

Kopiec

(2001). Why is sexual abuse declining? A survey of state child protection administrators. Child Abuse & Neglect, 25, 1139-1158.

18.

Langevin

Curnoe

Bain

(2000). A study of clerics who commit sexual offenses: Are they different from other sex offenders? Child Abuse & Neglect, 24, 535-545.

19.

Leclerc

Proulx

McKibben

(2005). Modus operandi of sexual offenders working or doing voluntary work with children and adolescents. Journal of Sexual Aggression, 11, 187-195.

20.

Magalhães

Taveira

Jardim

Santos

Matos

Santos

(2009). Sexual abuse of children. A comprehensive study of intra and extra-familial cases. Journal of Forensic and Legal Medicine, 16, 455-459.

21.

Mercado

C. C.

Jeglic

E. L.

Markus

K. A.

(2011). Sex offender management, treatment, and civil commitment: An evidence based analysis aimed at reducing sexual violence. Washington, DC: U.S. Department of Justice.

22.

Mercado

C. C.

Tallon

J. A.

Terry

K. J.

(2008). Persistent sexual abusers in the Catholic Church: An examination of characteristics and offenses patterns. Criminal Justice and Behavior, 35, 629-642.

23.

Moulden

H. M.

Firestone

Kingston

D. A.

Wexler

A. F.

(2010). A description of sexual offending committed by Canadian teachers. Journal of Child Sexual Abuse, 19, 403-418.

24.

Moulden

H. M.

Firestone

Wexler

A. F.

(2007). Child care providers who commit sexual offences. International Journal of Offender Therapy and Comparative Criminology, 51, 384-406.

25.

Pereda

Guilera

Forns

Gómez-Benito

(2009). The international epidemiology of child sexual abuse: A continuation of Finkelhor (1994). Child Abuse & Neglect, 33, 331-342.

26.

Perillo

A. D.

Laake

A. L. W.

Calkins

(2015). Understanding sexually abusive clergy as a unique offender subgroup: Risk-based comparisons across the course of offending. International Journal of Forensic Mental Health, 16, 58-68.

27.

Perillo

A. D.

Mercado

Terry

K. J.

(2008). Repeat offending, victim gender, and extent of victim relationship in Catholic Church sexual abusers. Criminal Justice and Behavior, 35, 600-614.

28.

Shakeshaft

(2004). Educator sexual misconduct: A synthesis of existing literature (U.S. Department of Education Document No. 2004-09). Washington, DC: U.S. Department of Education.

29.

Sullivan

Beech

A. R.

(2004). A comparative study of demographic data relating to intra- and extra-familial child sexual abusers and professional perpetrators. Journal of Sexual Aggression, 10, 39-50.

30.

Sullivan

Beech

A. R.

Craig

L. A.

Gannon

T. A.

(2011). Comparing intra-familial and extra-familial child sexual abusers with professionals who have sexually abused children with whom they work. International Journal of Offender Therapy and Comparative Criminology, 55, 56-74.

31.

Terry

K. J.

Mercado

C. C.

Perillo

(2008). Priests who abuse and were abused: Understanding victimization in the Catholic Church. Victims & Offenders, 3, 412-422.

32.

Thornton

(2002). Constructing and testing a framework for dynamic risk assessment. Sexual Abuse: A Journal of Research and Treatment, 14, 139-153.

33.

Turner

Briken

(2015). Child sexual abusers working with children—Characteristics and risk factors. Sexual Offender Treatment, 10(1),1-12.

34.

Turner

Rettenberger

Lohmann

Eher

Briken

(2014). Pedophilic sexual interests and psychopathy in child sexual abusers working with children. Child Abuse & Neglect, 38, 326-335.

35.

Turner

Rettenberger

Yoon

Klein

Eher

Briken

(2014). Risk assessment in child sexual abusers working with children. Sexual Abuse: A Journal of Research and Treatment, 28, 572-596.

36.

United States Conference of Catholic Bishops Safe Environment Work Group. (2007). Report on safe environment training. Origins, 37(1), 1-6.

37.

Wurtele

S. K.

(2012). Preventing the sexual exploitation of minors in youth-serving organizations. Children and Youth Services Review, 34, 2442-2453.