Please detail your particular refutations of the study in question. Did you find fault with the methodology? Please provide examples of those "holes" you have discerned, along with the evidence supporting your criticisms.
Your blanket dismissal does not constitute a refutation of any value. It is the equivalent of sticking your fingers in your ears and yelling "Nyah! Nyah! Nyah!".
OK – here goes. I will aim to use the same terminology as the study where possible to seek to avoid possible misinterpretation. All highlighting is mine, for emphasis, as is square-bracketed wording or comments:
From “Abstract (Background)”:
The aim of this study was to examine the recidivism rates for hands-on and hands-off sex offences in a sample of child pornography users using a 6 year follow-up design.
From “Background”:
The aim of this study was to analyze the characteristics of a sample of child pornography users and the proportion of those who subsequently re-offended with hands-on and hands-off sex offences.
Whilst not a flaw per se (one surely cannot argue with the
aim of a study!), the authors seem not to be sure what the exact aim of the study is. The more detailed description of the aim qualifies the abstracted aim by limiting the study to pre-offenders (child porn consumers). I will work with the qualified aim, as that is the one that the study clearly follows.
This qualified aim introduces an overriding major problem. The study only includes identified pre-offenders (child porn consumers). Moreover, hands-on child sex offences are measured only by conviction and investigation/charge rates:
According to a strict definition, recidivism was assumed only if new convictions subsequent to the index offence (child pornography consumption) were registered in the criminal records.
A broader definition of recidivism was introduced including convictions, as well as ongoing investigations and charges, as it is a well known fact that the use of criminal records frequently leads to an underestimation of especially sex and violent recidivism. Thus, in Switzerland, convictions of criminally non-responsible offenders, which often involve sex and violent offences, are not registered in the official criminal records. Also, the registration of sex and violent offences in the criminal records often takes several years due to longer investigation and appeal periods.
This raises a fundamental issue which I’ll come onto next, but the most blatant flaw (no, it’s so blatant that it must be called a “failing”), is that consumption of child pornography was only a
pre-requisite to selecting the study group. Hands-on child sex offences were only measured by looking at subsequent conviction and investigation/charge rates
with complete disregard for whether the offenders subsequently consumed child porn during the 6-year sex offence monitoring period. So, the aim of the study as described in “Abstract (Background)” should be more accurately worded:
“The aim of this study was to examine the recidivism rates for hands-on and hands-off sex offences in a sample of historical child pornography users using a 6 year follow-up design.”
The fundamental issue referred to above is, by definition the study completely excludes the category of hands-on sex offenders who have not been identified as such. It is widely acknowledged that the majority of hands-on child sex offences occur in the home by a relative of the victim. Given the nature of this crime it is reasonable to conclude that most hands-on child sex offences go unchecked because:
- Nobody other than the perpetrator and the victim know about the crime and the victim is generally not suitably placed to report the crime (assuming the victim even recognizes the wrong being done), and
- Even if the victim reports the wrongdoing (in all likelihood to another family member) it is reasonable to conclude that in most cases it will go no further and not be reported to the Authorities.
Indeed, the study essentially explicitly supports this:
… it is a well known fact that the use of criminal records frequently leads to an underestimation of especially sex and violent recidivism.
In short, when it comes to hands-on child sex offences the study relies on convictions for its offence data but does not investigate the incidence of convictions compared to non-convictions (Of course, it is all but impossible to obtain offence data from non-convictions, which raises the question as to whether a conclusive scientific study is even possible. The possible fact that it might not be is not, however, justification for accepting the findings of a “best attempt” study, which, by definition, if fundamentally flawed.). That in itself invalidates the statistical results (from “Abstract (Background)”):
4.8% (n = 11) of the study sample had a prior conviction for a sexual and/or violent offence, 1% (n = 2) for a hands-on sex offence, involving child sexual abuse, 3.3% (n = 8) for a hands-off sex offence and one for a nonsexual violent offence. When applying a broad definition of recidivism, which included ongoing investigations, charges and convictions, 3% (n = 7) of the study sample recidivated with a violent and/or sex offence, 3.9% (n = 9) with a hands-off sex offence and 0.8% (n = 2) with a hands-on sex offence.
Which can, of course, be re-written thus:
“4.8% (n = 11) of identified child porn offenders had a prior conviction for a sexual and/or violent offence, 1% (n = 2) for a hands-on sex offence, involving child sexual abuse, 3.3% (n = 8) for a hands-off sex offence and one for a nonsexual violent offence. When applying a broad definition of recidivism, which included ongoing investigations, charges and convictions, 3% (n = 7) of identified child porn offenders recidivated with a violent and/or sex offence, 3.9% (n = 9) with a hands-off sex offence and 0.8% (n = 2) with a hands-on sex offence.”
Of course, it is reasonable to conclude that a child porn offender is likely to alter his behaviour immediately following conviction, even following mere acquittal, arguably:
During the court proceedings, 95% (n = 217) of the sample confessed to having used child pornography, however, the evidence only led to convictions in 55% (n = 127) of the cases. Unlike current legal practice, saving illegal data in the temporary data cache of the computer was not enough for a conviction in 2002 – only if a permanent recording of the illegal material could be proven was it possible to convict the accused. Though only around half of the sample was actually convicted, there were sufficient grounds for assuming that all subjects in the study sample had consumed child pornography, seeing as they had registered with their personal credit cards and confessed to the allegations.
which, in the case of conviction, probably corresponds with a period of suspended sentence or at least probation (the study is silent on the sentencing status of those convicted).
With the objective of discrediting the study in terms of the broader aim of identifying a correlation
per se between child porn and hands-on child sex offences I could quite legitimately stop there, but I’ll go on.
The study sample is clearly inappropriate. From “Sample”:
Through a number of house searches in 2002, a Swiss police operation ("Operation Genesis") revealed over 400 persons suspected of having consumed illegal pornographic material via Landslide Productions Inc. Of these 400 people, only those suspected of child pornography consumption, were included in the study sample. This led to a sample size of N = 231.
… but from “Characteristics of the illegal material”:
19% (n = 43) of the sample were in possession of more than 5'000 files of illegal pornography [“illegal pornography” includes all types of illegal porn, not just child porn]. Forty percent (n = 93) of the subjects consumed only child pornography, the rest consumed other types of illegal pornography, such as pornography depicting sexual acts with animals, excrement, or involving brutality. One out of three subjects (33%, n = 77) consumed at least three types of illegal pornography.
Clearly there are too many factors influencing the study. Only those 40% of the sample who consumed only child porn can be considered to be a possible suitable study group. Interestingly and surprisingly, whilst the study included a stratified analysis separating illegal porn consumption between convicted and acquitted consumers, it did not take the opportunity to isolate the 40% of the study sample who
consumed only child porn. That seems like a gross oversight to me, and indeed seems suspicious.
... is generally suspect. From “Limitations”:
The investigated population could be severely biased, seeing as the users needed a credit card and sufficient knowledge of a foreign language (English) in order to gain access to the pornographic material. This could explain the rather elevated proportion of well educated subjects in the study sample and indicates that the investigated sample is most probably not representative for consumers of illegal pornography in Switzerland [not to mention other countries].
And from “Statistical Analysis”:
The aim of the stratified analysis was to determine whether a legal sanction led to a behavioral change regarding the consumption of illegal pornography
This seems like rather a misguided aim for a study entitled “The consumption of Internet
child pornography and violent and sex offending”. Consequently, the recidivism rates data presented in Table 1, which only show hands-on sex offences
collectively, are meaningless. There is no differentiation between the consumption of child porn specifically and illegal porn generally.
From “Criminal History”:
Before the police operation "Genesis" in 2002, 4.8% (n = 11) of the study sample had prior convictions for a sex and/or violent offence: Two subjects (1.0%) had prior convictions for hands-on sex offences involving child sexual abuse, 3.5% (n = 8) subjects had prior convictions for hands-off sex offences (possession/consumption of illegal pornography) and one subject had a prior conviction for a violent offence.
In other words
229 of the 231 men surveyed (99.1%) could arguably be categorized as “non-child sex offenders”
even before the 6-year monitoring period started. In a broader context 21 is a woefully inadequate sample size for identifying recidivism changes generally, especially when subdivided into three separate categories.
In summary:
At the absolute best, all the study does is present a statistical analysis of hands-on child sex offences
measured by registered conviction, investigation and charge rates by an unrepresentative sample of men within a specific social, cultural, ethical and legislative environment who,
in the past, happen to have consumed child porn to considerably varying degrees.
Indeed from “Background”:
Firstly, there are studies that examine the role of child pornography consumption on offending in samples of hands-on sex offenders. Following such an approach, Kingston, Fedoroff, Firestone, Curry and Bradford examined convicted hands-on sex offenders and found that the consumption of illegal pornography was a relevant risk factor, namely that those offenders who had consumed illegal pornography were more likely to re-offend – irrespective of their risk-level of recidivism.
And from “Discussion”:
The question of whether consumers of child pornography pose a risk for hands-on sex offences has not yet been answered.
Studies investigating hands-on sex offender populations support the assumption that the consumption of illegal pornography is a relevant risk factor for recidivism.