Two contrasting arguments have been made about child sex offenders’ proclivity to reoffend. In public and media discourse, child sex offenders are often constructed as compulsive recidivists who are virtually certain to reoffend. For example, in a second reading speech to the Legislative Council of South Australia about the Criminal Law (Sentencing) (Mandatory Imprisonment of Child Sex Offenders) Amendment Bill, one Parliamentarian described child sex offenders as ‘beings of a subhuman category...[they are]...the least rehabilitatable people’ (Bressington 2010).
Conversely, in the criminological literature, the opposite is often posited—that child sex offenders have low rates of recidivism compared with other types of offenders (see eg McSherry & Keyzer 2009; Minnesota Department of Corrections 2007).
It is certainly the case that many studies of child sex offenders have found low levels of recidivism (Doren 1998). Measuring sexual recidivism is, however, a challenging task (see Falshaw, Friendship &Bates 2003 for a discussion) and it is important to be aware of the limitations of these studies. There are a number of key decisions that researchers make when measuring the recidivism of child sex offenders that can impact the findings of studies. Two key decisions are the definition of recidivism and the period of time over which recidivism is measured.
In most studies of general reoffending, recidivism is defined as a reconviction for a new offence. As sexual offences are often not reported (Abel et al. 1987; Bates, Saunders & Wilson 2007) and sexual offending against children has one of the highest rates of attrition of any offence (ie a relatively small proportion of cases progresses successfully through the criminal justice system; Eastwood, Kift & Grace 2006), studies of child sex offender recidivism that rely on reconvictions as a measure of recidivism provide only ‘a diluted measure of true reoffense rates’ (Doren 1998: 99).
As a result, some studies of child sex offender recidivism have defined recidivism as an arrest or charge (rather than a conviction) for a new sexual offence. This approach is also limited, but is likely to provide a more accurate measure of recidivism than reconvictions. As Doren (1998: 101) argues
although some portion of the people charged with a new sexual crime may [have] been both innocent of that charge and of any other recidivating sexually predatory acts, this portion would likely be far smaller than the number of re-offenders who are never caught and charged.For a variety of reasons, recidivism studies usually follow up offenders over a short period, such as two or three years. While this is often necessary due to time and budget constraints, the longer a period over which recidivism in measured, the higher the rate of recidivism is likely to be (Tresidder, Homel & Payne 2009). While child sex offender studies often show low levels of recidivism, Salter (2003) argues that these studies obscure the reality that in the long term, rates of recidivism can be much higher (see also Bates, Saunders & Wilson 2007).
Studies that narrowly define recidivism and use short follow-up periods may therefore underestimate the rate of recidivism of child sex offenders (Moulden et al. 2009). Prentky et al.’s (cited in Doren 1998) study of recidivism rates among extrafamilial child sex offenders over a 25 year period used a new charge for a sex offence as the measure of recidivism. This study found that 52 percent of child sex offenders reoffended during the 25 year at-risk period. As Doren (1998: 101) argues, however, due to the limitations of recidivism studies on child sex offenders described above
the 52% recidivist figure should be considered as a conservative approximation of the true base rate for sex offense recidivism in previously convicted child molesters...[it]...represents the lowest approximation for extrafamilial child molester sexual recidivism.
As described above, the category of ‘child sex offender’ includes diverse offenders with diverse motivations, including those who meet the diagnostic criteria for paedophilia. It is important to recognise that within the broad offender category of child sex offenders, some subcategories of offenders are likely to be at greater risk of reoffending than others. As Petrunik and Deutschmann (2008: 500) argue:
some sex offenders—notably, extrafamilial offenders with male victims who meet clinical criteria for paraphilias, such as paedophilia or exhibitionism—do offend with high frequency over long periods.
Research by Prentky et al. (cited in Doren 1998) described above, measured the recidivism of extrafamilial child sex offenders. As discussed in more detail below, research shows that extrafamilial child sex offenders perpetrate offences against many more victims than intrafamilial offenders and should therefore not be considered representative of all child sex offenders.
The empirical literature therefore suggests that both the media’s insistence that child sex offenders are compulsive recidivists and criminologists’ counterargument that child sex offenders are unlikely to reoffend may be somewhat skewed. While better quality evidence is required on the question of child sex offender recidivism, the existing research literature indicates that some subgroups of child sex offenders have higher rates of recidivism than others. For example, those who offend against children in their own families have access to only a small number of children, thereby limiting opportunities for recidivism to occur. The competing claims outlined at the opening of this section—ie that all child sex offenders will reoffend/that there is a low recidivism rate among child sex offenders—may not be as mutually exclusive as they appear. The research literature indicates that among a subset of child sex offenders—those who target male victims outside of their family—reoffending in the long term is likely and far more likely than for child sex offenders who target female and/or family member victims.