Several points stand out from this study.
the low reconviction rate of serious
sexual offenders for another sexual offence is confirmed. Even after
six years it remains below 10 per cent and is not very much higher than this
when convictions for serious violent
crime are added into the equation. However, it was not insubstantial (about one in four within six years) amongst those who had committed extra-familial offences against children, although they made up just one-fifth of the sample.
It has to be recognized that those who were reconvicted of a sexual offence had, as judged from the sentences imposed, committed another grave offence. In other words, the general level of risk was relatively low, but the stakes were high. In the shorter time span -- compared with the old parole system -- that now exists under the Discretionary Conditional Release scheme between parole eligibility and the point at which the prisoner denied parole would otherwise be released from custody (i.e. between half-way and two-thirds of the sentence) very few of these serious sex offenders would have accrued another conviction for a serious crime of the same type.
This study found that only 1.2 per cent had been imprisoned again
for a sexual crime, and less than 5 per cent for a sexual or serious violent
crime, within two years of their release. These facts need to be more widely
recognized and disseminated if there is to be rational debate on this emotive
all but one of those subsequently reconvicted for a sexual offence were
identified by at least one panel member as a 'high risk'. On the other hand,
panel members considered more than half the sex offenders they reviewed to be
'high risk'. If sentencing policy for sex offenders were to be changed, in
line with the Halliday recommendations, so that those regarded as dangerous by
the courts could be retained in custody until the end of their sentence, it is
possible that the Parole Board will be very reluctant to take the risk, as
they see it, of releasing such prisoners earlier in their sentence. This study
suggests that, if this were the case, a substantial proportion of those
detained for longer terms in prison than is possible at present would be
likely to be 'false positives', even if the Board were to make use of an
actuarial risk assessment tool such as Static-99.
The evidence from this study suggests that Board members over-estimated, in particular, the risk of reconviction of sex offenders who had committed their crimes wholly within their own family unit as well as the risk of reconviction posed by 'deniers'.
Until recently, as we have shown in a study of parole decisions made in 1998 and 1999, those who denied their sexual offences were very rarely granted parole (Hood and Shute 2000: 29-30).
However, according to the most recent Report of the Parole Board, an internal study of sex-offender 'deniers' considered by the Board between 17 April and 27 July 2001 showed that 19 per cent were granted parole, a rate not much different from that for all sex offenders (21 percent).
This change in the Board's practice in relation to 'deniers' was attributed, in the main, to the decision reached by the Court of Appeal in the case of Oyston, in which it was held that a prisoner's application for parole should not be refused on the sole ground that he denied the offence, and to subsequent training of Board members (Parole Board 2001).
The research reported in this article should give some
comfort to the Board that its change in practice in relation to 'deniers' will
not be accompanied by a large rise in sexual reconvictions amongst the
relatively small proportion of 'deniers' who are now granted parole.
it is important to recognize that these findings should not be taken to be a particular criticism of members of the Parole Board and that the prisoners followed-up in this study were considered for parole before new procedures and directions were issued to the Board as a consequence of the implementation of the Criminal Justice Act 1991
The problem of 'false positives'
arises largely because of the difficulties of trying to
rare events, a difficulty that has haunted all efforts to identify in advance,
by clinical or actuarial means, those who subsequently turn out to be
'dangerous' (Centre for Sex Offender Management 2001). Attempts to predict
reconviction when the 'base rate' is low inevitably produce a high rate of
'false positives'. Nevertheless, it may be that parole assessments of the
risks posed by serious sexual or violent offenders could be somewhat improved
if the Parole Board were in the future to have available to it a new actuarial
instrument which takes into account both static and dynamic risk factors and
measures of actual reoffending rather than solely reconviction.
we think that our study may shed light on a puzzle discussed both by Marshall (1994) and by Friendship and Thornton (2001), each of whom have noted a substantial decline in the proportion of sex offenders reconvicted of a sexual offence in recent years: a decline that, according to Friendship and Thornton, 'appears to be unique to England and Wales'.
They followed-up more than 1,000
sex offenders released from sentences of four years' imprisonment or longer in
1992 and 1993.
This compares with a sexual reconviction rate of 12 per cent found by Thornton and Travers in their follow-up study of 313 offenders released from sentences of similar length in 1980, also followed-up for four years.
The explanation for the decline favoured by Friendship and
Thornton is that there has been a general decline from 38 per cent to 17 per
cent during the period 1981 to 1997 in the proportion of reported sexual
offences which resulted in a conviction or caution. In other words, it had
become more difficult to bring reported cases to book, either by clearing them
up or by securing a conviction. The decline in the proportion of sexual
offenders reconvicted of a sexual crime is, they believe, probably due to the
difficulties of securing convictions rather than a decline in actual sexual
reoffending by this group.
appears to be a plausible explanation, and it may provide part of the answer.
But it rests on the assumption that the proportion of sexual crimes committed
by sexual offenders released from a long-term prison sentence which resulted
in a conviction has fallen since 1980 to the same extent as the ratio of all
reported sex offences to those reported and cleared up by conviction. This
may or may not be the case.
There is another factor which may have played a part.
In our sample of cases, drawn from the early 1990s, 41 per cent of the prisoners had been convicted of an offence solely against a child within their own family circle. We have shown that none of the prisoners in this category who could be followed-up was subsequently convicted of a sexual offence within a four- and a six-year follow-up period, and this is not out of line with the literature generally, which has shown such offenders to have a much lower reconviction rate than those who victimize children outside of their family unit (see for example Greenberg et al. 2000).
It may be that the higher profile given to child sexual abuse in the last 20 years has increased the proportion of sex offenders serving long sentences in prison who have been convicted of an intra-familial sex offence against a person or persons under the age of 16.
Certainly there is evidence to suggest that among sex offenders in prison
during the 1990s a higher proportion had been convicted for victimizing
children than in the previous decade (Thornton and Travers 1992) .It is
probable therefore that the proportion of intra-familial offenders has
also increased. If this were so, one would expect the rate of reconvictions
for sex offences of sex offenders released from long-term custody to have
declined. The hypothesis that a change in the composition of the sex
prison population may be part of the explanation for the declining
reconviction rate is, in our view, worthy of further exploration.