J. DRUG EDUCATION, Vol. 37(4) 429-445, 2007
BARROOM AGGRESSION IN HOBOKEN, NEW JERSEY: DON’T BLAME THE BOUNCERS!
JAMES C. ROBERTS University of Scranton, Pennsylvania
ABSTRACT
Relying on a structured observation guide listing a large number of variables shown to be good predictors of aggression in bars by past researchers, trained observers spent a total of 444 hours collecting data in 25 licensed drinking establishments in Hoboken, New Jersey. Observations took place at two separate time periods, 7:30pm–10:30pm and 11:00pm–2:00am, on Thursday, Friday, and Saturday nights. Logistic regression analyses revealed the absence of bouncers and doormen as the strongest predictor of aggression in Hoboken barrooms. This is a particularly important finding considering the frequency with which bouncers and doormen are vilified in the available research on barroom aggression. Several practical prevention strategies aimed at reducing aggression in bars are proposed.
INTRODUCTION Research on the relationship between alcohol and aggression has traditionally focused on the characteristics of drinkers and the effects of alcohol with little consideration of the places or settings where alcohol-related aggression occurs. Only recently have researchers begun studying the characteristics of drinking settings and their contribution to incidents of naturally occurring alcohol-related aggression. Licensed drinking establishments are often at the center of this research as they are known to be high risk settings for aggressive incidents (Graham & Wells, 2001; Homel, Tomsen, & Thommeny, 1992; Stockwell, Lang, 429 Ó 2007, Baywood Publishing Co., Inc. doi: 10.2190/DE.37.4.f http://baywood.com
430 / ROBERTS
& Rydon, 1993). Research on barroom aggression is grounded in the belief that social contexts have a direct effect on aggressive incidents. Those studying aggression in bars do not discount various biological, psychological, psychopharmacological, or cultural explanations of alcohol-related aggression (see Graham, 1980; Graham et al., 1998; Graham, Schmidt, & Gillis, 1996; Graham, West, & Wells, 2000) but instead take for granted that individuals bring to drinking episodes a variety of unique individual characteristics, including unique propensities for aggression. Graham, La Rocque, Yetman, Ross, and Guistra (1980) conducted the first and most influential observational study of barroom aggression. Teams of observers spent three months and a total of 633 hours conducting unobtrusive observations in 185 licensed drinking establishments in Vancouver, British Columbia. Variables found to be correlated with aggression included state of intoxication, ventilation, noise level, theme, competitive activities, decorum expectations, staff attitude, and location of bar stations. Graham et al.’s (1980) study provided a model for future research in barrooms by identifying a number of potential predictors of alcohol-related aggression and providing several important insights regarding observational research in natural drinking settings. Homel et al. (1992) used insights from Graham et al.’s (1980) study of Vancouver barrooms as a starting point for their research in Sydney, Australia. Relying on detailed observation sheets listing a large number of variables to be observed, many of which were identified as predictors of aggression in the Vancouver study, teams of observers spent 300 hours conducting unobtrusive observations in 23 licensed drinking establishments known to be frequented by young people. An analysis of the data revealed groups of male strangers, low comfort, high boredom, high drunkenness, and aggressive and unreasonable bouncers and floor staff as the chief variables present during violent episodes. Like Graham et al. (1980), Homel et al. (1992) concluded that incidents of aggression were not necessarily the result of any single variable, but rather the subtle interaction in one time and place of several variables. Homel and Clark (1994) set out to reproduce findings from the 1989 Sydney study using more quantitative research methods. Teams of observers spent 300 hours conducting unobtrusive observations in 36 Sydney barrooms. As in the 1989 study, structured observation sheets guided the observations. An analysis of the data revealed the presence of Pacific Island bouncers, refusal of service, and overall server responsibility as the strongest predictors of physical violence. Male drunkenness and “drinking in rounds” were identified as significant predictors of non-physical aggression. Comparing findings from this study to those of the 1989 Sydney study, Homel and Clark (1994) state, “The results are on the whole consistent with the qualitative study, which suggested that drunkenness usually only leads to violence when other risk factors, such as aggressive bouncers or high levels of frustration due to a lack of food and comfortable seating, are also present” (p. 34).
BARROOM AGGRESSION AND BOUNCERS /
431
More recent observational studies of barroom aggression include Graham and Wells’ (2001) research in Ontario, Canada, which entailed 93 nights of observation in 12 bars known to be frequented by young people. Findings from this study support past research (Graham et al., 1980; Homel & Clark, 1994) linking permissive drinking environments to increased aggression in bars. Graham and Wells (2001) also suggest that “macho” drinking environments, such as those that feature violent sporting events on large-screen TVs, may set up expectations that aggressive behavior is expected and will be tolerated. Fox and Sobol (2000) provide one of the few observational studies of barroom aggression conducted in the United States. Their research entailed three months of structured observation in two bars located in a popular night recreation area of a medium sized northeastern industrial city. A major finding from this study was that a greater level of guardianship by the use of effective bouncers resulted in a lower amount of patron aggression. Interestingly, observers in Graham, Bernards, Osgood, Homel, and Purcell’s (2005) study of bars in Toronto, Canada found that staff members often became offenders when responding to incidents of patron aggression rather than serving in the preventive role of the guardian. The major aim of the present study was to use observational methodology to identify situational variables that predict aggression in bars so that practical prevention strategies aimed at reducing such incidents could be formulated. The present study was also an attempt to fill a void in the research, as few studies of barroom aggression have been conducted in the United States. METHODS Sample of Bars Hoboken, New Jersey served as an appropriate study site for several reasons, one of which being its high concentration of bars catering to young, college-aged patrons. In keeping with past research on barroom aggression, the present study focused specifically on licensed drinking establishments that functioned primarily as entertainment venues for young people. The total population of bars located in Hoboken’s southeast quadrant was included in the present study, with the exception of two venues that functioned primarily as restaurants. The southeast quadrant of Hoboken contains over half of the city’s licensed drinking establishments, as well as a major hub for public transportation that attracts customers both from within and outside the city. It should be noted that each of the bars included in this study (n = 25) were within reasonable walking distance from this transportation hub and were therefore capable of being visited by the same population of customers. Bars ranged in capacity from about 100 to over 900 patrons. While the primary entertainment in most of the bars was prerecorded music, several also attracted customers with televised sporting events, live bands, and disc jockeys. Bars included in this study catered primarily to young (i.e., 20s),
432 / ROBERTS
white, professionals and college students from Hoboken and the surrounding areas. Male patrons outnumbered female patrons in most Hoboken barrooms. Procedures Observations took place at two separate time periods, 7:30pm–10:30pm and 11:00pm–2:00am, on Thursday, Friday, and Saturday nights. Each bar, with the exception of one venue that was only open on Fridays and Saturdays, was visited on six separate evenings, allowing for observations on each day of observation at each of the possible time periods. In total, the study entailed 148 individual observation periods lasting a total of 444 hours. Barrooms were assigned to nights and times of observation using a random start, the final result being a structured observation schedule. The present study employed one primary observer who participated in all 148 observation periods. In order to measure the reliability of observer responses in completing the data collection instrument, a second observer was employed who participated in approximately one-third of the total number of observation periods. Once the observation schedule had been devised, indicating the days and times each bar was to be observed, the second observer was randomly assigned to 50 observation periods. The primary observer conducted observations alone during the remaining 98 observation periods. During the months leading up to the study, the two observers studied the contents of a structured observation guide containing a large number of variables found to be good predictors of barroom aggression by past researchers (Graham et al., 1980; Homel et al., 1992; Homel & Clark, 1994). They also studied guidelines for identifying and recording incidents of physical and non-physical aggression. As was the case in Homel and Clark’s (1994) study of Sydney barrooms, physical aggression was defined as deliberate unfriendly bumping, grabbing, pushing, punching, kicking, etc. Non-physical aggression included one-way abuse, heated arguments, and challenges and threats. Incidents involving mere horseplay were not recorded unless they escalated into more serious acts of physical or nonphysical aggression. During observation periods, observers circulated throughout the premises, paying particular attention to situational variables contained in the observation guide, as well as any incidents of aggression. They did not consume any alcoholic beverages while conducting observations but instead purchased beverages that closely resembled mixed drinks (e.g., sodas with lemons or limes, etc.). As in previous observational studies of violence in bars (Graham et al., 1980; Homel et al., 1992; Homel & Clark, 1994), only limited note taking was conducted while inside barrooms and was done so in places that would not draw any unwanted attention, the most common location being bathroom stalls. Written recordings taken inside barrooms were limited to items in the observation guide that required observers to make counts of particular phenomena, as well as brief descriptions of violent episodes.
BARROOM AGGRESSION AND BOUNCERS /
433
Immediately following each observation period, observers independently rated items in the observation guide and wrote out their final descriptions of aggressive incidents. For observation periods in which both observers were present, a third and final observation guide was prepared for analyses after differences in individual coding were resolved. The two observers maintained a high degree of inter-rater reliability throughout the study, with individual responses to items in the observation guide never varying to such a degree as to cause concern. For example, if individual coding of variables did vary, it rarely did so by more than a single response category (e.g., somewhat smoky vs. smoky). Further, observers maintained 100% agreement throughout the study when identifying incidents and types of aggression. RESULTS General Characteristics of Incidents In the 444 hours of observation, observers in the present study witnessed 37 incidents of nonphysical aggression, which was broken down into 14 incidents of one-way verbal abuse, 16 arguments, and 7 challenges/threats. They also witnessed 43 incidents of physical aggression. Of the 80 incidents of aggression observed in this study, 76.2% (60) involved males only, 13.8% (11) involved females only, and 10.0% (8) involved both males and females. As was expected, the majority of aggressive incidents (66.2%) involved patrons only. In total, 32.5% (26) of aggressive incidents involved confrontations between patrons and bar staff. Patrons were identified as the perpetrators of 62% (16) of these confrontations, with servers of alcohol being the most common victims. In the vast majority (95%) of the aggressive incidents, the level of drunkenness displayed by participants was high. The rate of aggression per 100 hours of observation (18.0) was significantly lower than in Homel and Clark’s (1994) Sydney study (34.0). It should be noted, however, that the sample of barrooms selected for the Sydney study included several premises known to be “high risk� for violence. A closer examination of the types of aggression recorded in this study revealed extremely skewed distributions of these variables. To eliminate problems associated with skewed distributions in the bivariate and multivariate analyses, a new variable was constructed to provide an overall measure of aggression. The variable AGGRESSION is a dichotomous variable recording whether or not any aggressive incidents, physical or nonphysical, occurred (it did during 42 observation periods). As mentioned previously, there were 148 individual observation periods in the present study, each of which entailed three hours of observation at a specific bar and on a specific day of the week. Further, while multiple incidents of aggression may have occurred during a single observation period, the analyses that follow are limited to an examination of variables that characterized
434 / ROBERTS
observation periods “with” aggression (one or more incidents) versus observation periods “without” aggression (no incidents). Bivariate Analyses Table 1 presents the relationship between AGGRESSION and several variables related to the physical and social environment of barrooms. As indicated, aggression was more likely to occur during observation periods in which bars were unattractive, uncomfortable, smoky, hot, noisy, crowded (high), difficult to move around in, and unclean. Aggression was also more likely to occur during observation periods in which tables, ledges, and bar surfaces were littered with trash, bottles, and glasses. Conditions within many Hoboken barrooms worsened as evenings progressed. Bars were typically hotter, smokier, more crowded, and less clean at later time periods. There were also more problems with aggression at later time periods. Table 1 also indicates that aggression was more likely to occur during observation periods taking place in bars other than “watering holes.” Watering holes, which represented the majority of bars included in this study, provided no real entertainment other than prerecorded music and the occasional dartboard or pool table. “Other” bar types, on the other hand, attracted customers through some specialized type of entertainment, such as live bands, disc jockeys, and televised sporting events. Dancing, a common feature of “other” bar types, also demonstrated a statistically significant relationship with AGGRESSION (p < .01). Table 2 presents the relationship between AGGRESSION and several variables related to alcohol service and consumption. As indicated, aggression was more likely to occur during observation periods in which bar staff regularly checked patrons’ identification at the front door. Some of the more stringent door policies were observed at barrooms that local law enforcement identified as being on probation for serving underage patrons. Aggression was also more likely to occur during observation periods in which customers frequently purchased rounds of alcoholic beverages. Aggression was witnessed in 74.1% of observation periods in which bouncers and doormen were present (N = 107) and drinking alcoholic beverages. This is significant considering that aggression was witnessed in only 8.8% of observation periods in which they were present (N = 107) and not drinking alcoholic beverages. Drinking by servers of alcohol was also problematic, as aggression was witnessed in 59.6% of observation periods in which servers of alcohol were observed drinking alcoholic beverages while working and in only 8.8% of those in which they were not. Table 2 also reveals a statistically significant relationship between “service of alcohol to obviously intoxicated customers” and AGGRESSION (p < .01). The average number of times obviously intoxicated customers were served alcohol was significantly higher during observation periods with aggression (181.26) than
BARROOM AGGRESSION AND BOUNCERS /
435
during observation periods with no aggression (42.30). While servers observed drinking on-the-job were responsible for a greater number of such incidents, the service of alcohol to already intoxicated customers was a problem in most bars and among most servers. Observers in the present study witnessed only a single incident in which an intoxicated patron was actually “cut off.” Relationships between independent variables were examined by running a series of cross-tabulations (chi-square tests). These analyses revealed statistically significant relationships between several variables described as “aversive environmental stimuli,” including: appearance of bar, smokiness in bar, temperature in bar, crowding (maximum capacity) in bar, movement throughout bar, overall comfort of bar, clearing of tables and ledges, clearing of bar surfaces, and overall cleanliness of bar. In order to address potential problems associated with multicollinearity in the multivariate analyses, a factor analysis was conducted to see if variables could be combined. This analysis resulted in the extraction of a single underlying factor, which explained 67.2% of the variability in the nine variables. This factor, labeled as “aversive environmental stimuli,” was substituted into the multivariate analyses. In preparation of the multivariate analyses, it was also necessary to dummy code the variable “bouncers and doormen drinking alcoholic beverages.” To account for observation periods in which bouncers and doormen were not present, this variable was dummy coded into: variable 1 = bouncers and doormen drinking (1 = yes, 0 = no) and variable 2 = bouncers and doormen not present (1 = not present, 0 = present). “Bouncers and doormen not drinking” was used as the reference category in the multivariate analyses. Multivariate Analyses In an attempt to isolate key predictors of AGGRESSION, a multivariate logistic regression analysis was carried out. Collinearity diagnostics were run on all independent variables going into the logistic regression model (i.e., variables demonstrating statistically significant relationships with AGGRESSION in the bivariate analyses) and variance inflation factors were examined in order to assess the level of multicollinearity among these variables. All of the variables included in the logistic regression analysis demonstrated VIFs of below 3.0, indicating that multicollinearity was not an issue. Table 3 presents the results for the logistic regression analysis examining the relative influence of the key independent variables on AGGRESSION. As indicated, four variables demonstrated significant unique effects on AGGRESSION: aversive environmental stimuli (p < .05), bouncers and doormen drinking alcoholic beverages (p < .01), bouncers and doormen not present (p < .01), and servers of alcohol drinking alcoholic beverages (p < .01). Of these variables, “bouncers and doormen not present” was the strongest predictor of AGGRESSION (Wald = 13.842). The odds of aggression when bouncers and
436 / ROBERTS
Table 1. AGGRESSION by Variables Related to the Physical and Social Environment (N = 148 Observation Periods)
N
AGGRESSION Yes %
Appearance of bar Attractive Somewhat unattractive Unattractive
66 46 36
10.6% 23.9 66.7
Smokiness in bar Mostly smoke free Somewhat smoky Smoky
91 34 23
15.4 35.3 69.6
Temperature in bar Cool Somewhat warm Hot
70 40 38
14.3 27.5 55.3
Crowding (maximum capacity) in bar Lowa Medium Highb
95 28 25
16.8 28.6 72.0
Movement throughout bar Easy to move about Somewhat difficult to move about Difficult to move about
76 35 37
11.8 34.3 56.8
Overall comfort of bar Comfortable Somewhat uncomfortable Uncomfortable
82 31 35
12.2 29.0 65.7
Clearing of tables and ledges Mostly kept clear Somewhat cluttered Cluttered
85 29 34
12.9 37.9 58.8
Clearing of bar surfaces Mostly kept clear Somewhat cluttered Cluttered
89 25 34
14.6 40.0 55.9
c2 (df) 36.67(2)**
27.55(2)**
20.36(2)**
29.62(2)**
25.48(2)**
34.57(2)**
26.77(2)**
22.62(2)**
BARROOM AGGRESSION AND BOUNCERS /
437
Table 1. (Cont’d.)
N Overall cleanliness of bar Mostly clean Somewhat unclean Unclean
AGGRESSION Yes %
c2 (df) 36.57(2)**
91 22 35
12.1% 36.4 65.7 23.24(1)**
Dancing No Yes
100 48
16.0 54.2
74 74
14.9 41.9
117 31
23.9 45.2
13.29(1)**
Time T1 = 7:30 pm–10:30 pm T2 = 11:00 pm–2:00 am Bar type Watering hole Other (e.g., band bars, sports bars, dance clubs, etc.)
5.43(1)*
aLow = approximately 1/3 bar’s maximum capacity. bHigh = at or exceeding bar’s maximum capacity.
*p < .05. **p < 01.
doormen were not present were about 36.342 times greater than when they were present and not drinking. This is a particularly important finding considering the frequency with which bouncers and doormen are vilified in the research on barroom aggression (Graham & Wells, 2003; Graves, Graves, Semu, & Sam, 1981; Homel et al., 1992; Lister, Hobbs, Hall, & Winlow, 2000; Wells et al., 1998). DISCUSSION Based on the findings presented in the results section, and in particular, the four variables that demonstrated significant unique effects on AGGRESSION in the multivariate analyses, several practical prevention strategies aimed at reducing aggression in bars are proposed. Aversive Environmental Stimuli Aversive environmental stimuli include a range of variables believed to provoke aggressive behaviors among bar patrons, including poor ventilation and
438 / ROBERTS
Table 2. AGGRESSION by Variables Related to Alcohol Service and Consumption (N = 148 Observation Periods)
N ID check No Yes
AGGRESSION Yes % Mean (SD)
c2 (df) t(df) 5.37(1)*
108 40
23.1% 42.5 4.62(1)*
Drink specials No Yes
94 54
22.3 38.9
Drinking in rounds Infrequent Frequent
68 80
14.7 40.0
Bouncers and doormen drinking alcoholic beverages (N = 107) No Yes
80 27
8.8 74.1
Servers of alcohol drinking alcoholic beverages No Yes
91 57
8.8 59.6
No. times obviously intoxicated customers served alcohol
42
181.26 (159.31)
11.57(1)**
45.65(1)**
44.60(1)**
5.45(47.35)**
*p < .05. **p < 01.
loud noise (Graham et al., 1980), crowded corridors, stairs, and doorways (Homel et al., 1992; Macintrye & Homel, 1997), low comfort due to lack of seating or food (Homel et al., 1992; Homel & Clark, 1994), excessive heat and smoke (Quigley, Leonard, & Collins, 2003), etc. In addition to triggering aggressive behaviors among bar patrons by causing feeling of discomfort and frustration, Wortley (2001) suggests that aversive environmental stimuli are problematic in that they provide situational cues that promote normally proscribed behavior. He likens the influence of these variables on patron behavior to the contagious effect of â&#x20AC;&#x153;Broken Windows.â&#x20AC;? Just as a single broken window left unattended may lead to further disorder and eventually increased criminal activity in some neighborhoods,
BARROOM AGGRESSION AND BOUNCERS /
439
Table 3. Logistic Regression Model: Key Independent Variables on AGGRESSION Variable
B
S.E.
Wald
P
–1.455
.831
3.069
.080
.233
Bar type
.918
.978
.882
.348
2.505
ID check
–.397
.992
.160
.689
.672
Aversive environmental stimuli
1.001
.494
4.097
.043
2.720
Dancing
1.563
1.001
2.436
.119
4.773
Bouncers and doormen drinking alcoholic beverages
3.010
.939
10.284
.001
20.282
Bouncers and doormen not present
3.593
.966
13.842
.000
36.342
Drink specials
–.295
.816
.131
.718
.745
.639
.798
.641
.423
1.894
2.171
.710
9.363
.002
8.769
.005
.005
.935
.333
1.005
–3.922
1.585
6.122
.013
.020
Time
Drinking in rounds Servers of alcohol drinking alcohol beverages No. times obviously intoxicated customers served alcohol (Constant)
Exp(B)
Model chi-square 96.277, p = .000 Nagelkerke R2 = .686
a single broken bottle left unattended may lead to further disorder and eventually increased problems with violence and aggression in some barrooms. While the presence of aversive environmental stimuli may be pervasive and problematic in high risk drinking settings, minor modifications of the barroom landscape may reduce their effect on aggressive incidents. For instance, problems associated with excessive heat may be avoided by installing air conditioning units or ceiling fans. Similarly, problems with excessive smoke may be addressed by installing “smoke eaters” or other air filtration systems. Homel, Carvolth, Hauritz, McIlwain, and Teague (2004) state that strategies aimed at
440 / ROBERTS
improving the overall level of comfort within bars are critical to reductions in rates of violence and aggression. A major component of Graham et al.’s (2004) “Safer Bars” initiative in Toronto, Canada was helping owners and managers identify ways to reduce environmental risks. Graham and her colleagues suggest that environmental changes for reducing aggression in bars should be easier to market to owners and managers than interventions that focus primarily on serving less alcohol. Drinking On-The-Job Surprisingly, the consumption of alcohol by bar staff has received little attention in past research on barroom aggression. Observers in the present study only added this phenomenon as a variable of interest after conducting a series of preliminary observations in Hoboken barrooms. The consumption of alcohol by servers and security personnel was problematic for two main reasons. First, alcohol consumption negatively impacted the ability of staff members to perform their respective duties. Security personnel observed drinking on-the-job failed to detect several minor altercations between customers that eventually erupted into full-blown aggressive incidents. Servers observed drinking on-the-job were more likely to serve alcohol to already intoxicated customers than those who were not. The consumption of alcohol by bar staff also contributed to the prevailing sense of “normlessness” that existed in many high risk drinking settings. According to Wortley (2001), “Situational factors can help distort moral reasoning processes and so permit individuals to engage in normally proscribed behavior” (p. 70). Further, he suggests that given the right circumstances, ordinary individuals are capable of great brutality. Much like the presence of various aversive environmental stimuli, the sight of bar staff openly drinking alcoholic beverages while working conveys to patrons that the establishment is a place where “anything goes.” The consumption of alcohol by bar staff in the present study was symptomatic of the lax attitude with which many owners and managers ran their bars. According to Homel et al. (1994), “The link between permissiveness and violence makes a lot of intuitive sense, since if managers have an ‘anything goes’ attitude it is not surprising that violence and aggression occur” (p. 28). Findings from the present study suggest that owners and managers can play a major role in restoring order in bars by setting and enforcing rules of conduct for bar staff and patrons. Rule setting is one of Clarke’s (1997) 16 opportunityreducing techniques used in situational crime prevention initiatives and involves, “. . . the introduction of new rules or procedures (and the clarification of those in place), which are intended to remove any ambiguity concerning the acceptability of conduct” (p. 24). Rules of conduct for bar staff should be written in employee handbooks and made available to all staff members. Rules of conduct for patrons should be posted in highly visible locations, such as over bar stations and on doors. According to Graham et al. (2006), “. . . setting and maintaining standards of
BARROOM AGGRESSION AND BOUNCERS /
441
behavior, including reducing rowdy behavior, may be an effective strategy not only for reducing the incidence but also the severity of aggression” (p. 1577). Lack of Security Staff The absence of security personnel was identified as the strongest predictor of aggression in Hoboken barrooms. Lack of security personnel closely parallels “lack of capable guardians” in Cohen and Felson’s (1979) well-known routine activities theory. According to Fox and Sobol (2000), “The routine activities of the bar scene often put likely victims and motivated offenders in close proximity without adequate protection measures. Thus, the convergence of offenders, victims, and an absence of capable guardians appears to increase the chances of some bar patrons’ victimization” (p. 446). While bouncers and doormen are in unique positions to significantly reduce aggression in bars, as they are recognized by many as being the primary agents of social control within barrooms (Hobbs, Lister, Hadfield, Winlow, & Hall, 2000), they are often blamed for encouraging and escalating such incidents. Wells et al. (1998) utilized observational and interview data to better understand security staff responses to incidents of alcohol-related aggression. Security staff responses in this study fell into four main categories: good, bad, ugly, and neutral. A major finding from this study was that “many security staff do not behave in a manner that discourages aggression or sets up nonviolent norms” (Wells et al., 1998, p. 829). Wells and her colleagues join a growing number of researchers (Graham et al., 2004; Graham, Jelley, & Purcell, 2005; Graham & Wells, 2003; Hauritz, Homel, McIlwain, Burrows, & Townsley, 1998; Homel et al., 1992; Homel & Clark, 1994; Homel, Hauritz, Wortley, McIlwain, & Carvolth, 1997; Lister, Hadfield, Hobbs, & Winlow, 2001; Pernanen, 1998) who suggest a strong need for better training and greater legal accountability of bouncers and doormen. Hobbs, Hadfield, Lister, and Winlow (2003) discuss an ad hoc system of occupational licensing for barroom employees that exists throughout the UK. A similar system exists in Hoboken that requires prospective staff members to register with the local Alcohol Beverage Control (ABC) board and submit to a thorough background investigation. Unlike the system described by Hobbs et al. (2003), the system that exists in Hoboken has operated with great success. While bouncers and doormen working in Hoboken barrooms did not always behave professionally, particularly when out of eyeshot of local law enforcement, they did not display the same aggressive tendencies observed in previous research (Graham et al., 2005; Graham & Wells, 2001; Wells et al., 1998). Further, while the ideal situation would be to have an adequate number of trained, professional bouncers and doormen patrolling barrooms, findings from the present study suggest that even the presence of a substandard security staff, or rather, a security staff that is behaving badly (e.g., drinking on-the-job) may be better than no security staff at all.
442 / ROBERTS
CONCLUSIONS When considering the aforementioned findings and strategies for preventing aggression in bars, one must keep in mind the limitations of the present study. One limitation of observational studies of barroom aggression in general is the inability to assess causality. For instance, it is plausible that the environments encountered by observers in these studies attract personalities that will be prone to violence. Thus, changing the environment will only be effective if it repels violence prone patrons. As mentioned previously, researchers studying aggression in bars take for granted the fact that individuals bring to drinking episodes a variety of unique individual characteristics, including unique propensities for aggression. Assessing individual characteristics of bar patrons, and in particular, individual propensities for aggression, is simply not possible under most circumstances, especially in studies utilizing unobtrusive observation techniques. Despite this obvious and unavoidable limitation of the research, analyses presented in this article support the interpretation that the barroom environment has an important impact on aggression. A specific limitation of the present study was the use of only one observer during 98 of the 148 observation periods. During observation periods with two observers, any discrepancies in individual coding were discussed and decisions on how to code items for the third and final observation guide were agreed upon. This type of cross validation was not possible during those observation periods with a single observer. However, as mentioned previously, observers in the present study maintained a high degree of inter-rater reliability throughout the present study. The consistency of observer responses to items contained in the observation guide was likely the result of the observer training and pilot-testing of the data collection instrument that preceded the study. When possible, however, future research should utilize multiple observers, not just to ensure that the individual coding of observation guides remains consistent, but also to better examine the many situational variables present in the barroom environment. Due to the homogeneity of barrooms (i.e., mostly watering holes) and patrons (i.e., primarily young and white) in the present study, findings from this study cannot be generalized to all types of drinking establishments or all populations of customers. Whether predictors of aggression identified in the present study hold up in barrooms catering primarily to non-white or older customers is a question that future research should attempt to address. Future research might also examine whether situational variables found to be good predictors of aggression in the present study hold up in large samples of sports bars, dance clubs, or band bars. The present study provides only limited evidence that they do. As most ethnographical studies of barrooms aggression have taken place outside the country, there is a real need for more studies of this kind to be replicated in other cities throughout the United States. A common set of variables shown to predict aggression in licensed drinking establishments, regardless of bar type, patron
BARROOM AGGRESSION AND BOUNCERS /
443
characteristics, or local, may be particularly useful in the development of strategies aimed a reducing such incidents. The present study offers additional evidence that aggression in bars is strongly related to situational variables found in these establishments. As suggested by past researchers (Homel et al., 1992), possibly the easiest and most practical approach to reducing problems with aggression in bars is regulating the routine activities of licensed drinking establishments, rather than attempting to modify the routine activities of bar patrons. Future research is needed to better understand the role of bouncers and doormen as guardians of the barroom environment. Findings from this study also suggest a strong need to further explore the impact of the absence of bouncers and doormen on barroom aggression. ACKNOWLEDGMENTS I gratefully acknowledge the Rutgers University School of Criminal Justice and the First Security Services Corporation for providing funding for the present study. The opinions expressed in this article are those of the author and do not necessarily reflect the views or policies of the funding sources. I am also forever grateful to Drs Ronald V. Clarke, Marcus Felson, Michael Maxfield, and Helene Raskin White for their insights, guidance, and feedback during this project. I would also like to acknowledge the second observer in this study, Kimberly Lutter, who served as my second pair of ears and eyes during the data collection. Finally, I would like to thank Phyllis Schultz for her kind and continuous support and assistance in researching all things related to alcohol and aggression. REFERENCES Clarke, R. V. (Ed.). (1997). Situational crime prevention: Successful case studies (2nd ed.). New York: Harrow and Heston. Cohen, L. E., & Felson, M. (1979). Social change and crime rate trends: A routine activity approach. American Sociological Review, 44, 588-608. Fox, J. G., & Sobol, J. J. (2000). Drinking patterns, social interaction, and barroom behavior: A routine activities approach. Deviant Behavior: An Interdisciplinary Journal, 21, 429-450. Graham, K. (1980). Theories of intoxicated aggression. Canadian Journal of Behavioral Science, 12, 141-158. Graham, K., Bernards, S., Osgood, D. W., Homel, R., & Purcell, J. (2005). Guardians and handlers: The role of bar staff in preventing and managing aggression. Addiction, 100, 755-766. Graham, K., Bernards, D., Osgood, D. W., & Wells, S. (2006). Bad nights or bad bars? Multi-level analysis of environmental predictors of aggression in late-night largecapacity bars and clubs. Addiction, 101, 1569-1580. Graham, K., Jelley, J., & Purcell, J. (2005). Training bar staff in preventing and managing aggression in licensed premises. Journal of Substance Use, 10, 48-61.
444 / ROBERTS
Graham, K., La Rocque, L., Yetman, R., Ross, T. J., & Guistra, E. (1980). Aggression and barroom environments. Journal of Studies on Alcohol, 41, 277-292. Graham, K., Leonard, K. E., Room, R., Wild, T. C., Pihl, R. O., Bois, C., & Single, E. (1998). Current directions in research on understanding and preventing intoxicated aggression. Addiction, 93, 659-676. Graham, K., Osgood, D. W., Zibrowski, E., Purcell, J., Gliksman, L., Leonard, K., et al. (2004). The effect of the Safer Bars programme on physical aggression in bars: Results of a randomized control trial. Drug and Alcohol Review, 23, 31-41. Graham, K., Schmidt, G., & Gillis, K. (1996). Circumstances when drinking leads to aggression: An overview if research findings. Contemporary Drug Problems, 23, 493-557. Graham, K., & Wells, S. (2001). Aggression among young adults in the social context of the bar. Addiction Research & Theory, 9, 193-219. Graham, K., & Wells, S. (2003). Somebody’s gonna get their head kicked in tonight: Aggression among young males in bars—A question of values? British Journal of Criminology, 43, 546-644. Graham, K., West, P., & Wells, S. (2000). Evaluating theories of alcohol-related aggression using observations of young adult bars. Addiction, 95, 847-863. Graves, T. D., Graves, N. B., Semu, V. N., & Sam, I. A. (1981). The social context of drinking and violence in New Zealand’s multi-ethnic pub settings. In T. C. Harford & L. S. Gaines (Eds.), Research Monograph No. 7. Social drinking contexts (pp. 103-120). Rockville, MD: NIAAA. Hauritz, M., Homel, R., McIlwain, G., Burrows, T., & Townsley, M. (1998). Reducing violence in licensed venues through community safety action projects: The Queensland experience. Contemporary Drug Problems, 25, 511-551. Hobbs, D., Hadfield, P., Lister, S., & Winlow, S. (2003). Bouncers: Violence and governance in the night-time economy. Oxford, UK/New York: Oxford University Press. Hobbs, D., Lister, S., Hadfield, P., Winlow, S., & Hall, S. (2000). Receiving shadows: Governance and liminality in the night-time economy. British Journal of Sociology, 51, 701-717. Homel, R., Carvolth, R., Hauritz, M., McIlwain, G., & Teague, R. (2004). Making licensed venues safer for patrons: What environmental factors should be the focus of interventions? Drug and Alcohol Review, 23, 19-29. Homel, R., & Clark, J. (1994). The prediction and prevention of violence in pubs and clubs. Crime Prevention Studies, 3, 1-46. Homel, R., Hauritz, M., Wortley, R., McIlwain, G., & Carvolth, R. (1997). Preventing alcohol-related crime through community action: The Surfers Paradise Safety Action Project. Crime Prevention Studies, 7, 35-90. Homel, R., Tomsen, S., & Thommeny, J. (1992). Public drinking and violence: Not just an alcohol problem. The Journal of Drug Issues, 22, 679-697. Lister, S., Hadfield, P., Hobbs, D., & Winlow, S. (2001). Accounting for bouncers: Occupational licensing as a mechanism for regulation. Criminal Justice, 1, 363-384. Lister, S., Hobbs, D., Hall, S., & Winlow, S. (2000). Violence in the night-time economy; bouncers: The reporting, recording and prosecution of assaults. Police and Society, 10, 383-402. Macintyre, S., & Homel, R. (1997). Danger on the dance floor: A study of interior design, crowding and aggression in nightclubs. In R. Homel (Ed.), Policing for prevention:
BARROOM AGGRESSION AND BOUNCERS /
445
Reducing crime, public intoxication and injury, crime prevention studies (Vol. 7, pp. 91-113). Monsey, NY: Criminal Justice Press. Pernanen, K. (1998). Prevention of alcohol-related violence. Contemporary Drug Problems, 25, 477-509. Quigley, B. M., Leonard, K. E., & Collins, R. L. (2003). Characteristics of violent bars and bar patrons. Journal of Studies on Alcohol, 64, 765-772. Stockwell, T., Lang, E., & Rydon, P. (1993). High risk drinking settings: The association of serving and promotional practices with harmful drinking. Addiction, 88, 1519-1526. Wells, S., Graham, K., & West, P. (1998). â&#x20AC;&#x153;The good, the bad, and the uglyâ&#x20AC;?: Responses by security staff to aggressive incidents in public drinking settings. Journal of Drug Issues, 28, 817-836. Wortley, R. (2001). A classification of techniques for controlling situational precipitators of crime. Security Journal, 14, 63-82.
Direct reprint requests to: James C. Roberts, Ph.D. University of Scranton Department of Sociology/Criminal Justice Scranton, PA 18510-4605 e-mail: robertsj7@scranton.edu