Home Office mum over mad March at the CRB 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 …
I wonder what the statistical chance or getting a wrong CRB Check is against getting run over outside your own home.
Would be nice to see NULabour doing something about that
I would imagine it would be a little more tricky than sticking up a speed camera...... and not really a vote winner either
Their concept of false positives and false negatives is hopelessly procedural. Surely their whole purpose is the protection of vulnerable individuals (improbable though that outcome may be)? A false positive means they did the paperwork wrong, and likewise a false negative means they let someone through who records showed was a risk. This is all about forms... Do we know that a false positive was really no risk? Of course, we don't.
The sooner this management-based bureaucracy is thrown out the better. Is there any evidence these partly-competent bureaucrats are having a positive effect in the world beyond their filing cabinets and flakey processes?
The fact they have no coherent overall management grasp of what they are doing, or how it is working, doesn't surprise me at all.
Seriously, if I was offered a sensitive job, and then my CRB check came back with a positive, does anyone really expect the employer to hang around for 21 days while I appeal it? Nope, they'd drop me like a greasy lead weight and employ the next sucker to walk through the door.
Likewise, if I did have a dodgy past but the check came back with a negative, who would notice? Who would appeal?
Although you might get some kind of indication of the number of false positives from the number of appeals, I assume that most people who pass when they shouldn't tend to keep quiet. Sure, some will re-offend, and some will turn out to be as bad as the rumours and innuendo listed against their name: but some will not.
I take issue. If it were "undoubtably important" then, quite obviously, before there was a CRB the world would have been going to hell in a handbasket, propelled netherwards by a foetid tsunami of evil criminal enterprise and all would be sweetness and light now.
Yeah, right.
The big new thing introduced by the CRB is that whereas in the old days you had to commit a crime* and be convicted by a court to be deemed a criminal, now you can be one just because someone's validation of a database fuzzy match is a sack of shit.
*Ok, that bit's optional.
Perhaps we should give them the benefit of doubt? Perhaps in March they just got a lot better at detecting the false positives and false negatives? This would just mean they were bad all the time but just have got better at realising how bad they are?
How do they actually work out their accuracy statistics anyway?
They could run a CRB check twice and consider it a mistake if they result don't match. This would detect cases where they pick up the wrong source record, but wouldn't detect at all issues where the source record is completely wrong.
It was the end of the financial year (FY), and they had not used up their error budget, so it had to be made up by extra errors that month.
Everyone knows that if you do not use your budget up in the current FY then you will lose it in the next FY by the amount you did not use in the last FY
Since they relaxed the requirement on people to produce harder to obtain Group 1 Documents (original Birth Certs, Passports etc) and now allow 5 Group 2's instead (stuff like payslips, utility bills, store card docs, tv license etc) its trivial to create a new identity for CRB checks anyway.
How do they measure the error rate?
How do they know they've caught all the errors?
What do they do to address the question of undetected errors when compiling their statistics?
How can they measure the undetected error rate?
How big do their error bars need to be when plotting the error rates?
On the off chance you are not really a troll:
"Surely their whole purpose is the protection of vulnerable individuals"
No. If that was the case then it would be OK to assassinate any non-vulnerable persons "just to be on the safe side".
Their main purpose is to protect the vulnerable, but with a duty to not screw up anyone's life who doesn't deserve it.
"A false positive means they did the paperwork wrong, and likewise a false negative means they let someone through who records showed was a risk."
And yet you fail to see the realities of this. Let me re-write it for you:
"A false positive means they unjustly tarred someone with being 'unsuitable' for a sensitive position and may as well just tattooed "paedo" on their forehead, and likewise a false negative means they put some vulnerable persons at an increased risk through failing to do the very thing they are there to do"
Importantly if there is [believed to be] a valid reason for a CRB then if you are seen to fail it then mud will stick very badly, and if you have a situation where there is a risk to one or more vulnerable people then there is a duty of care to make sure the bad guys don't pass through.
"Do we know that a false positive was really no risk? Of course, we don't."
Um, not the sharpest tool in the box, are you? Couple butties short of a picnic? Tinny short of a six-pack? The crb is not able to identify a risk through clairvoyance, it is supposed to check someone's background and make sure they have no previous, whether real or insinuated.
To use a "every man is a potential rapist" type argument is frankly absurd, your reasoning appears to be "if they fail the crb then they are a risk, and if they pass the crb they they are prolly a risk anyway - they have just not been caught yet". You'd never employ anyone
I have worked with people employing teachers in several secondary schools. There is a box to tick on the application form if you have anything dodgy to declare. I know for a fact that in the schools that I have experience with, if you tick the box, there is no way you will ever get a job there. It doesn't matter what you have done, no-one will ever read it. This practice is illegal but there is no way to detect it. The reason is that if you give a job to a potential peedo you may be held responsible and fired/prosecuted/held responsible, so you cover your back by suggesting that other candidates are clearly better.
If someone failed a CRB check, for what ever reason, unless they had been in the job for some time already, there is no way on earth that they would be employed/kept on. So 1570 false positives is 1570 new peedos. CRB people cant detect a false negative, so they are all false positives.
When the service is expanded to cover 10 times the number of people we will be making about 16000 new peedos every year. And with the enhanced CRB check, there is no way these people will ever work again. Thank god the government is keeping us all safe.
Chris, I don't think your fully understanding the situation.
The CRB data input bods mess up CRB checks on a regular basis by verifying someone else is CRB cleared. (ie, Mr Joseph Blags rather than Mr Joseph Blogs.) I know one person who had no less than three checks against other peoples identities performed even having informed the CRB in writing and with considerable detail of the problem.
Therefore, what they are defining as an error rate is either a false positive or a false negative. You see what they've spun their way out of counting? They aren't mentioning complete misidentification. How many times where there where someone IS authorised to do the job but they get a CRB check with someone elses ID.
If you did try and dig into it I think you'd get stonewalled, but I suspect the answer is "a lot". I'd be surprised it if was not several orders of magnitude higher than the admitted error rate by the offical stats.
AC for obvious reasons.
whoever gave the game away that March was dreadful, by emphasising the success of the previous 11 months, must be feeling the heat. Homer Simpson-like in his or her failure. Unless this was secreted into the press release, in which case kudos for the leak, kudos for making whoever made the statement look completely daft, and hope tracks have been well covered.
But I don't think it was a leak, I think this shows the Home Office way of doing things fleetingly biting itself on the arse.
surely this is disengenuous of the home office they have no idea of the actual error rate. Don't the numbers they refer to relate to the applications returned by them, to commercial submitting organisations, as containing errors. Alongside this the real undisclosed error rate relates to individuals wrongly being refused, for which they rely on complaints to identify these, given the nature of the process there will inevitably be less complaints than errors. And of course those wrongly passed for which one has to assume that there will not be a high conrresponding level of complaint !
If you read what I wrote you'll find I didn't say anything like:
"To use a "every man is a potential rapist" type argument is frankly absurd, your reasoning appears to be "if they fail the crb then they are a risk, and if they pass the crb they they are prolly a risk anyway - they have just not been caught yet"."
That argument and your rebuttal have nothing whatever to do with what I wrote.
You have made an argument I wasn't making and refuted it. That is your prerogative but you can't blame me. I'm not involved in your fantasy world.
Spoke to a teacher friend about this. He mentioned that staff (cleaners, janitors, cleaning staff, and teachers) all start looking for new places around that time of year and if there is a problem with thier CRB checks they tend to ignore it until they want to move then scream blue murder.
So, lots of school staff moving jobs suddenly find they can't and the "error" rate goes through the roof - is it just me or is the non technical, behavioural (brits dont complain unless they have to) reason never considered because .gov.uk lives in blame cultrue and this does not deflect blame
in the right direction?
Jacqui
Hi Jacqui,
I checked whether something like your friend's explanation could be to blame, but no.
As the footnote to the story says, there were 370,000 checks in March, compared to the monthly mean of 325,000. So a little above average, but nowhere near explaining more than 13 times the number of errors.
- Chris Williams
Consider:
In order to achieve processing targets staff may well defer dealing with more difficult cases until just that bit later,
This lag builds up throughout the year until March when there is a rush to get through, not only 45,000 extra cases, but a backlog of the awkward ones,
Those awkward cases are more likely to be error prone, need more work and are being dealt with at a particularly busy time.
A perfect storm for errors.
surely the CRB is CAPITA? like every other department? When such 'bodies' go wrong they talk about these Gov't bodies in the news as if they are part of the state but they are all privatised out to mostly Crapita arent they? THEY run the country, not the government. And this is the result.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
