Is it liquor outlets or urban density?
May 6th, 2012 at 10:49 am by David FarrarIan Steward in the SST reports:
Rates of serious violent crime double within 900m of a liquor outlet, a new study has found.
And the nationwide study has confirmed that the more liquor stores an area has, the more likely it is to have a higher rate of serious violent crime, regardless of poverty and other factors.
Now I think there is an obvious linkage, as alcohol is a factor in some violent crime.
Study lead author Peter Day said the study ranked the country’s 286 police station areas into five groups (quintiles) according to their rate of serious violent crime.
“The number of alcohol outlets consistently increased with increasing quintiles of serious violent offence rates,” the study said.
A more detailed analysis was performed using census “mesh blocks” where the country was divided into 41,393 small blocks representing about 100 people in each.
Using mapping software researchers were able to work out the median travel distance to a liquor outlet. Areas with the lowest rates of serious violence had to travel a median distance of 4.5km to the nearest off-licence. For the highest rates of serious violence, the median distance to an off-licence was just 1.1km.
Using the mesh block analysis, crime rates were calculated for distances from liquor outlets. On average nationwide the incidence of serious violent crime doubled once you got within 900m of a liquor outlet.
I do wonder though how much of this effect, is just because people are in more dense urban areas. As a comparison, what if you compared violent crime incidence to the distance to the nearest hairdresser? Would you also find there is more violence crime close to hairdressers?
Alac chief executive Gerard Vaughan said the current law did not take into account how many outlets were in an area when granting licences.
“If you want to set up a pub or a restaurant you need to get resource consent and demonstrate you’re a person of good character. There’s no consideration around `does this community need another licence?’,” he said.
The Alcohol Reform Bill will give local authorities the ability to set a local alcohol policy. I can understand the desire not to have a bottle store at every corner. However I hope that restaurants would not be declined on the basis of pre-existing restaurants.
The three spikes with the highest numbers of liquor outlets were Auckland central (447 alcohol licences), Wellington central (423) and Christchurch central (394), all of which had high crime rates.
Again, how much is this due to urban density?
May 6th, 2012 at 10:53 am
Indeed David, using mesh analysis I am sure you could conclude a link between crime and the availability of police stations.
It is the availability of police stations driving the crime
Vote:May 6th, 2012 at 10:57 am
I should hope that the study compares like with like, ie similar neighbourhoods where the difference is the number of liquor outlets. To do that, one would presumably look at suburban mesh blocks of similar income etc, but varied by number of liquor outlets.
Vote:May 6th, 2012 at 10:57 am
[...] Christchurch central (394), all of which had high crime rates
Obviously dated figures.
Vote:May 6th, 2012 at 11:01 am
I see from the study’s abstract that it was ‘controlling for area-level measures of social deprivation, Māori population, young males 15–29 years and population density’. So it would appear to be a valid conclusion and urban density doesn’t really enter into it.
Vote:May 6th, 2012 at 11:02 am
Oh come on – we need more liquor outlets like we need another hole in the head.
In Hamilton many dairies have closed and they have reopened as liquor outlets. Forget the studies – they should close half of them – or else put the cost of alcohol as seen in hospitals and thru the health system back on the manufacturers, importers and retailers.
That might get their attention.
Vote:May 6th, 2012 at 11:34 am
Where are the statistics? All of these reports should be regarded as complete crap until you can examine the source data and analysis.
I live within 900m of heaps of liquor outlets. Amount of violent crime = zero.
Vote:May 6th, 2012 at 11:36 am
A post where David Farrar studiously examines every corner of the room for the source of the elephant droppings whilst scrupulously avoiding noting the elephant in the middle of it…
[DPF: And you ignore where I said "Now I think there is an obvious linkage, as alcohol is a factor in some violent crime.".
It must be difficult for you, to never think research should be challenged]
Vote:May 6th, 2012 at 11:39 am
Oh look, he’s got form using this perverse “logic” to call for stuff to be banned:
“Call to stop junk food sales around schools…Peter Day from Canterbury University said junk food outlets are five and a half times more likely to be found within easy walking distance of a school than any other area.”
http://tvnz.co.nz/health-news/call-stop-junk-food-sales-around-schools-4007539
Vote:May 6th, 2012 at 11:40 am
Yes, fish. I mean, it took about 30 seconds via google to find out that the study controlled for urban density.
Vote:May 6th, 2012 at 11:41 am
fish_boy, you mean crims and thugs like to booze? I doubt it’s so much population density as the density of the local population.
Vote:May 6th, 2012 at 11:41 am
I see from the study’s abstract that it was ‘controlling for area-level measures of social deprivation, Māori population, young males 15–29 years and population density’.
The fond belief of social scientists that their correlation=causation errors can’t be errors because they’ve “controlled” for a few obvious factors is a major reason people tend to roll their eyes every time they read “a new study shows…”
I used to live in the central city and it had violent crime, riotous parties, obnoxious stereos, constant burglaries, run-down houses, lots of unemployed people and students, cars with no WoF or reg (mine, for instance) and, oh yes, also a couple of bottle stores within easy walking distance. These days I live in a nice leafy suburb which has none of those things except the occasional burglary – and yet somehow it never occurred to me to credit the absence of a bottle store within easy walking distance for the absence of all the other things. Oh, how the scales have fallen from my eyes!
Vote:May 6th, 2012 at 11:43 am
I would like to see comparisons between various suburban areas in one city, like say Ilam (which is dry and of high Student area intensity) to a nearby area which is not (like say Riccarton next door, also high student intensity).
Another valid context, would be areas of drinkability in a given mesh, ie beaches, parks, restaurants and night clubs, strip joints, etc compared with just suburban residential streets, as these the former would have exponentially higher alcohol crime incidents. So,, a mesh with more parks and restaurants is going to have higher alcohol crime than a residential-only sleepy suburb.
Vote:May 6th, 2012 at 11:45 am
This is an unbelievable load of rubbish. Why? Simple : “Correlation does not imply causation”. Sure, there might be some correlation between violent crime and liquor outlets, but so frigging what. Don’t they teach people this basic fact in NZ Universities?
I can understand simple minded journalists finding it hard to grasp basic tenets of common sense, but surely Universities are meant to know better. Apparently not.
Idiots.
Vote:May 6th, 2012 at 11:45 am
If alcohol-related crime suddenly increases in a suburb next door to an area that has newly gone ‘dry,’ that would be solid measurable evidence that availability of alcohol in retail contexts lifts alcohol-crime in that area.
Vote:May 6th, 2012 at 11:49 am
“My density has brought me to you”
- George McFly, Back To The Future
Vote:May 6th, 2012 at 11:50 am
It looks like that the analytic model they used is not dynamic but static. Well, I wound’t trust it. As DPF said , what if one compared violent crime incidence to the distance to the nearest hairdresser? I have no doubt that it would shows up the same result, ie, violent crime rate is highly correlated to densities of hairdressers.
DPF, I believe that urban density has got something to do with it. Perhaps researchers can try something dynamic as Square-lattice Ising Model which has been applied in other domains to explain emergence behavior in complex system. Economics Nobel Laureate Thomas Schelling applied the Ising Model to the problem of urban segregation in the 1970s and was able to explain the dynamics of segregation. Static model can’t reproduce such segregation dynamics. An animation of Schelling model is here and animation of Ising is here.
The study smells of nanny state to me. Besides that it is too simplistic. Mesh block analysis hasn’t caught up with advances in dynamic or agent-based modeling (which is far more realistic in terms of modeling a dynamic/complex system) than static modeling.
Vote:May 6th, 2012 at 11:50 am
beautox, most university science and engineering departments know better but many ideological ghettos especially in the social and environmental “sciences” do not.
Vote:May 6th, 2012 at 11:56 am
Oh look, he’s got form using this perverse “logic” to call for stuff to be banned:
Yes, and it’s the same face-palm idiocy. Junk food outlets very likely to be within easy walking distance of a school – er, yes… and where are schools found? Within easy walking distance of where people fucking live, you colossal tosspot. Every considered the possibility of two correlated factors both arising from a third?
Vote:May 6th, 2012 at 11:57 am
A nice alternative headline here: http://www.statschat.org.nz/2012/04/08/2242/
“Statistical crimes double near liquor stores”
Vote:May 6th, 2012 at 12:13 pm
So you are saying Len Brown is completely wrong in the spatial plan, David. Why not just say it?
Vote:May 6th, 2012 at 12:32 pm
I call on John Key’s government to chop funding to all Sociology Departments at every University in the country. The other option is to chop them completely from existence. Whatever remaining useful courses left from there are to be folded back into the Psychology Department and also the Statistics Department. As far as I see there is no special skills needed for Sociology because once they start doing data analysis, it is simply statistics, where statisticians are well suited for that role. Do statisticians need prior knowledge in Sociology in order to do such data analysis? Nope! One can hire a bunch of 3rd year statistics students to do the same analysis with no prior knowledge of Sociology at all. The behavioral studies in sociology can be pretty much covered in psychology, so it is logical to fold Sociology Departments into Statistics and Psychology.
There is no justification for the existence of Sociology Departments at our Universities. They are no different to TV, Film & Media Studies Departments in teaching bullsh*t courses such as pornography, gay theory, lesbian theory, etc… I say chop all Sociology Departments in the country to save money to be re-directed to other useful R&Ds in other departments.
Vote:May 6th, 2012 at 12:41 pm
From the report – onlinelibrary.wiley.com/doi/10.1111/j.1753-6405.2012.00827.x/full
“A number of area-level socio-demographic predictors of violent crime were also controlled for in the models. These predictors have been shown to be potential area-level predictors of crime in previous studies in New Zealand and include social deprivation,39 the proportion of Māori population,40 the proportion of young males 15–29 years and population density”
and
“To control for important population-level influences (confounders) on serious violent crime, the proportion of the total population who were Māori (15+ years), young males (15–29 years), population density (usual resident population per km2), and the NZ Deprivation index (NZDep2006) scores were derived for each police station area by aggregating meshblock level data. The Māori and young male population measures for each area were then categorised into quartiles, population density into tertiles, and NZDep2006 categorised into quintiles.”
and
“This study has limitations. First, given the ecological study design the association between alcohol outlet access and violent crime is strongly related to geographical scale. A different geographical scale may result in changes in the degree of association between alcohol outlet access and violent offending. The extent of causal inference one can conclude from the results is limited given the ecological study design. Second, alcohol outlet size and licence operating hours are important factors in the availability of alcohol and are strongly linked to alcohol-related harms however the relative size of premises and licence hours were not included in the available data for our analysis. Third, a number of confounders were not adjusted for in the analysis as data were not available including the level of policing which may influence reported crime, the level of alcohol sales and alcohol consumption. Fourth, we have used recorded violent offences data regardless of whether or not these were alcohol related as data linking alcohol involvement with individual offences were not available. However, given the high degree of alcohol involvement associated with violent offending, the data are likely to be highly indicative of alcohol-related violence. Finally, we have not included health outcomes such as Emergency Department and hospital admissions that result as a consequence of violent crime. The numbers of area-level violent offences are suggestive only of the geographic variation in the burden of alcohol-related harm on the health care system and health inequalities.”
There are fair few fudge factors and assumptions going in to bringing up these numbers
.
And if I understand their conclusion correctly there is approximately a 50% ‘adjusted’ greater occurence of violent crime depending on access to booze outlets.
Considering the limitations of their approach they are careful to class this as an association.
@Alan – yes statistically poorly done
Vote:May 6th, 2012 at 12:54 pm
According to Peter Day, Karori should be a real hotbed of violence.
Of course his comment:
Could equally be interpreted as, “perhaps violent crims like to have a beer after giving someone the bash”
What came first, the booze or the fist?
Vote:May 6th, 2012 at 1:22 pm
Perhaps most notorious was Michael Mann, creator of the Hockey Stick chart, who confessed “I am not a statistician.”
I wonder if this Peter Day could get a job as a climate scientist. He certainly seems to have the right qualifications.
Vote:May 6th, 2012 at 1:52 pm
wat dabney, you are on to it. His next paper will be on how global warming causes liquor store outlets to increase and thereby the violent crime rate. A funding bonanza awaits his team.
Vote:May 6th, 2012 at 3:46 pm
The upside of the equation, is of course, that all these areas have a much lower per capita rate of accident due to drunken driving as the pissheads can travel far less distance from their place of consumption to their place of supply and back again.
More intensive placement of alcohol outlets on the highways and byways of say Ekatahuna should reduce those “rural folks dying on rural roads” stats that the authorities keep boring us with between the useful Harvey Norman adverts.
Vote:May 6th, 2012 at 6:13 pm
Did anyone ask for the opinion of the Doctor of Bogans?
Vote:May 6th, 2012 at 6:16 pm
bla bla bla what you do is get ready to die when you are ready, and then you do it,
Vote:May 6th, 2012 at 11:30 pm
barry said
Oh come on – we need more liquor outlets like we need another hole in the head.
In Hamilton many dairies have closed and they have reopened as liquor outlets. Forget the studies – they should close half of them – or else put the cost of alcohol as seen in hospitals and thru the health system back on the manufacturers, importers and retailers.
That might get their attention.
I’m with Barry on this. This country has got an alcohol problem and is in denial about it. The study may be flawed, but it does control for population density as others have pointed out, and I think its conclusions are not completely without value. They also mesh with conclusions formed from my own long-term personal experience – a study entitled “real life” which alas has to date received no funding
Regards
Vote:Peter J
see http://www.sensiblesentencing.org.nz