For the 999th time: targets do not work!

Somehow, this idea that setting targets improves performance and standards across the public sector and the commercial world has gripped politicians, chief execs and managers in a haze of long running hysteria that is only matched by a belief in witchcraft from the 17th century.

And I really cannot be bothered to go over again, the arguments for why targets do not work. (But if you really want to know, here are links to my other blog posts or just read anything by Simon Guilfoyle or John Seddon. And if want to go back a bit, why not read about W. Edwards Deming too.)

So I refuse to get all hot under my collar about the latest 'scandal' about the police fiddling crime data. Anyone who has been involved with the police service (as I have been for 25 years) knows that crime data is often (to use a technical term) 'finagled'! But then so are education performance league tables, and health stats, and the government's own research into PbR etc etc etc...! Why? Because people want to put on a good show, don't like getting beaten up or bullied (or 'performance managed' as it often called), think it doesn't really matter anyway, everyone's doing it, because it isn't really lying to guild the truth a little and so forth. And the fact that every senior manager and politician knows this but then seeks to ignore it or pretend that it can be smoothed out with ethical audits of data collection is living in la la land.

OK. Maybe I am being a bit polemical here. But seriously, when parliamentary select committees start putting on severe expressions and wanting to expose this scandal in policing, and the Daily Mail publishes articles entitled "How we can't trust the crime figures: After Plebgate, now watchdog says police statistics are unreliable", I think I am allowed a bit of hyperbole myself!

So what to do? I have been trying to do my bit by a) exposing the mythology of targets and b) asking the good people at ONS if they could make the Crime Survey of England and Wales have a bit more geographical granularity. Please see my blog post here as a starting point. And then I have been involved in a conversation with them:

191113 Email to crimestatistics@ons.gsi.gov.uk: Would it be feasible to have this survey broken down into police areas? I look forward to your reply. Many thanks

251113 Reply: Theoretically, police force area data are available from the Crime Survey for England and Wales, although due to reductions in sample sizes, we do not feel estimates at this geography level are sufficiently robust and we would advise against using them. 

271113 My response back to them: Many thanks for your helpful response. As psychology graduate, I appreciate the issue of sample size and significance etc. Are you able to estimate how much extra resource would be required to increase the samples to a size whereby individual forces & PCCs could more reliably use the data to compare trends over time and across the country? If so, I would be keen to know what that resource would be. Thank you.

161213 Their reply: My apologies for the slow reply. The difficulty with considering the sample size is that it varies for different crime type estimates. In order to get an overall count of crime at force level, the sample size may not need to be a great deal larger than it is currently. However, in order to measure specific types of crime by force, the sample would need to be extremely large. We're investigating whether we can feasibly calculate some figures for you, but I hope this information is still useful. 

161213 My most recent email to them: Many thanks for your email and insights contained therein. Most useful. I am trying to establish the principle of whether NCS E&W data could be collected in such a way that local communities would be able to know about trends in crime in their areas that are not dependent on reporting crimes to the police (which we know to be an underestimate and unreliable for other reasons). This would allow people to judge how safe their area and critically appraise the performance of their Police & Crime Commissioner. Whilst the link between what PCCs do and local crime figures (as measured by your surveys) is at best tenuous, it is the key measure that the government used (in part) to justify the value of PCCs. Please continue to investigate the feasibility of whether such figures can be calculated. Thank you.

And that is where the trail ends for the time being. But I am hopeful that this could lead to something. But I am just me. What could be achieved if it were the Public Accounts Select Committee or Home Affairs Select Committee that were asking these questions of ONS? Has the Home Secretary ever asked such questions?

I wonder...


UPDATE 0912|160114: As if my magic, up pops my own PCC with this headlined story in Police Oracle. Force targets 'will improve performance' says PCC. Ah well...