We are developing the social individualist meta-context for the future. From the very serious to the extremely frivolous... lets see what is on the mind of the Samizdata people.

Samizdata, derived from Samizdat /n. - a system of clandestine publication of banned literature in the USSR [Russ.,= self-publishing house]

Violent crime in Britain

Here are the latest statistics on crime in Britain. Police statistics, according to this BBC report, show that violent crimes have gone up, while another survey shows that violent crimes are broadly stable. (The usual health warnings about statistics obviously apply). However you look at it, crime is high.

Regardless of what one thinks about the potentially civil libertarian worries about millions of CCTV cameras now scattered around the country, it hardly appears that they are very useful in deterring crime, which as far as I know, was the stated purpose for the things.

55 comments to Violent crime in Britain

  • D Anghelone

    Here in the US it’s been written that we tally violent crime by body count while you folk tally by incident. That is, three Brits killed in one incident will be tallied as one murder.

    True?

  • Richard Easbey

    as a fellow American, D. A., that doesn’t make any sense. Those would be VERY distorted and unreliable statistics…. at least, that’s the way it seems to me.

  • Midwesterner

    D Anghelone & Richard Easbey, from what I’ve read, true for Scotland but not for England and Wales. IIRC Dunblane was recorded as one homicide with multiple victims. They may have changed methods, since.

    And, as Americans, does that last figure of 24% of people were victims of at least one crime in a year sound amazingly high? It sure does to me.

  • Midwesterner

    I should have mentioned, they have (had?) a separate category for number of victims. Think of it as counting criminals v. counting victims.

  • Howard

    You have forgotten that most other sorts of crime is down, down, down. So cctv does work after all.

  • Quasi

    24% doesn’t sound too high for someone living in the inner-city of a major city but is unheard of most anywhere else in the US. Don’t forget that a small number of individuals in a handful of cities greatly skew US crime rates. Most of us behave ourselves like civilized humans.

  • D Anghelone

    And, as Americans, does that last figure of 24% of people were victims of at least one crime in a year sound amazingly high? It sure does to me.

    Lies, damned lies and too many yobs.

    This torpedoes any hope of my understanding the figures:

    “Under the National Crime Recording Standard established in 2002, low-level anti-social behaviour not previously included started being recorded as violent crime.”

    Anti-social behaviour? Samizdata.net could account for the rise. :-)

  • Agammamon

    Heh, acording to Charles Strauss in The Atrocity Archive, they’re really basilisk weapons to protect London against the return of the elder gods in 2007.

    Gonna’ turn any spawn of Cthulhu into stone if they get within LOS.

  • Midwesterner

    D Anghelone, if it wasn’t so serious I would be having quite a chuckle. As it is, your humor may be predictive.

    I went looking and found a definition of ‘antisocial behavior‘.

    1.6 Figures 1-4 also reveal some stronmg commonm elements of behariours that are included in definitions as anti-social. These include:

    noise;

    conflicts, including harassment, domestic violence and racist incidents;

    litter and rubbish dumping;

    graffiti and vandalism;

    uncontrolled pets;

    nuisance from vehicles, including parking and abandonment; and

    unkempt gardens.

    ‘unkempt gardens’?! In UK I would be quilty of a ‘violent’ crime!

  • Midwesterner

    Apologies for the typos, I had to manually re-type it.

  • Robert Alderson

    The figure of 24% of Britons being the victim of crime is from the British Crime Survey which is, essentially, an opinion poll which asks a large sample of people if they have been the victim of crime. This will be far greater than the number of people who bother to report an incident to the police.

  • Jim

    “However you look at it, crime is high.”

    Not true. Figure 1.3 in this summary from July 2005 of trends since 1995 shows that the experience of violent crime (as actually reported by people, that is, rather than as recorded by the police) is down 43% in a year.

    That is, violent crime has nearly halved. So no, it’s not high however you look at it. Something to celebrate, no?

  • Johnathan Pearce

    Howard says that other forms of recorded crime are down. Fine. But violent crime is up. So again, do the CCTVs work in the latter case? I’d say no.

    Jim, you say violent crime has nearly halved and the situation is something to celebrate. I must say that this contradicts pretty much most statistics I come across, not to men anectodal evidence. In a way I hope you are right since there are times when I think this country is turning into a nightmare. Anyway, think of the heavy spate of shootings in places like Nottingham in recent months, etc. Hardly reassuring.

  • Tim Sturm

    Have cameras had any effect on the proportion of crimes solved?

  • Jim

    “I must say that this contradicts pretty much most statistics I come across, not to men anectodal evidence.”

    Like what? And please don’t say police records, since we already know that they’ve changed how they record crime (see D Anghelone’s comment above). So, apart from that, what statistics exactly are you talking about and how are they more representative than an annual survey of 40,000 people?

    ” not to men anectodal evidence.”

    Says it all really.

  • Johnathan Pearce

    “And please don’t say police records,” says Jim. So what sort of data would you use then? Stuff gleaned from the back of an envelope?

    This story from the BBC on July 21:

    “Violent offences in England and Wales reached record levels in 2004-05 with police recording one million crimes, up 7 percent from the previous year”.

    That is a high figure, regardless of how you can try to parse the data or look at how it is recorded.

    “Police figures show 1,035,046 violent incidents against the person, excluding sexual offences and robberies,” the story continued.

    http://news.bbc.co.uk/1/hi/uk_politics/4700575.stm

    As you might say, that says it all, really.

  • kARL rOVE

    The British Crime Survey is not as impressive as Jim & al Guardian think. The obvious question is – why is crime mostly unreported? If the left-wing sociologists who run the BCS – whose figures may well be faked – were to ask, many people wd say because they’re too frightened.

  • Jim

    “So what sort of data would you use then? Stuff gleaned from the back of an envelope?”

    Good Lord. Must I repeat myself? I used data from the British Crime Survey, a nationally representative annual survey of 40,000 people. I already said this quite clearly above and linked to the actual document, yet you seem to think this is ‘gleaned from the back of an envelope’. Honestly, your wilful ignorance is breathtaking.

    “Violent offences in England and Wales reached record levels in 2004-05 with police recording one million crimes, up 7 percent from the previous year”

    How many times do people have to tell you that police records of recorded crime are not reliable indicators of trends before it sinks in?

    I can only conclude that either you’re a complete numbskull or you’re simply choosing to rely on whichever statistics happen to support your beliefs regardless of their quality. Which is it?

  • Johnathan Pearce

    Jim, keep your temper, please. You prefer to base your views on a sample of 40,000 people instead of recorded, hard data by the police, that is your privilege. As a guy said before you, there is evidence that a lot of crimes are not reported.

    Nothing would make me happier to think that we are living in an era of rapidly falling crime rates.

  • Jim

    “You prefer to base your views on a sample of 40,000 people instead of recorded, hard data by the police, that is your privilege.”

    You obviously have no understanding of statistics if you think that a sample of 40,000 is something to be sniffed at. That’s a massive sample, and completely representative. As for the “hard” data recorded by police, they change their definitions all the time so by definition the trends are not reliable.

    ” As a guy said before you, there is evidence that a lot of crimes are not reported.”

    That’s the whole damn point. The BCS picks up a *higher* level of crime than the police statistics, but this level has been *falling* over time because less crime has been committed.

    It might be an idea to actually read the document I linked to above before continuing to make such hopelessly ill-informed claims.

  • Johnathan Pearce

    Jim, let’s have another go (I read the item you linked to). If police statistics under-report crime, then, taken at face value, the numbers are even worse, presumably. Right? I don’t see how quoting the BCS data quite counters that.

    The BCS data may be taken from a representative sample of 40,000 people. (I dunno quite how “representive,” I haven’t time to look at the sampling methods). But consider this, out of a population of about 60 million persons, more than 1 million crimes were reported last year. Fact. That’s a lot of crimes, however once slices and dices it.

  • llamas

    If noone can agree on which set of statistics are true representations of crime trends in the UK – perhaps we should look at market-driven indicators. What are people spending their money on when it comes to crime?

    Burglar alarms
    Bolts and bars
    Sophisticated car alarms and the like
    (I’ve worked in auto theft in the US, and will say, without equivocation, that the standard locks and alarms now fitted in UK automobiles are among the most sophisticated in the world – streets ahead of what is normal in the US. These things are not done for amusement.)
    Private security

    When I was last in England (a year ago) I was told, quite unequivocally, that there were parts of town where I shouldn’t go at certain times of day. Apparently, random acts of violence, petty street crime, public drunkeness, gang activity and drug crime are now so endemic that parts of some towns have been effectively ceded to the goblins. If you’d have told me, 25 years ago, that it wasn’t safe for me to walk through downtown Nottingham or Derby on a Saturday night, I’d have laughed in your face. What would you say to me now? I’m no shrinking violet – I’ve lived in places like Brixton and Stockwell, also in the suburbs of Detroit, and the fear and concern I hear expressed by the residents of what used to be peaceful, provincial towns in the UK far exceeds anything I’ve ever heard in those less-than-salubrious places.

    Are all these people deranged, deluded by a chimera of increasing crime which does not actually exist? I don’t think so.

    Read your own newspapers. Every other day, it’s yet-another example of some ghastly act of senseless violence, more-and-more often with firearms, each one more-senseless than the last. Then go and look in the newspapers of 25 years ago, and see how many such incidents were reported. Random shootings in the streets, innocent citizens killed in the cross-fire? Virtually unheard-of, yet now it’s every week. Remember when WPC Fletcher was shot in St James’? The sheer outrage about such a crime lasted months. Now every PC on the street wears body armour – and it’s not to keep them warm.

    Arguie about statistics all you like. The real-world evidence speaks volumes, even if the exact proportions cannot be quantified.

    llater,

    llamas

  • Jim

    “If police statistics under-report crime, then, taken at face value, the numbers are even worse, presumably. Right?”

    Yes, when applied to the total level of crime. On violent crime it’s not quite as clear, as what the police call ‘violent crime’ might not be the same as what people call it (see the example of anti-social behaviour mentioned above – also, the police apparently count one man punching two other men as two violent crimes).

    “I don’t see how quoting the BCS data quite counters that.”

    The point I was making was not about the level of violent crime but about the trend, which is very significantly down.

    “That’s a lot of crimes, however once slices and dices it.”

    Yes, it is. And in 1995 there was a lot more, which was the point I was making. I’m certainly not pretending that we live in a crime-free land.

  • Daveon

    So again, do the CCTVs work in the latter case

    They appear to be very useful for identify terrorists after the fact, which, if they aren’t detering things will be a suitable bogeyman for the public.

    When I was last in England (a year ago) I was told, quite unequivocally, that there were parts of town where I shouldn’t go at certain times of day. Apparently, random acts of violence, petty street crime, public drunkeness, gang activity and drug crime are now so endemic that parts of some towns have been effectively ceded to the goblins.

    And???

    I remember my mother telling me that there were parts of London that she, as a girl and young working woman wouldn’t go it at certain times because of the risks. That was in the 1940s and 50s.

    There were parts of Manchester that were no go zones, especially Moss Side, when I worked at Salford Quays in the early 1990s.

    I can’t speak for Derby or Nottingham, (although I do recall friend’s moving to Beeston in about 1997 and finding it pretty vile and worrying about their safety), but personal safety no-go “zones” have been a fact of life as long as I can remember (37 years so far).

    What doesn’t get factored into the mix is that some of the “no go” zones my mother would talk about in East London, or my sister as a WPC in the 80s in Stoke Newington are now anything but.

    The question is how much worse things are now to 10,20 or more years ago. Not that there are areas where things are bad.

    Another point. You mention PC Fletcher, yes there was outrage, but that had a lot to do with letting the bugger walk and it being a shot from a foreign embassy.

    Police do wear body armour, but I suggest you Google on the number of officers killed in the line of duty in the UK – it is still a fantastically low number given all this apparent “crime”.

    I’m not trying to say that there are not problems, but there does seem to be a knee jerk “in my day…” reaction going on here.

  • Daveon

    violent incidents against the person

    I’ve been concerned for a long time about this “phrase” and it’s formal police use.

    A lot of the growth here matches the growth of expensive consumer electronics carried on the person which can be grabbed in a second and run off with.

    Yes, it’s still crime – but it’s an emotional way of describing something more mundane. There’s also the issue of crime tending to follow cost trends – burglary is down as the value of items to steal in the house tends to drop. Why nick a DVD player to flog to a man in a pub if a person can buy a legit DVD for £20 in a supermarket.

    People routinely carry a mobileworth (NB: Different to the amount you pay for one) £ 100s and ever increasingly a music player.

    All easy to steal as opportunitistic crime and easy to sell for a lot less than you’d buy one for.

    Not sure that has changed much either from people targetting cash and cheque book when I was a kid.

  • Johnathan Pearce

    Jim, we’ll have to agree to disagree. Let me be clear, I have no desire to exaggerate the problems of crime. If the data consistently shows a significant drop, then I’ll hoist a beer for whoever helps bring that about. As a libertarian I am certainly no cheerleader for “moral panics” on crime or other pathologies. Some crimes are clearly falling, others are flat. But the violence data bothers me.

    (Full disclosure: I was mugged badly, and twice, about five years ago).

  • Midwesterner

    Jim,

    I found this homicide data from the Home Office. I choose this data because it is based on police reports of violent gun crime, which has presumably been one of the most accurately and consistently recorded categories of crime. I believe it is one of the most resistant (but not totally immune) to spin or changes in record keeping. Therefore, I think it can be used to project general trends in overall crime. Llamas perceptions are probably accurate.

    http://www.homeoffice.gov.uk/rds/pdfs2/hosb0104.pdf(Link)

    Warning, it’s fairly big.

    Look at the graph on page 4. The data goes back to 1951 and clearly belies the trend you are describing.

    Look at page 8. You have to look at the total, not the convictions because cases pending is a huge factor. Curiously, if you look at the number of suspects, it trends flat while the number of homicides increases ten percent. The usual suspects are getting away with more crime.

    Look at page 9. Don’t just look at the number of offenses recorded, look at the number per million population. You want to talk trends, from ’85 to ’02-’03 it’s up almost double.

    Look at page 26. Firearms crimes are clearly trending up since 98/99. In that period, the non-airguns part of firearms crime (‘real’ guns) virtually doubled. In FIVE YEARS!

    Look on page 30. From ’98/’99 to ’02/’03 total ‘real’ gun offenses went from 864 to 2,179. Fatalities went from 49 to 80. Serious injuries went from 162 to 416. Slight injury went from 653 to 1,683.

    Look at the chart on page 31. The category of gun crime called ‘violence against the person’ more than doubled from ’92 to ’02/’03. It is a steady upward trend. Unlike robbery, it doesn’t fluctuate down. Only up.

    Look at page 32. The ‘real’ guns are used almost exclusively against people. The air guns specialize in crimes against property. That’s why I’m focusing on ‘real’ gun crimes.

    Look at Table 2.01. For all offenses in which firearms were used, excluding criminal damage, the year you picked, 1995, is 7,577. For ’02/’03 , the most recent data, it is 13,114. The figures in the columns of violence against the person, are similar.

    Look at Table 2.06. Crimes in which firearms caused injury, excluding air guns, went from 597 in 1992 to 2,179 in ’02/’03.

    To quote Truman (with a nod to D Anghelone), ‘There’s lies. There’s damn lies. And there’s statistics.’ I think this data is probably the least vulnerable to spin or manipulation.

  • karl rove

    Mr Pearce – the BCS is done by phone. Cheaper but less reliable than face-to-face. They claim 40000 subjects (not necessarily true). The cops record only 5 million crimes. But that’s a sample of several million.

    As a matter of interest, about your muggings – did anyone go to jail?

  • Johnathan Pearce

    Midwesterner: thanks very much for the stats. I think that settles part of the debate I had with Jim (though to be fair to Jim he may have a point about other trends in crime).

    Karl (!!) I was hit over the head in the first instance, blacked out and was robbed of my wallet, briefcase and house keys. In a very bad state. I had to take three days off work. In the second case I was also hit right outside my front door. Both incidents happened when I was living in Clapham, south London. I never want to go the place again.

    No-one was convicted as far as I know and the police told me that the chances of getting an arrest were small. I dunno if that is a typical outcome but my hunch suggest that it is.

    Tim Sturm earlier asked whether CCTV may help get convictions. I’ll have to pass on that one. Fraid I don’t know.

    Anyway, off for a relaxing evening. Thanks for the comments.

  • guy herbert

    llamas,

    The market also tells you that British people are among the most malnourished in the world (after Germans and Americans, of course), if you view their vitamin supplement consumption in the same light.

  • Midwesterner

    Actually, Guy, since you said malnourished, not undernourished, it’s not that far from possible. Double bacon cheeseburger with fries and soda (or the UK equiv.) isn’t the well rounded meal it’s cracked up to be. ;-)

  • anon

    Trust non of the statistics.
    If you want to know the levels of crime in the UK, you’ll have to get the documents and work them out yourselves.

    Take an example.

    I listened to a BBC radio report on firearms incidents.
    The number of incidents is increasing.
    Everybody in their right minds must think ‘Terrible! It’s not safe to leave the house any more’.

    Until it was also pointed out that the number of deaths and injuries caused by firearms was decreasing.

    Everybody now thinks that there are a bunch of cowboys out there, armed to the teeth, who can’t shoot straight.

    Guess again.

    Firearms incidents now includes reports of little boys running around playing with those plastic ‘Soft Air Guns’ – the ones that shoot those little plastic balls – which in turn is a pretty fair summation of the statistics, and the government policies on crime, as well as the government.

    Whoever said ‘there are lies,
    d****d lies,
    and statistics’

    Didn’t know the half of it

  • Midwesterner

    anon, you need to go back and read not just my post at 4:45, you need to go to the link I referenced. It contradicts very thoroughly all of the points you make.

    Air guns are recorded separately. Look at murder and violent assault. Feel free to ignore the air gun crime, I did.

    I was going to repeat my opening and closing remarks from that post but why don’t you just go back and read them.

  • Midwesterner

    An amendment to my 4:45 post.

    The change from ’round’ years to ‘split’ years, example 1992 v 2002/2003, is due to change in the recording period not the combination of two years numbers.

    This change occured in the 1996 to 1997/1998 accounting years.

  • zmollusc

    Hmmmm………
    If CCTV cameras really inconvenienced criminals, surely they would be shooting the things? Is there any evidence of this?

  • Jim

    Midwesterner,

    “clearly belies the trend you are describing.”

    Not quite sure what you mean by ‘belies’ but if you mean ‘contradicts’ then, no, it doesn’t. Homicides and firearm offenses together account for less than half a per cent of violent crime, which, whether you like it or not, is clearly down.

    Domestic violence, to take a specific example, accounts for a much larger proportion of violent crime, and according to the BCS that’s down by 60% since the mid-1990s – are you saying that’s irrelevant?

    Homicides and gun crime has gone up, and that’s worrying, but that quite clearly doesn’t mean that violent crime in total has also gone up. It hasn’t – it’s gone down, and nobody has yet managed to give me a good reason why the BCS statistics on this are wrong. There’s been a lot of vague hand-waving, of anecdotal evidence and the usual conspiracy theories – the usual Samizdata shite, in other words.

    I appreciate your taking the time to go through the stats on homicide and gun crime, and I don’t want to belittle the importance of the rising trend in these areas, which is certainly worrying. But you’re not actually contradicting what I said.

  • GCooper

    Jim writes:

    ” nobody has yet managed to give me a good reason why the BCS statistics on this are wrong.”

    Because they are the result of a bloody opinion survey.

    If you can’t understand why they are, essentially, no more reliable that the ‘secretaries say bosses don’t wash their underpants often enough’ or ‘husbands don’t do the housework’ nonsense that newspapers publish because they won’t employ journalists to find news, then there is little point in anyone trying to engage you in serious debate.

  • Midwesterner

    Jim.

    “Not quite sure what you mean by ‘belies’”

    M-W 11th collegiate ‘belie b : to present an appearance not in agreement with’

    “- are you saying that’s irrelevant?”

    I’m saying it was a survey written by people with a case to make. You can get a survey to say anything depending on how you write the questions. Not only that, it’s a very emotional type of crime, very subject to the feelings and interpretations of the victim. A simple MSM story might lead someone to believe that a remark was an assault, while a victim of serious assault may be too resigned or intimidated to report even battery. See my comment below on why I chose ‘real’ gun injuries.

    “There’s been a lot of vague hand-waving, of anecdotal evidence and the usual conspiracy theories – the usual Samizdata shite, in other words.”

    There wasn’t much of anything anecdotal in that report. You want anecdotal, look at the BCS. I understand it’s a telephone public opinion poll, er, crime poll!

    “I don’t want to belittle the importance of the rising trend in these areas”

    I selected them precisely because it’s difficult to re- or mis- interpret a bullet hitting someone as anything other than a bullet hitting someone. And they have probably the highest and most consistent report rate. In all other categories, the bullspin is amazing.

    I saw that trend data that you site when I’ve been out looking for info. I’ve been looking at enough data the last few months that I’m learning to spot data manipulation a little better. I frequently find data so heavily biased or filtered that it’s almost impossible to find the truth in it. Did you know that most official crime data doesn’t draw any distinction between killing somebody who’s trying to murder you, and killing someone to steal their wallet?

    Good luck trying to figure out fact from fiction. If you would like a single, simple example, check out my next post.

  • Midwesterner

    Anybody want a good example of how you can turn a flat trend into a really impressive improvement?

    http://www.homeoffice.gov.uk/rds/pdfs2/guncrime.pdf(Link)

    It’s just a one page PDF.

    Look at what the Research Development and Statistics Directorate (RDS) of the Home Office did with 800 offences in Sept ’01 fluctuating over to 800 offences in Mar ’03. They generated a bar graph that went from plus 50 to minus 10.

    Also, in the comments, did you catch that reference to 80 gun fatalities, ‘down from 95’? Makes things look really good, huh?

    Well, I found 5 years of that statistic. The five year trend is 49, 62, 72, 95, 80. My bet is the trend regresses in approximately that same pattern. Not only that, see Table 2b on page 30 of (61 page PDF) and you’ll see that in all of ‘non-air weapons’ (read real guns)where there was a year/value change, they picked the only one of 16 that went down! Hmmm…..

    When there is this much of an effort to manipulate the data, you cannot believe summaries. You cannot believe surveys. You have to go back to the most raw data you can find and then check even that for games with data collection.

    .

  • John K

    Police do wear body armour, but I suggest you Google on the number of officers killed in the line of duty in the UK – it is still a fantastically low number given all this apparent “crime”.

    It doesn’t surprise me in the slightest that few police officers get killed in the line of duty. The police rarely “fight” crime, they turn up after a crime has been committed, take statements, issue a crime number, and every so often arrest a suspect, who then has to be processed for about six hours back at the station.

    Very few criminals are so stupid as to want to kill a copper, there’s no point attracting that sort of heat. I remember being told once that on average more social workers are killed on duty than coppers. I don’t know if that’s still true, but it makes sense. They tend to work on their own, in rough areas, with disturbed “clients”, and have no body armour or weapons.

    I noticed that in the recent Birmingham riots one officer was “shot” by a BB gun. The press tried to hype it up, but it hardly counts as a bad night in Basra does it? Still, it’ll be worth six months on the sick for the copper.

  • Jim

    Gcooper,

    “Because they are the result of a bloody opinion survey.”

    The best way to find out how much crime people are experiencing is to ask them. The worst way is to ask the police. Frankly I’m surprised Samizdatistas are so trusting of the cops, but then again I suppose it’s another case of believing whoever is telling you what you want to hear. The police change how they measure crime all the time – including gun crime, but apparently that’s gospel truth to you. Meanwhile a survey of 40,000 people which hasn’t changed its methodology since 1981 gets dismissed as sophistry. What a collection of cranks this place is.

    Midwesterner,

    “to present an appearance not in agreement with”

    Well yes, it attempted to present that appearance, but in fact it wasn’t in disagreement at all, as I explained. You’re just cherrypicking the one or two trends out of all violent crime that happen to support your case, and declaring that everything else is fraudulent. For example,

    “You can get a survey to say anything depending on how you write the questions. Not only that, it’s a very emotional type of crime, very subject to the feelings and interpretations of the victim.”

    I see. People are obviously just making stuff up about their experience of domestic violence in vast numbers, and if that’s not enough, the BCS researchers are manipulating their answers. It’s all a vast conspiracy to convince people that domestic violence is falling!

    Please cop the fuck on. The BCS has asked the questions consistently since 1981 – in direct contrast to police recording. The data shows a consistent drop since 1995. The BCS has also used a self-completion questionnaire to measure inter-personal violence – that also shows a drop over time. Those are the facts. Now, for some reason you’d like to ignore that – whatever, I don’t really care why. But please don’t lecture me about scientific method because you appear to think that means “whatever method gives me the result I want to hear”.

    “I understand it’s a telephone public opinion poll, er, crime poll!”

    (a) No, it’s a face to face survey. I thought you knew about this stuff?
    (b) My god, they’re actually asking people what happened to them! You’re suggesting we should trust what Inspector Knacker has to say instead? I suppose instead of asking people what they think on any political issue we should just ask the politicians instead, eh?

    “Look at what the Research Development and Statistics Directorate (RDS) of the Home Office did with 800 offences in Sept ’01 fluctuating over to 800 offences in Mar ’03. They generated a bar graph that went from plus 50 to minus 10.”

    Yes, that shows year to year change. They also ‘generated’ a line graph showing an increase over the period. Both are informative and both are quite clearly visible. What’s your problem again? The RDS document you linked to in your first post sets out acres of firearms offences data in tables, so to suggest that they’re engaged in graphical sophistry is just plain dishonest. You’re the one twisting the facts, not them.

  • Jim

    Gcooper,

    “Because they are the result of a bloody opinion survey.”

    The best way to find out how much crime people are experiencing is to ask them. The worst way is to ask the police. Frankly I’m surprised Samizdatistas are so trusting of the cops, but then again I suppose it’s another case of believing whoever is telling you what you want to hear. The police change how they measure crime all the time – including gun crime, but apparently that’s gospel truth to you. Meanwhile a survey of 40,000 people which hasn’t changed its methodology since 1981 gets dismissed as sophistry. What a collection of cranks this place is.

    Midwesterner,

    “to present an appearance not in agreement with”

    Well yes, it attempted to present that appearance, but in fact it wasn’t in disagreement at all, as I explained. You’re just cherrypicking the one or two trends out of all violent crime that happen to support your case, and declaring that everything else is fraudulent. For example,

    “You can get a survey to say anything depending on how you write the questions. Not only that, it’s a very emotional type of crime, very subject to the feelings and interpretations of the victim.”

    I see. People are obviously just making stuff up about their experience of domestic violence in vast numbers, and if that’s not enough, the BCS researchers are manipulating their answers. It’s all a vast conspiracy to convince people that domestic violence is falling!

    Please cop the fuck on. The BCS has asked the questions consistently since 1981 – in direct contrast to police recording. The data shows a consistent drop since 1995. The BCS has also used a self-completion questionnaire to measure inter-personal violence – that also shows a drop over time. Those are the facts. Now, for some reason you’d like to ignore that – whatever, I don’t really care why. But please don’t lecture me about scientific method because you appear to think that means “whatever method gives me the result I want to hear”.

    “I understand it’s a telephone public opinion poll, er, crime poll!”

    (a) No, it’s a face to face survey. I thought you knew about this stuff?
    (b) My god, they’re actually asking people what happened to them! You’re suggesting we should trust what Inspector Knacker has to say instead? I suppose instead of asking people what they think on any political issue we should just ask the politicians instead, eh?

    “Look at what the Research Development and Statistics Directorate (RDS) of the Home Office did with 800 offences in Sept ’01 fluctuating over to 800 offences in Mar ’03. They generated a bar graph that went from plus 50 to minus 10.”

    Yes, that shows year to year change. They also ‘generated’ a line graph showing an increase over the period. Both are informative and both are quite clearly visible. What’s your problem again? The RDS document you linked to in your first post sets out acres of firearms offences data in tables, so to suggest that they’re engaged in graphical sophistry is just plain dishonest. You’re the one twisting the facts, not them.

  • Jim

    Oops, excuse the double-post.

  • Karl Rove

    Dim Jim

    The police come under the Home Office. The BCS comes under the Home Office.

    You’re wrong. It’s a phone poll. It’s doubtful if they really use 50k (as they now claim) – a 1000 face to face is just as accurate. And they don’t ask the obvious question – why is most crime not reported?

  • Jim

    Karl Rove,

    “You’re wrong. It’s a phone poll.”

    No, you’re wrong.

    Here’s the relevant passage from Crime in England and Wales 2004/05:
    “BCS METHODOLOGY
    BMRB Social Research carried out 45,120 face-to-face interviews with adults aged 16 or over living in private households in England and Wales. Interviews took place between April 2004 and March 2005. The sample was designed to be representative of private households, and of adults aged 16 and over living in private households.”

    Who’s dim now? Idiot.

  • Jim

    Oh and one more thing, “Karl”.

    “And they don’t ask the obvious question – why is most crime not reported?”

    Section 3.3. of the same document, Crime in England and Wales 2004/05:

    “Victims of crime were asked why they did not report incidents to the police (Table 3.03).
    • As in previous years, the most frequently mentioned reason for not reporting incidents was
    that victims perceived them to be too trivial, there was no loss or they believed that the police
    would or could not do much about them (71% of incidents). In one in five cases (20%), the
    victim felt the incident was a private matter to be dealt with themselves.
    For violent crime, however, a large proportion of the incidents were not reported because
    victims considered the issue to be a private matter and dealt with it themselves (41%). This
    may reflect the personal relationships involved with some of these offences …”

    So that’s two absolute howlers you managed to pack into a four-line comment. You have got to be one of the stupidest people I’ve ever encountered on Samizdata, and that’s saying something.

  • Midwesterner

    Jim, your faith in the benevolent political statisticians correcting those evil lying policemen’s records is touching.

    In truth, none can be trusted but I think this time the police records are the least tamperable. I explained in detail why, when faced with clearly contradictory appearances from two functionaries, I chose the one I did.

    As to your blowing off my example, no. One is raw data, one is ‘interpreted’ data. Raw, chronological, zero baseline data will always look similar in a graph. Selecting some particular hidden feature to be discovered and presented with the aid of a mathmatical formula, is another thing entirely. Especially when it generates such a lovely downward trend where no downward trend in fact exists in the actual numbers of crimes.

    And you didn’t mention their selectivety in refering to the police data, but picking out the only one of sixteen possibilities that went down. More correcting of artificial impressions from the police, no doubt.

    And, of course, when they selected those 45,120 people representive of all of England & Wales, there was no opportunity there. The millions represented in the police data is such a small sample.

    Jim, NONE can be trusted. I’ve layed out clearly why I chose the data from the police this time. Your faith in the BCS is your own, not mine.

  • Midwesterner

    A comment on Karl’s and Jim’s debate over survey method.

    My opinion is that, if a survey must be resorted to, a blind written survey is best. One that the actual survey as it was mailed out is also included with the data.

    A telephone survey is worse, offering the surveyor an opportunity to influence the respondant’s choices.

    A face to face, in my opinion, is the very worst. It is as un-double blind as you can get. No chance of the political hack conducting the interviews to influence the answers, is there?

    Just another opinion of mine. Your’s may be different, results may vary, some settling of contents may occur, etc.

  • SD

    I’m completely with Jim on this. Obviously no source is completely reliable but the BCS is much more accurate than official Home Office statistics. This isn’t because of fiddling or malfeasance, it simply reflects the nature of criminal statistics. The great problem with all figures of crime reported to or known to the police or criminal justice system more generally is that of the ‘dark figure’, the number of crimes that happen but are never reported and therefore don’t appear in the record. Unless you have some idea of what the reporting rate is statistics of this kind tell you very little about the underlying real rate of crime or the trend. An apparent increase or decline may be caused by a change in the reporting rate, a shift in the use of resources or even an increase in police efficiency. That is why the Thatcher government set up the BCS in 1981. What it did initially was to show that the increase in crime since the 1950s was real and not imaginary (as many academics had been claiming). It also revealed wide variations and changes in reporting rates. When it started the reporting rate for burglary was low, now it’s high, because of insurance requirements. The reporting rate for violent crime, particularly assault, is low, for the reasons Jim cites. The methodology of the BCS is about as robust as you can get for a survey of this kind (one aspect of this is that a face-to-face survey is vastly more reliable than either telephone or postal surveys). It’s done on the same basis as the census but doesn’t have the same non-compliance problem because it doesn’t have to be universal, just representative.

    It helps to put the current state of affairs into historical perspective. Recorded crime figures show an increasing trend from the 1760s up to the 1840s (when per capita rates reached a level that would not be passed until the 1980s). After 1850 there’s a decline which accelerates markedly after 1870, reaching a low point in the 1900s. Recorded crime rates then remain stable and low for most of the first half of the 20th century. After the mid 1950s they start an upward trend, which becomes much steeper after the later 1970s and reach a peak in the early 1990s. Since then there’s been a steady decline in crime generally but an increase in violent crime (which is actually a small part of total crime). What the BCS does is confirm both the rise after 1981 and the decline since 1995 (the police figures understate both). This pattern, of a fall in crime since the early 1990s, is found in every OECD country but Britain is unusual in having the pattern of a decline in property crime such as burglary (a big drop) and a rise in violent crime.

    Why these large shifts in behaviour take place is ultimately mysterious. It certainly can’t be down to public policy because the decline takes place at much the same rate in jurisdictions with very different policies.

  • I think SD is correct and the BCS is actually one of the more reasonable surveys of this kind. I also agree that one has to be very careful about drawing too many ‘obvious’ conclusions from the data and Howard’s remark that declines are down to CCTV is bizarre.

    For what it is worth, some years ago a fairly senior policeman with whom I was acquainted put it to me that the significant decline in burglary had nothing to do with CCTV or detection rates (which were actually declining) or convictions (ditto) but rather that as items like computers, DVD players and CD players had now become so inexpensive compared to steadily rising national incomes that even in quite ‘deprived’ areas, the ‘economics of crime’ simply made that sort of offence uneconomical. Why buy a stolen DVD player from some thief when you can get a new one is certain to actually work for the relatively trivial sum of £100?

    Make of that what you will.

  • D Anghelone

    SD: “Why these large shifts in behaviour take place is ultimately mysterious.”

    One factor not mysterious is the percentage of young people in an area. Higher percentage equals more crime.

    SD: “What it did initially was to show that the increase in crime since the 1950s was real and not imaginary (as many academics had been claiming).”

    Quite in line with my above contention. Mine and Billy Jeff Clinton’s, that is. His info surely came from the FBI or DOJ.

  • Johnathan Pearce

    Jim, I think Midwesterner’s focus on the homicide crime rate is particularly useful because as he says it deals with some of the distortions of data one gets with surveys and police reporting. I have no dog in this fight: if crime is down and can be proven to be so, then excellent news and bully for you.

  • SD

    Homicide is an interesting crime for long term trend indication because the reporting rate is fairly stable (criminologists estimate about 1 in 3 is reported or discovered). It doesn’t generally show the dramatic rises and declines that you find with many other kinds of crime. In the second half of the nineteenth century in the UK homicide figures show a steady but slow decline (per cap of course) while other kinds of crime, notably theft see a much steeper fall (apart from burglary which actually went up). Since 1950 there has been a rise but it’s much less dramatic than the increases in such offences as theft, robbery, and assault. It is pretty stable now and has been for some time, with some indication of a slow decline setting in.

    The big declines since 1995 in the UK have been in burglary, theft of and from cars, criminal damage, and theft in general (larceny as it’s actually called).

  • American crime is mostly down and has been for the last six years or seven years. Since we started three strikes laws. Crimes are dropping, and are at a level equal to the 1960s per capita.

    We actually have less crime in our cities now than European cities.

    You can thank the pols in the UK with their stupid gun control laws, plus the laws against self defense, for the rampant crime there. If they allowed everyone who wants to to have a gun and be able to defend their homes, crime would drop overnight.

  • ajay

    Although it’s true that murder is rather more easily definable than other offences, one should still be cautious about pointing at a decline in murder and attributing it to any policy or demographic change. Don’t forget that emergency medicine has come on in leaps and bounds over the last few years – a stabbing which was murder in 1970 could be only GBH today, thanks to better treatment.
    This also applies to using murder as a proxy for other crimes, for the reasons cited above – even if overall crime or violent crime were constant, one would expect to see a steady decline in murder rates. One would also expect an increase in the percentage (if not the absolute number) of murders committed with more immediately lethal weapons, such as firearms.