Examining the Roles of Victim–Perpetrator Relationship and Emotional Closeness in Judgments Toward a Depicted Child Sexual Abuse Case

Abstract

The current study investigated the impact that respondent gender, victim–perpetrator relationship, and the level of emotional closeness had on attributions in a hypothetical child sexual abuse case. A total of 160 university students read a hypothetical scenario depicting a female child sexually abused by an adult male. The perpetrator was either the victim’s biological father or her stepfather, with this relationship described as being either emotionally close or emotionally distant. Respondents read one of four (2 victim–perpetrator relationship × 2 emotional closeness) scenarios before completing 26 attribution items pertaining to credibility, blame, and severity. Principle components analysis yielded five factors, namely victim credibility, mother culpability, perpetrator culpability, assault severity, and victim culpability. Multivariate analysis of covariance—controlling for respondent (Caucasian vs. non-Caucasian) ethnicity—revealed, as predicted, significant main effects for respondent gender, victim–perpetrator relationship, and emotional closeness. In general, females assigned more provictim/ antiperpetrator/antimother attributions than males. Results were also suggested that both victim–perpetrator relationship and emotional closeness influence attributions made toward the victim, perpetrator, and nonoffending mother. Methodological issues and suggestions for future work are also discussed.

Keywords

child sexual abuse culpability credibility biological relatedness emotional closeness victim–perpetrator relationship gender differences

In recent years, there has been a flurry of academic interest in understanding perceptions of child sexual abuse (CSA). Several factors, such as victim gender, victim age, and victim–perpetrator relationship, are known to significantly impact on observers’ judgments of CSA perpetrators, CSA victims, and the victim’s nonoffending parents (e.g., Davies & Rogers, 2009; Graham, Rogers, & Davies, 2007).

Arguably the most robust finding is that women endorse more provictim and antiperpetrator views than men (e.g., Back & Lips, 1998; Graham et al., 2007; Rogers & Davies, 2007). One explanation for these gender differences is Shaver’s (1970) defensive attribution hypothesis (DAH). The DAH stipulates that individuals employ psychological strategies to defend themselves first from the possibility of ever being victimized (“harm avoidance”), and second to protect their own self-esteem should victimization ever occur (“blame avoidance”). Observers who see themselves as being characteristically similar to a given victim are less likely to attribute blame to that victim, as doing so would indirectly attack their own ego and self-esteem (Shaver, 1970). Thus, because of their shared gender, male observers may identify with, and defensively transfer blame away from, male CSA perpetrators and on to female victims (cf. Fergusson & Mullen, 1999). In contrast, female observers will identify more with CSA victims and attribute less victim blame in order to reduce their own feelings of personal vulnerability.

The nonoffending parents of CSA victims—particularly the mother—are also assigned a certain degree of blame for the child’s abuse. Waterman and Foss-Goodman (1984), for example, examined the levels of responsibility assigned to nonoffending parents in a depicted CSA case and found that while both parents were allocated some amount of responsibility, the mother was deemed more culpable than the father. More recent work suggests that maternal blame is greater when a CSA victim is younger (i.e., 10 years old as opposed to 15 years old), presumably because the former is assumed to be in need of more maternal protection and/or supervision (Rogers, Josey, & Davies, 2007). In sum, it seems that gender stereotypes about women being more nurturing and/or the main family caregiver also raise expectations about the role mother’s play in protecting children from harm (Back & Lips, 1998).

In general, sexual contact between genetically related kin is assumed maladaptive for species survival and thus, through nonconscious recognition of genetic relatedness, avoided in all animals (Lieberman, Fessler, & Smith, 2011). In contrast to this evolutionary model, other writers claim that sexual attraction between human relatives is curtailed merely by consciously acknowledged cultural taboos (e.g., Fraley & Marks, 2010; for a critique, see Lieberman et al., 2011).

Perceptions of CSA perpetrators also differ according to whether the perpetrator is or is not biologically related to the victim. In particular, it is society’s general prohibition toward biologically related kin showing sexual interest in each other that triggers greater negative evaluations about cases of incest compared to nonincestuous CSA (Eisenberg, Owens, & Dewey, 1987; Giles-Sims & Finkelhor, 1984). Considering incest cases to be more deplorable than nonincest cases could lead to perpetrators of incest being more negatively evaluated, and victims considered more traumatized than those in nonincest cases. Evidence backs up this possibility. Bornstein, Kaplan, and Perry (2007) found laypersons (i.e., undergraduate students and nonstudents) deemed CSA more severe and more traumatic for the victim—although just as likely to be repressed—if it was perpetrated by a parent (father or mother) versus a (male or female) babysitter with, in both cases, CSA deemed more severe/traumatic when the perpetrator was male. Interestingly, however, CSA committed by a father was also judged more common and more likely to reoccur than any other type of abuse (Bornstein et al., 2007). Moreover, and consistent with Bornstein et al.’s work, Davies and Rogers (2009) found CSA perpetration by a biological father was deemed more severe than CSA perpetration by a biologically unrelated family friend. Furthermore, and in line with their attitudes toward CSA, generally, women tend to hold more negative views of biologically related CSA (incest) than do men (Fessler & Navarette, 2004).

Given the perceived commonality of incestuous CSA and the research stated above, it seems plausible that children who report being sexually abused by their biological father will also be presumed less likely to lie and thus a more trustworthy and more credible witness than those who report being sexually assaulted by any other type of CSA perpetrator. Alternatively, it is possible that in the eyes of the public incestuous CSA is such a heinous crime (compared even to nonincestuous CSA) that any child who reports it must, for them, lack credibility.

Indeed, not all research indicates that victims will be more positively evaluated, or perpetrators negatively evaluated in incest, than nonincest cases. For example, Kelley (1990) found that child protection workers, nurses, and police officers deemed a sexually abusive father to be no more responsible, or worthy of equal punishment, than an abusive male neighbor. Furthermore, Finkelhor (1984) found that male health professionals underestimated the occurrence of father–daughter incest, and were more skeptical of the victim’s credibility, than female health professionals.

Inconsistencies within this area could be due to the samples being tested, with implications for how child welfare services respond to incestuous versus nonincestuous CSA. While others have failed to find differences in CSA blame attributions across various professional groups (e.g., psychologists vs. lawyers; Smith, 1996; schools counselors vs. schools principals vs. educational psychologists vs. schools teachers; Schlinder, 1998), Adams and Betz (1998) did find male counselors viewed incest to be less severe having fewer long-term negative effects, and were more doubting of the victim’s credibility, than their female counterparts, shows a lack of understanding about the severity of incest CSA.

Research might suggest a more complex picture than one solely related to biological relatedness. One further variable that could influence perceptions of incest versus nonincest CSA is the degree of emotional closeness that the victim and perpetrator shared before the abuse took place. Unpublished work by Pensinger (1996), for instance, suggests that (male) perpetrators of incest tend to feel less emotionally alienated than offenders of nonincestuous CSA. In terms of CSA perceptions, Davies and Rogers (2009) found CSA was deemed more severe, yet the abuser less culpable, when perpetrated by the child’s biological father as opposed to an unrelated, male, family friend. This, the authors speculated, suggests biological fathers are presumed more likely to partake in spontaneous as opposed to premeditated CSA by virtue of the greater emotional closeness shared with their victims (Davies & Rogers, 2009). In comparison, Reynolds and Birkimer (2002) found stepfather-perpetrated CSA was deemed more serious than neighbor-perpetrated CSA, suggesting that it is the perpetrator’s status as a father figure— rather than biological relatedness per se—that underpins judgments of CSA severity.

In sum, the significance of a biological element in paternity may be of less importance to CSA perceptions than the emotional closeness between the child and his/her² male caregiver (La Fontaine, 1990). If true, blame may be attributed to a perpetrator with an emotionally close relationship with the child because this may be seen as a greater betrayal of caregiver trust than when CSA is carried out by a perpetrator who is not emotionally close to the victim. To date, no studies have teased out the relative contributions of biological relatedness (vs. nonrelatedness) and emotional closeness (vs. distance) in their shaping of CSA perceptions. The current study seeks to address this omission.

The aim of the present study was to investigate the extent to which attributions of victim blame and credibility, perpetrator blame, nonoffending mother blame, and assault severity are influenced by both victim–perpetrator’s biological relationship and their emotional closeness in a depicted CSA case.³ In line with previous research (e.g., Rogers & Davies, 2007) and Shaver’s (1970) DAH the impact these relationship factors have on males’ versus females’ perceptions was also examined. The victim was depicted as a 13-year-old female who was sexually assaulted through nonconsensual genital fondling either by her biological father or her stepfather, with the victim–perpetrator relationship portrayed as either emotionally close or emotionally distant. The age 13 was chosen in line with previous studies, which have shown that attributions become more negative toward child victims, when the victim is pubertal or postpubertal (see, e.g., Davies & Rogers, 2009; Rogers & Davies, 2007; Rogers, Davies, Anderson, & Potton, 2011, for detailed discussions) and because 13 years is the modal age of female CSA victims (Fergusson & Mullen, 1999).

Several predictions were made. First, females were predicted to make more provictim (such as, considering the victim nonblameworthy and a more credible witness), more antiperpetrator judgments, be more blaming of the nonoffending parent (in this case, the victim’s biological mother (cf. Kelley, 1990; Suffoletta-Maierle, 2002), and deem the assault more severe than males regardless of condition. Second, respondents were expected to perceive CSA by the victim’s biological father to be more severe, though the perpetrator not necessarily more culpable, than CSA committed by the victim’s stepfather. In line with the incest taboo, it was further predicted that victims abused by their biological father would be considered more credible than those abused by their stepfather. Third, it was predicted that respondents would be particularly blaming toward the perpetrator and the victim’s nonoffending mother, when the perpetrator and victim had an emotionally close, as opposed to an emotionally distant, relationship. Finally, corresponding interaction effects were predicted with, for example, the assault deemed most severe and the perpetrator most blameworthy if CSA was committed by an emotionally close biological father. This is expected to be more pronounced among female respondents.

Method

Design

This study employed a 2 (respondent gender: male vs. female) × 2 victim–perpetrator relationship (biological father vs. stepfather) × 2 victim–perpetrator emotional closeness (emotionally close vs. emotionally distant) between- subjects design. Dependent measures comprised 26 attribution items relating to victim credibility, perpetrator, victim and nonoffending mother blame, and perceived assault severity. Conditions were randomly distributed across respondents with items presented in the same order throughout.

Respondents

Similar to previous studies of this kind, respondents comprised an opportunity sample of 160 undergraduate psychology students (80 males and 80 females) at a large university in the North West of England. Respondent age ranged from 18 to 45 years (M = 23.38 years; SD = 5.40 years; see below for further demographic details).

Materials

CSA vignettes

Each vignette was approximately 450 words in length and detailed a hypothetical CSA case in which a female child was sexually assaulted through genital fondling by a male perpetrator. The latter was either the victim’s biological father or her stepfather and had either an emotionally close or an emotionally distant relationship with the victim. In all conditions, the perpetrator was described as working from home and taking care of the child while the mother was out with friends. In the emotionally close condition, the perpetrator maintained a pleasant relationship and was actively involved with the victim’s interests, such as watching her sports sessions at school on a regular basis. In the emotionally distant condition, the victim–perpetrator relationship was described as unpleasant, where victim and perpetrator did not spend quality time together as part of their everyday activities. While details of the victim’s character and abuse context were written in the third person, details of the actual CSA situation were presented in the first-person account as part of the victim’s statement to the police. An example vignette can be found in the appendix.

Attribution measures

The 26 attribution items—adapted from Davies and Rogers (2004)—assessed perceptions of blame assigned to the perpetrator, the victim, and victim’s nonoffending mother, perceived assault severity, and the perceived credibility of the victim’s statement. All items were rated on a 7-point Likert-type scale ranging from 1 “strongly disagree” to 7 “strongly agree.” Initially, some items reflected a provictim/antiperpetrator/antimother/more severe standpoint with others reflecting the opposite. Prior to analysis, all items were (re)coded to reflect the former stance.

Procedure

A total of 160 questionnaires were distributed to an opportunity sample during breaks from scheduled psychology lectures at the University of Central Lancashire. Questionnaire booklets were handed out to potential participants personally by F.P. Volunteers were asked to read the instructions and vignette and then to complete all questionnaire in their own time without conferring. No payment or other incentive was given for participation. All 160 questionnaires were returned and used in analysis with respondents encouraged to keep a detachable debrief sheet containing various relevant helpline telephone numbers. All aspects of the study conferred to School and hence British Psychological Society (BPS) requirements.

Results

Demographics

As noted earlier, there was an even split of male and female respondents (80 each) with the entire sample averaging just over 23 years of age (M = 23.38 years; SD = 5.40 years). In addition, the majority of respondents were either Caucasian (46.3%; comprising 67 White British, 4 White Irish, and 3 White other nationals) or Asian (also 46.3%; comprising 46 Indian, 21 Pakistani, 1 Bangladeshi, and 2 other Asian nationals) with the remainder being of either Afro-Caribbean (3.8%), mixed race (1.9%) or “other” ethnic origin (1.9%). For simplicity, respondent ethnicity was dichotomized into Caucasian (46.3%) versus non-Caucasian (53.7%) categories. No other demographic data were collected.

Principal Components Analysis

Principal components analysis (PCA) was conducted on the 26 items using SPSS version 17.0. Kaiser’s normalization was used to limit the extracted factors to those with an eigenvalue greater than 1.0 with factor loadings greater than 0.40 suppressed. Priority was given to highest loadings with closely matched items (loadings <0.1) dropped from subsequent analyses. The resultant five-factor model explained 69.93% of the total variance in attribution scores. Details of initial eigenvalues, variance explained, and factor loadings for all factors are presented in Table 1.

Table 1.

PCA Initial Eigenvalues, Percentage Variance, and Factor Loadings for Each Factor

	Factor loadings
Factor and items	F1	F2	F3	F4	F5
Victim credibility
Eigenvalue 9.35: variance 22.13%
25. Dependable witness	0.88
07. Reliable witness	0.86
10. Trustworthy account	0.82
12. Victim’s statement credible	0.81
21. Victim memory	0.80
18. Victim’s statement accurate	0.78
05. Victim truthful	0.77
02. Victim no blame^a	0.42			0.41
Mother culpability
Eigenvalue 4.57: variance 17.62%
22. Mother responsibility		0.93
19. Mother should foresee		0.89
20. Mother blame		0.88
15. Mother guilt		0.84
06. Mother protection		0.83
09. Mother right to leave daughter		0.75
Perpetrator culpability
Eigenvalue 1.98: variance 11.77%
11. Perpetrator advantage			0.83
04. Perpetrator responsibility			0.76
08. Perpetrator abused role			0.69
24. Perpetrator blame^b			0.54		0.42
Assault severity
Eigenvalue 1.23: variance 10.74%
16. Upsetting experience				0.81
26. Police seriousness				0.70
14. Victim sympathy				0.62
03. Victim trauma^a	0.44			0.49
13. Perpetrator guilt^a			0.44	0.48
Victim culpability
Eigenvalue 1.05: variance 7.67%
23. Victim fault					−0.81
01. Victim blame					−0.68

Item omitted due to double-factor loadings.

Higher loading prioritized (n = 160).

As Table 1 illustrates, seven items relating to the victim’s trustworthiness, accuracy, and dependability as a witness loaded onto Factor 1. This had a very high internal reliability (α = 0.95) and was subsequently labeled “victim credibility.” Six items reflecting blame ascribed to the victim’s nonoffending mother loaded onto Factor 2, which also had very high internal reliability (α = 0.93) and was therefore labeled “mother culpability.” Factor 3 comprised four items, relating to perpetrator blame, guilt, responsibility, and abuse of the caregiver’s role. Factor 3 had high internal reliability (α = 0.81) and was thus labeled “perpetrator culpability.” Similarly, Factor 4 comprised three items associated with perceptions of assault seriousness and the likely impact this would have for the victim. This also had acceptable internal reliability (α = 0.75) and was labeled “assault severity.” Finally, two items relating to victim blame and fault loaded on to Factor 5, which had moderate but acceptable internal reliability (α = 0.65) and was labeled “Victim Culpability.” These factor names are consistent with those in previous CSA attribution studies (e.g., Davies & Rogers, 2009). Preliminary checks found five cases that were outliers across two or more factors. These were removed from the data set (final n = 155).

Overall, respondent age was unrelated to all five factors. In contrast, significant respondent (Caucasian vs. non-Caucasian) ethnicity differences subsequently found for mother culpability, t(150) = 2.84; p = .005, with Caucasian participants (M = 4.85; SD = 1.76) assigning higher blame scores to the nonoffending mother than non-Caucasians (M = 4.07; SD = 1.65). Conversely, Caucasian participants (M = 1.28; SD = 0.54) assigned lower victim culpability scores, t(150) = −2.16; p = .032 than their non-Caucasian counterparts (M = 1.55; SD = 0.92). Future analyses will control for respondents’ (Caucasian vs. non-Caucasian) ethnicity. Finally, factor intercorrelations found no evidence of multicollinearity (all r < .47).

Multivariate Analysis of Covariance

A 2 (respondent gender) × 2 (victim–perpetrator relationship) × 2 (emotional closeness) between-subjects multivariate analysis of covariance (MANCOVA)—controlling for respondent’s (Caucasian vs. non-Caucasian) ethnicity—was performed across the five factors. Adjusted means and standard errors are given in Table 2.

Table 2.

Mean Factor Scores^a Across Victim–Perpetrator Relatedness, Their Emotional Closeness, and Respondent Gender (Controlling for Respondents’ Caucasian vs. non-Caucasian Ethnicity)

		Biological father						Stepfather						All
	Resp.	Close		Distant		All		Close		Distant		All		Close		Distant		All		Sig.
Factor	Gender	M	(SE)	M	(SE)	M	(SE)	M	(SE)	M	(SE)	M	(SE)	M	(SE)	M	(SE)	M	(SE)	Effects
Victim credibility	Male	6.10	(0.18)	4.99	(0.20)	5.54	(0.13)	5.17	(0.19)	4.80	(0.18)	4.98	(0.13)	5.63	(0.13)	4.90	(0.13)	5.26	(0.09)	B***
	Female	6.96	(0.18)	5.88	(0.18)	6.42	(0.12)	5.79	(0.18)	5.95	(0.19)	5.87	(0.13)	6.37	(0.13)	5.92	(0.13)	6.14	(0.09)	C***
	All	6.53	(0.13)	5.43	(0.13)	5.98	(0.09)	5.48	(0.13)	5.38	(0.13)	5.43	(0.09)	6.00	(0.09)	5.41	(0.09)	5.70	(0.06)	G***
																				B×C***
Mother culpability	Male	2.05	(0.31)	4.42	(0.35)	3.23	(0.24)	4.94	(0.33)	4.61	(0.32)	4.77	(0.23)	3.50	(0.23)	4.51	(0.24)	4.00	(0.17)	B*
	Female	4.63	(0.32)	5.58	(0.32)	5.11	(0.22)	4.96	(0.32)	4.28	(0.34)	4.62	(0.23)	4.80	(0.23)	4.93	(0.23)	4.86	(0.16)	C*
	All	3.34	(0.22)	5.00	(0.24)	4.17	(0.16)	4.95	(0.23)	4.45	(0.23)	4.70	(0.16)	4.15	(0.16)	4.72	(0.16)	4.43	(0.11)	G***
																				B×C***
																				B×G***
Perpetrator culpability	Male	6.55	(0.12)	6.41	(0.14)	6.48	(0.09)	6.61	(0.13)	6.26	(0.12)	6.44	(0.09)	6.58	(0.09)	6.34	(0.09)	6.46	(0.06)	G**
	Female	6.90	(0.13)	6.51	(0.12)	6.71	(0.09)	6.59	(0.12)	6.86	(0.13)	6.72	(0.09)	6.74	(0.09)	6.69	(0.09)	6.71	(0.06)	B×C×G*
	All	6.72	(0.09)	6.46	(0.09)	6.59	(0.06)	6.60	(0.09)	6.56	(0.09)	6.58	(0.06)	6.66	(0.06)	6.51	(0.06)	6.59	(0.04)
Assault severity	Male	6.65	(0.13)	6.54	(0.15)	6.59	(0.10)	5.96	(0.14)	6.45	(0.13)	6.20	(0.10)	6.31	(0.10)	6.49	(0.10)	6.40	(0.07)	B*
	Female	6.76	(0.14)	6.51	(0.13)	6.63	(0.09)	6.47	(0.13)	6.77	(0.14)	6.62	(0.10)	6.61	(0.10)	6.64	(0.10)	6.63	(0.07)	G*
	All	6.70	(0.09)	6.53	(0.10)	6.61	(0.07)	6.22	(0.10)	6.61	(0.10)	6.41	(0.07)	6.46	(0.07)	6.57	(0.07)	6.51	(0.05)	B × C**
Victim culpability	Male	1.82	(0.16)	1.98	(0.18)	1.90	(0.12)	1.16	(0.17)	1.62	(0.16)	1.39	(0.12)	1.49	(0.12)	1.80	(0.12)	1.64	(0.09)	G**
	Female	1.05	(0.17)	1.22	(0.16)	1.13	(0.11)	1.29	(0.16)	1.35	(0.17)	1.32	(0.12)	1.17	(0.12)	1.28	(0.12)	1.23	(0.08)	B × G**
	All	1.43	(0.12)	1.60	(0.12)	1.51	(0.08)	1.22	(0.12)	1.49	(0.12)	1.35	(0.08)	1.33	(0.08)	1.54	(0.08)	1.43	(0.06)

Adjusted means. Range: 1 “strongly disagree” to 7 “strongly agree”; higher factor scores reflect a more provictim/antiperpetrator/antimother/severe stance. Sig.: respondent gender (G), biological relatedness (B), emotional closeness (C), and subsequent interaction effects at the *p < .05, **p < .01, and ***p < .001 levels (two-tailed: n = 155).

MANCOVA revealed respondent ethnicity to be a significant multivariate covariate, Wilk’s λ = 0.91; F(5,139) = 2.64; p = .026; η² = 0.09 with Caucasians generally making more provictim/antiperpetrator/antimother/more severe attributions than non-Caucasians. In addition, highly significant multivariate main effects were found for respondent gender, Wilk’s λ = 0.66; F(5,139) = 14.14; p < .001; η² = .34, victim–perpetrator relationship, Wilk’s λ = 0.82; F(5,139) = 6.15; p < .001; η² = .18, and emotional closeness, Wilk’s λ = 0.80; F(5,139) = 6.92; p < .001; η² = .20. Significant two-way multivariate interaction effects were also found for both respondent gender × victim–perpetrator relationship, Wilk’s λ = 0.81; F(5,139) = 6.24, p < .001; η² = .18, and victim–perpetrator relationship × emotional closeness, Wilk’s λ = 0.80; F(5,139) = 6.82; p < .001; η² = .20. No other significant multivariate effects were forthcoming.

Post hoc univariate ANCOVA confirmed respondent (Caucasian vs. non-Caucasian) ethnicity to be a significant covariant of just one factor, namely mother culpability, F(1,144) = 11.53; p = .001; η² = .08, with Caucasians (M = 4.85; SD = 1.76) judging the victim’s nonoffending mother more culpable than non-Caucasians (M = 4.07; SD = 1.65). Having controlled for respondent ethnicity, numerous significant univariate effects were found for each of the five separate factors (see Table 2 for adjusted means).

First, significant main effects were found across respondent gender for all five factors, with males perceiving the victim to be less credible, F(1,143) = 45.06; p < .001; η² = 0.24, and more culpable, F(1,143) = 11.89; p = .001; η² = 0.08, both the perpetrator F(1,143) = 7.90; p = .006; η² = 0.05, and the victim’s nonoffending mother, F(1,143) = 13.42; p < .001; η² = 0.09, less culpable, and the assault less severe, F(1,143) = 5.26; p = .023; η² = 0.04 than females.

Second, significant differences across victim–perpetrator relationship were found for three factors, with respondents judging the victim more credible F(1,143) = 18.74; p < .001; η² = 0.12, her nonoffending mother less culpable, F(1,143) = 5.27; p = .023; η² = 0.04, and the assault more severe, F(1,143) = 4.40; p = .038; η² = 0.03, when CSA was perpetrated by the victim’s father as opposed to her stepfather.

Third, a significant emotional closeness effect was found for two factors with respondents rating the victim more credible F(1,143) = 21.42; p < .001; η² = 0.13, and the girl’s nonoffending mother less culpable, F(1,143) = 6.24; p = .014; η² = 0.04, when CSA was perpetrated by a father figure (i.e., a male caregiver regardless of biological relatedness) who was emotionally close rather than emotionally distant.

Fourth, highly significant victim–perpetrator relationship × emotional closeness interactions were found for three factors, namely victim credibility, F(1,143) = 14.67; p < .001; η² = 0.09, mother culpability, F(1,143) = 21.60; p < .001; η² = 0.13, and assault severity, F(1,143) = 8.41; p = .004; η² = 0.06. Results for subsequent post hoc simple effects analysis (via ANCOVA controlling for respondents’ Caucasian versus non-Caucasian ethnicity and with α adjusted to 0.025)⁴ are summarized in Table 3. As Table 3 shows post hoc tests confirmed the victim was deemed more credible and her nonoffending mother less culpable, when CSA was perpetrated by an emotionally close, as opposed to emotionally distant, biological father. The assault was also judged more severe in this context too. Finally, the victim was deemed more credible, the mother less culpable, and the assault more severe when the girl was sexually abused by an emotionally close biological father as opposed to stepfather.

Table 3.

Summary of Two- and Three-Way Post Hoc Simple Effects (Via ANCOVA Controlling for Respondents’ Caucasian Vs. Non-Caucasian Ethnicity) for All Factors

Factor	Interaction		Comparison		Adj. means	F	p	Sig.	Direction of effect
Victim credibility	B × C	01.	Biological father	Close vs. distant	6.53 vs. 5.43	F(1,73) = 24.95	<.001	***	Victim more credible if close
		02.	Close	Biological vs. step	6.53 vs. 5.48	F(1,75)) = 31.65	<.001	***	Victim more credible if biological
Mother culpability	B × C^a	01.	Biological father	Close vs. distant	3.34 vs. 5.00	F(1,73)) = 17.75	<.001	***	Mother less culpable if close
		02.	Close	Biological vs. step	3.34 vs. 4.95	F(1,75)) = 17.26	<.001	***	Mother less culpable if biological
	B × G	01.	Biological father	Male vs. female	3.23 vs. 5.11	F(1,73)) = 24.40	<.001	***	Mother less culpable if male
		02.	Male	Biological vs. step	3.23 vs. 4.77	F(1,71)) = 20.80	<.001	***	Mother less culpable if biological
	C × G	01.	Close	Male vs. female	3.50 vs. 4.80	F(1,75)) = 11.52	.001	**	Mother less culpable if male
		02.	Male	Close vs. distant	3.50 vs. 4.51	F(1,71) = 7.66	.007	**	Mother less culpable if close
Perpetrator culpability	B × C × G^b	01.	Stepfather and distant	Male vs. female	6.26 vs. 6.44	F(1, 35) = 8.62	.006	**	Perpetrator Less culpable if male
Assault severity	B × C	01.	Stepfather	Close vs. distant	6.22 vs. 6.61	F(1,73) = 6.52	.013	*	Assault less severe if close
		02.	Close	Biological vs. step	6.70 vs. 6.22	F(1,75) = 9.98	.002	**	Assault more severe if biological
	B × G	01.	Stepfather	Male vs. female	6.20 vs. 6.62	F(1,73) = 5.98	.017	*	Assault less severe if male
		02.	Male	Biological vs. step	6.59 vs. 6.20	F(1,71) = 6.33	.014	*	Assault more severe if biological
Victim culpability	B × G	01.	Biological father	Male vs. female	1.90 vs. 1.13	F(1,73) = 18.02	<.001	***	Victim less Culpable if male
		02.	Male	Biological vs. step	1.90 vs. 1.39	F(1,71) = 5.24	.025	*	Victim less culpable if biological

Biological relatedness (B); emotional closeness (C); respondent gender (G). Higher factor scores reflect a more provictim/antiperpetrator/antimother/severe stance. Alpha adjusted to 0.025 except for ^awhere α set at 0.017 and ^bwhere α set at 0.0083. Sig. effects at *p < .05, **p < .01, and ***p < .001 levels (two-tailed). No two-way interactions for perpetrator culpability.

Fifth, a significant victim–perpetrator relationship × respondent gender interactions was found for both mother culpability, F(1,143) = 19.55; p < .001, η² = 0.12, and victim culpability, F(1,143) = 8.63; p = .004; η² = 0.06, ratings. Simple effects analysis revealed that males judged both the victim and her nonoffending mother to be less culpable when CSA was perpetrated by the girl’s biological father rather than by her stepfather. Male respondents also viewed the victim and her mother to be less culpable for CSA perpetrated by a biological father, and CSA perpetrated by the victim’s stepfather to be less severe, than did female respondents.

Finally, a significant three-way interaction was found for just one factor, namely perpetrator culpability, F(1,143) = 5.69; p = .018; η² = 0.04. Examination of line graphs suggested the need for six post hoc comparisons. Simple effects analysis (again via ANCOVA controlling for respondents’ Caucasian versus non-Caucasian ethnicity but this time with α set to 0.0083) confirmed that males viewed the emotionally distant stepfather to be more culpable than did females. No other significant effects were found.

Discussion

Results revealed broad support for predictions. Consistent with expectations females made more pro-victim attributions, deemed the perpetrator more culpable, and the assault more severe than did males. These trends are consistent with previous research (e.g., Back & Lips, 1998; Bottoms & Goodman, 1994; Graham et al., 2007; Rogers & Davies, 2007) and lend further support for Shaver’s (1970) DAH. It appears men are more inclined to identify with, and so defensively attribute blame away from, a male perpetrator of (child) sexual assault; with females, in contrast, more disposed to make direct comparisons between themselves and the (child) victim.

Of particular interest was the finding that women also judged the victim’s nonoffending mother to be more culpable for her daughter’s CSA than did men (cf. Back & Lips, 1998; Waterman & Foss-Goodman, 1984), particularly when the perpetrator was the child’s biological father. It is possible female respondents merely endorsed stronger gender role stereotypes about the role women as the family protector (cf. Rogers et al., 2007). Alternatively, females may have identified more with the victim’s nonoffending mother and so, in line with Shaver’s (1970) notion of defensive “harm avoidance,” perhaps assigned more blame to the mother as a means of psychologically distancing themselves from the possibility of similar abuse ever happening to their own child in the future. This is a speculative suggestion and certainly worth of future investigation.

Future research should examine further the reasons why blame is attributed to nonoffending mothers. In real-world cases of CSA, nonoffending mothers (indeed, nonoffending parents generally) may also suffer negative consequences such as depression, guilt, and self-blaming following their child’s sexual abuse (e.g., Hernandez et al., 2009) and if subsequently blamed by society for allowing their child to suffer, will themselves be subjected to a form of secondary victimization. This is particularly true of high-profile CSA cases (such as those that also involve child abduction and/or murder) where negative attitudes toward the victim’s parents are often presented through as least some sections of the media (see Davies, 2011, for a longer discussion in relation to child cases in the media). In short, in CSA cases, investigating the blame issued to nonoffending parents can be in every way as damaging as blaming the victim, and for the same reasons should be considered worthy of more research.

In the current study, CSA by the biological father was, as predicted, deemed more severe than identical abuse perpetrated by the stepfather. In support of with previous work (Davies & Rogers, 2009), it seems respondents recognized incest may have more serious long-term negative consequences for the victim (e.g., through the additional loss of a trusted caregiver) than nonincestuous CSA (see, e.g., Fergusson & Mullen, 1999). In addition, the victim’s biological father was deemed just as culpable for the abuse as the stepfather, suggesting that some culpability was assigned to the perpetrator regardless of the biological relatedness between him and his victim. Given that natural fathers are seen to be less culpable for CSA against their own child than strangers are for CSA against someone else’s child (Davies & Rogers, 2009), it seems that in some situations, perpetrator culpability is assigned according to factors other than the perpetrator’s biological relatedness to his victim. Considering the emotional closeness of the victim–perpetrator relationship could play an important part in shaping these attributions.

Contrary to expectations, current findings failed to support the incest taboo (Giles-Sims & Finkelhor, 1984) with the depicted victim deemed more—not less—credible when abused by her biological father as opposed to her stepfather. It seems respondents felt the child had less reasons to lie and make false allegations of CSA perpetrated by her biological father. This may say more about perceptions of step relationships, particularly when a teenage girl is involved (see, e.g., Fine, Coleman, & Ganong, 1998). More research is needed to explore this line of enquiry further.

Several interesting findings emerged in relation to manipulations in the victim–perpetrator’s level of emotional closeness. First, contrary to expectations, the victim’s nonoffending mother was deemed less—rather than more—blameworthy when CSA was perpetrated by an emotionally close, rather than distant, father figure. One interpretation is that respondents felt the mother should have been aware of this emotional distance and subsequently monitored this relationship more closely. That is, mother-blame ratings could reflect the extent to which respondents perceived the victim’s nonoffending mother as having neglected her child’s safety and welfare needs (cf. Rogers, Wczasek, & Davies, 2011) Again, this line of enquiry necessitates further research, though the degree to which nonoffending parents are seen as not innocent for various reasons is subject to media attention in real-world, high-profile child cases (see Davies, 2011). Second, and again contrary to predictions, the perpetrator was considered equally blameworthy regardless of whether he had an emotionally close or distant relationship with the victim. This suggests that the emotional quality of the victim–perpetrator relationship is of little importance when determining perpetrator blame, at least in cases of CSA.

While emotional closeness had no impact on attributions of victim culpability, the victim was deemed more credible when she was abused by an emotionally close—rather than emotionally distant—father figure. Like CSA perpetrated by a biological father, it seems respondents believe victims have less reasons to lie and falsely accuse a perpetrator with whom they have an emotionally close relationship. For instance, it could be respondents believed that a 13-year-old girl who felt neglected by her (biological or step) father figure is more capable of making a false accusation of CSA against him in some, albeit extreme, attention-seeking ploy.

These findings were supported by subsequent interactions between the victim–perpetrator’s biological relatedness and emotional closeness. First, the victim was deemed more credible, and the victim’s nonoffending mother less culpable, when CSA was perpetrated by an emotionally close, as opposed to emotionally distant, biological father.

The implication here is that respondents believed the child had little reason to falsely accuse and that her mother had little reason to suspect, a loving, natural father of sexually abusing his own daughter.

Until now no research has investigated specific nuances in respondents’ perceptions of CSA when both the biological relatedness and emotional closeness of the victim and perpetrator are systematically varied. While the current study provides some insight into perceptions toward a specific case, future research needs extending to consider similar factors in a range of contexts (e.g., with male CSA victims and/or involving different types of CSA, cf. Davies & Rogers, 2009; Graham et al., 2007). Furthermore, in real-world cases, negative judgments that vary depending on the victim’s apparent emotional closeness to her perpetrator need to be challenged in treatment and legal settings. In the worst-case scenario, such negative judgments can mean that justice goes undone and victims are left uncared for.

Given that, generally, men made more negative attributions toward the victim and her nonoffending mother than women, it is of little surprise that these trends continued when both biological relatedness and emotional closeness manipulations were also taken into account. Specifically, males judged both the victim and her nonoffending mother to be less culpable, and the assault more severe, when CSA was perpetrated by the victim’s biological father and additionally in the case of mother culpability, also when the victim and perpetrator shared an emotionally close relationship.

Methodological Issues, Implications, and Conclusion

Unlike almost all previous studies of this nature (e.g., Davies & Rogers, 2009; Graham et al., 2007; Rogers & Davies, 2007), the current sample was not predominantly of Caucasian ethnicity but included a large proportion of Asian (mainly Indian or Pakistani) respondents. This was a reflection of the snowball sampling technique adopted here. No previous research using a United Kingdom sample has investigated ethnicity differences in CSA attributions, so examining of respondent ethnicity as a covariate is another novel aspect of the current study.⁵ Current findings suggest Caucasian participants blamed the victim’s nonoffending mother more than did their non-Caucasian counterparts. These trends may reflect cross-cultural differences in the way (predominantly) White British versus Indian/Pakistani respondents view either the sexuality of teenage girls and/or the role of women within the Asian family unit.

For instance, differences in attitudes toward the role mothers play in child-rearing practices across ethnicity may impact on the level of blame they receive if ever their child is harmed (Jambunathan, Burts, & Pierce, 2000). CSA in a culture where the mother has a more central, controlling, and proactive role in promoting child welfare is likely to afford even greater maternal blame compared to CSA in a culture where the mother’s protective function is not so clearly prescribed. While current findings provide a preliminary indication of respondent ethnicity differences in CSA attributions (specifically, toward nonoffending mothers), more work is needed to consider cross-cultural variations in CSA attributions more thoroughly both within and across national boundaries.

A second methodological issue concerns the recruitment of (predominantly final year) undergraduate psychology student many of whom would have had some knowledge of forensic psychology and/or attributional processes. One might reasonably assume that some knowledge about, for example, attribution theories, or studies that utilize vignette methodologies, would influence their perceptions rendering current findings somewhat lacking in generalizability. However, despite—or perhaps even because of—their academic background, clear and specific differences were found in how current respondents viewed the victim, perpetrator, and the nonoffending mother in a hypothetical CSA scenario. Indeed, it can be argued that respondents’ level of psychological training had no effect on how they perceived this particular case. Although research does need to be extended to other (e.g., general population/community) samples (cf. Graham et al., 2007), the fact that psychology students may make biased attributional judgments is of important consideration in its own right. As noted earlier, many mental health and welfare professionals will maintain more negative views toward incest over nonincestuous CSA survivors (e.g., Kelley, 1990). As many (psychology) students are likely to foster careers that involve contact with CSA and other sexual assault survivors (e.g., counseling, social work, the police, and educational services), current findings suggest that those about to enter these professions might be prone to similar prejudices as those without specific psychological training. Thus, another practical application of the current study is to help educate potential welfare workers in understanding these CSA myths.

To conclude, the present study offers novel findings in how United Kingdom respondents of various ethnic groups judge child victims of sexual abuse following experimental manipulations in both the degree of (biological) relatedness, and the degree of emotional closeness between a 13-year-old, female victim and male perpetrator of CSA. Advances in CSA attribution research both here and elsewhere have shown how lay perceptions can vary depending on small manipulations in victim, perpetrator, and assault characteristics. Such findings offer information to those working with individuals affected by CSA and highlight how various factors can influence how people think and feel about victims, and ultimately how victims are treated after their abuse. Future work should endeavor to not only continue this important line of inquiry, but to further educate the wider population as well as welfare professionals about how damaging negative judgments can be to CSA victims and their families.

Footnotes

Appendix: Vignette

Emma is a 13-year-old girl with a bubbly personality. She is very friendly and popular among her family and friends. She enjoys school, and her term reports are always exceptional. She is very active and is part of the school netball team. She attends practice at least three times a week.

Emma is an only child and lives with her mother, Amy and stepfather, David. He is self-employed and for the majority of the time works from home. He is mainly responsible for the school runs as Amy works long hours.

Emma and David have never/always maintained a pleasant relationship. They enjoy/hardly spend(ing any) time together, and David often/never attends her sports practices and tournaments. Emma feels she can/not approach David, when needing help for school homework, and often/rarely go on excursions and days out.

One day, Emma was somewhat reserved. He mother noticed that this was not like her and asked what was wrong. Emma seemed reluctant to tell her, but eventually broke down and told her what happened. Her mother was shocked, and immediately took Emma to the police station. The following is Emma’s statement, as noted by the police:

My mum went out that evening . . . she always does on a Saturday with her friend, and me and David just stayed in.

I was just in my room reading a book for school, when David came in, and he asked me if I wanted anything to eat. I said no, and so he just came in and sat next to me on the bed. He put his arm around me and asked me about the book, and I told him how I am really enjoying it and would like him to read it. He scanned through the book with me and he got really close to me.

David moved his hand downwards and he touched me here (points to her genitals). I felt scared and really uncomfortable. I knew it was wrong and I just froze and couldn’t get myself to look at him. I started to cry but David didn’t seem to care and he kept his hand there for ages. Then the doorbell rang downstairs and David quickly got up and left the room. I just wished that my mum was there.

Emma’s mother cannot believe what has happened and the police investigation is currently taking place.

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) received no financial support for the research, authorship, and/or publication of this article.

Notes

Bios

Michelle Davies is a senior lecturer in psychology at the University of Central Lancashire, United Kingdom. Her research interests include male rape and child sexual abuse, attitudes towards violence, fear of crime, and sexuality issues.

Fehmida Patel is a doctoral student in clinical psychology at Lancaster University, United Kingdom. The current study was conducted as part of her undergraduate work at the University of Central Lancashire, United Kingdom.

Paul Rogers is a senior lecturer in psychology at the University of Central Lancashire, United Kingdom. His research interests include beliefs in the paranormal, attitudes towards child sexual abuse, and social cognitive processes.

References

Adams

E. M.

Betz

N. E.

(1993). Gender differences in counselors’ attitudes toward attributions about incest. Journal of Counselling Psychology, 40, 210-216.

Back

Lips

H. M.

(1998). Child sexual abuse: Victim age, victim gender and observer gender as factors contributing to attributions of responsibility. Child Abuse & Neglect, 22, 1239-1252.

Bornstein

B. H.

Kaplan

D. L.

Perry

A. R.

(2007). Child abuse in the eyes of the beholder: Lay perceptions of child sexual and physical abuse. Child Abuse & Neglect, 31, 375-391.

Bottoms

B. L.

Goodman

G. S

(1994). Perceptions of children’s credibility in sexual assault cases. Journal of Applied Social Psychology, 24, 702-732.

Davies

(2011). Working with adult and child victims of crime. In Davey

(Ed.), Introduction to applied psychology (Chap. 13, pp. 251-265). Chichester, UK: Wiley.

Davies

Rogers

(2004). Attributions towards victims and perpetrators in a child sexual abuse case: Roles of respondent, perpetrator, and victim gender. Forensic Update, 79, 17-23.

Davies

Rogers

(2009). Perceptions of blame and credibility toward victims of childhood sexual abuse: Differences across victim age, victim-perpetrator relationship, and respondent gender in a depicted case. Journal of Child Sexual Abuse, 18, 78-92.

Eisenberg

Glynn-Owens

Dewey

M. E.

(1987). Attitudes of health professionals to child sexual abuse, Child Abuse & Neglect, 11, 109-116.

Fergusson

D. M.

Mullen

P. E.

(1999). Childhood sexual abuse: An evidence based perspective. London, UK: Sage.

10.

Fessler

D. M. T.

Navarette

(2004). Third party attitudes toward sibling incest: Evidence for Westermarck’s hypotheses. Evolution & Human Behaviour, 25, 277-294.

11.

Fine

M. A.

Coleman

Ganong

L. H.

(1998). Consistency in perceptions of the step-parent role among step-parents, parents and stepchildren. Journal of Social & Personal Relationships, 15, 810-828.

12.

Finkelhor

(1984). (Ed.). Child sexual abuse: New theory and research. New York, NY: Free Press.

13.

Fraley

Marks

M. J.

(2010). Westermarck, Freud, and the incest taboo: Does familial resemblance activate sexual attraction? Personality & Social Psychology Bulletin, 36, 1202-1212.

14.

Giles-Sims

Finkelhor

(1984). Child abuse in stepfamilies. Journal of Family Relations, 33, 407-413.

15.

Graham

Rogers

Davies

(2007). Attributions in a hypothetical child sexual abuse case: Roles of abuse type, family response and respondent gender. Journal of Family Violence, 22, 733-745.

16.

Hernandez

Ruble

Rockmore

McKay

Messam

Harris

Hope

(2009). An integrated approach to treating non-offending parents affected by sexual abuse. Social Work in Mental Health, 7, 533-555.

17.

Jambunathan

Burts

D. C.

Pierce

(2000). Comparisons of parenting attitudes among five ethnic groups in the United States. Journal of Comparative Family Studies, 31, 395-406.

18.

Kelley

S. J.

(1990). Responsibility and management strategies in child sexual abuse: A comparison of child protective workers, nurses, and police officers. Child Welfare: Journal of Policy, Practice, and Program, 69, 43-51.

19.

La Fontaine

(1990). Child sexual abuse. Cambridge, MA: Polity Press.

20.

Lieberman

Fessler

D. T.

Smith

(2011). The relationship between familial resemblance and sexual attraction: An update on Westermarck, Freud, and the incest taboo. Personality & Social Psychology Bulletin, 37, 1229-1232.

21.

Pensinger

(1996, November). Breaking the incest taboo: The role of emotional alienation, cognitive distortions, and proximity of relationship in child sexual abuse. Dissertation Abstracts International, 57, 3418.

22.

Pollard

(1992). Judgements about victims and attackers in depicted rapes: A review. British Journal of Social Psychology, 31, 309-326.

23.

Reynolds

L. L.

Birkimer

J. C.

(2002). Perceptions of child sexual abuse: Victim and perpetrator characteristics, treatment efficacy, and lay vs. legal opinions of abuse. Journal of Child Sexual Abuse, 11, 53-73.

24.

Rogers

Davies

(2007). Perceptions of credibility and attributions of blame towards victim in a childhood sexual abuse case: Gender and age factors. Journal of Interpersonal Violence, 22, 566-584.

25.

Rogers

Davies

Anderson

Potter

(2011). Does norm theory explain negative perceptions of child sexual abuse victims? Journal of Interpersonal Violence, 26, 1911-1933.

26.

Rogers

Josey

Davies

(2007), Victim age, attractiveness and abuse history as factors in the perception of a hypothetical child sexual abuse case. Journal of Sexual Aggression, 13(2), 121-137.

27.

Rogers

Wczasek

Davies

(2011). Attributions of blame in a hypothetical internet grooming case: Roles of victim naivety, parental neglect & respondent gender. Journal of Sexual Aggression, 17, 196-214.

28.

Schindler

(1998, July). School professionals’ attributions of blame and related attitudes concerning father-daughter incest. Dissertation Abstracts International Section A, 59, 0096.

29.

Schoenberg

N. E.

Ravdal

(2000). Using vignettes in awareness and attitudinal research. International Journal of Social Research Methodology, 3, 63-74.

30.

Shaver

K. G.

(1970). Defensive attribution: Effects of severity and relevance on the responsibility assigned for an accident. Journal of Personality & Social Psychology, 14, 101-113.

31.

Sleed

Durrheim

Kriel

Solomon

Baxter

(2002). The effectiveness of the vignette methodology: A comparison of written and video vignettes in eliciting responses about date rape. South African Journal of Psychology, 32, 21-28.

32.

Smith

(1996, November). Allegations of child sexual abuse: Judgments of credibility by psychologists and judges. Dissertation Abstracts International, 57, 3424.

33.

Suffoletta-Maierle

(2002, June). Assigning responsibility to nonoffending mothers in child sexual abuse cases: The effects of a brief educational intervention. Dissertation Abstracts International, 62, 5980.

34.

Waterman

C. K.

Foss-Goodman

(1984). Child molesting: Variables relating to attribution of fault to victims, offenders, and non-participating parents. The Journal of Sex Research, 20, 329-349.