Send email for updates
Palk G.R.M., Davey J.D., Freeman J.E.
Police Practice and Research: 2010, 11(1), p. 5–15.
Unable to obtain a copy by clicking title? Try asking the author for a reprint by adapting this prepared e-mail or by writing to Dr Palk at firstname.lastname@example.org.
Australia has been trying a novel way to curb alcohol-related disorder – banning late-night drinking venues admitting customers during the final few opening hours. The aim is to prevent disturbance-generating movements between bars. In one very distinctive area it may have worked, but in others the evidence is weak.
Summary Recently several Australian jurisdictions have tried to control alcohol-related violence and disorder by preventing patrons from entering or re-entering late-night on-licensed premises for a specific period prior to their closure, though customers who remain inside may continue to drink until closing time. Referred to as a 'lockout', this tactic aims to prevent altercations associated with patrons travelling between late-night drinking venues. Lockouts do not target drinking as such or even where drink is consumed, but aim during lockout hours to eliminate one of the mechanisms which help convert intoxication in to violence and disorder.
On 1 April 2004 the Gold Coast City Council which administers a major tourism and night-time entertainment area in the state of Queensland in Australia implemented a lockout policy for all nightclubs trading after 3am, meaning none of these premises could admit new or returning customers from then until the 5am closing time. However, customers could remain in the clubs where they were at 3am and continue drinking.
Coincidentally, in the same area from the 1st to the 28th March 2004 a pilot study was conducted immediately preceding the lockout, in preparation for the substantive study which took place roughly over the five weeks (from midnight 29 March to midnight 2 May 2004) immediately after the lockout was implemented. Both involved 'first response' police officers recording details of their attendance at incidents in the area, including where each incident occurred and whether they believed it was alcohol-related. The pilot, intended to enable officers to become proficient in the recording procedure, provided a pre-lockout baseline of the frequency of alcohol-related incidents requiring police attendance, against which could be benchmarked the frequency noted in the substantive study after the lockout had been put in place. Differences in the time periods and perhaps too in recording procedures led the researchers to focus not on the absolute number of incidents, but on the proportions thought alcohol-related and/or which occurred in and around licensed premises. The assumption was that if these proportions fell it would be indicative of the intended crime/disorder reduction impact of the lockout.
Of all recorded incidents whether or not alcohol-related, before the lockout policy 4.6% were in and around licensed premises. After the lockout this fell significantly to 3.4%.
Regardless of where they had occurred, before the lockout alcohol-related incidents accounted for 26% of the total. After the lockout this fell significantly to 23%, and the greatest reductions were in the late evening and early morning, and in particular during the two hours (3am to 5am) when the lockout was in force. Before the lockout policy had been implemented 51% of all incidents during these hours were alcohol-related; after the lockout the proportion fell significantly to 39%. There was no evidence that alcohol-related incidents had simply been displaced to other time periods.
Some categories of alcohol-related incidents were more affected than others. Before the lockout policy had been implemented 41% of disturbances/disputes within licensed premises This was the description given in a later paper but seemingly referring to the same data, the featured study called them "general" disturbances. were judged alcohol-related, afterwards, 34%. The corresponding proportions for street disturbances were 73% and 61% and for sexual offences 57% and 23%. However, the corresponding proportions for traffic offences and for stealing and property offences were virtually unchanged, and while the proportion of offences against the person judged alcohol-related fell by 11%, the drop did not meet the criterion for statistical significance.
Findings of the featured study suggest that lockout policies can reduce alcohol-related crime and other incidents requiring police attendance in and around the licensed premises affected by the lockout. After the lockout was implemented, alcohol-related incidents and incidents in and around licensed premises fell as proportions of the totals and particularly so during peak drinking hours and days. Disturbances and sexual incidents related to drinking fell significantly as proportions of the totals. These changes are likely to be attributable to the lockout, as trends in offence rates remained relatively stable during the study period, and levels of alcohol-related violence have remained stable for a lengthy period.
It is not possible to eliminate the possibility that the changes in recorded incidents noted by the study were influenced by other factors such as time of year, weather, police activities, and changes in the management of licensed premises, but none of these seem likely to have been major influences. It is however possible that as the study progressed officers became better at recording alcohol-related incidents.
commentary The type of lockout policy evaluated by the study seems to have been tried almost exclusively in Australia but is potentially applicable wherever a concentration of late-night drinking venues results in intoxicated customers moving between them, creating opportunities for disturbance-generating encounters; at the time of the study the Gold Coast venues affected by the lockout were concentrated in just two streets of the Surfers Paradise area. Clearly there is a risk that movements and incidents will be displaced to the hours before the lockout, but in the featured study this was not the case. Neither does it seem that altercations escalated due to would-be customers trying to gain entry during the lockout hours and being refused. Queensland police noted reduced assaults and calls related to public drunkenness which they attributed to the Gold Coast lockout, data cited in support of its extension to the city of Brisbane.
However, the strength of the evidence from this and other studies is weak. The researchers took advantage of a coincidence between their study period and the lockout, but this also means the study was not intended as an evaluation of the lockout policy. In particular it seems highly likely that (as intended) the pilot period led to improved or more extensive recording of incidents by the police. During the pilot 951 incidents were recorded per week; during the substantive study while the lockout was in operation the number increased to 1799. The number of alcohol-related incidents too increased from 246 a week to 406 and the number inside licensed premises rose from 34 a week to 48. This apparently worsening situation was transformed in to an apparent success for the lockout policy by converting the raw numbers in to proportions of the total number of such events whether alcohol-related or not. Effectively the authors assumed that changes in absolute numbers of events were due to recording practices, so if the proportion judged alcohol-related fell, this meant that in reality the number had too. Without this assumption they would have had to conclude that the lockout was counterproductive.
Reporting procedure changes could also have affected the proportion of incidents judged by the officers to be related to drinking or which occurred in and around licensed premises, raising the possibility that changes in these proportions were wrongly attributed to the lockout. However, the timing of the greatest changes in the relevant proportions, and the type of incidents most affected, all point to a specific impact on alcohol-related crime and disorder rather than a general change in reporting. For example, traffic offences might remain unaffected because the lockout probably mainly restricted pedestrian movements between nearby premises, and stealing and property offences are not as closely associated with intoxication as physical and sexual aggression.
The later paper from the same authors expanded on the Gold Coast results and added those from similar lockouts and similar studies in two nightlife areas in Brisbane. In those areas generally there were no statistically significant differences in alcohol-related incidents in general or during lockout hours. Among the factors which might account for the non-impact in those areas in contrast to the Gold Coast were more punctilious recording in Brisbane after the lockouts were implemented, the relative dispersion of late-night venues in Brisbane, and better preparation and enforcement infrastructure on the Gold Coast. The results do however at best suggest that the impact of lock-outs is highly dependent on the context, and at worst that the Gold Coast results were due to factors other than the lockout and/or were a methodological artefact. If, as the authors suggest, recording procedure changes might account for non-impact in Brisbane, so too might they account for the apparent impacts on the Gold Coast.
The same paper reported on interviews with staff at the affected venues, police, politicians and licensing law administrators. Their perceptions were that the lockouts had led to benefits but also some problems. Licensed premises staff felt it helped them exclude the most troublesome customers without affecting revenue. Some commentators identified the downsides of the pre-3am rush to switch venues including violence on the streets and long queues, and some initial disputes after 3am as door staff refused entry.
Another research team has for the first time investigated the long-term impact of a lockout on emergency department attendances. Alcohol-related attendances (drunkenness and assault) seen at peak drinking times at local emergency departments from 1999 to 2009 in the Australian city of Ballarat (where a lockout was introduced in 2003) were compared with another Australian area without a lockout policy. The results suggested at best a short-term impact, though even this might have been due to other factors, ending with a greater rate of incidents per 10,000 of the population in the lockout city.
A review Personal communication from Peter G Miller, August 2013. of lockout research including the featured study found the evidence largely inconclusive. The reviewers noted some apparently positive impacts on assault and sexual assault, but confidence in these was undermined by limited evaluation designs and especially by the difficulty of isolating the effect of the lockout from other simultaneous interventions, especially policing changes including a heightened presence, more patrols, and use of foot patrols.
Given these results it remains an open question whether lockouts are on balance an effective crime/disorder reduction tool, especially where late-night venues are dispersed rather than concentrated within easy walking distance.
Thanks for their comments on this entry in draft to Peter G. Miller of Deakin University in Australia. Commentators bear no responsibility for the text including the interpretations and any remaining errors.
Last revised 06 August 2013. First uploaded 30 July 2013
Give us your feedback on the site (two-minute survey)
Open Effectiveness Bank home page
Add your name to the mailing list to be alerted to new studies and other site updates
HOT TOPIC 2017 Controlling alcohol-related crime and disorder
DOCUMENT 2010 Alcohol in our lives: curbing the harm
DOCUMENT 2016 Modern Crime Prevention Strategy