Extending the Hours of Operation of Alcohol Serving Establishments

Abstract

On May 11, 2009, the City of San Marcos, Texas adopted an ordinance that allowed local bars to extend their hours of operation from midnight until 2 a.m. The objective of this initiative, which went into effect on June 4, 2009, was to minimize the problems that arise from the consumption of alcohol. The present investigation uses interrupted time series autoregressive integrated moving average [ARIMA]) modeling techniques to analyze weekly data to assess the impact of the ordinance, along with a limited increase in patrol strength in the downtown district, on the number of calls for service involving verbal disturbances, physical disturbances, public intoxication, and driving under the influence of alcohol. Overall, the interrupted time series analyses indicate that neither the city council’s nor the police department’s strategies for mitigating some of the negative consequences of the consumption of alcoholic beverages in bars and taverns were successful. The implications of these findings for lessening alcohol-related conflicts between bar patrons and other residents of the community are discussed.

Keywords

interrupted time series social policy alcohol-related incidents

Introduction

Legal authorities often react to citizen complaints about disorder and crime proactively. Typical responses include, but are not limited to, aggressive patrol (Novak, Hartman, Holsinger, & Turner, 1999; Smith, 2001; Weiss & Freels, 1996), sting operations (Langworthy, 1989; Langworthy & LeBeau, 1992), the criminalization, through the passage of appropriate legislation, of the offending behavior (Simpson, Bouffard, Garner, & Hickman, 2006; Webster, Vernick, & Hepburn, 2002; White, 2003), and changes in law enforcement practices and/or administrative policies (Holmes, Daudistel, & Taggart 1992; Katz, Webb, & Schaefer, 2001; Villveces et al., 2000; White, 2000).

Occasionally, municipalities embrace less conventional methods to deal with public demands for elected officials and the police to do something about an emerging social problem. The present investigation examines the effectiveness of one such attempt to do so in a small, southwestern city.

In May of 2009, the city council of San Marcos, Texas adopted an ordinance extending the hours for the sale of alcoholic beverages from midnight until 2 a.m. In response to the new policy, 29 establishments licensed to sell alcohol applied for and received permits for on-premises consumption during the extended hours (City of San Marcos, 2010). It is worth noting that many of the city’s bars are located in a central downtown district. Approximately half (14) of the bars receiving extended hours permits were located downtown, primarily in a popular entertainment district known as “The Square.”

The primary objective of the extension was to decrease late night and early morning complaints arising from after-hours house parties. The underlying logic supporting the measure was that, prior to the ordinance, bar patrons would leave the bars at midnight and then disperse to houses or apartments around town where they would continue drinking until the early morning hours.

The extended hours ordinance was one of two strategies implemented by local authorities in an attempt to decrease alcohol-related disturbances and to ease tensions between college students (who were thought to comprise a majority of the guests at the after-hours parties) and the more established members of the community (City of San Marcos, 2010). Concomitant to the 2-hr extension of alcohol sales, the police increased, for approximately one year, downtown patrols from two officers to four or five officers at closing time (George, 2010b).

The city took several measures in order to manage and monitor the effects of the ordinance. At the request of the city council, the city manager assembled the Extended Bar Hours Task Force, which has been monitoring the effects of the ordinance since 2009. The task force implemented a system for tracking police calls for service for offenses commonly associated with alcohol consumption. Those offenses included, but were not limited to, driving while intoxicated, public intoxication, minor in possession of alcohol, physical disturbances, and verbal disturbances (City of San Marcos, 2010).

In an innovative attempt to minimize civilian complaints about noxious behaviors that often arose from the after-hours consumption of alcohol, the City of San Marcos adopted two policy initiatives. The expectation was that by coupling an extension in the hours of operation of alcohol serving establishments with an increase in patrol strength in the area of the city (downtown) that contained the majority of bars and taverns, many problems relating to the consumption of alcohol could be avoided, thereby reducing calls for service throughout the city (City of San Marcos, 2010).

The purpose of this investigation is to assess the impact of these dual strategies on alcohol-related calls for service to the police. Of particular interest is determining whether or not these initiatives lessened the quantity of verbal and physical disturbance complaints or if they inadvertently exacerbated the legal problems associated with the consumption of alcohol.

Research Relating to the Effects of Extending Drinking Hours

We found no prior studies that examine the influence of increasing the hours of operation of alcohol serving, commercial establishments on the incidence of verbal or physical disturbances. However, we did locate a number of studies that assess the impact of allowing bars to close at a later time on a variety of other outcome measures.

Two studies examine the influence of extending bars hours on driving under the influence of alcohol. The first inquiry assesses the impact of the enactment of a Minnesota law that extends, by 1-hr (from 1 a.m. to 2 a.m.), the closing time of alcohol serving bars and restaurants on the number of police stops for driving under the influence of alcohol within a single, unnamed locality (Bouffard, Bergeron, & Bouffard, 2007). The interrupted time series analyses reveal that the implementation of the legislative initiative led to a permanent increase (as indicated by the parameter estimates from the autoregressive integrated moving average [ARIMA], zero-order transfer function equation) in police stops for driving under the influence.

The second article assesses the impact of the May 1, 1996 amendment to Ontario, Canada’s Liquor Licence Act that allows bars, as well as other alcohol serving businesses, to extend their sales for 1-hr (from 1 a.m. to 2 a.m.) on arrests for drunk driving in two Ontario cities (Vingilis, McLeod, Mann, & Seeley, 2008). Contrary to what Bouffard et al. (2007) find, Vingilis et al. (2008) report that their zero-order transfer equations show that the 1-hr extension in drinking hours produces a permanent decline in drunk-driving arrests in both Windsor and London, Ontario.

Vingilis and colleagues also examine the effect of the change in Ontario’s Liquor Licence Act on several other outcome series. Unfortunately, the results from these analyses provide little, if any, guidance for our investigation.

Vingilis and associates first estimate the effects of the liberalization of drinking hours on motor vehicles fatalities within the entire province of Ontario (Vingilis et al., 2005). They report, based on the specification and estimation of zero-order transfer function models, that the 1-hr extension of alcohol sales has no appreciable impact on the total number of motor vehicle fatalities or on motor vehicle fatalities involving alcohol.

Vingilis’ next evaluation of the modification of the Liquor Licence Act focuses on its effect on motor vehicle fatalities in Windsor, Ontario, and Detroit (which is located across the border from Windsor). Contrary to what they report for the province as a whole, Vingilis et al. (2006) find that extending the drinking hours produces a significant increase in motor vehicle fatalities in the city of Windsor. Interestingly, the interrupted times series analyses also suggest that the additional hour of alcohol sales results in a decline in motor vehicle fatalities in Detroit (Vingilis et al., 2006).

Vingilis and associates also model the influence of the 1-hr extension of alcohol sales on serious, but nonfatal, injuries from motor vehicle accidents and other causes within Ontario (Vingilis, McLeod, Stoduto, Seeley, & Mann, 2007). The results from the zero-order transfer function equations indicate that the change in Ontario’s Licence Act has no effect on the level of serious injuries, produces a significant decline in the level of serious injuries from other causes between the hours from 11 p.m. to 12 a.m. and 1 a.m. to 2 a.m., but leads to a significant increase in the level of serious injuries from other causes between the hours of 2 a.m. and 3 a.m. (Vingilis, McLeod, Stoduto, & Mann, 2007).

Lastly, Vingilis et al. (2008) assess the impact of the hour extension of alcohol sales on assault arrests within the cities of London and Windsor, Ontario. The results vary by the level of temporal aggregation. For the entire temporal period under investigation, 11 p.m. to 4 a.m., the intervention had no impact on assaults in either city. However, the authors report that, for London, there is a significant decrease in assaults for Thursday through Saturday, from 1 a.m. to 2 a.m. For Windsor, they report that this is a significant increase in assaults for Thursday through Saturday, from 2 a.m. to 3 a.m.

In sum, our review of previous examinations of the impact of extending the closing time for commercial establishments involved in the sale of alcoholic beverages on various legal outcomes is not particularly helpful. We can find no discernible pattern to the published findings (Bouffard et al., 2007; Vingilis et al., 2006; Vingilis et al., 2007; Vingilis et al., 2008). The only inference we can draw from this body of research is that we need to be particularly careful to root our research design decisions in theory and accepted practices and not on matters of convenience.

Method

The Data

This study uses ARIMA techniques to model the impact of the two strategies designed to reduce calls for service arising from the behavior of patrons leaving bars at closing time. The time series, which were created from unpublished calls for service, kindly provided by the San Marcos, Texas police department, consisted of 209 weekly observations. These count data include all incidents reported to the police, regardless of their source (civilian complaint or police observation).

The first week for each of the outcome series begins June 1, 2007, approximately two years prior to the onset of the extension of operating hours for bars. The last weekly observation ends May 31, 2011, approximately two years after the extension of operating hours for bars and one year after the cessation of increased patrols in the downtown area of the city. Hence, we have a sufficient number of observations before and after the two interventions to assess their impact on calls for service (cf. Cochran, Chamlin, & Seth, 1994; Singer & McDowall, 1988).

Outcome Series

As we discussed above, the primary motivation for extending the hours of operation for bars and increasing the number of patrols in the downtown portion of the city was to reduce the number of citizen complaints stemming from the consumption of alcohol. In particular, the city council and the chief of police hoped to substantially reduce the number of calls for service relating to verbal and physical disturbances (City of San Marcos, 2010). Hence, we examine the impact of the contemporaneous interventions on the weekly count of calls for service for verbal disturbances and physical disturbances, respectively.

A conceivable, though unintended, consequence of allowing retail establishments to sell alcoholic beverages for an additional 2-hr is an increase in the levels of public intoxication and driving under the influence of alcohol. Consequently, we also examine the influence of the contemporaneous interventions on calls for service for public intoxication and driving under the influence of alcohol.

Interventions

We created two dummy series to model the impact of the city council’s efforts at reducing citizen complaints arising from the disruptive behavior of individuals after the consumption of alcohol in bars and taverns. The first series is coded zero for the weeks prior to the implementation of the ordinance allowing alcohol serving retail businesses to stay open until 2 a.m. and the increase police patrols of the downtown area (from June 7, 2007 through June 3, 2009) and one thereafter. The second series is coded zero for the weeks prior to the termination of the increased police patrols (from June 7, 2007 through May 26, 2010) and one thereafter.

To clarify, the first intervention captures the joint influence of the city council’s and the police department’s strategies for reducing calls for service relating to the consumption of alcohol. The inclusion of the second series into the model specifications allows us to distinguish between the influence of the extension of the operating hours to 2 a.m. from that of the augmentation of the police presence in the downtown section of the city.

Estimation Procedures

We use ARIMA interrupted time series analytic techniques to model the two interventions on the outcome series. Since ARIMA procedures are well established in the criminal justice literature (Chamlin, Myer, Sanders, & Cochran, 2008; Langworthy, 1989; Loftin, Heumann, & McDowall, 1983; Singer & McDowall, 1988), we offer a limited discussion of its most salient features.

A fundamental concern associated with the evaluation of the efficacy of legislative (the city council’s bar hours extension ordinance) and/or administrative (the increase the number of patrol officers in the downtown, bar district) initiatives is distinguishing their effects from other social processes that may be influencing an outcome series. ARIMA techniques, unlike simple pre and postintervention mean or percentage difference tests, explicitly take into account the potentially confounding effects of other causal mechanisms and, as a consequence, allow one to assess the change in the level of any outcome series independently of ongoing stochastic processes (McDowall, McCleary, Meidinger, & Hay, 1980).

In brief, an ARIMA interrupted transfer function model consists of two parts. The first, the “noise” component, uses information from prior observations of an outcome series to model the systematic variation (autocorrelation) within the outcome series. By applying the appropriate seasonal and nonseasonal differencing, along with the estimation of the appropriate seasonal and nonseasonal autoregressive and moving average parameters (prewhitening), one can separate the confounding influences of other causal processes from those associated with intervention.

Once a satisfactory noise component is identified and estimated, the intervention component is added to the transfer function equation. If the inclusion of a dummy series for the intervention (coded zero for the period prior to the onset of the intervention and coded one beginning with the observation in which the intervention occurs and thereafter) increases the explanatory power of the model above and beyond that provided by the noise component (Granger causality), then we can conclude that the intervention significantly affects the outcome series (Granger, 1980; McDowall et al., 1980).

Another advantage of ARIMA modeling techniques over simple pre and postintervention change scores is that they allow one to examine the functional form of the relationship between an intervention series and an outcome series. Crude mean and/or percentage difference tests assume that the effect of an invention is well represented as an abrupt, permanent change in the level of the outcome series (at least for the remainder of the observations for a given time series). While one can estimate this functional form (as a zero-order transfer function) using ARIMA modeling techniques, one can also examine the relative fit of competing adjustment models. It is possible that the effect of an intervention gradually reaches a new level or that the effect is instantaneous but short-lived (often reflecting a publicity effect). A first-order transfer function can be estimated to model the former pattern of change in the level of a series, while a pulse function can be estimated to model the latter (McDowall et al., 1980).

ARIMA model building is an iterative process. By successively estimating the noise and intervention components, and subjecting them to a number of diagnostic tests, a final transfer model can be derived. For the statistical details involved in the identification and estimation of the noise and intervention components of ARIMA interrupted times series models we refer the reader to readily available published sources (McCleary & Hay, 1980; McDowall et al., 1980).

Analytic Strategy and Expectations

The analyses proceed as follows. Recall that the increase in patrol strength was limited to the downtown portion of San Marcos. Consequently, we conduct separate analyses for downtown San Marcos and the remainder of the city. Insofar as the impact of this initiative is limited to that part of the city, the failure to disaggregate the data prior to executing the interrupted time series analyses is likely to obfuscate the impact of the increase in, as well as the cessation of, patrol strength on the outcome series.

Based on the results from prior interrupted time series analyses of the impact of bar hours on drunk-driving-related outcome series (Bouffard et al., 2007; Vingilis et al., 2008) we anticipate that the effects of the two interventions will be immediate and permanent (at least for the postintervention periods under investigation). Hence, we decided to specify and estimate zero-order transfer function equations.¹

To recap, the downtown area of the city is affected by two policy initiatives, the extension of operating hours for alcohol serving establishments and the increase in officers on patrol. Consequently, we specify and estimate two zero-order transfer function equations to assess the impact of these interventions on the four outcome series. The first equation contains a single intervention series (coded zero from June 7, 2007 to June 3, 2009, and one thereafter) to capture the influence of the simultaneous implementation of the citywide ordinance extending alcohol sales and the increased police presence in the downtown area on a dependent series. In order to distinguish between the effects of the city council ordinance from those of the increase in patrol strength, we introduce a second intervention (coded zero from June 7, 2007 to May 26, 2010, and one thereafter) to denote the cessation of added patrols in the downtown portion of the city.

The parameter estimates from the ARIMA analyses, show the effects of the two interventions on calls for service. If the first intervention produces a significant change in the level of an outcome series, but the second does not, then it would be fair to conclude that the effect of the first intervention results from the extension in drinking hours rather an increase in patrol strength. However, if both interventions significantly affect an outcome series, then determining their relative effects will depend on the direction and magnitude of the parameter estimates from the second transfer function equation. For example, if the first intervention produces a significant decrease in the level of an outcome series, while the second series has a significant positive effect on that same outcome series, then it seems likely that the initial reduction in calls for service reflects the deterrent effect of an increase in patrol strength on the behavior of bar and tavern patrons. In contrast, if the first intervention produces a significant decrease in the level of an outcome series and the second series also evidences a negative association with that same outcome series, then it seems most reasonable to infer that the citywide ordinance led to the decrease in calls for service and that the increase in patrol strength led to an increase in calls for service (perhaps because citizens may have felt that, with the increased police presence, their complaints would be dealt with more effectively).

The predictions regarding the downtown area do not apply to the remainder of the city. Since the increase in the number of patrol officers was limited to the downtown district, the first intervention reflects a single change in policy, the onset of increased operating hours for alcohol serving, retail businesses, while the second intervention becomes superfluous and need not be included in the model specifications.²

Results

Descriptive Analyses

Table 1 presents the mean number of alcohol-related calls for service by area of the city and intervention period. We begin our discussion of changes in the mean level of complaints with the downtown portion of San Marcos (see Panel A). Recall that this area of San Marcos was affected by two distinct policy initiatives, the extension of bar hours and the temporary increase in patrol strength. Hence, we estimate and compare the mean number of complaints for three portions of the outcome series. The first column contains the mean level of calls for service for the period prior to the implementation of the dual interventions. The second column reports the mean level of calls for service for the period between the implementation of the city council ordinance and the increase in patrol strength and the cessation of the increase in patrol strength. The third column presents the mean level of calls for service after the return to the preintervention number of patrol officers.

Table 1.

Mean Number of Alcohol-Related Calls for Service by Area of the City and Intervention Period (N = 209 weeks).

Panel A: Downtown San Marcos
	Preinterventions	Between first and second intervention	Post-second intervention
Verbal disturbances	1.88	2.92	2.83
Physical disturbances	0.58	0.78	1.24
Public intoxication	1.39	2.51	2.86
Driving under the Influence of alcohol	0.35	0.61	0.83
N	104	51	54
Panel B: Remainder of the city
	Preintervention	Postintervention
Verbal disturbances	21.38	21.31
Physical disturbances	8.48	7.19
Public intoxication	5.16	5.47
Driving under the influence of alcohol	3.87	2.88
N	104	105

Inspection of Panel A would seem to suggest that neither intervention had the desired effect on alcohol-related calls for service in downtown San Marcos. The mean level of complaints, for each of the outcome series, increased after the execution of the dual initiatives (see column two of Table 1), while the return of the preintervention number of patrol officers led to a further increase in the mean level of complaints for three of the four outcome series (see column three, Table 1).

Since the increase in the number of patrol officers was limited to the downtown area, we report the mean alcohol-related complaints for the remainder of the city for two time periods. Thus, column one of Panel B reports the mean level of complaints for each of the outcome series prior to the implementation of the 2-hr extension of alcohol sales, while column two presents those for the postintervention observations.

The effects of the city council ordinance for the remainder of San Marcos are mixed. The extension of operation hours for alcohol serving businesses had no appreciable effect on the level of verbal disturbances, led to an increase in the mean level of public intoxication complaints, but led to a decrease in the mean level of physical disturbances and driving under the influence of alcohol (DUI) complaints (compare columns one and two).

Taken together, the descriptive analyses would seem to indicate that the dual interventions failed to mitigate conflicts between bar patrons and other members of the community. However, such a conclusion must be viewed with caution. It is possible that the reported changes in the mean level of verbal and physical disturbances, as well as incidents of public intoxication and driving under the influence of alcohol may be confounded by ongoing causal processes that began prior to the implementation of the extended hours ordinance and/or the temporary increase in the number of patrol officers in the downtown district. Therefore, we turn to the ARIMA models, which explicitly take into account, by means of the prewhitening procedure, any ongoing systematic processes that could confound the influence of the interventions on each of the dependent series (McCleary & Hay, 1980).

Interrupted Time Series Analyses

To recap, the primary impetus for the passage of the ordinance extending the sale of alcoholic beverages from midnight to 2 a.m. was to reduce the number of disturbance complaints stemming from the continued consumption of alcohol at after-hours house parties, especially, but not exclusively, in the downtown district of San Marcos (City of San Marcos, 2010). Therefore, we begin our discussion of the results with a presentation of the interrupted time series analyses for this area of the city.

The results are clear. At least one of the two policy initiatives significantly affects each of the four outcome series. However, contrary to the intent of the local authorities, the implementation of these interventions led to an increase, rather than a decrease, in calls for service in downtown San Marcos.

Consider the interrupted time series analyses for calls for service involving verbal disturbances (see Panel A of Table 2). The first intervention, which captures the joint influence of the extension in operating hours and the increase in patrol strength, positively affects the level of verbal disturbances (ω_1o = .50, p < .01). Note further that the second intervention, which denotes the return of patrol activity to its prior level, is insignificant. Taken together, these findings suggest that the city council ordinance, rather than the change in patrol strategy, led to the increase in citizen complaints about verbal disturbances.

Table 2.

Final Transfer Function Models for Downtown San Marcos, Texas.

Panel A: Verbal disturbances
Model 1: First intervention only
Zero-order transfer function model: LgY_t = ω_1oI_t + b_o
Q = 29.29	df = 26	p < .45
where: I_t = 0 for June 7, 2007 to June 3, 2009, and I_t = 1 thereafter
Coefficient	t-Statistic
b₀ = 0.27	t = 2.51*
ω_1o = 0.50	t = 3.26**
Model 2: Both interventions
Zero-order transfer function model: LgY_t = ω_1oI_t + ω_2oI_t + b_o
Q = 26.84	df = 25	p < .37
where: I_t = 0 for June 7, 2007 to May 26, 2010, and I_t =1 thereafter
Coefficient	t-Statistic
b₀ = 0.27	t = 2.51*
ω_1o = 0.40	t = 2.12*
ω_2o = 0.19	t = 0.88
Panel B: Physical disturbances
Model 1: First intervention only
Zero-order transfer function model: LgY_t = ω_1oI_t + b_o
Q = 21.25	df = 26	p < .73
where: I_t = 0 for June 7, 2007 to June 3, 2009, and I_t = 1 thereafter
Coefficient	t-Statistic
b₀ = −1.29	t = −9.71***
ω_1o = 0.71	t = 3.77***
Model 2: Both interventions
Zero-order transfer function model: LgY_t = ω_1oI_t + ω_2oI_t + b_o
Q = 22.41	df = 25	p < .62
where: I_t = 0 for June 7, 2007 to May 26, 2010, and I_t =1 thereafter
Coefficient	t-Statistic
b₀ = −1.29	t = −9.78***
ω_1o = 0.40	t = 1.98*
ω_2o = 0.19	t = 2.38*
Panel C: Public intoxication
Model 1: First intervention only
Zero-order transfer function model: Lg(1-B)Y_t = ω_1oI_t + 1/(1- θ₁B)
Q = 29.60	df = 26	p < .29
where: I_t = 0 for June 7, 2007 to June 3, 2009 and I_t = 1 thereafter
Coefficient	t-Statistic
θ₁ = −0.87	t = −25.17***
ω_1o = 1.35	t = 2.27*
Model 2: Both interventions
Zero-order transfer function model: Lg(1-B)Y_t = ω_1oI_t + ω_2oI_t + 1/(1- θ₁B)
Q = 29.54	df = 25	p < .25
where: I_t = 0 for June 7, 2007 to May 26, 2010 and I_t =1 thereafter
Coefficient	t-Statistic
θ₁ = −0.87	t = −25.04***
ω_1o = 1.34	t = 2.27*
ω_2o = 0.08	t = 0.12
Panel D: Driving under the influence of alcohol
Model 1: First intervention only
Zero-order transfer function model: Lg(1-B)Y_t = ω_1oI_t + b_o
Q = 22.99	df = 26	p < .64
where: I_t = 0 for June 7, 2007 to June 3, 2009, and I_t = 1 thereafter
Coefficient	t-Statistic
b₀ = −1.59	t = −12.86***
ω_1o = 0.65	t = 3.72**
Model 2: Both interventions
Zero-order transfer function model: LgY_t = ω_1oI_t + ω_2oI_t + b_o
Q = 23.62	df = 25	p < .55
where: I_t = 0 for June 7, 2007 to May 26, 2010, and I_t =1 thereafter
Coefficient	t-Statistic
b₀ = −1.59	t = −12.87***
ω_1o = 0.50	t = 2.34*
ω_2o = 0.28	t = 1.13

Notes: Lg = Natural logarithm transformation. b₀ = Constant. θ = Moving average parameter. B = Backward shift operator. ω_1o = Zero-order input parameter estimate for the first intervention.ω_2o = Zero-order input parameter estimate for the second intervention. Q = Test statistic for the null hypothesis that the model residuals are distributed as white noise. I_t = Intervention series.

p < .05. **p < .01. ***p < .001.

It was necessary to transform the verbal disturbance series by its natural logarithm to induce variance stationarity prior to the estimation of the transfer function equations (see Panel A). Thus, the parameter estimate for the first intervention, ω_1o = .50, is in the log metric and, consequently, is not directly interpretable. However, this coefficient can be expressed as the percent change in the level of the outcome series as a result of the intervention with the following formula: percent change = (e^(ω) −1)100; where e^(ω) is the antilog for the parameter estimate for the zero-order transfer function (McCleary & Hay, 1980, pp. 174-175). Applying this equation to the coefficient for the effect of the first intervention on verbal disturbances reveals that the change in the hours of operation led to a 64% increase in verbal disturbances for the remainder of the period under investigation.

Consistent with the results for verbal disturbances, the first intervention directly affects the quantity of physical disturbances (ω_1o = .71, p < .001). However, unlike what we reported with respect to verbal disturbances, both parameter estimates are found to positively affect the level of physical disturbance complaints (see Panel B). Specifically, the first intervention yields a 54% increase, while the second intervention produces a further increase of 72% in physical disturbances.

Obviously, since the goal of the interventions was to reduce, not increase, physical disturbances, we did not anticipate this pattern of findings. Nonetheless, the interpretation of these results is relatively straightforward. In brief, they indicate to us that the increase in the number of patrol officers helped to suppress some of the deleterious effects of the 2-hr extension of alcohol sales. That is to say, the added police presence probably served to deter some of the bar patrons from engaging in disorderly behaviors. However, when the patrol strength returned to its preintervention level (the second intervention), the full effect of the city ordinance came into force (as indicated by the additional 72% increase in physical disturbance complaints).

Overall, the findings for the alcohol-related outcome series are comparable to those reported for verbal disturbances. Consistent with the interrupted time series analyses of verbal disturbances, the first intervention is associated with an increase in (ω_1o = 1.35, p < .05), while the second intervention has no effect on, the volume of public intoxication complaints. Similarly, the first intervention positively (ω_1o = .65, p < .01) affects the number of DUI complaints, while the return of downtown patrol strength to preintervention levels is unrelated to the level of DUI complaints.

There is one difference worth noting concerning the impact of extending the hours of operation for bars and taverns on the two alcohol-related time series as opposed to its effect on the verbal disturbance time series. The magnitude of the impact of the citywide ordinance on the former two series is substantially greater than its impact on the latter series. Specifically, the influence of extending commercial sales of alcoholic beverages on public intoxication complaints is approximately four and one-half times greater than it is for verbal disturbance complaints (285% vs. 64%). Its relative impact on DUI complaints is smaller but still substantial, approaching one and one-half times greater than its influence on verbal disturbance complaints (92% vs. 64%).

Although the extension of the hours of operation for bars and taverns did not produce a reduction in calls for service within the downtown area, it is still possible that it did achieve its objective in the remainder of the city. To determine the differential impact of the citywide ordinance on the downtown subsection and the rest of city, we examined the influence of the onset of the citywide alcohol sales ordinance on the four outcome series for the remainder of San Marcos. Since the increase in patrol strength was limited to the downtown area, there is no reason to conduct a second analysis to model the termination of this administrative policy. Hence, Table 3 contains the results from ARIMA analyses of the single intervention, the citywide ordinance, on each of the dependent series.

Table 3.

Final Transfer Function Models for the Effects of the Citywide Ordinance on the Remainder of San Marcos, Texas.

Panel A: Verbal disturbances
Zero-order transfer function model: LgY_t = ω_1oI_t + b_o
Q = 32.22	df = 26	p < .19
where: I_t = 0 for June 7, 2007 to June 3, 2009 and I_t = 1 thereafter
Coefficient	t-Statistic
b₀ = 3.04	t = 18.25***
ω_1o = −0.01	t = −0.22
Panel B: Physical disturbances
Zero-order transfer function model: Y_t = ω_1oI_t + b_o
Q = 23.43	df = 26	p < .61
where: I_t = 0 for June 7, 2007 to June 3, 2009, and I_t = 1 thereafter
Coefficient	t-Statistic
b₀ = 8.48	t = 29.27***
ω₁ = −0.29	t = −3.04**
Panel C: Public intoxication
Zero-order transfer function model: Lg(1-B)Y_t = ω_1oI_t + 1/(1- θ₁B)
Q = 30.60	df = 26	p < .25
where: I_t = 0 for June 7, 2007 to June 3, 2009 and I_t = 1 thereafter
Coefficient	t-Statistic
θ₁ = −0.90	t = −27.34***
ω_1o = 1.86	t = 1.60
Panel D: Driving under the influence of alcohol
Zero-order transfer function model: Y_t = ω_1oI_t + b_o
Q = 32.12	df = 26	p < .19
where: I_t = 0 for June 7, 2007 to June 3, 2009 and I_t = 1 thereafter
Coefficient	t-Statistic
b₀ = 0.84	t = 11.79***
ω_1o = 0.60	t = 6.01***

Notes: Lg = Natural logarithm transformation. b₀ = Constant. θ = Moving average parameter. B = Backward shift operator. ω_1o = Zero-order input parameter estimate for the intervention. Q = Test statistic for the null hypothesis that the model residuals are distributed as white noise. I_t = Intervention series.

p < .05. **p < .01. ***p < .001.

The results are disappointing. There is no detectable pattern to the findings presented in Table 3. For the remainder of the city, the 2-hr extension of commercial sales of alcoholic beverages generates a small increase of approximately one additional DUI complaint every 2 months (ω_1o = .60, p < .001), a decline of a single physical disturbance complaint every 3 months (ω_1o = −.29, p < .01), and has no effect on either verbal disturbance or public intoxication complaints.

In sum, the interrupted time series analyses indicate that neither the city council’s nor the police department’s strategies for mitigating some of the negative consequences of the consumption of alcoholic beverages in bars and taverns were successful.

Discussion

In order to decrease late night and early morning complaints arising from after-hours parties, the city council of San Marcos Texas adopted an ordinance that allowed local bars to extend their hours of operation from midnight until 2 a.m. Prior to the implementation of the ordinance on June 4, 2009, bar patrons would leave the bars at midnight and then disperse to various locations around town to continue drinking alcoholic beverages.

The extended hours ordinance was part of a combined strategy to decrease disturbances and to ease tensions between college students (the perceived source of the disturbance problem) and more permanent members of the community. Concomitant to the 2-hr extension of alcohol sales, the police increased, for approximately one year, downtown patrols from two officers to four or five officers at closing time (George, 2010b). The expectation was that coupling the extended hours initiative with an increase in the number of police in the downtown area would result in a reduction of alcohol-induced calls for service throughout the city (City of San Marcos, 2010; Miller, 2008).

Clearly, the objectives of the city council and police department were not met. Neither the 2-hr extension of operations for bars and restaurants nor the increase in patrol strength in the downtown district served to lessen or contain calls for service relating to the closing of drinking establishments. Rather, the results from the interrupted time series analyses indicate that the citywide ordinance in all likelihood exacerbated the problem, increasing the volume of calls for service with respect to public intoxication in the downtown area as well as the number of driving under the influence of alcohol complaints throughout the city. Increasing the size of the police patrols in the downtown area from two officers to four or five had no appreciable impact on calls for service involving verbal disturbances, public intoxication, or driving under the influence of alcohol.

To be sure, not all of the findings were antithetical to the goals of the local authorities. As we noted above, the increase in patrol strength evidenced a small, but significant, deterrent effect on physical disturbances within the downtown area, while the citywide ordinance had a negative impact on physical disturbances in the remainder of the city. Nonetheless, the overall pattern of findings is most consistent with the conclusion that the dual interventions failed to achieve their primary objectives: the reduction, if not the outright elimination, of some of the more onerous problems associated with the retail sales of alcoholic beverages.

One should not infer from these results that the underlying reasoning that led to the extended hours ordinance is inherently flawed. It is possible that the ordinance failed to achieve its larger purpose (the reduction of alcohol-induced calls for service) because it accomplished its more immediate one (the reduction of the consumption of alcohol at after-hours parties throughout the city).

As might be expected, the relaxation of the hours of operation for bars and taverns generated an appreciable increase in the sale and consumption of alcoholic beverages (George, 2010a). While this may have led to a decline in the number of after-hours parties, we suspect that it simultaneously led to an increase in the number of intoxicated individuals needing to leave the downtown area and return to their respective homes at 2 a.m. (instead of at midnight). If such were the case, it should be of little wonder that the statistical analyses indicate that the overall effect of the ordinance was to increase, rather than decrease, drinking-related calls for service.

At the risk of being paternalistic, we believe that it is naive to expect individuals to act responsibly after they have been allowed to drink intoxicating beverages for an additional 2-hr. This is not to say that the bar extension ordinance cannot help minimize conflicts, arising from after-hours parties, between bar patrons and local residents. What we are suggesting is that this strategy needs to be supplemented by an effective means of returning bar patrons to their residences after closing time. Specifically, we propose that the city, perhaps in conjunction with the university, establish an evening and late night bus service to and from the downtown bars and taverns. While it may be unrealistic to expect people to make good choices about getting home after a night of drinking, they may be willing to make more prudent decisions about transportation prior to beginning the evening’s festivities.

Footnotes

Acknowledgements

We would like to thank our colleague Beth Sanders for suggesting that we pursue this project.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

Author Biographies

Mitchell B. Chamlin is Professor of Criminal Justice at Texas State University-San Marcos. His research focuses on elaborating and falsifying macro-criminological theory.

Sarah E. Scott is a Doctoral Teaching Assistant at Texas State University-San Marcos. Her research interests included evaluating programs related to substance use and abuse, particularly substance abuse treatment programs in the criminal justice system.

References

Bouffard

Bergeron

Bouffard

(2007). Investigating the impact of extended bar closing times on police stops for DUI. Journal of Criminal Justice, 35, 537-545.

Chamlin

Myer

Sanders

Cochran

(2008). Abortion as crime control. Criminal Justice Policy Review, 19, 135-152.

Cochran

Chamlin

Seth

(1994). Deterrence or brutalization? An impact assessment of Oklahoma’s return to capital punishment. Criminology, 32, 107-134.

City of San Marcos. (2010). Report on extended bar serving hours. San Marcos, TX: H. Williams & J. Barger.

George

(2010a, January 24). San Marcos bars benefit from later hours, but more arrests seen. Austin American Statesman. Retrieved from http://www.statesman.con/mews/local/san-marcos-bars-benefit-from-later-hours-but-193649.html

George

(2010b, September 26). San Marcos DWI arrests up since bar hours changed. Austin American Statesman. Retrieved from http://www.statesman.com/news/local/san-marcos-dwi-arrests-up-since-bar-hours-939137.html

Granger

(1980). Testing for causality: A personal viewpoint. Journal of Economic Dynamics and Control, 2, 329-352.

Holmes

Daudistel

Taggart

(1992). Plea bargaining policy and state district court caseloads: An interrupted time series analysis. Law & Society Review, 26, 139-160.

Katz

Webb

Schaefer

(2001). An assessment of the impact of quality-of-life policing on crime and disorder. Justice Quarterly, 18, 825-876.

10.

Langworthy

(1989). Do stings control crime? An evaluation of a police fencing operation. Justice Quarterly, 6, 29-45.

11.

Langworthy

LeBeau

(1992). The spatial distribution of sting targets. Journal of Criminal Justice, 6, 541-551.

12.

Loftin

Heumann

McDowall

(1983). Mandatory sentencing and firearms violence: Evaluating an alternative to gun control. Law & Society Review, 12, 287-318.

13.

McCleary

Hay

Jr. (1980). Applied time series analysis for the social sciences. Beverly Hills, CA: SAGE.

14.

McDowall

McCleary

Meidinger

Hay

Jr. (1980). Interrupted time series analysis. Beverly Hills, CA: SAGE.

15.

Miller

(2008, September 25). Bar hours: How late? San Marcos Daily Record. Retrieved from http://www.sanmarcosrecord.com/local/x1169225578/Bar-hours-how-late

16.

Novak

Hartman

Holsinger

Turner

(1999). The effects of aggressive policing of disorder on serious crime. Policing: An International Journal of Police Strategies & Management, 22, 171-194.

17.

Simpson

Bouffard

Garner

Hickman

(2006). The influence of legal reform on the probability of arrest in domestic violence cases. Justice Quarterly, 23, 297-316.

18.

Singer

McDowall

(1988). Criminalizing delinquency: The deterrent effects of the New York Juvenile Offender Law. Law & Society Review, 22, 521-535.

19.

Smith

(2001). Police-led crackdowns and cleanups: An evaluation of a crime control initiative in Richmond, VA. Crime & Delinquency, 47, 60-83.

20.

Villaveces

Cummings

Espitia

Koepsell

McKnight

Kellermann

(2000). Effect of a ban on carrying firearms on homicide rates in 2 Colombian cities. Journal of the American Medical Association, 283, 1205-1209.

21.

Vingilis

McLeod

Mann

Seeley

(2008). A tale of two cities: The effect of extended drinking hours in licensed establishments on impaired driving and assault charges. Traffic Injury Prevention, 9, 527-533.

22.

Vingilis

McLeod

Seeley

Mann

Voas

Compton

(2006). The impact of Ontario’s extended drinking hours on cross-border cities of Windsor and Detroit. Accident Analysis and Prevention, 38, 63-70.

23.

Vingilis

McLeod

Seeley

Mann

Beirness

Compton

(2005). Road safety impact of extended drinking hours in Ontario. Accident Analysis & Prevention, 37, 549-556.

24.

Vingilis

McLeod

Stoduto

Seeley

Mann

(2007). Impact of extended drinking hours in Ontario motor-vehicle collision and non-motor-vehicle collision injuries. Journal of Studies on Alcohol and Drugs, 68, 905-911.

25.

Webster

Vernick

Hepburn

(2002). Effects of Maryland’s law banning “Saturday night special” handguns on homicides. American Journal of Epidemiology, 155, 405-412.

26.

Weiss

Freels

(1996). The effects of aggressive policing: the Dayton traffic enforcement Experiment. American Journal of Police, 15, 45-64.

27.

White

(2001). Controlling police decisions to use deadly force: Reexamining the importance of administrative policy. Crime & Delinquency, 47, 131-151.

28.

White

(2003). Examining the impact of external influences on police use of deadly force over time. Evaluation Review, 27, 50-78.