Original URL: http://www.theregister.co.uk/2010/03/26/csord_review/

Sarah's Law review skewed by handpicked sample

Home Office research raises questions

By Jane Fae Ozimek

Posted in Government, 26th March 2010 11:26 GMT

Home Office jubilation over the "success" of its sex offender disclosure scheme may be premature amid yet more evidence of the Home Office twisting research to suit its own agenda.

That is the conclusion of sharp-eyed blogger, Hawktalk, who also questioned whether the way the scheme works in practice might create problems for wholly innocent individuals through the creation of misleading audit trails and over-zealous public officials applying the rule that "there’s no smoke without fire".

The Child Sexual Offender Review Disclosure (CSORD) – popularly referred to as "Sarah’s Law" – was piloted last year in four police forces: Hampshire, Warwickshire, Cambridge and Cleveland.

CSORD allows parents and those with responsibility for children to obtain information about the past (sexual) history of specific adults in contact with their children. Information would be provided at the discretion of the police, and those receiving it would be required to maintain what they learned in confidence. The focus is intended to be on disclosure in respect of convicted sex offenders.

A review of the scheme (pdf) by Home Office researchers was published earlier this month. The focus of the review was on the process of the CSORD – as opposed to the overall impact of the scheme – and was based on very limited data sets. According to the report:

The number of enquiries and subsequent disclosures under the scheme were smaller than initially anticipated: a total of 585 enquiries were made and 315 proceeded with as applications. Twenty-one disclosures (4% of applications) were made under the remit of the pilot with a further 11 applications resulting in non-pilot disclosures, i.e. information not relating to convictions for child sex offences.

When it came to looking at how the scheme had worked, the researchers were able to review just 159 applications – or just over half of those investigated by the police. This reduction in numbers appears to be because the police determined which cases to forward to the review body for evaluation, which raises questions about the soundness of the sampling procedure and possible bias introduced into the research by police forces.

If alarm bells were not ringing at that point, some additional concern should have been raised when it was revealed that the researchers were able to talk with just 43 of the individuals responsible for those applications. While the researchers inserted the usual caveats about concern due to sample size, the Home Office has nonetheless declared itself impressed with a scheme that is likely to make major inroads into the civil liberties of UK citizens – and signs are that they are likely to be rolling it out in future.

Hawktalk draws attention to a number of flaws in the scheme that appear not even to have crossed the researchers’ radar. Table 7 of the Home Office report shows that 54 per cent of the individuals investigated did not have access to the children named in the application. In other words, over half of the applications should have fallen at the first hurdle.

However, the police process throws up an intriguing anomaly, as the report reveals. After an initial application is made, step two is for police to trawl the PNC, sex offender and local criminal intelligence data in respect of the subject. Step four – after preliminary trawls have been made and possible flags set on these databases registering the fact of a check – is to decide whether or not the application is valid and falls within the scope of the CSORD.

On the same page, the researchers play down the finding that 18 per cent of applicants had applied to the police despite "having no particular concerns".

In other words - exactly as critics have claimed - it is possible for a sex offender check to be raised on the basis of simple curiosity. Assuming that the PNC and other official bases retain an audit trail of checks made, the fact that an individual has been checked will go on their record – ready to trip that individual up should they ever wish to apply for a job working with children in the future.

So, someone who has no history as a sex offender could generate a profile in the system as a result of repeated baseless inquiries - just as credit checks become a "fact" on credit databases even when there is no other source of concern over an individual's creditworthiness.

Last word to our blogger:

No data protection analysis or privacy impact assessment was made during the pilot with respect to disclosure. I believe the research does not provide a reliable basis for extending the scheme nationwide and that the disclosure procedures adopted by the police in relation to disclosure are flawed. Instead of a national roll-out (which risks being disastrous if it drives sex offenders underground), a more in-depth pilot is needed; one which includes some very important data protection considerations.

®