Saturday, April 27, 2013

Lewisham is the UK's least peaceful place


Well it certainly was last week when I read the BBC report that a new UK peace index had worked out out that that I live in the country's least peaceful borough.

The index consists of 5 separate indicators:
  • Homicides per 100,000 people
  • Violent Crimes per 100,000 people
  • Weapons Crime per 100,000 people
  • Public Disorder Offences per 100,000 people (used as a proxy for fear of crime, rather than using data from the Crime Survey of England and Wales, which isn't (as far as I know) robust to local authority level, though it's certainly more accurate than recorded crime figures)
  • Police Officers per 100,000 people (not including Policy Community Support Officers (PCSOs).)
So all parts of the index apart from Police officer numbers relate to recorded crime statistics, which are not known for their reliability. The UKPI report claims that while recorded crime figures underestimate total crime, they do so in a uniform way across areas. They don't provide a reference to back this up, and I'm not convinced myself.

Each of these 5 indicators is given a score from one to 5 depending on its value and then the indicators are weighted and combined to produce the overall index where 1 if low and 5 is high. The weights are as follows: Homicide 26.7%,  Violent Crime 26.7%, Weapons Crime 13.3%, Public Disorder 13.3%, Police Officers 20.0%. So a fifth of the index is accounted for by police officer numbers. It isn't clear whether a high number of officers is considered good or bad, but it's worth noting that Lewisham borough contains Britain's largest police station, and this police station is located in Lewisham Central ward, which the report notes accounts for the majority of the crime recorded in Lewisham... Alarm bells are ringing already!

The method used to work out the index is pretty standard, but most other indices used in social research either have a lot more indicators feeding into them (eg the indices of deprivation have 38 separate indicators) or are based on a range data, eg Townsend and Carstairs scores. The Townsend score is made up of four variables from the Census:
  • Unemployment as a percentage of those aged 16 and over who are economically active. 
  • Non-car ownership, as a percentage of all households. 
  • Non-home ownership as a percentage of all households. 
  • Household overcrowding. 
Carstairs, developed to avoid the use of households as the denominator, uses the following four Census variables:
  • low social class
  • lack of car ownership
  • overcrowding
  • male unemployment

What you'll notice about these indices is that they contain variables that differ from each other in what they are trying to measure, though all cover different aspects of deprivation. The variables going into the peace index all relate to different types of crime. Nothing else (apart from the odd choice of police officer numbers) is considered.

The other thing that annoyed me about the peace index was its undue focus on the top and bottom areas. Lewisham and four other areas are picked out and discussed in detail. Lewisham's score was 4.529, Tower Hamlets 4.360. Lewisham's neighbouring boroughs of Southwark and Greenwich scored 4.314 and 4.002 respectively, yet these areas are not discussed at all, and no indication is given as to the confidence we can put on the figures. Confidence intervals are not shown, if they were calculated at all, so we have no idea whether or not the figure of 4.529 is statistically significantly different from the 4.494 scored by the second placed borough, Lambeth. My guess would be it is not, and one thing I would be willing to bet good money on is that these numbers will vary a lot by area on a year to year basis.

Still, people love a league table, reliable or not, so the press and public were all over this story last week and of course Lewisham council's very own rentaquote was on hand to call for the Mayor's resignation over the figures. I'd argue that this is a disproportionate reaction and the Lewisham is no worse than other inner London areas. One of our local MPs, Heidi Alexander, has a more measured approach arguing that, "for most people, most of the time, Lewisham is a safe place to live." That doesn't mean that things shouldn't be done to tackle youth and gang-related crime, but these initiatives need to cross artificial borough-based boundaries and look across London for solutions, and this is where league tables simply do not help. Organisations such as the Jimmy Mizen foundation are working tremendously hard to put a stop to youth violence and these simply don't get a mention in the peace index report, which prefers to start its section on Lewisham with a reference to something which happened in 1977. By all means use data to start a debate, but at least make sure it's robust, and that you do your research into what's going on the ground now, before pontificating!

If this index is produced again next year, lets see just how different the figures are year-on-year before singling out Lewisham for criticism!