558941 research-article2014
SAXXXX10.1177/1079063214558941Sexual AbuseGriffee et al.
Article
Etiological Risk Factors for Sibling Incest: Data From an Anonymous Computer-Assisted Self-Interview
Sexual Abuse: A Journal of Research and Treatment 1–40 © The Author(s) 2014 Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/1079063214558941 sax.sagepub.com
Karen Griffee1, Sam Swindell2, Stephen L. O’Keefe3,4, Sandra S. Stroebel3, Keith W. Beard4, Shih-Ya Kuo2, and Walter Stroupe2
Abstract Retrospective data from 1,821 women and 1,064 men with one or more siblings, provided anonymously using a computer-assisted self-interview, were used to identify risk factors for sibling incest (SI); 137 were participants in SI. In order of decreasing predictive power, the risk factors identified by the multiple logistic regression analysis included ever having shared a bed for sleeping with a sibling, parent–child incest (PCI), family nudity, low levels of maternal affection, and ever having shared a tub bath with a sibling. The results were consistent with the idea that SI in many families was the cumulative result of four types of parental behaviors: (a) factors that lower external barriers to sexual behavior (e.g., permitting co-sleeping or co-bathing of sibling dyads), (b) factors that encourage nudity of children within the nuclear family and permit children to see the parent’s genitals, (c) factors that lead to the siblings relying on one another for affection (e.g., diminished maternal affection), and (d) factors that eroticize young children (e.g., child sexual abuse [CSA] by a parent). Thirty-eight of the 137 SI participants were participants in coerced sibling incest (CSI). In order of decreasing predictive power, risk factors for CSI identified by multiple logistic regression analysis included ever having shared a bed for sleeping with a brother, PCI, witnessing parental physical fighting, and family nudity. SI was more likely to have 1Concord
University, Athens, WV, USA Virginia State University, Institute, WV, USA 3Marshall University Graduate College, South Charleston, WV, USA 4Marshall University, Huntington, WV, USA 2West
Corresponding Author: Karen Griffee, Associate Professor of Psychology, Division of Social Sciences, Concord University, P.O. Box 1000, Athens, WV 24712, USA. Email: [email protected]
Downloaded from sax.sagepub.com at Bobst Library, New York University on November 14, 2015
2
Sexual Abuse
been reported as CSI if the sibling had touched the reporting sibling’s genitals, and less likely to have been reported as CSI if the siblings had shared a bed. Keywords sibling incest, brother–sister incest, brother–brother incest, sister–brother incest, sister–sister incest, maternal affection, nudity, parental child sexual abuse, risk factors Convenience samples have provided estimates of 15% for all sexual contacts between siblings in girls (Finkelhor, 1981; Greenwald & Leitenberg, 1989) and 10% in boys (Finkelhor, 1981). Atwood (2007) obtained an estimated prevalence of 17% for brother–sister incest (BSI), but after excluding cases of sexual contact with a relative that were wanted and with a peer and including cases in the denominator that were not at risk because they had no siblings, Russell (1986) obtained estimates of only 2.0% for BSI and 0.3% for sister–sister incest (SSI). BSI can have problematic long-term effects on the sisters (Stroebel, O’Keefe, Beard, Kuo, Swindell, & Stroupe, 2013; Tidefors, Arvidsson, Ingevaldson, & Larsson, 2010). Sister–brother incest (SBI) can also have problematic long-term effects on the brothers (O’Keefe et al., 2014). SSI (Stroebel, O’Keefe, Griffee, Kuo, Beard, & Kommor, 2013) and brother–brother incest (BBI; Beard et al., 2013) also have problematic effects on the involved siblings. The known adverse effects of all four types of sibling incest (SI) emphasize the importance of prospective prevention. Claims by some authors based on purported tests of the 19th-century theory of Westermarck (e.g., Antfolk, Karlsson, Bäckström, & Santtila, 2012; Antfolk, Lieberman, & Santtila, 2012; Bevc & Silverman, 1993, 2000; Lieberman, Tooby, & Cosmides, 2003, 2007; see Silverman & Bevc, 2005, for review) that there is an innate biological resistance to incestuous sexual relationships because of close association in the nuclear family, disgust at the thought of incest, or biological relatedness are controversial. The theory is at odds with recent evidence showing that individuals who grew up in the communal kibbutz education system were actually often attracted to their peers and rarely developed a sexual aversion toward them (Shor & Simchai, 2009). Furthermore, the theory of Westermarck is at odds with historical evidence that incestuous marriages became common place in societies that tolerated brother–sister marriages (Roman Egypt in the first to third centuries AD) or promoted brother–sister and father–daughter marriages (Persia from the fifth century BC to the seventh century AD; Scheidel, 1996). Based on papyrus census records from Roman Egypt from first to third centuries AD, it has been estimated that brother–sister marriages occurred in about one third of families with marriageable opposite-sex siblings with the brother older than the sister (Scheidel, 1996). Conditioned inhibition of sexual behavior between the siblings (for reviews of conditioned sexual preferences, see Hoffman, 2012; Pfaus et al., 2012) and critical period learning of sexual behavior (Griffee, O’Keefe, Beard, Young, Kommor, & Linz, 2014; Griffee, O’Keefe, Beard, Young, Kommor, Linz, Swindell, & Stroebel, 2014) working together provide an alternative
Downloaded from sax.sagepub.com at Bobst Library, New York University on November 14, 2015
3
Griffee et al.
explanation for inhibition of marriages between children reared together attributed to the Westermarck effect without the need to invoke any special evolved mechanism to reduce mating between siblings. Such conditioning would be predicted theoretically by parental efforts to prevent sex between the siblings. The term “negative imprinting” has been used to describe the effect (Shepher, 1971). Programs currently in place to prevent CSA have focused on teaching children to protect themselves against CSA (see Babatsikos, 2010; Hunt & Walsh, 2011; Kenny, Capri, Thakkar-Kolar, Ryan, & Runyon, 2008; Kenny & Wurtele, 2012 for reviews). Unfortunately, by the time children are old enough to assimilate information about how to protect themselves against SI, they may already have become victims. SI often starts when the siblings are very young, and SI is often experienced by both the involved siblings as voluntary (Beard et al., 2013; Caffaro & Conn-Caffaro, 1998; O’Keefe et al., 2014; Stroebel, O’Keefe, Beard, et al., 2013; Stroebel, O’Keefe, Griffee, et al., 2013). Another, potentially more fruitful approach to prevention would be to identify etiological risk factors for SI (using valid epidemiological methods based on statistically comparing data from SI cases with that from valid controls) and then make this information available to families so that preventative measures can be implemented within families before the incest has occurred. Such information could be presented to both mothers and fathers as empowering families in the prevention of incest. Despite a computerized search in November 2012 of the Medline, PsychINFO, and PsychARTICLES databases using Academic Search Premier for the term “incest” in the abstract, no valid epidemiological studies that used statistical techniques, SI cases, and controls were identified. Due to the dearth of data in the more recent literature on risk factors for SI, we were forced to turn to the older literature. Other clinical studies of SI participants and studies of normal sexual behaviors of children and adolescents have suggested a number of possible risk factors for SI not derived from valid epidemiological studies that used statistical techniques. Furthermore, definitions of SI differed among the many (non-epidemiologic) studies of SI. Some differentiated SI from normative “sex play” between siblings (e.g., Bank & Kahn, 1982; Caffaro & ConnCaffaro, 1998; Russell, 1986) whereas others (including the present article) included any kind of sexual behavior between siblings within the definition of SI (e.g., Araji, 1997; Beard et al., 2013; Finkelhor, 1980, 1981; Stroebel, O’Keefe, Beard, et al., 2013; Stroebel, O’Keefe, Griffee, et al., 2013). There are at least four levels where professionals who treat incest offenders might identify risk factors for SI, and to avoid confusion, we will identify and name the four most obvious levels. First, there are the etiological factors that actually predispose or protect against the occurrence of SI. Second, there are the intra-familial factors that increase or decrease the likelihood that SI will be detected by the siblings’ parents. Third, there are the factors that increase or decrease the chance that the parents will actually report the case of SI to the authorities. Finally, there are the factors that increase or decrease the likelihood that the SI offenders will re-offend. When it comes to preventing the occurrence of SI or treating incest offenders, it is the etiological factors that are the most useful to know (Offord & Kraemer, 2000).
Downloaded from sax.sagepub.com at Bobst Library, New York University on November 14, 2015
4
Sexual Abuse
One systematic study that made an attempt to identify risk factors for SI (Caffaro & Conn-Caffaro, 1998) was based on face-to-face clinical interviews of 29 adult SI survivors referred for study by therapists and facilities specializing in CSA, but because there was no control group or statistical analysis, identification of the risk factors was based solely on the authors’ clinical opinions. Furthermore, no measures of odds ratios (ORs) or adjusted relative risk were possible. Caffaro and Conn-Caffaro (1998) suggested risk factors for sibling assault and incest trauma included parental unavailability, a sexualized environment in the home, lack of sex education, lack of parental supervision, lack of other supportive relationships, siblings standing in for absent parents, power imbalance among siblings (due to developmental, physical, or intellectual differences), unclear individual and generational boundaries, unusually frequent and ready access to a victim (e.g., room sharing), a victim’s dependence on an older and more powerful sibling, a history of victimization, and parental favoritism. When affection from parents is scarce or unavailable, siblings often turn to one another for affection, and this can lead to incestuous behaviors (Bank & Kahn, 1982). Families that were more permissive about parental and child nudity and that modeled greater physical intimacy reported observing more sexual behaviors in their young children than families that emphasized modesty and that were less physically intimate (Friedrich, Grambsch, Broughton, Kuiper, & Beilke, 1991; Sandfort & CohenKettenis, 2000). A few adults anecdotally looking back at their own experiences with SI have identified bed sharing, being bathed together with a sibling, skinny dipping with siblings, and casual household nudity as experiences that led to their involvement in SI (Symonds, Mendoza, & Harrell, 1981). Young victims of incest with a parent or sibling often become eroticized and actively seek sex from other children and alternative adult care givers (Araji, 1997; Beard et al., 2013; O’Keefe et al., 2014; Salter et al., 2003; Stroebel et al., 2012; Stroebel, O’Keefe, Beard et al., 2013; Stroebel, O’Keefe, Griffee, et al., 2013; Yates, 1982, 1990), implying that being a victim of incest perpetrated by a parent or sibling could be a risk factor for SI. As reviewed by Araji (1997), there was an extensive literature that also contains clinical-opinion-based “risk factors” for children and adolescents becoming sexually aggressive children, and we were able to find one case-control epidemiological study of risk factors for becoming a child, adolescent, or adult CSA perpetrator given the fact that a given child or adolescent (himself) had been the victim of CSA (Salter et al., 2003). It is possible that some of the factors identified by Araji (1997) and Salter et al. (2003) were second-, third-, or fourth-level risk factors that increased the likelihood of the SI coming to the attention of the authorities—rather than first-level etiological risk factors for SI—because the data were obtained from cases that had come to the attention of the authorities. SI, like father–daughter incest (FDI), is seldom reported to the police (Atwood, 2007; Carlson, Maciol, & Schneider, 2006; Russell, 1986). Furthermore, SI is seldom reported to parents either, principally because each sibling believes that she or he will be punished (Atwood, 2007; Carlson et al., 2006; Laviola, 1992; Russell, 1986). This means the best way to accurately identify cases of SI is to obtain the data from the siblings themselves (Caffaro & Conn-Caffaro, 1998) in a way that permits them to
Downloaded from sax.sagepub.com at Bobst Library, New York University on November 14, 2015
5
Griffee et al.
disclose the information without concern that the disclosure could harm either their siblings or themselves. Suitable controls must be identified in the same population from which the SI cases were obtained. The present study was part of a larger study on human sexuality presented to participants as a “cradle to the grave” study on human sexuality. The method of recruiting participants employed in the present study minimized factors that might select for or against women and men who had been involved in SI. Obtaining the data from unselected individuals (some of whom happened to disclose to the computer questionnaire that that they had been involved in SI) maximized the chances that etiological risk factors for SI would be identified. We tested the following hypotheses: Hypothesis 1: Some behaviors within nuclear families constitute risk factors for SI. As detailed above, based on the few applicable studies in the literature, we were able to identify some potential etiological risk factors for SI. 1. Bed sharing (Symonds et al., 1981). 2. Siblings being bathed together (Symonds et al., 1981). 3. Parentally approved nuclear family nudity behaviors (Friedrich et al., 1991; Sandfort & Cohen-Kettenis, 2000). 4. Low parental affection (Bank & Kahn, 1982). 5. Eroticization resulting from being a victim of parental incest (Araji, 1997; Salter et al., 2003; Stroebel et al., 2012; Stroebel, O’Keefe, Beard, et al., 2013; Yates, 1982, 1990). 6. A sexualized family environment resulting from parental sexual behaviors not entirely hidden from the children (Araji, 1997; Caffaro & Conn-Caffaro, 1998). 7. Witnessing violence within the nuclear family (Araji, 1997; Salter et al., 2003). Because many cases of SI never come to the attention of either the parents or the authorities, we also tested the following hypotheses: Hypothesis 2: Certain behaviors of the incestuous sibling dyad or conditions within nuclear families increase the chances that the SI will be identified. Hypothesis 3: Some behaviors within nuclear families increase the chances that SI will occur between two adult siblings.
Method Participants All 1,821 female participants and all 1,064 male participants were above the age of 18 and gave informed consent using forms approved by the relevant institutional review
Downloaded from sax.sagepub.com at Bobst Library, New York University on November 14, 2015
6
Sexual Abuse
board. The participants were recruited from a population consisting mainly of undergraduate and graduate college students from six mid-sized, mid-Atlantic college campuses using bulletin board postings and announcements in classes between 2002 and 2011. The present article is the seventh in a series of articles on incest utilizing data sets that included many of the participants in the present article to demonstrate the effects of FDI (Stroebel et al., 2012), BSI (Stroebel, O’Keefe, Beard, et al., 2013), SSI (Stroebel, O’Keefe, Griffee, et al., 2013), BBI (Beard et al., 2013), and SBI (O’Keefe et al., 2014) on those involved or to identify risk factors for FDI (Stroebel, Kuo, et al., 2013). To obtain a wider base and to increase age, education, and life-experience diversity, we also recruited university faculty and staff and individuals from the same general population of the mid-Atlantic United States who had already completed their education. We included all individuals who were 18 years old or older who were willing to participate in the study after being provided with information about the risks and benefits of the study. The announcement processes were slightly different on each of the campuses because we relied on the individual professors to make announcements to their classes. In some cases, the investigators were invited into classes to make the announcements. To increase the number of sexual minority individuals who participated in the study, we attended area “Pride” parades and Lesbian, Gay, Bisexual, and Transgender (LGBT) picnics and worked with LGBT churches and other LGBT-friendly organizations. Members of these groups were also invited to participate. Whenever individuals agreed to participate, we encouraged them to bring along friends to participate or to encourage their friends to participate at a later time (snowball recruiting). All participants were unpaid, but many of the students received credit from their professors in psychology, social work, and criminal justice courses. All participants were volunteers. The median age of the 1,821 female participants was 21 years (M = 25.0, SD = 9.7, range: 18-78 years). The median age of the 1,064 male participants was 21.5 years (M = 26.6, SD = 11.8, range: 18-86 years). The 2,885 participants had the following distribution for their level of education: high school only 3.8%, some college 63.3%, bachelor’s degree 14.5%, master’s degree 4.9%, and doctoral-level degree 2.3%.
Computer-Assisted Self-Interview (CASI) The anonymous survey was administered using a CASI program, S-SAPE1 (©2002 S-SAPE, LLC, P.O. Box 11081, Charleston, West Virginia 25339). The actual items that elicited the data used in this research are provided in the appendix. The factor analysis, and the reliability analysis of the nudity scales and several additional scales in a larger database that included all of the present participants can be found in Griffee, O’Keefe, Beard, Young, Kommor, and Linz, et al. (2014). Items 1 to 18 are similar to those that can be found in O’Keefe et al. (2014) and Stroebel, Kuo, et al. (2013). Content validity. All items for the survey were carefully reviewed and approved by four doctoral-level experts in the field: a psychiatrist trained in psychotherapy, a social worker trained in psychotherapy, a psychologist, and a gynecologist who was also trained in psychology.
Downloaded from sax.sagepub.com at Bobst Library, New York University on November 14, 2015
7
Griffee et al.
Procedure Surveys were administered in university computer laboratories that had up to 45 computers in a room, and sufficient space between participants so they could not see others’ computer screens. Anonymity was protected by randomized filing of the encrypted results in a hidden file filled with fake data as well as simultaneous filing of many fake decoy lines. Decoding was performed on the file containing all respondent’s randomly filed encrypted data. During a 10-min orientation, respondents were informed of these protections to their anonymity and that the S-SAPE1 computerized anonymous survey instrument was designed to obtain a history of sexual experiences and behaviors. They were again informed that they could withdraw from the study at any time they wished. No reference to incest was made during promotion of the study. During the orientation provided before the participants started entering their data into the computers, they were informed about the operation of the screen that presented the sexual behavior items and the sexual behavior sub-items. They were informed that for each behavior, they would be asked whether a brother or sister or a mother or a father had been a partner. Furthermore, they were specifically instructed to include step fathers or other father figures within the definition of “father.” This instruction was reinforced in Item 3. Selection of participants for the study. From the records of all potential participants available in the database, we selected two sub-sets of potential participants: (a) all 1,821 women who were not transsexual and who had at least one sibling and (b) 1,064 men who were not transsexual and who had at least one sibling.
Results Our study included any kind of sexual behavior between siblings (even showing and looking at genitals) within the definition of SI. For the purpose of the study, no distinction was made between full, half, and step siblings. The detailed breakdown of the 137 SI cases by type (opposite-sex vs. same-sex) and sex of the reporting participant is presented in Table 1. There were a total of seven cases who participated in both opposite-sex and same-sex SI. As can be seen in Table 2, cross-tabulation of the two sexes of reporting participants with the two types of SI (opposite-sex and samesex) resulted in four sub-sets of SI: BSI, SSI, SBI, and BBI. Also, as can be seen by the numbers in the subsequent text outlining the logistic regressions, the victims who participated in both opposite-sex and same-sex SI were included in all the appropriate logistic regressions that included the sex of the reporting participants. The percentages of participants in each of these sub-sets who participated in the behavior (coded “1”) or who were coded “1” in the other predictors are presented in Columns 1 to 6 of Table 2 based on all participants with data for that variable. The percentages in Columns 1 to 6 are useful only for showing the percentage of the group that participated in the behavior (coded “1”) or that was coded “1” in the other predictors. Such data do not provide information about predictive power. There were 45 cases of FDI, 10 cases of father–son incest (FSI), 25 cases of mother–daughter
Downloaded from sax.sagepub.com at Bobst Library, New York University on November 14, 2015
8
Sexual Abuse
Table 1. Sibling Incest Cases and Controls. Participants Sibling incest cases (all) Opposite-sex (BSI and SBI)a Same-sex (SSI and BBI)a Both opposite-sex and same-sexb Controls (no sibling incest) Total participants Sibling incest cases (coerced by sibling)d Opposite-sex (BSI and SBI)a Same-sex (SSI and BBI)a Both opposite-sex and same-sexb Controls (no coerced sibling incest) Total participants
Female
Male
Total
86 54 37 5 1,735c 1,821 24 21 3 0 1,797 1,821
51 27 25 1 1,013c 1,064 14 7 8 1 1,050 1,064
137 81 62 6 2,748 2,885 38 28 11 1 2,847 2,885
Note. BSI = brother–sister incest; SBI = sister–brother incest; SSI = sister–sister incest; BBI = brother– brother incest; SI = sibling incest; PCI = parent–child incest; MDI = mother–daughter incest; MSI = mother–son incest; FDI = father–daughter incest; FSI = father–son incest. aThe numbers are the same as the number coded “1” in the respective logistic regressions. bThe numbers are for the number of overlapping cases between BSI and SSI or SBI and BBI for females and males, respectively. cThe numbers of controls for each of the logistic regressions are provided in the text. dThe victims of sex coerced by a sibling is a sub-set of the 137 sibling incest cases.
incest (MDI), and 10 cases of mother–son incest (MSI) based on Items 1 through 4 (see Table 2 for percent within each sub-set).
Ages at Which Exposure to the Shared Bed and the Shared-Tub Risk Factors First Occurred in Participants in the Various Types of SI The median ages participants recalled first sharing a bed or a tub bath with a partner of the same sex as the SI partner ranged from 3 to 5 years of age, and the median age that they first recalled engaging in either voluntary or coerced sex of any kind age with a partner of the same sex as the SI partner ranged from 8 to 10 years of age (Table 3). The majority of participants who reported sharing tub baths or beds for sleeping did so with siblings who were within 4 years of their own age, but there was also some sharing of beds and tub baths with siblings who were more than 4 years older or younger. We designed the three measures of affection in the nuclear family to reflect the behavioral displays of affection that a young child would easily be able to observe and remember: Paternal Affection, Maternal Affection, and Parental Modeling of Affection for one another. Maternal affection was significantly lower in the SI group than in the 2,748 controls with mean values of 2.55 versus 2.69 (p = .003), but there were no
Downloaded from sax.sagepub.com at Bobst Library, New York University on November 14, 2015
9
Griffee et al. Table 2. Binary Sibling Incest Predictor Prevalence in Sibling Incest Victims and Controls. Female participants
Male participants
All participants
1
2
3
3
5
6
Column label
BSI
SSI
SBI
BBI
SI
Controls
n Victim of CSA by (Items 1-4)a A parent (PCI) Mother (MDI, MSI) Father (FDI, FSI) Victim of physical abuse by a parent (Items 19-20)a,c Victim of verbal abuse by a parent (Items 21-22)a,c Shared a bed witha A sibling Sister Brother Shared a tub bath witha A sibling Sister Brother New man in the homea Witnessed parentala Coitus (Item 17) Physical fighting (Item 18) Genital petting (Item 16) Mother’s attitudes (Item 23)a Sex is dirty and filthy Sex is healthy Able to talk about sex Father’s attitudes (Item 24)a Sex is dirty and filthy Sex is healthy Able to talk about sex
54
37
27
25
137
2,748
16.7%b 7.4% 16.7% 17.9%
8.1%b 0.0% 8.1% 20.8%
7.4%b 3.7% 7.4% 17.4%
12.0%b 8.0% 8.0% 30.4%
10.9%b 5.8% 9.5% 21.2%
1.9%b 1.0% 1.5% 6.2%
33.3%
45.8%
52.2%
47.8%
43.3%
21.0%
63.0% 25.9% 51.9%
97.3% 97.3% 16.2%
78.4% 63.0% 25.9%
84.0% 4.0% 84.0%
76.6% 46.0% 43.8%
30.2% 22.0% 13.7%
53.7% 9.3% 50.0% 38.9%
78.4% 75.7% 10.8% 37.8%
63.0% 48.1% 22.2% 22.2%
68.0% 8.0% 68.0% 20.0%
65.0% 34.3% 38.0% 32.1%
24.9% 15.9% 11.2% 23.4%
9.3% 5.6% 0.0%
5.4% 2.7% 0%
11.1% 0.0% 0.0%
4.0% 4.0% 0.0%
8.0% 3.6% 0.0%
3.9% 1.2% 1.2%
27.8% 68.5% 44.4%
18.9% 81.1% 51.4%
22.3% 74.1% 44.4%
24.0% 64.0% 16.0%
22.6% 67.9% 40.9%
17.8% 69.6% 49.7%
7.4% 70.4% 13.0%
13.5% 56.8% 8.1%
7.4% 77.8% 37.0%
8.0% 64.0% 16.0%
8.0% 76.1% 16.8%
11.6% 73.0% 24.0%
Column number
Note. BSI = brother–sister incest; SSI = sister–sister incest; SBI = sister–brother incest; BBI = brother– brother incest; SI = sibling incest; PCI = parent–child incest; MDI = mother–daughter incest; MSI = mother–son incest; FDI = father–daughter incest; FSI = father–son incest. aPercentages were calculated from the predictors in the sibling-incest logistic regression, which were coded “1” when participants had engaged in the behavior. bColumn percent who engaged in the behavior within the group. The n represented by the percent can be calculated from the column total (at top). cData were available only for the first 1,936 participants.
Downloaded from sax.sagepub.com at Bobst Library, New York University on November 14, 2015
10
Sexual Abuse
Table 3. Median Earliest Ages That Participants Shared Beds or Tub Baths or Engaged in Sex With Female and Male Sibling Partners. Female participants
BSI
Group n
SSI n
Mdn
Shared a bed Female 5 Male 3 Shared a tub bath Female 4 3 Male Sexual behavior Female 8 Male 8
All participants
Male participants SBI n
BBI n
SI n
Controls n
n
54
Mdn
37
Mdn
27
Mdn
25
Mdn
137
Mdn
2,748
41a 32
3.5 10
36a 19
5 4
18a 13
11 5
8a 31
5 4
98a 82
6 10
1,381a 1,126
31 30
3 4
31 11
3 4
15 9
6 2
5 17
3 3
79 65
3 4
854 652
25 54
7 11
37 24
8 8
27 13
14 10
12 25
8 9
95 110
14 15
842 1,075
Note. BSI = brother–sister incest; SSI = sister–sister incest; SBI = sister–brother incest; BBI = brother– brother incest; SI = sibling incest; ASI = adult SI. aSome n values are smaller than the group n because medians were calculated on those who had participated in the behavior.
statistically significant differences between the means for Paternal Affection or Parental Modeling of Affection. We used three items that applied equally to male and female sexual anatomy to measure how far the incest had progressed before the 137 SI participants had reached 18 years of age based on the study participants having specifically entered the fact that they had engaged in the behavior with a sibling using the appropriate sub-item: 48.2% had touched their sibling’s genitals, 42.3% had their genitals touched by their sibling, and in 29.9% of the cases, both kinds of genital touching had occurred. The siblings had approximated their genitals in 18.2% of the cases. Among the 54 opposite-sex SI cases reported from the sister’s point of view, the SI had progressed to intra-vaginal coitus with her brother in 5 (9.3%) of the cases at a median of 10 years of age. All five cases were reported by the sister as coerced by the brother, but in two cases involving brothers in two different age groups, the sister also reported having participated in coitus with her brother voluntarily. Among the 27 opposite-sex SI cases reported from the brother’s point of view, the SI had progressed to intra-vaginal coitus with his sister in 7 (25.9%) of the cases at a median of 13 years of age. Two of the cases were reported as coerced by the brother, but in all seven cases, including one case with sisters in two different age groups, the brother (also) reported participating voluntarily. There was no significant difference in the rates of intra-vaginal coitus reported by the
Downloaded from sax.sagepub.com at Bobst Library, New York University on November 14, 2015
11
Griffee et al. Table 4. Non-Parametric Correlations Among Sibling Incest and Seven Predictors. 1 1. Sibling incest (0/1) 2. Shared bed (0/1) 3. Shared tub bath (0/1) 4. Parental CSA (0/1) 5. Family Nudity Scale 6. Paternal Nudity Scale 7. Maternal Nudity Scale 8. Maternal affection
1
2 .211***
.211*** .192*** .117***
1 .679*** −.013
3 .192***
4 .117***
5
6
7
8
.104***
.101***
.085***
−.054**
.679***
−.013
.189***
.091***
.195***
.026
1
−.004
.203***
.125***
.202***
.024
.057***
.083***
.041*
−.004
1
−.093***
.104***
.189***
.203***
.057***
1
.502***
.876***
.090***
.101***
.091***
.125***
.083***
.502***
1
.235***
.020
.085***
.195***
.202***
.041*
.876***
.235***
1
.095***
.026
.024
−.093***
.090***
.020
.095***
1
−.054**
*p < .05. **p < .01. ***p < .001.
two sexes, χ2(1, 137) = 1.62, ns. The partner-brother had inserted his penis into the rectum of the respondent-brother in 3 (12%) of the 25 BBI cases at a median of 13 years of age, and in 3 of the 25 cases, the respondent-brother reported having inserted his penis into his brother’s rectum at a median of 13 years of age. In two BBI cases, the bothers had taken on both roles.
Hypothesis Exploration Some behaviors within nuclear families constitute risk factors for SI. We used multiple logistic regressions to obtain ORs adjusted for the effects of other variables in the model. Over-fitting was a concern in the construction of these regression models given the effective N (137 participants in SI) for our SI-dependent variables and the effective Ns (54 participants in BSI, 37 participants in SSI, 27 participants in SBI, and 25 participants in BBI) for the sub-set models for predicting these types of SI (Babyak, 2004). Accordingly, we implemented the following suggestions of Babyak (2004): First, we decided to restrict the number of predictors in the model to a maximum that resulted in a ratio of at least 10 cases/predictor with each of the variables in the models representing a different type of risk factor. Where there were several possible choices of variables to represent the same type of risk factor, we chose the variable with the lowest correlation with other predictors in the model (Table 4) and the highest predictive power for the dependent variable. Second in a process similar to constructing scales from correlated variables derived from responses to a series of individual items to create and analyze a construct, we combined potential predictors within predictor classes: We combined four father–child nudity variables (scored 0/1) into a single father–child nudity scale (with possible scores ranging from 0 to 4), hereafter referred to as the Paternal Nudity Scale. We combined four mother–child nudity variables into
Downloaded from sax.sagepub.com at Bobst Library, New York University on November 14, 2015
12
Sexual Abuse
Table 5. Predictive Power of Scales for Sibling Incest as Measured by the Score Statistic at Step 0. Female participants Column number Column label n
Item Rangea
Family Nudity 5-12 Scale Paternal Nudity 5-8 Scale Maternal Nudity 9-12 Scale Maternal affection 13 Parental modeling 15 of affection
Male participants
All participants
1
2
3
3
4
5
6
7
BSI
SSI
SBI
BBI
SI
ASI
Controls
SI
54
37
27
25
137
11
2,748
137
0-8
12.0***b 10.8***b 15.3***b
9.9**b 41.8***b 0.2 ns 1.53 ± 1.82c 2.58 ± 2.15***c,d
0-4
22.8***
7.2**
13.0***
6.1*
0-4
5.1*
7.9**
11.1***
1-3 1-3
12.6*** 0.3 ns
1.7 ns 2.6 ns
0.4 ns 0.0 ns
37.2***
10.7*** 21.4*** 0.7 ns 2.0 ns
8.8** 4.0*
4.3*
0.36 ± 0.82 0.81 ± 1.21***d
0.4 ns 1.17 ± 1.46 1.77 ± 1.51*** 9.9** 2.69 ± 0.53 2.55 ± 0.61**d 0.0 ns 2.19 ± 0.68 2.07 ± 0.72*
Note. BSI = brother–sister incest; SSI = sister–sister incest; SBI = sister–brother incest; BBI = brother–brother incest; SI = sibling incest; ASI = adult SI. aRange of possible values. bScore statistic from Step 0 of the logistic regression for the dependent variable. cScale score for participants (M ± SD). dp values based on the rank-F test because of significant heterogeneity of variance in the untransformed data. *p < .05. **p < .01. ***p < .001.
a single mother–child nudity scale, hereafter referred to as the Maternal Nudity Scale. We combined all eight parent–child nudity variables (four father–child and four mother–child variables) into a single Family Nudity Scale (with possible scores ranging from 0 to 8). As can be seen in column 4 in Table 5, the Family Nudity Scale had a higher predictive power for SI than either the Paternal Nudity Scale or the Maternal Nudity Scale. We first combined the four partner-age specific 0/1 dummy variables indicating whether or not a bed for sleeping had been shared with a sister in that partner-age group into a single 0/1 dummy variable indicating whether or not a bed for sleeping had been shared with a sister of any age, and we combined the analogous four variables into a single 0/1 dummy variable indicating whether or not a bed for sleeping had been shared with a brother of any age. Furthermore, we also combined the above two variables into a single 0/1 dummy variable indicating whether or not a bed for sleeping had been shared with a sibling of any age. We also combined the analogous variables indicating whether or not a tub bath had been shared with a brother or a sister into three analogous 0/1 dummy variables indicating whether a tub bath had been shared with a sister, a brother, or a sibling, respectively. We also combined the two 0/1 dummy variables indicating whether the participant had been a victim of CSA by their father or mother, respectively, into a single 0/1 dummy variable parent–child incest (PCI) variable indicating whether or not a participant had been a victim of CSA by one or more parents.
Downloaded from sax.sagepub.com at Bobst Library, New York University on November 14, 2015
13
Griffee et al.
Table 6. Power of Binary Predictors of Sibling Incest as Measured by the Score Statistic at Step 0 of the Logistic Regression Analyses. Female participants Column number
Male participants
All participants
1
2
3
4
5
6
7
Column label
BSI
SSI
SBI
BBI
SI
SI rate (%)
ASI
n
54
37
27
25
137
137
11
41.9***a 17.0*** 47.4*** 8.5**
3.4 nsa 0.5 ns 4.6* 6.9**
4.4*a 10.1** 1.7 ns 3.4 ns
16.1***a 11.8*** 11.8*** 19.7***
46.1***a 25.7*** 44.2*** 34.1***
22.1b 22.9 23.6 16.3
0.2 nsa 0.1 ns 0.2 ns 0.2 ns
3.4 ns
8.8**
10.3***
8.6**
28.2***
10.5
2.6 ns
17.8*** 0.2 ns 30.8***
46.7*** 51.7*** 1.2 ns
40.9*** 43.6*** 6.7**
49.2*** 0.3 ns 58.7***
128.7*** 48.0*** 91.8***
11.2 9.9 13.7
0.0 ns 0.8 ns 0.3 ns
14.8*** 2.0 ns 37.5*** 2.9 ns
32.8*** 36.4*** 0.4 ns 1.3 ns
34.5*** 31.3*** 7.5** 0.0 ns
37.8*** 0.7 ns 48.7*** 0.2 ns
106.6*** 31.7*** 85.9*** 5.5*
11.5 9.7 14.5 6.4
0.1 ns 0.3 ns 0.7 ns
3.4 ns 5.2* 0.6 ns
0.1 ns 0.5 ns 0.4 ns
3.4 ns 0.4 ns 0.4 ns
0.0 ns 1.2 ns 0.4 ns
5.8* 5.7* 1.7 ns
9.4 12.8 0.0
1.0 ns 6.2* 0.1 ns
5.6* 3.3 ns 1.7 ns
0.0 ns 0.4 ns 0.5 ns
0.2 ns 0.0 ns 0.0 ns
0.5 ns 1.0 ns 6.5*
2.0 ns 0.7 ns 4.1*
6.0 4.6 3.9
5.1* 5.5* 4.4*
1.2 ns 0.3 ns 0.2 ns
0.0 ns 0.8 ns 1.2 ns
0.1 ns 0.3 ns 0.0 ns
0.0 ns 1.6 ns 4.7*
2.0 ns 0.2 ns 3.7 ns
3.3 4.6 3.4
0.0 ns 1.2 ns 0.0 ns
Victim of CSA by (Items 1-4) A parent (PCI) Mother (MDI, MSI) Father (FDI, FSI) Victim of physical abuse by a parent (Items 19-20)c 6,693 Victim of verbal abuse by a parent (Items 21-22)c Shared a bed withb A sibling Sister Brother Shared a tub bath with A sibling Sister Brother New man in the home Witnessed parental Coitus (Item 17) Physical fighting (Item 18) Genital petting (Item 16) Mother’s attitudes (Item 23) Sex is dirty and filthy Sex is healthy Able to talk about sex Father’s attitudes Sex is dirty and filthy Sex is healthy Able to talk about sex
Note. BSI = brother–sister incest; SSI = sister–sister incest; SBI = sister–brother incest; BBI = brother–brother incest; SI = sibling incest; ASI = adult sibling incest; PCI = parent–child incest; MDI = mother–daughter incest; MSI = mother– son incest; FDI = father–daughter incest; FSI = father–son incest. *p < .05, **p
SAXXXX10.1177/1079063214558941Sexual AbuseGriffee et al.
Article
Etiological Risk Factors for Sibling Incest: Data From an Anonymous Computer-Assisted Self-Interview
Sexual Abuse: A Journal of Research and Treatment 1–40 © The Author(s) 2014 Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/1079063214558941 sax.sagepub.com
Karen Griffee1, Sam Swindell2, Stephen L. O’Keefe3,4, Sandra S. Stroebel3, Keith W. Beard4, Shih-Ya Kuo2, and Walter Stroupe2
Abstract Retrospective data from 1,821 women and 1,064 men with one or more siblings, provided anonymously using a computer-assisted self-interview, were used to identify risk factors for sibling incest (SI); 137 were participants in SI. In order of decreasing predictive power, the risk factors identified by the multiple logistic regression analysis included ever having shared a bed for sleeping with a sibling, parent–child incest (PCI), family nudity, low levels of maternal affection, and ever having shared a tub bath with a sibling. The results were consistent with the idea that SI in many families was the cumulative result of four types of parental behaviors: (a) factors that lower external barriers to sexual behavior (e.g., permitting co-sleeping or co-bathing of sibling dyads), (b) factors that encourage nudity of children within the nuclear family and permit children to see the parent’s genitals, (c) factors that lead to the siblings relying on one another for affection (e.g., diminished maternal affection), and (d) factors that eroticize young children (e.g., child sexual abuse [CSA] by a parent). Thirty-eight of the 137 SI participants were participants in coerced sibling incest (CSI). In order of decreasing predictive power, risk factors for CSI identified by multiple logistic regression analysis included ever having shared a bed for sleeping with a brother, PCI, witnessing parental physical fighting, and family nudity. SI was more likely to have 1Concord
University, Athens, WV, USA Virginia State University, Institute, WV, USA 3Marshall University Graduate College, South Charleston, WV, USA 4Marshall University, Huntington, WV, USA 2West
Corresponding Author: Karen Griffee, Associate Professor of Psychology, Division of Social Sciences, Concord University, P.O. Box 1000, Athens, WV 24712, USA. Email: [email protected]
Downloaded from sax.sagepub.com at Bobst Library, New York University on November 14, 2015
2
Sexual Abuse
been reported as CSI if the sibling had touched the reporting sibling’s genitals, and less likely to have been reported as CSI if the siblings had shared a bed. Keywords sibling incest, brother–sister incest, brother–brother incest, sister–brother incest, sister–sister incest, maternal affection, nudity, parental child sexual abuse, risk factors Convenience samples have provided estimates of 15% for all sexual contacts between siblings in girls (Finkelhor, 1981; Greenwald & Leitenberg, 1989) and 10% in boys (Finkelhor, 1981). Atwood (2007) obtained an estimated prevalence of 17% for brother–sister incest (BSI), but after excluding cases of sexual contact with a relative that were wanted and with a peer and including cases in the denominator that were not at risk because they had no siblings, Russell (1986) obtained estimates of only 2.0% for BSI and 0.3% for sister–sister incest (SSI). BSI can have problematic long-term effects on the sisters (Stroebel, O’Keefe, Beard, Kuo, Swindell, & Stroupe, 2013; Tidefors, Arvidsson, Ingevaldson, & Larsson, 2010). Sister–brother incest (SBI) can also have problematic long-term effects on the brothers (O’Keefe et al., 2014). SSI (Stroebel, O’Keefe, Griffee, Kuo, Beard, & Kommor, 2013) and brother–brother incest (BBI; Beard et al., 2013) also have problematic effects on the involved siblings. The known adverse effects of all four types of sibling incest (SI) emphasize the importance of prospective prevention. Claims by some authors based on purported tests of the 19th-century theory of Westermarck (e.g., Antfolk, Karlsson, Bäckström, & Santtila, 2012; Antfolk, Lieberman, & Santtila, 2012; Bevc & Silverman, 1993, 2000; Lieberman, Tooby, & Cosmides, 2003, 2007; see Silverman & Bevc, 2005, for review) that there is an innate biological resistance to incestuous sexual relationships because of close association in the nuclear family, disgust at the thought of incest, or biological relatedness are controversial. The theory is at odds with recent evidence showing that individuals who grew up in the communal kibbutz education system were actually often attracted to their peers and rarely developed a sexual aversion toward them (Shor & Simchai, 2009). Furthermore, the theory of Westermarck is at odds with historical evidence that incestuous marriages became common place in societies that tolerated brother–sister marriages (Roman Egypt in the first to third centuries AD) or promoted brother–sister and father–daughter marriages (Persia from the fifth century BC to the seventh century AD; Scheidel, 1996). Based on papyrus census records from Roman Egypt from first to third centuries AD, it has been estimated that brother–sister marriages occurred in about one third of families with marriageable opposite-sex siblings with the brother older than the sister (Scheidel, 1996). Conditioned inhibition of sexual behavior between the siblings (for reviews of conditioned sexual preferences, see Hoffman, 2012; Pfaus et al., 2012) and critical period learning of sexual behavior (Griffee, O’Keefe, Beard, Young, Kommor, & Linz, 2014; Griffee, O’Keefe, Beard, Young, Kommor, Linz, Swindell, & Stroebel, 2014) working together provide an alternative
Downloaded from sax.sagepub.com at Bobst Library, New York University on November 14, 2015
3
Griffee et al.
explanation for inhibition of marriages between children reared together attributed to the Westermarck effect without the need to invoke any special evolved mechanism to reduce mating between siblings. Such conditioning would be predicted theoretically by parental efforts to prevent sex between the siblings. The term “negative imprinting” has been used to describe the effect (Shepher, 1971). Programs currently in place to prevent CSA have focused on teaching children to protect themselves against CSA (see Babatsikos, 2010; Hunt & Walsh, 2011; Kenny, Capri, Thakkar-Kolar, Ryan, & Runyon, 2008; Kenny & Wurtele, 2012 for reviews). Unfortunately, by the time children are old enough to assimilate information about how to protect themselves against SI, they may already have become victims. SI often starts when the siblings are very young, and SI is often experienced by both the involved siblings as voluntary (Beard et al., 2013; Caffaro & Conn-Caffaro, 1998; O’Keefe et al., 2014; Stroebel, O’Keefe, Beard, et al., 2013; Stroebel, O’Keefe, Griffee, et al., 2013). Another, potentially more fruitful approach to prevention would be to identify etiological risk factors for SI (using valid epidemiological methods based on statistically comparing data from SI cases with that from valid controls) and then make this information available to families so that preventative measures can be implemented within families before the incest has occurred. Such information could be presented to both mothers and fathers as empowering families in the prevention of incest. Despite a computerized search in November 2012 of the Medline, PsychINFO, and PsychARTICLES databases using Academic Search Premier for the term “incest” in the abstract, no valid epidemiological studies that used statistical techniques, SI cases, and controls were identified. Due to the dearth of data in the more recent literature on risk factors for SI, we were forced to turn to the older literature. Other clinical studies of SI participants and studies of normal sexual behaviors of children and adolescents have suggested a number of possible risk factors for SI not derived from valid epidemiological studies that used statistical techniques. Furthermore, definitions of SI differed among the many (non-epidemiologic) studies of SI. Some differentiated SI from normative “sex play” between siblings (e.g., Bank & Kahn, 1982; Caffaro & ConnCaffaro, 1998; Russell, 1986) whereas others (including the present article) included any kind of sexual behavior between siblings within the definition of SI (e.g., Araji, 1997; Beard et al., 2013; Finkelhor, 1980, 1981; Stroebel, O’Keefe, Beard, et al., 2013; Stroebel, O’Keefe, Griffee, et al., 2013). There are at least four levels where professionals who treat incest offenders might identify risk factors for SI, and to avoid confusion, we will identify and name the four most obvious levels. First, there are the etiological factors that actually predispose or protect against the occurrence of SI. Second, there are the intra-familial factors that increase or decrease the likelihood that SI will be detected by the siblings’ parents. Third, there are the factors that increase or decrease the chance that the parents will actually report the case of SI to the authorities. Finally, there are the factors that increase or decrease the likelihood that the SI offenders will re-offend. When it comes to preventing the occurrence of SI or treating incest offenders, it is the etiological factors that are the most useful to know (Offord & Kraemer, 2000).
Downloaded from sax.sagepub.com at Bobst Library, New York University on November 14, 2015
4
Sexual Abuse
One systematic study that made an attempt to identify risk factors for SI (Caffaro & Conn-Caffaro, 1998) was based on face-to-face clinical interviews of 29 adult SI survivors referred for study by therapists and facilities specializing in CSA, but because there was no control group or statistical analysis, identification of the risk factors was based solely on the authors’ clinical opinions. Furthermore, no measures of odds ratios (ORs) or adjusted relative risk were possible. Caffaro and Conn-Caffaro (1998) suggested risk factors for sibling assault and incest trauma included parental unavailability, a sexualized environment in the home, lack of sex education, lack of parental supervision, lack of other supportive relationships, siblings standing in for absent parents, power imbalance among siblings (due to developmental, physical, or intellectual differences), unclear individual and generational boundaries, unusually frequent and ready access to a victim (e.g., room sharing), a victim’s dependence on an older and more powerful sibling, a history of victimization, and parental favoritism. When affection from parents is scarce or unavailable, siblings often turn to one another for affection, and this can lead to incestuous behaviors (Bank & Kahn, 1982). Families that were more permissive about parental and child nudity and that modeled greater physical intimacy reported observing more sexual behaviors in their young children than families that emphasized modesty and that were less physically intimate (Friedrich, Grambsch, Broughton, Kuiper, & Beilke, 1991; Sandfort & CohenKettenis, 2000). A few adults anecdotally looking back at their own experiences with SI have identified bed sharing, being bathed together with a sibling, skinny dipping with siblings, and casual household nudity as experiences that led to their involvement in SI (Symonds, Mendoza, & Harrell, 1981). Young victims of incest with a parent or sibling often become eroticized and actively seek sex from other children and alternative adult care givers (Araji, 1997; Beard et al., 2013; O’Keefe et al., 2014; Salter et al., 2003; Stroebel et al., 2012; Stroebel, O’Keefe, Beard et al., 2013; Stroebel, O’Keefe, Griffee, et al., 2013; Yates, 1982, 1990), implying that being a victim of incest perpetrated by a parent or sibling could be a risk factor for SI. As reviewed by Araji (1997), there was an extensive literature that also contains clinical-opinion-based “risk factors” for children and adolescents becoming sexually aggressive children, and we were able to find one case-control epidemiological study of risk factors for becoming a child, adolescent, or adult CSA perpetrator given the fact that a given child or adolescent (himself) had been the victim of CSA (Salter et al., 2003). It is possible that some of the factors identified by Araji (1997) and Salter et al. (2003) were second-, third-, or fourth-level risk factors that increased the likelihood of the SI coming to the attention of the authorities—rather than first-level etiological risk factors for SI—because the data were obtained from cases that had come to the attention of the authorities. SI, like father–daughter incest (FDI), is seldom reported to the police (Atwood, 2007; Carlson, Maciol, & Schneider, 2006; Russell, 1986). Furthermore, SI is seldom reported to parents either, principally because each sibling believes that she or he will be punished (Atwood, 2007; Carlson et al., 2006; Laviola, 1992; Russell, 1986). This means the best way to accurately identify cases of SI is to obtain the data from the siblings themselves (Caffaro & Conn-Caffaro, 1998) in a way that permits them to
Downloaded from sax.sagepub.com at Bobst Library, New York University on November 14, 2015
5
Griffee et al.
disclose the information without concern that the disclosure could harm either their siblings or themselves. Suitable controls must be identified in the same population from which the SI cases were obtained. The present study was part of a larger study on human sexuality presented to participants as a “cradle to the grave” study on human sexuality. The method of recruiting participants employed in the present study minimized factors that might select for or against women and men who had been involved in SI. Obtaining the data from unselected individuals (some of whom happened to disclose to the computer questionnaire that that they had been involved in SI) maximized the chances that etiological risk factors for SI would be identified. We tested the following hypotheses: Hypothesis 1: Some behaviors within nuclear families constitute risk factors for SI. As detailed above, based on the few applicable studies in the literature, we were able to identify some potential etiological risk factors for SI. 1. Bed sharing (Symonds et al., 1981). 2. Siblings being bathed together (Symonds et al., 1981). 3. Parentally approved nuclear family nudity behaviors (Friedrich et al., 1991; Sandfort & Cohen-Kettenis, 2000). 4. Low parental affection (Bank & Kahn, 1982). 5. Eroticization resulting from being a victim of parental incest (Araji, 1997; Salter et al., 2003; Stroebel et al., 2012; Stroebel, O’Keefe, Beard, et al., 2013; Yates, 1982, 1990). 6. A sexualized family environment resulting from parental sexual behaviors not entirely hidden from the children (Araji, 1997; Caffaro & Conn-Caffaro, 1998). 7. Witnessing violence within the nuclear family (Araji, 1997; Salter et al., 2003). Because many cases of SI never come to the attention of either the parents or the authorities, we also tested the following hypotheses: Hypothesis 2: Certain behaviors of the incestuous sibling dyad or conditions within nuclear families increase the chances that the SI will be identified. Hypothesis 3: Some behaviors within nuclear families increase the chances that SI will occur between two adult siblings.
Method Participants All 1,821 female participants and all 1,064 male participants were above the age of 18 and gave informed consent using forms approved by the relevant institutional review
Downloaded from sax.sagepub.com at Bobst Library, New York University on November 14, 2015
6
Sexual Abuse
board. The participants were recruited from a population consisting mainly of undergraduate and graduate college students from six mid-sized, mid-Atlantic college campuses using bulletin board postings and announcements in classes between 2002 and 2011. The present article is the seventh in a series of articles on incest utilizing data sets that included many of the participants in the present article to demonstrate the effects of FDI (Stroebel et al., 2012), BSI (Stroebel, O’Keefe, Beard, et al., 2013), SSI (Stroebel, O’Keefe, Griffee, et al., 2013), BBI (Beard et al., 2013), and SBI (O’Keefe et al., 2014) on those involved or to identify risk factors for FDI (Stroebel, Kuo, et al., 2013). To obtain a wider base and to increase age, education, and life-experience diversity, we also recruited university faculty and staff and individuals from the same general population of the mid-Atlantic United States who had already completed their education. We included all individuals who were 18 years old or older who were willing to participate in the study after being provided with information about the risks and benefits of the study. The announcement processes were slightly different on each of the campuses because we relied on the individual professors to make announcements to their classes. In some cases, the investigators were invited into classes to make the announcements. To increase the number of sexual minority individuals who participated in the study, we attended area “Pride” parades and Lesbian, Gay, Bisexual, and Transgender (LGBT) picnics and worked with LGBT churches and other LGBT-friendly organizations. Members of these groups were also invited to participate. Whenever individuals agreed to participate, we encouraged them to bring along friends to participate or to encourage their friends to participate at a later time (snowball recruiting). All participants were unpaid, but many of the students received credit from their professors in psychology, social work, and criminal justice courses. All participants were volunteers. The median age of the 1,821 female participants was 21 years (M = 25.0, SD = 9.7, range: 18-78 years). The median age of the 1,064 male participants was 21.5 years (M = 26.6, SD = 11.8, range: 18-86 years). The 2,885 participants had the following distribution for their level of education: high school only 3.8%, some college 63.3%, bachelor’s degree 14.5%, master’s degree 4.9%, and doctoral-level degree 2.3%.
Computer-Assisted Self-Interview (CASI) The anonymous survey was administered using a CASI program, S-SAPE1 (©2002 S-SAPE, LLC, P.O. Box 11081, Charleston, West Virginia 25339). The actual items that elicited the data used in this research are provided in the appendix. The factor analysis, and the reliability analysis of the nudity scales and several additional scales in a larger database that included all of the present participants can be found in Griffee, O’Keefe, Beard, Young, Kommor, and Linz, et al. (2014). Items 1 to 18 are similar to those that can be found in O’Keefe et al. (2014) and Stroebel, Kuo, et al. (2013). Content validity. All items for the survey were carefully reviewed and approved by four doctoral-level experts in the field: a psychiatrist trained in psychotherapy, a social worker trained in psychotherapy, a psychologist, and a gynecologist who was also trained in psychology.
Downloaded from sax.sagepub.com at Bobst Library, New York University on November 14, 2015
7
Griffee et al.
Procedure Surveys were administered in university computer laboratories that had up to 45 computers in a room, and sufficient space between participants so they could not see others’ computer screens. Anonymity was protected by randomized filing of the encrypted results in a hidden file filled with fake data as well as simultaneous filing of many fake decoy lines. Decoding was performed on the file containing all respondent’s randomly filed encrypted data. During a 10-min orientation, respondents were informed of these protections to their anonymity and that the S-SAPE1 computerized anonymous survey instrument was designed to obtain a history of sexual experiences and behaviors. They were again informed that they could withdraw from the study at any time they wished. No reference to incest was made during promotion of the study. During the orientation provided before the participants started entering their data into the computers, they were informed about the operation of the screen that presented the sexual behavior items and the sexual behavior sub-items. They were informed that for each behavior, they would be asked whether a brother or sister or a mother or a father had been a partner. Furthermore, they were specifically instructed to include step fathers or other father figures within the definition of “father.” This instruction was reinforced in Item 3. Selection of participants for the study. From the records of all potential participants available in the database, we selected two sub-sets of potential participants: (a) all 1,821 women who were not transsexual and who had at least one sibling and (b) 1,064 men who were not transsexual and who had at least one sibling.
Results Our study included any kind of sexual behavior between siblings (even showing and looking at genitals) within the definition of SI. For the purpose of the study, no distinction was made between full, half, and step siblings. The detailed breakdown of the 137 SI cases by type (opposite-sex vs. same-sex) and sex of the reporting participant is presented in Table 1. There were a total of seven cases who participated in both opposite-sex and same-sex SI. As can be seen in Table 2, cross-tabulation of the two sexes of reporting participants with the two types of SI (opposite-sex and samesex) resulted in four sub-sets of SI: BSI, SSI, SBI, and BBI. Also, as can be seen by the numbers in the subsequent text outlining the logistic regressions, the victims who participated in both opposite-sex and same-sex SI were included in all the appropriate logistic regressions that included the sex of the reporting participants. The percentages of participants in each of these sub-sets who participated in the behavior (coded “1”) or who were coded “1” in the other predictors are presented in Columns 1 to 6 of Table 2 based on all participants with data for that variable. The percentages in Columns 1 to 6 are useful only for showing the percentage of the group that participated in the behavior (coded “1”) or that was coded “1” in the other predictors. Such data do not provide information about predictive power. There were 45 cases of FDI, 10 cases of father–son incest (FSI), 25 cases of mother–daughter
Downloaded from sax.sagepub.com at Bobst Library, New York University on November 14, 2015
8
Sexual Abuse
Table 1. Sibling Incest Cases and Controls. Participants Sibling incest cases (all) Opposite-sex (BSI and SBI)a Same-sex (SSI and BBI)a Both opposite-sex and same-sexb Controls (no sibling incest) Total participants Sibling incest cases (coerced by sibling)d Opposite-sex (BSI and SBI)a Same-sex (SSI and BBI)a Both opposite-sex and same-sexb Controls (no coerced sibling incest) Total participants
Female
Male
Total
86 54 37 5 1,735c 1,821 24 21 3 0 1,797 1,821
51 27 25 1 1,013c 1,064 14 7 8 1 1,050 1,064
137 81 62 6 2,748 2,885 38 28 11 1 2,847 2,885
Note. BSI = brother–sister incest; SBI = sister–brother incest; SSI = sister–sister incest; BBI = brother– brother incest; SI = sibling incest; PCI = parent–child incest; MDI = mother–daughter incest; MSI = mother–son incest; FDI = father–daughter incest; FSI = father–son incest. aThe numbers are the same as the number coded “1” in the respective logistic regressions. bThe numbers are for the number of overlapping cases between BSI and SSI or SBI and BBI for females and males, respectively. cThe numbers of controls for each of the logistic regressions are provided in the text. dThe victims of sex coerced by a sibling is a sub-set of the 137 sibling incest cases.
incest (MDI), and 10 cases of mother–son incest (MSI) based on Items 1 through 4 (see Table 2 for percent within each sub-set).
Ages at Which Exposure to the Shared Bed and the Shared-Tub Risk Factors First Occurred in Participants in the Various Types of SI The median ages participants recalled first sharing a bed or a tub bath with a partner of the same sex as the SI partner ranged from 3 to 5 years of age, and the median age that they first recalled engaging in either voluntary or coerced sex of any kind age with a partner of the same sex as the SI partner ranged from 8 to 10 years of age (Table 3). The majority of participants who reported sharing tub baths or beds for sleeping did so with siblings who were within 4 years of their own age, but there was also some sharing of beds and tub baths with siblings who were more than 4 years older or younger. We designed the three measures of affection in the nuclear family to reflect the behavioral displays of affection that a young child would easily be able to observe and remember: Paternal Affection, Maternal Affection, and Parental Modeling of Affection for one another. Maternal affection was significantly lower in the SI group than in the 2,748 controls with mean values of 2.55 versus 2.69 (p = .003), but there were no
Downloaded from sax.sagepub.com at Bobst Library, New York University on November 14, 2015
9
Griffee et al. Table 2. Binary Sibling Incest Predictor Prevalence in Sibling Incest Victims and Controls. Female participants
Male participants
All participants
1
2
3
3
5
6
Column label
BSI
SSI
SBI
BBI
SI
Controls
n Victim of CSA by (Items 1-4)a A parent (PCI) Mother (MDI, MSI) Father (FDI, FSI) Victim of physical abuse by a parent (Items 19-20)a,c Victim of verbal abuse by a parent (Items 21-22)a,c Shared a bed witha A sibling Sister Brother Shared a tub bath witha A sibling Sister Brother New man in the homea Witnessed parentala Coitus (Item 17) Physical fighting (Item 18) Genital petting (Item 16) Mother’s attitudes (Item 23)a Sex is dirty and filthy Sex is healthy Able to talk about sex Father’s attitudes (Item 24)a Sex is dirty and filthy Sex is healthy Able to talk about sex
54
37
27
25
137
2,748
16.7%b 7.4% 16.7% 17.9%
8.1%b 0.0% 8.1% 20.8%
7.4%b 3.7% 7.4% 17.4%
12.0%b 8.0% 8.0% 30.4%
10.9%b 5.8% 9.5% 21.2%
1.9%b 1.0% 1.5% 6.2%
33.3%
45.8%
52.2%
47.8%
43.3%
21.0%
63.0% 25.9% 51.9%
97.3% 97.3% 16.2%
78.4% 63.0% 25.9%
84.0% 4.0% 84.0%
76.6% 46.0% 43.8%
30.2% 22.0% 13.7%
53.7% 9.3% 50.0% 38.9%
78.4% 75.7% 10.8% 37.8%
63.0% 48.1% 22.2% 22.2%
68.0% 8.0% 68.0% 20.0%
65.0% 34.3% 38.0% 32.1%
24.9% 15.9% 11.2% 23.4%
9.3% 5.6% 0.0%
5.4% 2.7% 0%
11.1% 0.0% 0.0%
4.0% 4.0% 0.0%
8.0% 3.6% 0.0%
3.9% 1.2% 1.2%
27.8% 68.5% 44.4%
18.9% 81.1% 51.4%
22.3% 74.1% 44.4%
24.0% 64.0% 16.0%
22.6% 67.9% 40.9%
17.8% 69.6% 49.7%
7.4% 70.4% 13.0%
13.5% 56.8% 8.1%
7.4% 77.8% 37.0%
8.0% 64.0% 16.0%
8.0% 76.1% 16.8%
11.6% 73.0% 24.0%
Column number
Note. BSI = brother–sister incest; SSI = sister–sister incest; SBI = sister–brother incest; BBI = brother– brother incest; SI = sibling incest; PCI = parent–child incest; MDI = mother–daughter incest; MSI = mother–son incest; FDI = father–daughter incest; FSI = father–son incest. aPercentages were calculated from the predictors in the sibling-incest logistic regression, which were coded “1” when participants had engaged in the behavior. bColumn percent who engaged in the behavior within the group. The n represented by the percent can be calculated from the column total (at top). cData were available only for the first 1,936 participants.
Downloaded from sax.sagepub.com at Bobst Library, New York University on November 14, 2015
10
Sexual Abuse
Table 3. Median Earliest Ages That Participants Shared Beds or Tub Baths or Engaged in Sex With Female and Male Sibling Partners. Female participants
BSI
Group n
SSI n
Mdn
Shared a bed Female 5 Male 3 Shared a tub bath Female 4 3 Male Sexual behavior Female 8 Male 8
All participants
Male participants SBI n
BBI n
SI n
Controls n
n
54
Mdn
37
Mdn
27
Mdn
25
Mdn
137
Mdn
2,748
41a 32
3.5 10
36a 19
5 4
18a 13
11 5
8a 31
5 4
98a 82
6 10
1,381a 1,126
31 30
3 4
31 11
3 4
15 9
6 2
5 17
3 3
79 65
3 4
854 652
25 54
7 11
37 24
8 8
27 13
14 10
12 25
8 9
95 110
14 15
842 1,075
Note. BSI = brother–sister incest; SSI = sister–sister incest; SBI = sister–brother incest; BBI = brother– brother incest; SI = sibling incest; ASI = adult SI. aSome n values are smaller than the group n because medians were calculated on those who had participated in the behavior.
statistically significant differences between the means for Paternal Affection or Parental Modeling of Affection. We used three items that applied equally to male and female sexual anatomy to measure how far the incest had progressed before the 137 SI participants had reached 18 years of age based on the study participants having specifically entered the fact that they had engaged in the behavior with a sibling using the appropriate sub-item: 48.2% had touched their sibling’s genitals, 42.3% had their genitals touched by their sibling, and in 29.9% of the cases, both kinds of genital touching had occurred. The siblings had approximated their genitals in 18.2% of the cases. Among the 54 opposite-sex SI cases reported from the sister’s point of view, the SI had progressed to intra-vaginal coitus with her brother in 5 (9.3%) of the cases at a median of 10 years of age. All five cases were reported by the sister as coerced by the brother, but in two cases involving brothers in two different age groups, the sister also reported having participated in coitus with her brother voluntarily. Among the 27 opposite-sex SI cases reported from the brother’s point of view, the SI had progressed to intra-vaginal coitus with his sister in 7 (25.9%) of the cases at a median of 13 years of age. Two of the cases were reported as coerced by the brother, but in all seven cases, including one case with sisters in two different age groups, the brother (also) reported participating voluntarily. There was no significant difference in the rates of intra-vaginal coitus reported by the
Downloaded from sax.sagepub.com at Bobst Library, New York University on November 14, 2015
11
Griffee et al. Table 4. Non-Parametric Correlations Among Sibling Incest and Seven Predictors. 1 1. Sibling incest (0/1) 2. Shared bed (0/1) 3. Shared tub bath (0/1) 4. Parental CSA (0/1) 5. Family Nudity Scale 6. Paternal Nudity Scale 7. Maternal Nudity Scale 8. Maternal affection
1
2 .211***
.211*** .192*** .117***
1 .679*** −.013
3 .192***
4 .117***
5
6
7
8
.104***
.101***
.085***
−.054**
.679***
−.013
.189***
.091***
.195***
.026
1
−.004
.203***
.125***
.202***
.024
.057***
.083***
.041*
−.004
1
−.093***
.104***
.189***
.203***
.057***
1
.502***
.876***
.090***
.101***
.091***
.125***
.083***
.502***
1
.235***
.020
.085***
.195***
.202***
.041*
.876***
.235***
1
.095***
.026
.024
−.093***
.090***
.020
.095***
1
−.054**
*p < .05. **p < .01. ***p < .001.
two sexes, χ2(1, 137) = 1.62, ns. The partner-brother had inserted his penis into the rectum of the respondent-brother in 3 (12%) of the 25 BBI cases at a median of 13 years of age, and in 3 of the 25 cases, the respondent-brother reported having inserted his penis into his brother’s rectum at a median of 13 years of age. In two BBI cases, the bothers had taken on both roles.
Hypothesis Exploration Some behaviors within nuclear families constitute risk factors for SI. We used multiple logistic regressions to obtain ORs adjusted for the effects of other variables in the model. Over-fitting was a concern in the construction of these regression models given the effective N (137 participants in SI) for our SI-dependent variables and the effective Ns (54 participants in BSI, 37 participants in SSI, 27 participants in SBI, and 25 participants in BBI) for the sub-set models for predicting these types of SI (Babyak, 2004). Accordingly, we implemented the following suggestions of Babyak (2004): First, we decided to restrict the number of predictors in the model to a maximum that resulted in a ratio of at least 10 cases/predictor with each of the variables in the models representing a different type of risk factor. Where there were several possible choices of variables to represent the same type of risk factor, we chose the variable with the lowest correlation with other predictors in the model (Table 4) and the highest predictive power for the dependent variable. Second in a process similar to constructing scales from correlated variables derived from responses to a series of individual items to create and analyze a construct, we combined potential predictors within predictor classes: We combined four father–child nudity variables (scored 0/1) into a single father–child nudity scale (with possible scores ranging from 0 to 4), hereafter referred to as the Paternal Nudity Scale. We combined four mother–child nudity variables into
Downloaded from sax.sagepub.com at Bobst Library, New York University on November 14, 2015
12
Sexual Abuse
Table 5. Predictive Power of Scales for Sibling Incest as Measured by the Score Statistic at Step 0. Female participants Column number Column label n
Item Rangea
Family Nudity 5-12 Scale Paternal Nudity 5-8 Scale Maternal Nudity 9-12 Scale Maternal affection 13 Parental modeling 15 of affection
Male participants
All participants
1
2
3
3
4
5
6
7
BSI
SSI
SBI
BBI
SI
ASI
Controls
SI
54
37
27
25
137
11
2,748
137
0-8
12.0***b 10.8***b 15.3***b
9.9**b 41.8***b 0.2 ns 1.53 ± 1.82c 2.58 ± 2.15***c,d
0-4
22.8***
7.2**
13.0***
6.1*
0-4
5.1*
7.9**
11.1***
1-3 1-3
12.6*** 0.3 ns
1.7 ns 2.6 ns
0.4 ns 0.0 ns
37.2***
10.7*** 21.4*** 0.7 ns 2.0 ns
8.8** 4.0*
4.3*
0.36 ± 0.82 0.81 ± 1.21***d
0.4 ns 1.17 ± 1.46 1.77 ± 1.51*** 9.9** 2.69 ± 0.53 2.55 ± 0.61**d 0.0 ns 2.19 ± 0.68 2.07 ± 0.72*
Note. BSI = brother–sister incest; SSI = sister–sister incest; SBI = sister–brother incest; BBI = brother–brother incest; SI = sibling incest; ASI = adult SI. aRange of possible values. bScore statistic from Step 0 of the logistic regression for the dependent variable. cScale score for participants (M ± SD). dp values based on the rank-F test because of significant heterogeneity of variance in the untransformed data. *p < .05. **p < .01. ***p < .001.
a single mother–child nudity scale, hereafter referred to as the Maternal Nudity Scale. We combined all eight parent–child nudity variables (four father–child and four mother–child variables) into a single Family Nudity Scale (with possible scores ranging from 0 to 8). As can be seen in column 4 in Table 5, the Family Nudity Scale had a higher predictive power for SI than either the Paternal Nudity Scale or the Maternal Nudity Scale. We first combined the four partner-age specific 0/1 dummy variables indicating whether or not a bed for sleeping had been shared with a sister in that partner-age group into a single 0/1 dummy variable indicating whether or not a bed for sleeping had been shared with a sister of any age, and we combined the analogous four variables into a single 0/1 dummy variable indicating whether or not a bed for sleeping had been shared with a brother of any age. Furthermore, we also combined the above two variables into a single 0/1 dummy variable indicating whether or not a bed for sleeping had been shared with a sibling of any age. We also combined the analogous variables indicating whether or not a tub bath had been shared with a brother or a sister into three analogous 0/1 dummy variables indicating whether a tub bath had been shared with a sister, a brother, or a sibling, respectively. We also combined the two 0/1 dummy variables indicating whether the participant had been a victim of CSA by their father or mother, respectively, into a single 0/1 dummy variable parent–child incest (PCI) variable indicating whether or not a participant had been a victim of CSA by one or more parents.
Downloaded from sax.sagepub.com at Bobst Library, New York University on November 14, 2015
13
Griffee et al.
Table 6. Power of Binary Predictors of Sibling Incest as Measured by the Score Statistic at Step 0 of the Logistic Regression Analyses. Female participants Column number
Male participants
All participants
1
2
3
4
5
6
7
Column label
BSI
SSI
SBI
BBI
SI
SI rate (%)
ASI
n
54
37
27
25
137
137
11
41.9***a 17.0*** 47.4*** 8.5**
3.4 nsa 0.5 ns 4.6* 6.9**
4.4*a 10.1** 1.7 ns 3.4 ns
16.1***a 11.8*** 11.8*** 19.7***
46.1***a 25.7*** 44.2*** 34.1***
22.1b 22.9 23.6 16.3
0.2 nsa 0.1 ns 0.2 ns 0.2 ns
3.4 ns
8.8**
10.3***
8.6**
28.2***
10.5
2.6 ns
17.8*** 0.2 ns 30.8***
46.7*** 51.7*** 1.2 ns
40.9*** 43.6*** 6.7**
49.2*** 0.3 ns 58.7***
128.7*** 48.0*** 91.8***
11.2 9.9 13.7
0.0 ns 0.8 ns 0.3 ns
14.8*** 2.0 ns 37.5*** 2.9 ns
32.8*** 36.4*** 0.4 ns 1.3 ns
34.5*** 31.3*** 7.5** 0.0 ns
37.8*** 0.7 ns 48.7*** 0.2 ns
106.6*** 31.7*** 85.9*** 5.5*
11.5 9.7 14.5 6.4
0.1 ns 0.3 ns 0.7 ns
3.4 ns 5.2* 0.6 ns
0.1 ns 0.5 ns 0.4 ns
3.4 ns 0.4 ns 0.4 ns
0.0 ns 1.2 ns 0.4 ns
5.8* 5.7* 1.7 ns
9.4 12.8 0.0
1.0 ns 6.2* 0.1 ns
5.6* 3.3 ns 1.7 ns
0.0 ns 0.4 ns 0.5 ns
0.2 ns 0.0 ns 0.0 ns
0.5 ns 1.0 ns 6.5*
2.0 ns 0.7 ns 4.1*
6.0 4.6 3.9
5.1* 5.5* 4.4*
1.2 ns 0.3 ns 0.2 ns
0.0 ns 0.8 ns 1.2 ns
0.1 ns 0.3 ns 0.0 ns
0.0 ns 1.6 ns 4.7*
2.0 ns 0.2 ns 3.7 ns
3.3 4.6 3.4
0.0 ns 1.2 ns 0.0 ns
Victim of CSA by (Items 1-4) A parent (PCI) Mother (MDI, MSI) Father (FDI, FSI) Victim of physical abuse by a parent (Items 19-20)c 6,693 Victim of verbal abuse by a parent (Items 21-22)c Shared a bed withb A sibling Sister Brother Shared a tub bath with A sibling Sister Brother New man in the home Witnessed parental Coitus (Item 17) Physical fighting (Item 18) Genital petting (Item 16) Mother’s attitudes (Item 23) Sex is dirty and filthy Sex is healthy Able to talk about sex Father’s attitudes Sex is dirty and filthy Sex is healthy Able to talk about sex
Note. BSI = brother–sister incest; SSI = sister–sister incest; SBI = sister–brother incest; BBI = brother–brother incest; SI = sibling incest; ASI = adult sibling incest; PCI = parent–child incest; MDI = mother–daughter incest; MSI = mother– son incest; FDI = father–daughter incest; FSI = father–son incest. *p < .05, **p
