Feeds

Home Office mum over mad March at the CRB

A secret criminal record of errors

3 Big data security analytics techniques

The work of the Criminal Records Bureau (CRB) is undoubtedly important and often contentious.

So when on Monday this week it emerged that the error rate at the CRB had multiplied by a factor of 2.3 from a total of 680 in 2007/8 to 1,570 in 2008/9, it warranted further investigation.

Those error totals include false positives (people blocked from positions who should not have been) and false negatives (people allowed to take up positions they should not have been).

The Home Office issued an interesting statement alongside the figures. It said: "For the first 11 months of 2008-09 the CRB's accuracy rate exceeded the previous years. However, a dip in the quality in the final month brought down the overall figure."

Apparently intended to cast the 2008/9 figures in a positive light, with some rough calculations the statement in fact suggested a worryingly sudden collapse in the CRB's performance.

In 2007/8 the CRB ran 3.4 million background checks. In 2008 it ran 3.9 million, an increase by a factor of about 1.15. The 2.3 times increase in the number errors therefore couldn't be a simple function of more checks passing through CRB systems.

According to the Home Office statement, the proportion of errors from April 2008 through February 2009 was actually better than the overall rate for the previous year, which was 0.02 per cent.

So, at the end of those 11 months, assuming a constant number* of checks per month, there were fewer than 715 errors.

But by the end of March there had been 1,570, so there were at least 855 errors in one month. That compares to an average of fewer than 65 per month for the preceding 11 months.

For monthly error count to explode by order of magnitude, surely something must have gone drastically wrong.

We asked the Home Office press office for details. Perhaps there had been a technical glitch, new software, new staff or processes? Whatever the cause of the CRB's disastrous month, surely it was in the public interest to disclose it, and what has been done to ensure this "dip" is not repeated?

Apparently not.

After we sent him our sums, a spokesman said: "Not one single factor can be attributed to this dip in the quality in the final month.

"Disclosure quality and accuracy is heavily reliant on the information supplied by the individual and is subject to fluctuations."

The implicit suggestion of his second sentence - that in one month more than 10 times more CRB check applications than normal contained dodgy identity data - is simply not credible.

But he also said not one single factor caused the "dip". Fair enough. So what were the factors? The Home Office refused to provide this information, while admitting its answers were "not helpful" and pointing out that it is possible to appeal against a CRB decision, if a complaint is lodged within 21 days.

Our attempts to extract an answer as to what went wrong at the CRB in March over the course of several telephone conversations on Monday and Tuesday were consistently and aggressively refused.

As a final effort to uncover why the CRB got it so wrong all of a sudden, we'll try a Freedom of Information Act request. In the meantime and possibly forever - assuming the Home Office really doesn't want anyone to know - the lesson of this story appears to be: Don't apply for a job in March. ®

*A reasonable assumption for these purposes. The Home Office told us there were 370,000 checks in March 2008, only slightly above the monthly mean for the year of 325,000.

Combat fraud and increase customer satisfaction

More from The Register

next story
Lavabit loses contempt of court appeal over protecting Snowden, customers
Judges rule complaints about government power are too little, too late
Don't let no-hire pact suit witnesses call Steve Jobs a bullyboy, plead Apple and Google
'Irrelevant' character evidence should be excluded – lawyers
Record labels sue Pandora over vintage song royalties
Companies want payout on recordings made before 1972
EFF: Feds plan to put 52 MILLION FACES into recognition database
System would identify faces as part of biometrics collection
Edward Snowden on his Putin TV appearance: 'Why all the criticism?'
Denies Q&A cameo was meant to slam US, big-up Russia
Systems meltdown plunges US immigration courts into pen-and-paper stone age
Massive outage could last four weeks, sources claim
Ex-Tony Blair adviser is new top boss at UK spy-hive GCHQ
Robert Hannigan to replace Sir Iain Lobban in the autumn
Judge halts spread of zombie Nortel patents to Texas in Google trial
Epic Rockstar patent war to be waged in California
US Supreme Court supremo rakes Aereo lawman in oral arguments
Antenna-array content streamers: 'Ruling against us could dissipate the cloud'
prev story

Whitepapers

Mobile application security study
Download this report to see the alarming realities regarding the sheer number of applications vulnerable to attack, as well as the most common and easily addressable vulnerability errors.
3 Big data security analytics techniques
Applying these Big Data security analytics techniques can help you make your business safer by detecting attacks early, before significant damage is done.
The benefits of software based PBX
Why you should break free from your proprietary PBX and how to leverage your existing server hardware.
Securing web applications made simple and scalable
In this whitepaper learn how automated security testing can provide a simple and scalable way to protect your web applications.
Combat fraud and increase customer satisfaction
Based on their experience using HP ArcSight Enterprise Security Manager for IT security operations, Finansbank moved to HP ArcSight ESM for fraud management.