I’m now ensconced in my new house in Surry Hills, yet oddly I find myself spending my spare time wading through, of all things, the 2010 SAPOL report on Alcohol and Crime, focused on Hindley Street, which I’ve been looking at because I’ve been thinking a lot lately about how policy makers and agencies compile data to inform particular regulatory methodologies, and how the methodologies themselves impact on the data.
That sounds pretty dull. But, in more blunt terms, I’ve been thinking about the old adage of ‘Lies, damn lies and statistics’, and about how the culture of an organisation might influence the way they choose how they study a problem, and what parts of it they focus.
The SAPOL report is interesting in that it notes:
[…] the fundamental issue for the community of South Australia is whether or not the status quo should be allowed to prevail in the Adelaide CBD. If the status quo is unacceptable, then what is required is fundamental reform designed to reduce the social and economic costs of alcohol misuse in the night-time economy of the Adelaide CBD.
Which makes sense and I agree with the sentiment, it just seems like their response is “We should address the status quo of violence on Hindley Street by increasing the status quo of regulatory and policing action.”
The thesis of the SAPOL report is that decreased trading hours, and a decrease in the total number of venues selling booze, will solve the problem, which is the same regulatory logic and we’ve been employing it since the Gin Acts. Indeed, as Shane Homan underlines in his amazing book The Mayor’s A Square, it’s the same logic that drove Six O’Clock Closing. The problem, according to Homan, is that Six O’Clock Closing didn’t actually reduce the amount of alcohol being consumed. It just packed it into a smaller volume of time, and warped the social rituals that usually surround alcohol consumption.
The same argument could be made about Hindley Street. By restricting opening hours and putting in place a licensing system that is insurmountable for smaller venues, Adelaide hasn’t become safer. Since the 1997 Liquor Licensing Act I think – and SAPOL seems to agree – it’s gotten worse. Arguably, that’s because instead of dealing with the social rituals that surround alcohol consumption – and booze has always been a social drug – the status quo logic assumes alcohol is the problem, and if people are denied access to it they’ll behave better.
It’s a comfortingly simple logic of cause and effect. You could write it up mathematically like this:
People + Alcohol = Trouble
Thus the answer is:
People –Alcohol = No Trouble.
If Six O’Clock Closing had worked, I’d believe it, but it didn’t so I started wondering what else you could add to the equation that might suggest an actual, rather than rhetorical, diversion from the status quo. So I started looking for other indicators that corresponded to violent and anti-social behaviour.
And as luck would have it, I found one.
I have found a single statistical feature, easily measured, that is present in almost 80% of incidences of violence. Using the same logic as the SAPOL report, I am pretty sure that if we shift our focus to this feature, we can reduce violence on Hindley Street by a minimum of 78% and potentially more than 80%.
What is this statistical marvel?
Before revealing it, let me explain my logic.
As I’ve said, in reading the Alcohol and Violence report, it’s notable that they only measure two main themes: (1) Alcohol present in incidents of violent and anti-social behaviour, and (2) statistics related to the number of sources through which alcohol could be purchased.
Accordingly, they note that 58% of victim-reported crime was alcohol related, and that the portion of offenses against the person in the CBD involving alcohol is 62% compared to 41% more generally.
That sounds like logical support for the connection between alcohol and violence, but from a research perspective it means that 42% of victim-reported crime isn’t alcohol related. And, whilst an increase of around ten percent between the CBD and suburbs is notable, that could just as easily be written off based on the higher concentration of people in the city compared to say, Findon.
As the SAPOL report notes, there’s was a 22% increase in people working in the city from 89,000 to 108,000 between 2001 and 2006 and the student population went up from 50,000 in 2000 to 81,000 in 2007. If there’s more people in a smaller area of land, on a purely mathematical level, there’s more chance of incidents regardless of whether alcohol is present or not.
So that got me wondering. Of that 58%, how much was caused by alcohol compared to how frequently was it simply ‘related’? And what’s related? A glass of wine two hours earlier? Or they had the beer in their hand whilst they were kicking someone?
Then I began to wonder if there was a single statistical indicator that had a higher appearance in incidence of violence than alcohol’s 58%.
And there is. Indeed, there are numerous statistical features more prevalent in incidences of violence than alcohol. For example, something like 60% of people have brown hair, so if you only measured ‘incidence of violence’ and ‘number of people with brown hair responsible for violence’, with no reference to other indicators, you could prove roughly the same thing as SAPOL’s done in their 2010 report in relation to alcohol. You could show that 60% of incidences of violence are produced by people with brown hair.
Does that prove people with brown hair or more violent? Well. No. If you took the total number of people with brown hair, the proportion of them involved in incidences of violence might be statistically the same as those with any other colour of hair. Which would indicate that brown hair might be present, but it might not be the cause, or at least not the principle cause.
Similarly, if you were to measure the total number of people with alcohol in their system at any one time, and then measure their rate of offense in relation to, say, teetotallers or people who’d spent the afternoon drinking coffee, you could then compare the rate of offending. If drunks turned out to be more violent than tea drinkers you could say there was a link. But as it is, that data isn’t available. Depending on how the data is gathered, it could just be that 58% of the population has alcohol in their system at any given point in time and that they offend at the same rate as anyone else. That’s unlikely, but it could be the case.
Either way, the 58% statistic didn’t entirely convince me, hence I kept looking for other statistical evidence of factors present in incidences of violence. Particularly I was interested in whether other cultural or social factors were also present, because it seemed like they might be as influential, if not more so, in causing violent and anti-social behaviour.
SAPOL actually seems to agree with this methodology up to a point, declaring:
These alcohol-related problems in the Adelaide CBD have their roots in a range of complex cultural, environmental and legislative issues.
Except then they don’t actually measure those ‘complex cultural, environmental and legislative issues’. All they do is declare:
There are some clear potential steps forward. The national and international evidence concerning the harmful impacts of extended trading hours of alcohol sales is overwhelming and irrefutable.
Again, this would make sense if (a) 42% of violent incidents hadn’t occurred without the presence of alcohol (b) they listed their evidence or (c) we hadn’t already tried this approach before, both in relation to Hindley Street since the current Act came in 1997, and from 1919 to 1967 with the Six O’Clock Swill. As an indicator of South Australia’s reputed reputation for innovation, we were the first state to introduce the Swill and the last to admit it didn’t work. It was repealed under Dunstan.
In support of my enthusiasm for re-thinking the status quo of licensing policy, there was a report back in 2005 conducted by Allen’s Consulting Group, entitled Assessment of the Impact of Liquor Licensing Reforms, which found that alcohol consumption hadn’t changed much at all between 1997 and 2005, concluding that there was a “growing consensus that many entrenched liquor regulations do little to curb the alcohol-related harm for which they were designed to do.” That backs up Homan’s argument that the Swill just compacted drinking into a shorter period of time, rather than actually reducing the amount being drunk.
Allen’s Consulting go on to cite an Australian Economic Review report which found:
The most basic problem with existing regulations concerning [alcohol] consumption-related problems is that they are not closely geared to the true nature of the problems.
When an instrument cannot be targeted closely to the persons giving rise to some social problem it becomes less effective and may even cause a reduction in the welfare of society.
And, following that, a report from the Victorian State Government in 1998, reviewing changes to their licensing act, noted that the “period in which the number of licenses has increased very substantially, per capita consumption has not increased and, indeed, has probably undergone a small decline.”
Similarly, the Allen’s report reviews ABS data from 1990 to 2001, particularly the National Health Survey, which found the incidence of ‘high risk’ alcohol consumption behaviour was at the same level in 2001 as was recorded in 1990, at an average of 10.8%. I thought that was interesting, because if you could show a shift in rates of violence connecting to an increase in alcohol consumption that might better prove a casual link between alcohol and violence.
And there is some data in that respect through the 2005 NHS summary of results which indicates alcohol consumption had gone up between 1995 and 2005, rising from an average 8.2% percent of the population who had ‘risky’ levels of alcohol consumption to 13.4% in 2005. On the upside, the Apparent Alcohol Consumption Dataset offered by the ABS suggests it’d started going down again, and by 2010 it was down by 1.1% in comparison to 2001.
Does that connect to violence? Well, the Crime Victimisation Datasets indicate the rate of assaults in 2003 was 16,006 according to the ABS, in 2011 it was 16,234, which isn’t a huge increase. If I’ve done my maths right it’s about a 1.4% increase, no where near the 5.2 increase in ‘risky’ levels of consumption. If the relationship between alcohol consumption and violence was easily defined, then you’d probably expect the crime rates to have dropped in relation to the drop in ‘risky’ levels of alcohol consumption. I’m not a demographer, but it looks kind of ambiguous.
So, again, it seems like we need a better array of indicators; something that offers a wider scope than simply “Incidences of Violence” measured against “Incidences of Violence in Which Alcohol is Related” because that system doesn’t show the cause of violence very well. It just shows one factor.
In specific relation to Hindley Street, Table 7 of the 2012 Crime Victimisation Dataset throws another spanner in the works, listing the most common place by far to be a victim of physical assault, (with 143,100 incidences reported) as one’s own home (with 58.1% of incidences involving alcohol). Second is at work, with 68,500 (48.5%) and Entertainment venues produce 63,900. So a night out on the town is actually less risky than staying at home. Okay.
Regardless, I plodded through the alcohol and crime stats waiting to see if some other indicator consistently came up in incidences of violence with a rate of higher than the 58% SAPOL lists for incidents involving alcohol.
And then I found it. The Holy Grail. There is a single indicator that is undeniably and overwhelmingly present in incidences of violence.
I was looking at the 2011 Recorded Crimes Dataset when I discovered it. A whopping 78% percent of offenses were committed by men. Masculinity is a far greater factor in incidences of violence than alcohol.
The Gender Indicators report from 2012 reinforces the same point, concluding:
In 2010-11, men were up to one-and-a-half times more likely be the victims of physical or threatened assault or robbery than women. Men that experienced physical or face-to-face threatened assault were less likely to tell the police about the incident than were women.
Males were nearly four times more likely to commit offences intended to cause injury, more than six times more likely to commit robbery and more than 28 times more likely to commit sexual assault. However, the male victimisation rate for sexual assault was about one-sixth that of females in 2011.
That means that, as a man – either sober or drunk – I am statistically highly likely to both beat someone up and get beaten up. My biggest statistical threat isn’t the presence of alcohol. It’s the presence of my own gender.
If we apply the same limited set of indicators contained within the SAPOL report, there’s a much, much greater onus of proof that we need harsh regulatory systems in place surrounding masculinity than there is in relation to booze. Indeed, if we simply alter the equation SAPOL use:
People + Booze = Violence
And change it to:
People + Masculinity = Violence
We can produce regulations that much more effectively cut down on crime. Obviously, we can’t ban all men that easily. Personally, I know I’d raise objections. But, having recognised I’m my own biggest threat, I think I’d have to concede that based on the strength of that data, there’s greater grounds to license masculinity, restrict the hours in which its available, and to not provide masculinity to anyone who is already male or under the age of 18 then there is to do the same things with alcohol.
I’m being facetious. But the point is the same. Taken as a single indicator, there is greater proof that masculinity is a contributing factor to violence than there is for alcohol. I’m being facetious because it’s pretty obvious that masculinity alone doesn’t produce violence. Maybe I’m being naïve, but most men aren’t violent thugs and simply banning or regulating the entire gender fails to address the ‘complex cultural, environmental and legislative issues’ that surround male violence; the cultural glorification of violence, male mental health issues, alienation of labour, work/life balance, or any of hundreds of potential contributing factors.
If we wouldn’t make that simplification with a higher indicator of violence like masculinity, I’m wondering why we think it is okay to so with alcohol consumption? After all, we’ve tried treating the link between alcohol and violence (and I do believe there is one) out of cultural context for centuries and it hasn’t’ work so, to paraphrase SAPOL, maybe it’s time we questioned not only the status quo of CBD drinking culture but the status quo behind the methodologies that have tried, failed, and continue to fail, in altering drinking culture, as distinct from simply limiting the supply of alcohol.