A Criminal Careers Typology of Child Sexual Abusers

Abstract

We present a criminal careers typology of child sexual abusers constructed in terms of their offending persistence (persistent vs. limited) and versatility (specialized vs. versatile). Analyses were conducted on the official records of 362 convicted offenders, 213 of whom also provided confidential self-report data on their personal and offending histories. Forty-one percent of the sample were currently serving sentences for their first sexual offense conviction(s) but had at least one prior conviction for a nonsexual offense (limited/versatile); 36.4% had no previous convictions of any kind (limited/specialized); 17.8% had prior convictions for sexual and nonsexual offenses (persistent/versatile); and 4.8% had prior convictions for sexual offenses only (persistent/specialized). These four groups differed on a range of personal and offense-related variables, including abuse histories, sexual orientation, age at first sexual contact with a child, number of victims, duration of sexual involvement with victims, victim gender, and whether victims were familial or nonfamilial. These differences suggest the need to adopt different treatment and prevention strategies that target the specific characteristics of each group.

Keywords

sexual offenders sexual abuse typology criminal careers

The idea that offenders have criminal careers that parallel conventional careers has proved a useful analogy for criminologists examining individual patterns of offending over time. Conventional careers have a beginning and an end, and individuals may move in and out of particular occupations as circumstances change. Criminal careers researchers, then, are interested in when and how offenders commence their involvement in crime, the type and frequency of crimes they commit, and their desistance from criminal activity. Criminal careers research can help identify critical risk and protective factors along the developmental pathways of crime, leading to more effective crime prevention policies and practices. The present study considers the applicability of the criminal careers model to child sexual abusers. It presents a typology based on variations in offending persistence and versatility, arguing that child sexual abusers are a heterogeneous group and that their different patterns of offending require different types of prevention strategies.

Offending Persistence and Versatility

Two key variables in criminal careers research are offending persistence and versatility. Persistence refers to the extent of an offender’s criminal involvement over time. That offenders vary markedly in the frequency and duration of their offending has been of enduring interest to criminologists (see DeLisi & Piquero, 2011, for a recent review). In particular, it has been recognized for some time that a small group of persistent offenders is responsible for a disproportionately large amount of crime and that these offenders might therefore warrant special theoretical and criminal justice attention (e.g., Wolfgang, Figlio, & Sellin, 1972). These so-called career criminals—or life-course-persistent offenders as Moffitt (1993) labels them—commit offenses frequently, over an extended period. According to Moffitt, signs of antisocial commitment in these offenders are evident in early childhood, and their criminal dispositions are more or less stable over their life-course.

In contrast to career criminals, most offenders’ involvement in crime is transitory, often linked to specific stages of development or particular life events (Piquero, 2004). Many of these offenders may be what Moffitt (1993) refers to as adolescence-limited, although others may commence and cease their criminal involvement as adults. Their propensity to commit crime is not stable and individuals may move in and out of criminal involvement as circumstances change. The antisocial commitment of these offenders is at most ambivalent, and many offenders in this category have otherwise conventional value orientations. Research has shown that casual involvement in criminal behavior is so ubiquitous—with up to 80% of young males having some contact with police for delinquent behavior—that not engaging in crime at some time or other is statistically abnormal (Farrington, Ohlin, & Wilson, 1986).

Versatility refers to the extent to which offenders commit a variety of offenses on the one hand, or specialize in particular crimes on the other. The weight of empirical evidence indicates that most offenders are generalists rather than specialists (DeLisi & Piquero, 2011; Gottfredson, 2006). An implication of offending versatility is that fewer separate theories are required to explain different types of crime. Based on evidence of offending versatility, Gottfredson and Hirschi (1990) proposed a general theory of crime. They argued that the propensity to commit crime is widely distributed in the community, and that the basic cause of criminal behavior is universal—an absence of restraint. Criminal behavior is intrinsically rewarding and requires no special motivation or pathology, while criminal acts themselves generally demand little in the way of specialized skills or experience. Offenders do not learn to commit crimes, at least not initially; rather, they fail to learn not to commit them. Accordingly, low self-control can manifest in a wide range of behaviors. The offender who succumbs to the temptation to steal may also be easily provoked to assault, engage in irresponsible sexual behavior, and so forth.

Others have argued that there is an interaction between offending versatility and persistence. Offending persistence may affect criminal versatility in two opposing ways. On the one hand, specialization may increase as offenders become more experienced. According to this prediction, offenders refine their offending techniques over time and become more selective in the types of offenses that they commit, choosing crimes that experience has shown deliver the greatest returns (Blumstein, Cohen, Das, & Moitra, 1988). On the other hand, specialization may be associated with offending naiveté. Moffitt (1993) has argued that persistent offenders have ingrained antisocial attributes that produce relatively unrestrained offending behavior. They therefore commit a wider range of crimes across more diverse situations than do other offenders. However, occasional offenders are sensitive to situational contexts and so are more restricted in the scope of their offending. Youth may, for example, shoplift with their friends while otherwise continuing to conform to family or school rules.

One possibility, of course, is that both explanations of the relationship between persistence and versatility are valid, but that they describe separate categories of offenders. Conceptually, an interaction between persistence and versatility yields four possible categories. Limited (i.e., not persistent) and specialized offenders will have a low criminal involvement and focus on a narrow range of offenses (like Moffitt’s, 1993, adolescence-limited shoplifters). Limited and versatile offenders will also have a low criminal involvement but will not be selective in their offending choices (as for low-volume offenders according to Gottfredson & Hirschi’s, 1990, control theory). Persistent and specialized offenders will be career criminals who have a favored modus operandi (consistent with Blumstein et al., 1988). Finally, persistent and versatile offenders will have extensive criminal careers involving a diverse range of offenses (as for high-volume offenders according to control theory). It is the viability of this taxonomy applied to child sexual abusers that will be examined in this study.

Criminal Careers of Sexual Abusers

In these debates about offending persistence and versatility, the criminal careers literature has historically paid little attention to sexual offending, and even less to child sexual abuse offending in particular. Rape has often been included as an offense of interest—and is generally found to occur within a versatile offending pattern—but has typically been classified as a violent rather than a sexual offense (Blumstein et al., 1988; Piquero, 2000). Child sexual abuse offending has historically been all but ignored in criminal careers research.

For its part, the clinical literature on sexual offending betrays little knowledge of the criminal careers model. Explanations of child sexual abuse offending have instead been typically formulated within a sexual deviance model (e.g., Laws & O’Donohue, 1997). Child sexual abusers have traditionally been assumed to possess motivations that are pathological and long-standing, and that separate them from other offenders. To the extent that one can draw implications about criminal careers from this literature, child sexual abusers would be generally characterized as persistent and specialized in their offending patterns.

A number of studies, however, have found surprisingly high levels of offending versatility and low levels of sexual offending persistence among child sexual abusers (Hanson & Bussiere, 1998; Harris, Smallbone, Dennison, & Knight, 2009; Lussier, LeBlanc, & Proulx, 2005; Meithe, Olsen, & Mitchell, 2006; Pritchard & Bagley, 2000; Simon, 1997, 2000; Smallbone & Wortley, 2000, 2001, 2004a, 2004b; Soothill, Francis, Sanderson, & Ackerley, 2000). For example, Smallbone and Wortley (2000, 2001) examined the official records of 323 convicted child sexual abusers, 182 of whom provided detailed self-report data on their psychosocial/psychosexual histories and offending behaviors. Smallbone and Wortley reported that only 21% of their sample had previous convictions for sexual offenses, while 39% had prior convictions for property offenses and 23% for nonsexual violent offenses. Of those offenders with previous convictions, 82% were first convicted for a nonsexual offense. Furthermore, the mean self-reported age of first sexual contact with a child was 32 years, and there was little evidence of other paraphilic interests among the sample that would indicate an established pattern of sexual deviance. These findings suggest that many child sexual abusers are not persistent specialists, but rather, for them child sexual abuse is a limited or occasional activity that forms part of a more general involvement in crime. For these offenders, the problem may have less to do with the existence of some special motivation to sexually abuse children than with a more general problem involving the failure to restrain their behavior in the face of particular opportunities and temptations.

On the surface, these findings provide support for Gottfredson and Hirschi’s (1990) general theory of crime. However, before settling on this interpretation of the Smallbone and Wortley’s studies, some further consideration of these data is warranted. Even more striking than the low level of persistent specialization found was the variation in offending patterns in the sample. For example, the average number of nonsexual offense convictions was 7.8, but 32% had no such convictions while one offender had 225. Similarly, the mean number of self-reported child victims was 5.6, but 49% of the sample reported just one victim while one offender reported over 300. And while most offenders did not abuse their first victim until their early 30s, the age range for the first sexual abuse incident was 12 to 63 years. In other words, the presentation of averaged findings masks important differences among offenders. Many child sexual abusers may well be occasional generalists, but the sample also contains a nontrivial minority of persistent sexual predators (and more than likely other types as well). The questions of whether child sexual abusers are limited or persistent, or specialists or generalists, may therefore not be meaningful. Rather, there appears to be a number of distinct types of sexual abusers’ criminal careers patterns. The way forward is to identify and describe these multiple criminal-career paths of child sexual abusers.

The aim of the current study is to apply a criminal careers model to child sexual abusers. The study is an extension of Smallbone and Wortley’s (2000, 2001, 2004a, 2004b) earlier studies, involving additional participants and fresh analyses not previously reported. In the present study we examine the extent to which the offending patterns of child sexual abusers can be described along offending persistence and versatility dimensions. While there are numerous existing typologies of child sexual abusers, usually these are based either on the offender’s presumed sexual motivations (e.g., Groth & Birnbaum, 1978; Knight, Carter, & Prentky, 1989) or the characteristics of their victims (e.g., Bagley & Pritchard, 2000). A common theme in these typologies is the distinction between offenders with enduring sexual interests in children (preferential or fixated offenders) and those whose problematic sexual interests and behavior are more transitory (situational or regressed offenders). In some respects, this distinction parallels our proposed persistent/limited distinction. General antisocial or aggressive features have been recognized in typological schemes with adult-victim sexual offenders (e.g., Knight et al., 1989), but existing typologies have paid little attention to variations in general criminality among child sexual abusers. There is therefore no parallel in existing typologies to our proposed versatile/specialized distinction. Our aim is to construct a typology based on an empirical analysis of offending-career patterns, focusing on persistence (versus limited) and versatility (versus specialized) dimensions.

Method

Participants

The official records were accessed of 362 adult male prisoners in Queensland (Australia) who at the time of data collection were serving a sentence for sexual offenses involving victims under 16 years of age. Of these prisoners, 213 agreed to provide self-report data (respondents) on their offending behavior. Respondents did not differ from those who declined to provide self-report data on any demographic or offense history variable, with the exception of education level (respondents were more educated).

Measures

Official data

Official conviction data for all 362 offenders were recorded from offenders’ correctional files. For the current study, the relevant details were the number of prior and current sexual offense convictions, the number of prior and current nonsexual offense convictions, age at first conviction, age at current conviction, and current sentence length.

Self-report data

Respondents completed a 368 item self-report questionnaire (see Smallbone & Wortley, 2000, for a full description). Of relevance for the current study, they were asked whether they had been sexually abused as a child (yes or no), their current sexual orientation (whether they were attracted to women, to men, to both women and men, or to neither women nor men), and the total number of victims they had abused.

Offenders were asked how often they had experienced “intense, sexually arousing urges, fantasies or behavior” involving exhibitionism, fetishism, public masturbation, frotteurism, sexual masochism, sexual sadism, transvestic fetishism, voyeurism, telephone scatologia, necrophilia, and zoophilia. Simple, plain-language descriptions were used for each paraphilia (e.g., “showing your penis to a stranger,” “masturbating alone with objects like shoes, socks, underwear, etc.,” and so on). Responses for each item were initially coded as 1 = never, 2 = once or twice, or 3 = three or more times over a period of at least 6 months. For our present purposes, we dichotomized responses as either 1 = never or 2 = at least once or twice. Of the 11 individual paraphilic interests, zoophilia and necrophilia were discarded because too few offenders acknowledged any interest in these. Remaining scores were summed to produce a nine-item multiple paraphilic interests’ scale. Internal consistency of the scale proved to be adequate (Cronbach’s alpha = .74).

In addition, offenders were asked about the details of their first sexual contact with a child (irrespective of whether they had been charged with that offense), including their age at the time, the age of their victim, the victim’s gender, their relationship with the victim (originally coded as familial, known nonfamilial, or stranger), the number of sexual contacts that they had with the victim, and the length of time over which the sexual contact took place.

Procedure

All prisoners who were identified as child sexual abusers were approached individually by a member of the research team and invited to complete the self-report questionnaire. They were told that their participation would not benefit them with regard to their current sentence and nor would they be penalized should they choose not to participate. Prospective participants were asked to sign a consent form, and were provided with an information sheet which emphasized that while information could be provided anonymously, participants could provide their names to be contacted at a later date for follow-up contact. Prospective participants were assured that self-report information would be kept confidential, that no identifying information would be revealed outside the research team, and that records of names would be destroyed following cessation of the data collection phase.

Results

Results are presented in three sections. First, the classification of offenders in terms of the persistence and versatility dimensions is described. Second, official conviction data are analyzed with respect to the classification. Third, self-report data are analyzed with respect to the classification.

Classification of offenders

All offenders were classified along two dimensions—their persistence in sexual offending and their versatility in offending generally—on the basis of their official records. Persistence was operationalized by dichotomizing participants as having/not having prior sexual offense convictions (persistent/limited); versatility was operationalized by dichotomizing participants as having/not having previous convictions for nonsexual offenses (versatile/specialized). Note that “limited” here does not carry with it the implication—as is the case with Moffitt’s (1993) use of the term—that these offenses are largely confined to adolescence. In fact, as we have discussed, one of the noteworthy features of child sexual abuse offending is its relatively late onset. Furthermore, since persistence is defined only with respect to sexual offending, offenders who are classified as limited may in fact have extensive nonsexual offending criminal histories.

Cross-tabulation of persistence and specialization measures yielded the following four offender categories:

limited/specialized (no previous sexual convictions/no previous nonsexual convictions),

limited/versatile (no previous sexual convictions/previous nonsexual convictions),

persistent/specialized (previous sexual convictions/no previous nonsexual convictions), and

persistent/versatile (previous sexual convictions/previous nonsexual convictions).

The breakdown for the sample is shown in Table 1. Of the 362 records accessed, previous conviction histories were missing in 8 cases. Of the remaining 354 cases, the largest category is limited/versatile (41.0%), that is, offenders who are currently serving sentences for their first sexual offense conviction(s) but have at least one prior conviction for a nonsexual offense. Almost as many participants (36.4%) had no previous convictions of any kind (limited/specialized), while 17.8% had prior convictions for both sexual and nonsexual offenses (persistent/versatile). What might be thought of as the archetypal sexual abuser—persistent/specialized—in fact accounted for fewest participants (4.8%).

Table 1.

Prior Sexual by Nonsexual Offense Convictions.

		Past nonsexual convictions
		No (specialized)	Yes (versatile)	Total
Past sexual convictions	No (limited)	129 (36.4%)	145 (41.0%)	274 (77.4%)
Past sexual convictions	Yes (persistent)	17 (4.8%)	63 (17.8%)	80 (22.6%)
	Total	146 (41.2%)	208 (56.8%)	354

Official Conviction Data

Means (SDs) and ranges for official conviction data are shown in Table 2. Missing values meant that a further 5 cases were eliminated from analyses, leaving 349 cases (127 limited/specialized, 17 persistent/specialized, 144 limited/versatile, and 62 persistent/versatile). Over half the sample (56.8%) had previous convictions for nonsexual offenses, while less than a quarter (22.6%) had previous sexual offense convictions. For persistent offenders, the mean number of prior sex offense convictions was 4.00 (SD = 4.51), with no significant difference between persistent/specialized and persistent/versatile. For versatile offenders, the mean number of previous nonsexual convictions was 10.72 (SD = 22.83), with no significant difference between the limited/versatile and persistent/versatile groups. In terms of first convictions, by definition all specialized offenders were first convicted for a sexual offense while limited/versatile offenders were first convicted for a nonsexual offense. For the persistent/versatile group, 25.4% had a first conviction for a sexual offense, while a further 15.9% had first convictions for both sexual and nonsexual offenses (and thus 58.7% had first convictions unrelated to sexual offending). Overall, among offenders with at least one previous conviction, 14.8% were first convicted of a sexual offense, 4.5% for both sexual and nonsexual offenses, and 79.7% for a nonsexual offense.

Table 2.

Conviction Data for Prior Sexual by Nonsexual Convictions (SDs in Brackets).

Conviction history	Prior nonsex convictions
	No		Yes
	Prior sex convictions		Prior sex convictions
	No	Yes	No	Yes
Age at first conviction (years)	43.15 (12.68)	31.88 (11.80)	23.52 (9.21)	24.82 (10.85)
Age at present conviction (years)	43.15 (12.68)	42.35 (13.04)	39.78 (12.01)	44.05 (10.81)
Present sentence (months)	75.74 (53.01)	86.17 (43.12)	79.71 (54.62)	113.35 (92.61)
Present sex convictions	8.23 (7.94)	13.76 (12.59)	7.66 (8.19)	7.81 (7.30)
Present nonsex convictions	0.55 (1.55)	1.12 (3.35)	2.21 (5.26)	1.96 (4.25)

These data were entered as dependent variables in MANOVA (SPSS Windows 20), with previous sexual convictions and previous nonsexual convictions used as dichotomous independent variables in a 2 × 2 design. These results are summarized in Table 3. There was a significant multivariate interaction involving significant univariate results for age at first conviction and the number of current sexual convictions. For age at first conviction, inspection of the relevant means (Table 2) indicates that limited/specialized offenders were disproportionately older than other offenders. Post hoc tests show significant differences between the limited specialized and each of the other three groups (vs. persistent/specialized, t(142) = 3.42, p = .001; vs. limited/versatile, t(269) = 14.61, p < .001; vs. persistent/versatile, t(187) = 9.69, p < .001). Furthermore, persistent/specialized offenders were significantly older at first conviction than either limited/versatile offenders, t(159) = 3.44, p = .001, or persistent/versatile offenders, t(77) = 2.33, p = .022. Thus, in addition to the interaction effect, versatile offenders were generally younger at their first conviction than were specialized offenders.

Table 3.

Summary of MANOVA for Conviction History.

Effect	df	F	Significance	η²
Prior sex convictions	5 (341)	7.77	<.001	.11
Age at first conviction	1	8.89	.003	.03
Present sentence	1	5.61	.022	.02
Present sex convictions	1	5.33	.018	.02
Prior nonsex convictions	5 (341)	28.28	< .001	.31
Age at first conviction	1	64.85	<.001	.12
Present sex convictions	1	7.03	.008	.02
Present nonsex convictions	1	3.92	.049	.01
Prior sex × Nonsex convictions	5(341)	5.49	<.001	.07
Age at first conviction	1	14.21	<.001	.04
Present sex convictions	1	4.81	.029	.01

For the number of current sexual convictions, the means indicate that persistent/specialized offenders had disproportionately more convictions than the other groups. Post hoc analysis revealed significant differences between persistent/specialized offenders and each of the other three groups (vs. limited/specialized, t(144) = 2.51, p = .013; vs. limited/versatile, t(160) = 2.77, p = .006; vs. persistent/versatile, t(78) = 2.56, p = .012). There were no other significant differences.

There were also significant multivariate main effects for prior sexual convictions and prior nonsexual convictions. These involved two additional significant univariate results not otherwise accounted for by the significant interactions. Offenders with prior sexual convictions received significantly longer sentences than offenders with no prior sexual convictions, and offenders with a prior nonsexual conviction had significantly more current nonsexual convictions than offenders with no prior nonsexual convictions.

Importantly, there were no significant differences between the groups for age at present conviction. This means that other differences among offending categories cannot be simply explained in terms of different stages in the offenders’ criminal careers. For example, it does not seem that the limited/specialists are merely early-stage persistent/specialists (although we cannot rule out the possibility that some limited/specialists may reoffend in the future). On average persistent/specialists had their first conviction for a sexual offense more than 11 years earlier than the first convictions of limited/specialists.

Self-Report Data

Of the cases described in the first analysis, 213 offenders agreed to complete a confidential self-report questionnaire. Due to the cumulative impact of missing values, the analyses in this section involved 177 cases (62 limited/specialized, 10 persistent/specialized, 76 limited/versatile, and 29 persistent/versatile). A summary of the self-report data is shown in Table 4.

Table 4.

Self-Report Sexual Histories (SDs in Brackets).

Characteristic	Prior nonsexual convictions
	No		Yes
	Prior sex convictions		Prior sex convictions
	No	Yes	No	Yes
Sexually abused as a child	43.5%	90.0%	56.6%	70.0%
Paraphilic interests	1.27 (2.33)	1.40 (2.01)	1.71 (2.91)	1.17 (1.75)
Sexual orientation
Attracted to women	80.6%	30%	85.5%	72.4%
Attracted to men	6.5%	10%	5.3%	10.3%
Attracted to men and women	9.7%	50%	9.2%	13.8%
Not attracted to men or women	3.2%	10%	1.3%	3.4%
Total number of victims	2.85 (3.49)	9.50 (6.98)	2.99 (4.59)	5.38 (6.08)
Offender age at first child sex contact	33.05 (10.72)	28.00 (12.75)	34.50 (10.87)	32.17 (12.36)
Age of first victim	11.74 (2.80)	11.40 (3.20)	10.96 (3.28)	10.52 (2.89)
Gender of first victim
Male	17.7%	60.0%	15.8%	48.3%
Female	82.3%	40.0%	84.2%	51.7%
Relationship with first victim
Lived with child	58.1%	30.0%	65.8%	34.5%
Knew child	38.7%	70.0%	27.6%	51.7%
Stranger	3.2%	0	6.6%	8.2%
Number of contacts with first victim
1 time	12.9%	10.0%	32.9%	41.4%
2-3 times	14.5%	30.0%	19.7%	20.7%
4-5 times	17.7%	0	3.9%	13.8%
6-10 times	19.4%	20.0%	15.8%	10.3%
11-20 times	16.1%	10.0%	11.8%	3.4%
21-50 times	32.2%	20.0%	9.2%	0
>50 times	16.1%	10.0%	6.6%	10.3%
Length of contact with first victim
0-1 days	12.9%	10.0%	34.2%	48.3%
1-7 days	4.8%	20.0%	7.9%	3.4%
1-4 weeks	9.7%	60.0%	5.3%	13.8%
1-2 months	4.8%	10.0%	1.3%	3.4%
2-6 months	12.9%	0	13.2%	10.3%
6-12 months	8.1%	10.0%	13.2%	6.9%
>1 year	46.8%	50.0%	25.0%	27.6%

To simplify further analyses, the number of categories for some variables was reduced. For sexual orientation, “attracted to men,” “attracted to men and women,” and “not attracted to men or women” were collapsed into a single category, creating a dichotomous “exclusively heterosexual/not exclusively heterosexual” variable. For relationship with first victim, “known nonfamilial” and “stranger” were collapsed to create a familial/nonfamilial variable.

The distribution of the number of victims and the multiple paraphilic interest scores were markedly positively skewed, and log transformation of these data was therefore used for subsequent analyses.

MANCOVA was performed, with the 10 sexual history variables entered as the dependent set, and prior sexual offense convictions and prior nonsexual convictions as the independent factors. Because participants’ sexual histories are likely to be related to the period of time available to them to offend, age at current conviction (taken from official records) was entered as a covariate. Results of the MANCOVA are displayed in Table 5. There was a significant multivariate interaction involving significant univariate results for sexual orientation and total number of victims. For sexual orientation, means indicate that persistent/specialized offenders were disproportionately less likely to be exclusively heterosexual. Post hoc analyses indicated that persistent/specialized offenders were less likely to be exclusively heterosexual than limited/specialists, t(70) = 3.62, p = .001; limited/versatile t(84) = 4.46, p < .000; and persistent/versatile, t(37) = 2.50, p = .017. There were no significant differences between persistent/versatile offenders and either of the limited offender group. Persistent/specialists abused significantly more victims than limited/specialized offenders, t(70) = 4.75, p < .000, and limited/versatile offenders, t(84) = 3.95, p < .000, but not persistent/versatile offenders. Furthermore, the persistent/versatile group abused more victims than limited/specialists, t(86) = 2.51, p = .014, and limited/versatile, t(103) = 2.18, p = .032.

Table 5.

MANCOVA for Self-Reported Sexual Histories.

Effect	df	F	Significance	η²
Age at current conviction	10 (163)	22.27	<.001	.47
Offender age at first child sex contact	1	145.29	<.001	.41
Number of contacts with first victim	1	9.93	.002	.06
Length of contact with first victim	1	6.10	.015	.05
Prior sex convictions	10(163)	4.32	<.001	.22
Sexually abused as a child	1	9.34	.003	.05
Sexual orientation	1	16.06	<.001	.10
Total number of victims	1	21.84	<.001	.12
Offender age at first child sex contact	1	11.03	.001	.04
Gender of first victim	1	20.63	<.001	.13
Relationship with first victim	1	9.15	.003	.04
Prior nonsex convictions	10 (163)	3.07	.001	.18
Sexual orientation	1	8.65	.004	.07
Total number of victims	1	4.48	.036	.03
Number of contacts with first victim	1	7.40	.007	.03
Length of contact with first victim	1	7.61	.006	.03
Prior sex × Nonsex convictions	10 (163)	1.98	.038	.13
Sexual orientation	1	5.27	.023	.05
Total number of victims	1	5.25	.023	.03

There were significant main effects for prior sexual convictions and prior nonsexual convictions, involving additional univariate results to those already described for the interaction effect. Offenders with prior sexual convictions were more likely to have been sexually abused as a child, to have been younger when they abused their first victim, to have first abused a male victim, and to have first abused a nonfamilial victim. Offenders with prior nonsexual convictions were more likely to have fewer sexual contacts with their first victim, and to abuse that victim over a shorter period of time. Familial abuse was associated with no prior sexual convictions. In other words, it seems that offenders with prior nonsexual convictions are less sexually involved with their victims irrespective of their relationship with the victim.

The covariate, age at current conviction, was significant at the multivariate level. Offender age at current conviction was significantly associated with their age at their first sexual contact. This is not surprising since many first-time sexual offenders are likely to have committed their offense relatively recently. These older, first-time offenders were also more likely to have extended sexual relationships with their victim.

Groups did not significantly vary in their paraphilic interests. Separate correlation analyses indicated that the paraphilic interest scale was related to a number of other significant predictor variables, and it seems that these variables accounted for the variance in the analysis. Groups also did not vary significantly on age of first victim. This confirmation of the null hypothesis strengthens the basis for making comparison among groups. That is, the victims for each group are equally physically immature, indicating that none of the groups are misclassified borderline rapists.

Discussion

Significant differences among the four offender categories are summarized in Table 6. Each of the four categories has distinctive features and the content of each category describes a conceptually coherent type. Categories have been arranged from left to right in order of increasing possession of problematic characteristics.

Table 6.

Summary of significant differences among offender categories (darker shading indicates interactions).

	Limited/Specialized	Limited/Versatile	Persistent/Versatile	Persistent/Specialized
Prior sexual convictions
Prior nonsexual convictions
Current nonsexual convictions
Current sexual convictions
Early involvement with CJS
Long current sentences
History of sexual abuse
Non-heterosexual orientation
Many victims
Early sexual contact with child
Male victims
Nonfamilial
Multiple contacts with victim
Extended relationship with victim

Note. Darker shading indicates interactions. CJS = Criminal Justice System.

The first category, limited/specialized offenders, contained the second-largest number of offenders (36.4%). This is their first conviction for any offense and they are generally the most conventional in their profile. They are generally heterosexual; they tend not to have been sexually abused as a child; they are relatively old at the time of their first sexual contact with a child; they usually offend within the family; they have relatively few victims; but they tend to abuse their victims repeatedly over an extended period of time. The picture here is of a caregiver or other authority figure who abused a position of trust and had ongoing access to their victim. Despite the repetitive nature of their offending, these offenders may not have entrenched sexual preferences for children, at least not initially. Most have waited until their early thirties before they first offended and then they have abused children to whom they had ready access. Indeed, their access to children is no doubt a function of their increasing age and assumption of child-care roles. For these offenders, more effective supervision of children in the home and in institutional settings (schools, churches, etc.) may be sufficient to deter them (Smallbone, Marshall, & Wortley, 2008; Wortley & Smallbone, 2006). Results suggest that many of these offenders may not require specialized sexual offender treatment programs, although individualized assessment and case formulations may indicate otherwise for some.

The next and largest category (41%) was the limited/versatile offender. These are offenders who were serving sentences for their first sexual offense conviction but who had at least one prior conviction for a nonsexual offense. These offenders share many characteristics with limited/specialist offenders. However, reflecting their more general involvement in crime, they are more likely to have current nonsexual convictions and to have had an early contact with the criminal justice system. More intriguingly, while these offenders were similarly likely to offend within a family setting and to have relatively few victims, they had fewer sexual contacts with their victims and maintained sexual contact over a shorter period when compared with limited specialists. The reasons for this difference between the two groups are unclear but there are a number of possibilities. It may be that limited versatile offenders are less adept at concealing their abuse and are detected earlier; they may be more likely to be a peripheral family member (for example, an uncle) with a more restricted access to the victim; they may have more unstable lifestyles—either because they move more frequently between partners or they go to prison—that disrupt their access to particular victims; or they may be less interested in developing an emotional relationship with the child. Taken together, their offending patterns are those of the opportunist. Their criminal versatility suggests a generalized failure to inhibit self-gratifying urges; their relative lack of sexual offending persistence suggests sexual ambivalence rather than ingrained sexual deviance. As with limited specialist offenders, their offending is mediated by opportunities, and similarly, reducing opportunities for them to offend may be effective in many cases. Rather than specialized sexual offender treatment programs, these offenders may benefit more from offender programs that target more general offense-related factors such as antisocial beliefs and attitudes, social problem-solving, empathy, and so on. Evaluations of cognitive skills training, for example, have shown similar benefits for sexual and nonsexual offenders, without specifically targeting sexual offending (Robinson, 1995).

The next category, persistent/versatile, accounted for 17.8% of the sample. These offenders had prior convictions for sexual and nonsexual offenses. Like the limited versatile offenders, they had involvement with the criminal justice system at an early age and are likely to have current nonsexual convictions. However, in other respects there are marked differences in the profile of the persistent versatile offender and the preceding types. They have longer current sentences, they report a higher incidence of sexual abuse as a child, they have more victims, they report an earlier age for their first sexual contact with a child, they are more likely to have male victims, and they are more likely to abuse a nonfamilial victim. Like the limited versatile offender, but in contrast to the limited specialist offender, they sexually abuse their victims on relatively few occasions and over a relatively short period. Thus, they are more likely to be serial offenders, with more victims but less sustained sexual contact with individual victims. Reducing opportunities for them to offend may protect specific children, but they are likely to move on to other targets as circumstances allow, and may at best be slowed down by thwarting specific opportunities. Because they began offending at a young age, they may benefit from early intervention programs. Treatment programs should address sexual-offense-specific and more general criminogenic needs.

The final category comprises persistent specialists, describing offenders who have been convicted exclusively of sexual offenses on multiple occasions. These offenders constituted the smallest category (4.8%). Like the persistent versatile offender and in contrast with the limited categories, they are serving longer current sentences; they are more likely to have a history of being sexually abused as a child; they report earlier sexual contact with a child; they are more likely to abuse male victims; and they are more likely to abuse nonfamilial victims. However, they differ from persistent versatile offenders in a number of important ways. They are less likely to have current nonsexual convictions but likely to have more current sexual convictions; their involvement in the criminal justice system begins at a later age, although still earlier than for limited specialists; they are more likely to have a non-heterosexual orientation; they report more victims; and like limited specialists they are more likely to have frequent and extended sexual contact with their victims. Thus, while like the persistent versatile type they are more likely to be serial, nonfamilial offenders, they seem more interested in maintaining a relationship with their victim. These offenders will be the most difficult to deter in their sexual offending and are the most in need of specialized sexual offender treatment and risk management.

There are a number of qualifications to be made about the current typology. From a methodological standpoint, we acknowledge that our measures of persistence and versatility—based as they were on official convictions—are rather crude. Undoubtedly, offenders commit crimes for which they are not convicted. Thus, in our limited specialist group, for example, there will be misclassified offenders who have quite extensive sexual offending histories, and indeed the self-report data confirm this. Similarly, the official records and self-report data used to describe the content of each offender type are vulnerable to underreporting in the first case and self-serving distortion in the second. In this regard, of course, our study is no less reliable than any other empirical study in this area. In fact, we would argue that by reporting both sources of data, we are able to show a basic congruence between the two types of data that is not shown in most other studies. It is worth noting, too, that purer measures of persistence and versatility may produce even clearer differences among the groups.

Another potential methodological problem is low power associated particularly with the small size of the persistent specialist group. Our total sample sizes for the analyses of official and self-report data are more than adequate for MANOVA. However, minimum cell sizes are usually recommended to be around 20 (Simmons, Nelson, & Simonsohn, 2011). In the present case our smallest group by some margin was the persistent specialists—we could allocate only 10 of these for the analysis of self-report data. This in itself is intriguing, given that this is the group that most closely fits the popular stereotype of the persistent, sexually disturbed abuser. The problem with small cell sizes is of course that it increases the chance of not finding an effect that is actually there (reduced power). It is therefore arguably all the more striking that we found conceptually meaningful differences even with this small-sized group as part of the analysis. Readers should nevertheless keep in mind the small effect sizes obtained for some analyses.

From a conceptual standpoint, we know that distinctions between categories are not hard and fast, and that offenders may not fit neatly into a given categorization. A good typology should have as few discrete categories as possible, while at the same time having categories that are highly specific and homogeneous. However, all typologies are attempts to simplify complex reality, and in the process of simplification, some richness may be lost. In our typology, there will be some variations within categories and slippage between them. For example, there may be some limited specialists who have quite entrenched deviant sexual urges, but who have managed for the most part to resist hands-on offending. Similarly, persistent offenders may not always engage in premeditated grooming of victims but may also commit opportunistic offenses if the occasion arises. It is also possible that as their offending careers develop over time, some offenders will move out of the categories to which they have been assigned here. Again, however, any misclassifications are most likely to have suppressed differences between the groups.

Notwithstanding these qualifications, we believe that our typology provides a useful way to conceptualize variations among child sexual abusers. Understanding variations in criminal careers of child sexual abusers in turn offers the potential for a more carefully targeted delivery of prevention and treatment strategies.

Footnotes

Declaration of Conflicting Interests

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

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was funded by grants from the Queensland Crime Commission and the Criminology Research Council.

References

Bagley

Pritchard

(2000). Criminality and violence in intra- and extra-familial child sex abusers in a 2-year cohort of convicted perpetrators. Child Abuse Review, 9, 264-274.

Blumstein

Cohen

Das

Moitra

S. D.

(1988). Specialization and seriousness during adult criminal careers. Journal of Quantitative Criminology, 4, 303-345.

DeLisi

Piquero

A. R.

(2011). New frontiers in criminal careers research, 2000-2011: A state-of-the-art review. Journal of Criminal Justice, 39, 289-301.

Farrington

D. P.

Ohlin

L. E.

Wilson

J. Q.

(1986). Understanding and controlling crime: Toward a new research strategy. New York, NY: Springer-Verlag.

Gottfredson

M. R.

(2006). The empirical status of control theory in criminology. In Cullen

Wright

Blevins

(Eds.), Taking stock: The status of criminological theory (pp. 77-100). New Brunswick, NJ: Transaction.

Gottfredson

M. R.

Hirschi

(1990). A general theory of crime. Stanford, CA: Stanford University Press.

Groth

A. N.

Birnbaum

H. J.

(1978). Adult sexual orientation and attraction to underage persons. Archives of Sexual Behavior, 7, 175-181.

Hanson

R. K.

Bussiere

M. T.

(1998). Predicting relapse: A meta-analysis of sexual offender recidivism studies. Journal of Consulting and Clinical Psychology, 66, 348-362.

Harris

D. A.

Smallbone

Dennison

Knight

R. A.

(2009). Specialization and versatility in sexual offenders referred for civil commitment. Journal of Criminal Justice, 37, 37-44.

10.

Knight

R. L.

Carter

D. L.

Prentky

R. A.

(1989). A system for the classification of child molesters: Reliability and applications. Journal of Interpersonal Violence, 4, 3-23.

11.

Laws

D. R.

O’Donohue

(1997). Introduction: Fundamental issues in sexual deviance. In Laws

D. R.

O’Donohue

(Eds.), Sexual deviance: Theory, assessment and treatment (pp. 1-21). New York, NY: Guilford.

12.

Lussier

LeBlanc

Proulx

(2005). The generality of criminal behavior: A confirmatory factor analysis of the criminal activity of sexual offenders in adulthood. Journal of Criminal Justice, 33, 177-189.

13.

Meithe

T. D.

Olsen

Mitchell

(2006). Specialization and persistence in the arrest histories of sex offenders. Journal of Research in Crime & Delinquency, 43, 204-229.

14.

Moffitt

T. E.

(1993). Adolescence-limited and life-course-persistent antisocial behavior: A developmental taxonomy. Psychological Review, 100, 674-701.

15.

Piquero

A. R.

(2000). Frequency, specialization and violence in offending careers. Journal of Research in Crime & Delinquency, 37, 392-418.

16.

Piquero

A. R.

(2004). Somewhere between persistence and desistance: The intermittency of criminal careers. In Maruna

Immarigeon

(Eds.), After crime and punishment: Pathways to offender reintegration (pp. 102-125). Portland, OR: Willan.

17.

Pritchard

Bagley

(2000). Multi-criminal and violent groups amongst child sex offenders: A heuristic typology in a 2-year cohort of 374 men in two English counties. Child Abuse & Neglect, 24, 579-586.

18.

Robinson

(1995). The impact of cognitive skills training on post-release recidivism among Canadian Federal offenders. Ottawa, Ontario: Correctional Service Canada.

19.

Simmons

J. P.

Nelson

L. D.

Simonsohn

(2011). False-positive psychology: Undisclosed flexibility in data collection and analysis allows presenting anything as significant. Psychological Science, 22, 1359-1366.

20.

Simon

(1997). Do criminal offenders specialize in crime types? Applied & Preventive Psychology, 6, 35-53.

21.

Simon

(2000). An examination of the assumptions of specialization, mental disorder, and dangerousness in sex offenders. Behavioral Sciences & the Law, 18, 275-308.

22.

Smallbone

S. W.

Marshall

W. L.

Wortley

(2008). Preventing child sexual abuse: Evidence, policy and practice. Cullompton, UK: Willan Publishing.

23.

Smallbone

S. W.

Wortley

R. K.

(2000). Child sexual abuse in Queensland: Offender characteristics and modus operandi. Brisbane, Australia: Queensland Crime Commission.

24.

Smallbone

S. W.

Wortley

R. K.

(2001). Child sexual abuse: Offender characteristics and modus operandi (Australian Institute of Criminology Trends and Issues in Crime and Criminal Justice, No. 193). Canberra: Australian Institute of Criminology.

25.

Smallbone

S. W.

Wortley

R. K.

(2004a). Criminal versatility and paraphilic interests among adult males convicted of sexual offenses against children. International Journal of Offender Therapy and Comparative Criminology, 48, 175-188.

26.

Smallbone

S. W.

Wortley

R. K.

(2004b). Onset, persistence and versatility of offending among adult males convicted of sexual offenses against children. Sexual Abuse: A Journal of Research and Treatment, 16, 285-298.

27.

Soothill

Francis

Sanderson

Ackerley

(2000). Sex offenders: Specialists, generalists—Or both? British Journal of Criminology, 40, 56-67.

28.

Wolfgang

M. E.

Figlio

Sellin

(1972). Delinquency in a birth cohort. Chicago, IL: University of Chicago Press.

29.

Wortley

R. K.

Smallbone

S. W.

(2006). Applying situational principles to sexual offending against children. In Wortley

Smallbone

(Eds.), Situational prevention of sexual offenses against children. Crime prevention studies (pp. 7-35). Monsey, NY: Criminal Justice Press.