convenience store security at the millennium - … · convenience store security at the millennium....
TRANSCRIPT
NATIONAL
ASSOCIATION OF
CONVENIENCE
STORES
R
NACS Resource Book
Convenience StoreSecurity
at the MillenniumRosemary J. Erickson, Ph.D.
Convenience StoreSecurity
at the Millenniumby
Rosemary J. Erickson, Ph.D.President
Athena Research CorporationSan Diego, CA
February, 1998
All rights reserved. This publication may not be reproduced, stored in information or retrieval systems, or transmitted in whole or inpart, in any form or by electronic, mechanical, photocopying, recording or otherwise, without written permission of the publishers.
Printed in the United States of America.
Published by the National Association of Convenience Stores
1605 King StreetAlexandria, Virginia 22314-2792
(703) 684-3600FAX (703) 836-4564
E-MAIL: [email protected]
Other NACS Convenience Store Security Resources
Convenience Store Security:Executive Summary• Crime Distribution in the Industry• Effectiveness of Certain Security Techniques• Analysis of Rape & Homicides in Convenience StoresThe findings of three distinct research projects are captured in this report: the crime census for the industry, the circumstances ofhomicide and rape in the industry and an assessment of deterrence measures such as multiple-staffing, interactive television and bul-let resistant barriers. Published in 1992, this report reveals that 80 percent of all stores are crime-free in any 12-month period––a find-ing that holds true today. Law enforcement and government officials use this report and its findings in their consideration of securitymeasures.
Order #R1018CNACS Member Price: $20Non-Member Price: $50
Gainesville Convenience Store Ordinance:Findings of Facts, Conclusions and Recommendations• Firsthand Study of Gainesville’s Ordinance• The Impact of Other Deterrent Measures• Analysis of Multiple-Staffing EffectivenessNACS commissioned former Washington, D.C., Police Chief Jerry V. Wilson to review the Gainesville, Florida, two-clerk ordinance in1990. His conclusion: the reduction in convenience store robberies in Gainesville was not due to the two-clerk rule but, in fact, thebasic robbery deterrence measures developed by NACS. This report is particularly useful for operators who are facing multiple clerkordinances.
Order #Q1022CNACS Member Price: $20Non-Member Price: $30
The Store Safety Issue:Facts for the Future and Action, Not Reaction• Resource Book with Industry Research• Videotape on Safety Strategies• Reproducible Charts for City Council/
Law Enforcement MeetingsThis collection of facts and figures assists convenience store operators who need to discuss the store safety issue with local and statelawmakers. Includes Action, Not Reaction video and resource book titled The Store Safety Issue.
Order #Q1005CNACS Member Price: $25Non-Member Price: $35
Florida Version – reviews convenience store safety issues specific to Florida.Order #Q1036CNACS Member Price: $25Non-Member Price: $35
To order these materials, contact the NACS Information Center at (703) 684-3600.
i
ABSTRACT
This paper reviews the literature on the state of the knowledge about the use of twoclerks and other security measures to prevent robbery and violence in convenience storesand service stations. Dr. Rosemary J. Erickson, a sociologist, has studied the problem ofconvenience store security for over 25 years and undertook the task in order to drawtogether in one place what is known about the issue. The paper begins with the earlyresearch from the 1970s, conducted by Western Behavioral Sciences Institute and fundedby the Department of Justice, which resulted in Southland's involvement, as the owner andoperator of 7-Elevens. This research led to Southland's implementation of the programand their support of research in the area, including two large scale studies of armedrobbers in prisons conducted by the Athena Research Corporation. The history alsoincludes the role of the National Association of Convenience Stores (NACS) in fundingresearch.
In addition, the research in Florida in the area of the relative effectiveness of twoclerks as a safety measure is discussed. Also, the research conducted by the government,particularly the National Institute for Occupational Safety and Health (NIOSH), ishighlighted, as well as the legislative efforts on the state and federal levels, includingFederal OSHA. In addition to research and legislative initiatives, implementation effortsare included. The report also discusses the efforts by academics to evaluate and analyzethe results of the studies. The results of research are surprisingly consistent on theeffectiveness of certain measures in reducing robbery and violence, including:
• cash control• escape routes• lighting• visibility, and• training.
There is less agreement on the deterrent value of cameras, however, and moreresearch needs to be done to establish the relative effectiveness of other factors,including:
• two clerks• bullet-resistant barriers• guards, and• closing at night.
Results reported include the fact that since the risk of injury is apparently notreduced with one versus two clerks on duty, then even if robberies are reduced by havingtwo clerks, it is not enough to offset the number who would be injured, by having two clerkson duty. Research that is currently underway, including the UCLA study, funded by NIOSH,
ii
is also described, and the importance of establishing the relationship between crime in theneighborhood with homicide and prior robberies with homicide is discussed.
PREFACE
This report brings together the information on two clerks and other securitymeasures in convenience stores. It is not about the victims of crimes in conveniencestores, their families, or the efforts of victim's rights groups, working on their behalf. It israther about the scientific research on the subject. Such scientific reporting is in no wayintended to take away from the terrible suffering that victims, their families, their coworkers,managers and store owners have suffered when a tragic event has occurred. It is notalways easy to explore the issue from both sides and to wait for the science. Withoutquestion, all parties involved are united in their desire to stem the violence, but the meansof reaching that goal are where differences have occurred. This report is about thosedifferences.
This paper was reviewed by several individuals, including Dr. John Howard, M.D.,Chief of the Division of Occupational Safety and Health, Department of IndustrialRelations, State of California; Dr. Corrie Peek-Asa, Southern California Injury PreventionResearch Center at the UCLA School of Public Health, where she is co-director of theNIOSH-funded Workplace Violence Prevention Project; Dr. Harlan Amandus, SeniorResearch Scientist Batelle Institute in Arlington, Virginia; and Dr. Robert Figlio, Professorof Sociology, University of California, Riverside.
iii
CONTENTS
Abstract ............................................................................................................................ iPreface ............................................................................................................................ iiContents ......................................................................................................................... iiiList of Tables .................................................................................................................. ivINTRODUCTION .............................................................................................................1RESEARCH AND IMPLEMENTATION............................................................................3
WBSI Research and Southland Implementation ..................................................3WBSI Research .........................................................................................3Southland Implementation .........................................................................5
The Athena Studies ..............................................................................................5NACS Research and Implementation ...................................................................9
NACS Research.........................................................................................9NACS Implementation..............................................................................11
Florida Research and Implementation................................................................12Florida Research......................................................................................12Florida Implementation ............................................................................15Evaluation of Gainesville .........................................................................16
Government Research........................................................................................18NIOSH Surveillance and Conferences.....................................................18NIOSH Nine-State Study..........................................................................22NIOSH Virginia Study ..............................................................................23
LEGISLATIVE AND GOVERNMENT INITIATIVES .......................................................24Cal-OSHA Model Program..................................................................................24Federal-OSHA Draft Guidelines .........................................................................26
Internal vs. External Violence ..................................................................28Larger Social Context ..............................................................................28Validated vs. Unvalidated Research ........................................................29
Other Federal Initiatives .....................................................................................30FBI ...........................................................................................................30Congress..................................................................................................30
THE STATE OF THE KNOWLEDGE ............................................................................32Robbery vs. Homicide.........................................................................................32Customers vs. Clerks..........................................................................................33Exposure Data ....................................................................................................33Overview.............................................................................................................34
Robbers Who Injure .................................................................................36RESEARCH UNDERWAY.............................................................................................38
UCLA Study ........................................................................................................38CONCLUSIONS ............................................................................................................40REFERENCES ..............................................................................................................42INDEX............................................................................................................................47
iv
ABOUT THE AUTHOR..................................................................................................50
List of Tables
Table 1 Security Measures Related to Robbery: WBSI & Athena Studies ............8Table 2 Homicide Motives.....................................................................................10Table 3 Security Measures Related to Robbery: NACS Studies .........................11Table 4 Security Measures Related to Robbery: Florida Studies ........................15Table 5 Robbery and Homicides at Convenience Stores in Florida .....................18Table 6 Security Measures Related to Robbery: Overview .................................37
1
INTRODUCTION
Two clerks . . . two dead people.(A homicide detective, when asked the effect of having two clerks on duty, instead of one)
Two clerks . . . make me feel safer.(Almost any clerk you talk to)
This paper is for the seriousstudent of the two-clerk issue and otherconvenience store security measures. Itis not for someone whose mind is madeup on the issues, but for someone who isinterested in the etiology of one of themost divisive issues--the two-clerk issue--faced by the convenience store industry,government, and the public in the pasttwo decades. The history of two clerks isreally the history of a schism betweengovernment and industry; between thepublic sector and the private sector; andbetween employers and employees.
The contentious nature of the fightbegan with Gainesville in the 1980s. Prior to that time, government andindustry were actually working together tosolve the problem of crime inconvenience stores, a collaboration whichhad begun in the early 1970s, with theWestern Behavioral Sciences Institute(WBSI) study, funded by the NationalInstitute of Justice, and characterized byits unique cooperation of government,industry and a non-profit researchorganization. More will be said about theWBSI study and about the Gainesvillestudy, but the point to be made here is
that up until the time of the Gainesvillestudy, the effort could be characterizedas largely cooperative, not conflictual.The two-clerk issue, more than any othersecurity issue, has separated the partiesinvolved. The goals of this paper are to:
• review the history of the issue;
• describe the research done by industryand government;
• discuss the legislative initiatives; and
• outline future directions.
At the risk of overstating andoversimplifying the issues, the divisionsare basically these:
vv Employees (along with victim's rightsgroups), for the most part, believe that industrydoes not want two clerks because of the cost.
vv Industry, for the most part, does not believein putting on two clerks because they do notknow if it actually increases safety.
vv Government, as such, is not clearly on oneside or the other, but there have beenlegislative efforts at all levels--city, county,state and federal--to mandate two clerks.
2
Curiously, in legislative efforts torequire two clerks, the issue sometimeswins and sometimes loses, so it is by nomeans unanimously viewed as a positivesolution to the problem of security inconvenience stores when put to the vote. But first some definitions and then, in thenext section, the history of the two-clerkissue in the context of convenience storesecurity will be discussed.
The industry refers primarily to theconvenience store industry, though withthe combining of service stations andconvenience stores in the 1980s, thedistinctions are often blurred. Estimatessuggest that convenience stores in thiscountry number around 90,000.
The two-clerk issue refers to theidea, often recommended, of having twoclerks on duty, rather than one, especiallyduring the late night hours fromapproximately 11 P.M. to approximately 5A.M. in convenience stores and servicestations.
3
RESEARCH AND IMPLEMENTATION
As early as the 1970s,convenience store clerks wererecognized as being at risk for violence atwork. From 1968 to 1973, robbery inchain stores (not just convenience stores)had increased by 167%, while therobbery rate had gone up 39% in generalduring the same time period (Crow andBull, 1975). As a result, the conveniencestore industry faced a combination ofchallenges at that time:
n an increase in robbery;
n a reduction in other targets, sincepublic transportation had moved to a"no-change" rule to reduce theirrobberies;
n increased exposure due to having moreconvenience stores; and
n increased exposure, due to increasinghours from the original hours of 7 A.M.to 11 P.M. to being open 24-hours aday.
WBSI Research and SouthlandImplementation
To study the problem, a grant wassought, and received, by the WesternBehavioral Sciences Institute (WBSI)from the Law Enforcement AssistanceAdministration (LEAA), National Instituteof Justice (NIJ), with cooperation from theSouthland Corporation, parent companyof 7-Elevens.
WBSI Research
The WBSI team included socialscientists and ex-armed robbers, and thepurpose of the research was to test outnew techniques to prevent robbery andviolence. With innovative techniques,including behavioral changes, physicalchanges and employee training, theintervention measures were implementedin 60 experimental stores, which wereclosely matched, on a stratified randombasis, with 60 control stores. In acombined effort with industry andgovernment, the stores selected for studywere the 7-Eleven stores.
Southland had been using ex-armed robbers in training seminars formanagement and employees. This projectprovided the opportunity to test outdeterrence methods through a classicexperimental design in a field setting. Itwas, and remains, the only such large-scale experiment on the subject that hasbeen done to date. The approachincluded the unexplored strategy ofemphasizing prevention by alteringconditions at the scene of the crime, sothat the robber would not attempt therobbery. Those physical changes, nowsomewhat common, were new at that timeand included the following programcomponents:
qq reducing the amount of cash available;
4
qq posting signs regarding low cash;qq clearing the windows to make the
register visible;
qq repositioning the cash register up front;
qq increasing lighting, inside and out;
qq eliminating escape routes; and
qq training the employees in violence androbbery prevention techniques.
For the study, the rationale behindthe procedures was to look at the storesfrom the robbers' point of view and thendevise countermeasures to dissuadethem (Crow and Bull, 1975). Somepreliminary consideration on the part ofthe robber is, therefore, assumed. At thescene, the characteristics of the site andthe behavior of the store personnel, canlead the robber to reconsider and theneither 1) proceed with the robbery; 2) wait until conditions become favorable forthe robbery; or 3) give up the robbery andleave the site. The intent of theprevention procedures is to build into thesite those characteristics possessed bystores that are seldom robbed and toeliminate features found in stores that arefrequently robbed.
Certain procedures (good externallighting and clear windows) provide thestore with an external image which candiscourage a potential robber. Otherfeatures (an alert clerk or a blockedescape route) are designed to furtherinhibit a robber's plans. Still otherfeatures (signs posted in the store anddirect verbal and non-verbal
communication from the clerk) aredesigned to influence those who actuallyenter the store with the intention ofrobbing it. From the robber's perspective,an ideal robbery in a convenience storewould include the followingconsiderations:
rr Be sure there is money available.
rr Optimize the take\risk ratio.
rr Be persuasive.
rr Avoid disruptions.
rr Get the money quickly.
rr Avoid being seen by anyone but thevictim.
rr Avoid being recognized.
rr Get away quickly and easily.
From the point of view of robberyprevention, the countermeasures weredevised to address the following:
tt Persuade the robber there is littlemoney available.
tt Maximize the perceived risks for therobber.
tt Maximize the probability of the robberybeing witnessed.
tt Convince the robber he may berecognized.
tt Alter escape routes or provideobstacles to a quick and easy exit.
At WBSI, Dr. Erickson coordinatedthe project data collection and dataanalysis, and two points are particularly
noteworthy about that time. In all of theproject meetings, discussions with thedepartment of Justice representatives or
5
others did not include the idea of testingtwo-clerks as a deterrence measure.Secondly, the concept of CrimePrevention Through EnvironmentalDesign (CPTED) or EnvironmentalDesign (ED) (Jeffery, 1971), was notdiscussed. The WBSI study design wasa separate development, unrelated to ED. Even though the project would later becalled the first test of CPTED, it was infact coincidental with CPTED.
The WBSI experiment, institutingthe robbery and violence preventionmeasures, resulted in a 30% reduction inrobberies in experimental stores, overcontrol stores, during the experimentalperiod. There were significantly fewerrobberies in the experimental stores thanin the control stores. The resultssupported the concept that robbers do infact select their targets, and that physicaland behavioral changes at the site cansignificantly reduce robberies.
Southland Implementation
While the experiment wassuccessful, what remained to be seenwas whether the results of the experimentcould be applied with equal success on alarge scale over time. The SouthlandCorporation cooperated in a one-yearfollow-up in Houston, their most difficultgeographic area at that time.
After the year of implementation,the Southland Corporation decided toadopt the program company-wide in theirstores, numbering over 5,000--a numberwhich would eventually grow to over7,000 stores. Security managers wereadded to aid in the implementation and
the monitoring of the program, andcomputerized systems for tracking lossprevention data were instituted.
After 12 years of implementing therobbery deterrence program in the field,7-Eleven stores experienced a 65%decrease in robberies nationwide from1975 to 1986 (Crow, Erickson and Scott,1987). That decrease has held for twenty years, in spite of the fact thatrobbery has gone up elsewhere duringthat time period (Erickson, 1996c). Afterthe WBSI experiment and subsequentSouthland implementation, it wasconcluded that the five most importantfactors for reducing robberies inconvenience stores were the following:
_ Money
_ Escape Route
_ Anonymity
_ Interference
_ Police Patrols
The Athena Studies
Ten years after the original WBSIstudy was completed, the study wasupdated by conducting interviews witharmed robbers to see if the same factors,above, were still the most important torobbers when selecting a site to rob. Because armed robbers were usedoriginally in the 1970s in selecting thecomponents for the program, they wereused again as a source of information for
6
the updated research (Crow, Ericksonand Scott, 1987).
With a grant from the SouthlandCorporation to Athena ResearchCorporation (Athena), interviews wereconducted in five state prisons--Illinois,California, New Jersey, Louisiana andTexas. By this time, Crow and Ericksonhad left WBSI and formed their owncompany--Athena. In the 1985 survey,robbers were asked, among other items,to rate eleven factors as to how importanteach factor was in the decision to rob.The results, listed below, revealed thatthe factors were ranked in this order, withthe amount of money being the mostimportant to the robbers and videorecording systems the least important:
1. Amount of money
2. Escape routes
3. Anonymity
4. Interference
5. Active police patrol
6. Armed clerk
7. Number of clerks
8. Number of customers
9. Camera system
10. Alarm system
11. Video recording system
Because the results showed thetop factors to be essentially the same asat the time of the original study, nosubstantive changes were made to theoriginal intervention program. In otherwords, the intervening decade had notappreciably changed what robbers looked
for when robbing a store in the 1980s. Dr.Crow's early research in the area wasbrought to an end with his death in 1989,but the research continued. In an effortto respond to the changes of the decadefollowing the 1985 prison study, the studywas repeated again in 1995.
With support from the SouthlandCorporation, Athena conducted a surveyof robbers to determine what they look forwhen robbing a store. Armed robbers,recently incarcerated, were interviewed,as before, to obtain their most currentthinking. This time, the study wasconducted in 20 prisons in three states--Washington, Maryland and Texas(Erickson, 1996b; Erickson and Stenseth,1996).
In this study, 310 armed robbers,from street muggers to bank robbers, toldwhy they do it, how they do it and whypeople get hurt. Findings are comparedwith the earlier prison survey. The 1985prison study had been criticized for notseparating out the findings forconvenience store robbers, so specialattention was given to how conveniencestore robbers differ from other robbers. Some of the primary differences betweenconvenience store robbers and otherrobbers are that:
tt Convenience store robbers commit a lotmore robberies (13 compared to 8, onaverage).
tt They are more likely to live closer to thesite.
tt They are less likely to hurt someone.
7
tt They are not as easily deterred.
The top two considerations for allof the robbers in the survey in 1995 whenthey plan to rob a convenience store were1) the escape route, and 2) the amountof money. These were the same top twoconsiderations as in 1985, but in reverseorder. Three new deterrence items wereadded to the survey list in 1995. Theywere armed guards, which were rankedthird by the robbers; bullet-resistantbarriers, which were ranked eighth; andunarmed guards which were ranked last.
One-half of the top deterrencemeasures that they consider importanthave to do with being armed (armedguard, police or armed clerk), as shownon the list, rank-ordered below. The newitems on the list are followed by anasterisk:
1. Escape routes
2. Amount of money
3. Armed guard*
4. Anonymity
5. Police
6. Armed clerk
7. Interference
8. Bullet-resistant barrier*
9. Number of clerks
10. Alarm
11. Number of customers
12. Camera
13. Video
14. Unarmed guards*
Having two clerks ranked in thebottom half of the list, along with bullet-resistant barriers, cameras, alarms andvideos. The number of people in thestore is apparently of little concern to therobbers, primarily because of their use of guns. With a gun, 81% of the robberssay that they would take on two or morepeople. With a partner and a gun, 86%say they would take on two or morepeople, or an average of 11 people. Findings from other research indicate thatguns are used in at least three-fourths ofthe robbery\homicides in commercialrobberies (Erickson, 1991 and 1995a;NIOSH, 1996).
An important difference betweenthe 1985 and the 1995 robbers is that ittakes more money now than ten yearsearlier for a robber to be willing to rob alocation. In 1985, 90% would rob for$150; in 1995, only 40% were willing torob for that amount. The problem is thatthe robbers expect $200, on average,from a convenience store, so 45% wouldbe willing to rob anyway, thinking they willget that much.
Robbers rob primarily for themoney, but they also rob for otherreasons, including the thrill, power,anger, peer-pressure or being high. Their advice to victims included:
tt giving up the money
tt keeping hands in sight
tt not resisting
tt not making any sudden moves
8
tt not talking
tt not staring
tt not trying to be a hero.
Long-term deterrence measures,such as being identified by cameras,appeared to have little impact on therobbers. Their concern, they said, is withgetting caught, trapped or shot at the timeof the robbery. Eighty-three percent ofthose surveyed did not think they wouldbe caught. This was in spite of the factthat they were in prison now, and almosthalf had been in prison before. Thismeans that efforts to lengthen prison timeor make the time harder will have littledeterrent value because the robbers donot think they will be caught anyway. Infact, only 25% of robbers are caught (FBI,1996).
Since the amount of money is soimportant to the robbers, they have tobelieve that they will not get enoughmoney to make it worth their while, andthe risk has to be great enough to insurethat it is not worth their while. Implications of the research are thatfuture efforts need to be on:
uu reducing the money;
uu making sure the robber knows that there is less money before he robs;
uu blocking escape routes; and
uu increasing the risk to the robbers ofgetting caught.
Table 1 summarizes the results ofthe WBSI and Athena studies. On thefirst five measures listed--cash control,escape routes, lighting, visibility andtraining--a positive relationship wasestablished in all three studies. The
other measures listed either were notexamined or did not result in a positiverelationship, The exception is armedguards which showed a positiverelationship in the 1995 survey of armedrobbers.
Table 1: Security Measures Related To Robbery
WBSI & Athena Studies
Measures Crow &Bull1975
Crowet al.1987
Erickson1996
cash control + + +
escape routes + + +
lighting + + +
visibility + + +
training + + +
# of clerks
# of customers
cameras\video
alarms
armed guards +
unarmed guards
BRBs
location
hours
prior robberies
9
mirrors
selling gas
+=positive relationship
The WBSI and Athena research,as well as Southland's implementation ofthe measures, have indicated theimportance and success of theseparticular security measures in the stores.NACS Research andImplementation
While research efforts werecontinuing by Athena and Southland, theNational Association of ConvenienceStores (NACS) undertook their ownresearch and implementation asdiscussed in the following sections.
NACS Research
Representing convenience storesnationwide, NACS began funding theirown research efforts in the late 1980swith three studies, conducted byresearchers Schreiber (1991), Figlio(1991), and Erickson (1991). The firstwas a study to establish the base crimerates and distribution patterns forhomicide, sexual assaults, robbery andother crimes of violence in conveniencestores (Schreiber, 1991). Study resultsestimate that approximately 100 workersa year are murdered while working inconvenience stores (Schreiber, 1991).The second study analyzed the effects ofmultiple clerks, bullet-resistant barriersand interactive television on robberyrates in convenience stores (Figlio,1991).
A third study, which wasconducted by Erickson (1991), examinedthe risk and circumstances of homicide toconvenience store employees nationwide.The latter study was based on 79homicides which occurred in 1989 and1990. Information was obtained from thestores and companies where thehomicides were committed and frominterviews with the homicide detectives.The findings included a number of factswhich were not previously known aboutconvenience store homicide.
It was found, for example, thatthere was more than one person presentin almost one-third of the homicides, notincluding the multiple perpetrators, andthere was more than one perpetrator inalmost half (46%) of the cases. All of theperpetrators were male. The majority ofthe cases appeared to be gratuitous andsenseless in nature with no signs ofresistance. A previous study, in Chicago,found that resistance accounted for 82%of commercial robbery killings, andvictims who resisted were 49 times morelikely to be killed than those whocooperated (Zimring and Zuehl, 1986).
There was evidence of resistancein only 16% of these cases, no doubtbecause the clerks have been trained togive up the money and not resist. The
10
majority were found to be stranger-to-stranger, robbery-related homicides. Infact, two-thirds of the homicides wererobbery-related, which is fairly consistentwith NIOSH statistics. In 1993, 75% of allworkplace homicides were robbery-related, compared to 9% in the generalpopulation (NIOSH, 1996). Other findingsfrom the homicide study included(Erickson, 1991):
nn In almost half of the robbery homicides, less than $50.00 was
taken.
nn Sixty-nine percent of the murdersoccurred between 11 P.M. and 7 A.M.
nn Handguns were used in 71% of thecases.
Homicides were classified bymotives or causes, based on reports andinterviews with homicide detectives,according to the categories found inTable 2 (Erickson, 1995a and 1996a).
11
Table 2: Homicide MotivesN=79
Motives Number Percent
Gratuitous\Robbery 36 46
Gratuitous\Non-Robbery 15 19
Accidental\Resistance\Robbery
13 16
Acquaintance\Non-Domestic
7 9
Acquaintance\Domestic 5 6
Don't Know 3 4
Total 79 100
Almost two-thirds of the homicideswere gratuitous in nature. Only 15% wereacquaintance-related.
A follow-up study, supported byNACS, conducted by Erickson with CAPIndex, is seeking to determine if a patternexists with convenience store homicides,based upon the:
_ characteristics of the stores;
_ characteristics of the neighborhood;
_ number of previous robberies.
Previous research has shown thatpast robberies may predict futurerobberies, but it is not known whetherrobberies predict homicide (Crow andBull, 1975; Erickson, 1991; Schreiber,1991). That is, if a store has had morethan one robbery, they may haveadditional ones. If they have not had anyrobberies, they probably will not have. Infact, nearly 80% of stores, in a particular
year, do not have any robberies. In 1990,only 13% of convenience storesexperienced one robbery, and only sevenpercent had two or more that year(Schreiber, 1991). The research questionis whether prior robberies lead to murder;that is, whether a location is more likely tohave a murder if it has experiencedrobberies at that location. At this time,data are insufficient on whether robberiespredict homicide.
Characteristics of the store includewhat security measures they have ineffect, whether they have the basicrobbery deterrence program in place,including: clear windows, lighting, lowcash, signage, and employee training. Inaddition to that, other storecharacteristics would include whetherthey have video systems, alarms, guards,bullet-resistant glass or multiple-staffingand whether they close at night.
12
The characteristics of theneighborhood include crime in the areaand measures of social disorganization,such as that used in the CAP Index(Figlio, 1991). This includes the majorcrimes of homicide, rape, robbery,aggravated assault, burglary, larceny andmotor vehicle theft, along with severaldemographic and other economiccharacteristics, including physicalhousing characteristics, and populationmobility.
The 1991 Figlio study, supportedby NACS, found that two clerks on duty,particularly during the night shift, had apositive effect on the robbery rate inpreviously robbed stores. There was alsosome evidence that closed circuittelevision (CCTV) with a monitor in thefront, where potential perpetrators andothers can see themselves, may showpromise as a robbery deterrent. In apreliminary study, with one year ofresults, a reduction in robbery was found. Based on 81 stores, the robbery ratesdecreased from 1.27 robberies per storeper year to 0.59 per stores a year later,representing a 54% one year robbery ratereduction (Figlio, 1991). In general,cameras have not, in the past, proven tobe a deterrent (Erickson, and Crow, 1980;Crow and Erickson, 1984).
Data concerning bullet-resistantbarriers were insufficient, in the Figliostudy, because fewer than 1% of thelocations had them. A follow-up to thatstudy is currently underway, by Figlio,with continued funding by NACS.
Table 3 summarizes the results ofthe three studies that have been
undertaken by NACS, all of which founda relationship between prior robberiesand subsequent robberies. The positiverelationship indicates that the measuremay decrease robbery.
Table 3: Security Measures Related To Robbery
NACS Studies
Measures Schreiber1991
Figlio1991
Erickson1991
cash control
escape routes
lighting
visibility
training
# of clerks +
# of customers
cameras\video +
alarms
armed guards
unarmed guards
BRBs
location
hours
prior robberies * * *
mirrors
selling gas
+=positive relationship*=relationship
NACS Implementation
13
Twelve years after the NIJgovernment-funded study by WBSI waspublished, NACS introduced the robberyand violence prevention program to itsmembers. The program was largelybased on the WBSI study results andSouthland's experience (Hunter andJeffery, 1992; Reiss and Roth, 1993). Southland had already made theirmaterials available throughout thepreceding decade and shared them withthe industry, but the NACS publicationmade the program and materialsavailable on a wider scale.
While there are no statisticsavailable on the extent of theimplementation of the programnationwide, it is generally believed thatthe large chains of convenience storesand service stations do implement thebasic robbery and violence deterrenceprogram and the recommended securitymeasures. However, even ten years afterNACS introduced the program and over20 years after the initial research was firstpublished on the security measures, theimplementation is by no means universal.
NACS continues their support ofresearch to understand the problem; theyupdate their training materials regularly;and they make the training materialsavailable to the industry.
Florida Research andImplementation
At the same time that the researchand implementation efforts were beingundertaken by WBSI, by Athena, byNACS and by Southland, as discussed in
the previous section, a major state-wideeffort was underway in Florida. The city ofGainesville, Florida, and the state ofFlorida, along with researchers in theirstate universities, were working hand-in-hand on research and legislation effortsfor convenience store security asdiscussed in the next section.
Florida Research
In the Annual Review of PublicHealth, Kraus, Blander and McArthur(1995) evaluated the research, includingthat done on convenience store security,and that of the Florida researchers.Similarly, in a proposal to NIOSH forstudying security measures in theworkplace, Kraus (1995), evaluatedresearch efforts on the issue of securitymeasures in convenience stores,including that undertaken in Florida. Thisresearch is summarized in the followingparagraphs.
Swanson
In 1986, Swanson reported oninterviews with 65 inmates in Florida stateprisons and found the most appealingfactors to a robbery included these:
n no customers
n store in a remote area
n only one clerk on duty
n no back room
n female clerks
n no alarm
n easy access\get-away
14
n obstructed windows
n lots of cash
n type of safe
The least appealing factors to therobbers in Swanson's survey were:
qq many customers
qq heavy traffic in front of store
qq two or more clerks
qq a back room
qq male clerk
qq alarms
qq limited escape routes
qq clear visibility into the store
qq stores that sell gasoline
qq one-way mirrors
From these factors, Swansonconcluded that the primary elements indeterring robbery are:
rr work-shifts with more than one clerk onduty
rr limited hours of operation
rr visible security cameras
rr time release safes, and
rr being located near other twenty-fourhour stores.
Duffala
Features concerning the locationas a determinant of vulnerability torobbery were studied by Duffala (1976). While none of the factors alone had asignificant impact on vulnerability, theinteraction of these factors were found tobe significant:
qq being located within two blocks of amajor street
qq light trafficqq few surrounding commercial activities
qq being located in a residential and/orvacant land use area.
Jeffery, Hunter & Griswold
In 1987, Jeffery, Hunter andGriswold examined 34 conveniencestores in Tallahassee, Florida, from anenvironmental perspective and followedthem for four and a half years todetermine store characteristics related torobbery. They then distinguished tensignificant factors for robbery deterrence:
tt location of cash register in the center ofthe store
tt more than one mirror
tt less than two blind corners in the store
tt two or more clerks on duty
tt windows clear of obstructions
tt location off a major street
tt good exterior lighting
tt gas pumps located in front
tt not located near single familydwellings, and
tt heavy vehicular traffic.
Hunter
15
Hunter, in 1988, developed a listof 14 characteristics that are associatedwith vulnerability to robbery thatcomplimented the previous list of Jeffery,
Hunter and Griswold. Added to thisprevious list were:
n concealed access or escape routes
n security devices
n cash handling procedures
n location near other commercialactivities open night hours.
A sample of 110 Florida storeswere surveyed and two variables werefound to be significant at the statewidelevel: 1) having a concealed access, and2) the presence of gas pumps. Goodcash handling procedures werenegatively related to robbery in the study(Hunter, 1988). Despite what seemed tobe a negative finding Hunter laterrecommended good cash handlingprocedures, arguing that when used inconjunction with the other preventiontechniques, cash handling would have apositive effect (Hunter and Jeffery, 1992).
In 1990, Hunter reassessed 26 ofthe original stores and found a generaldecline in robberies of 24%, with somestores showing decreases of as much as86%, while others had increases inrobberies of as much as 50%. Only fiveof the original ten variables were shownto be significant deterrents, however. They were:
t location of cashier
tt two or more clerks
tt no easy access or escape route
tt evening commercial activity, and
tt gas pumps in front.
Hunter & Jeffery
Hunter and Jeffery (1992) in theircase study for Clarke's book onSituational Crime Prevention, reiteratemuch of what was presented above, fromKraus and others, but also add to it, in thefollowing ways. Hunter and Jeffery notethat the original WBSI study, previouslymentioned, has had the most impact uponsubsequent attempts to preventconvenience store robbery and wasindeed the first application of crimeprevention through environmental design-CPTED (Jeffery, 1971) to the crime ofrobbery. The WBSI findings, they say,have influenced the entire conveniencestore industry, and many of theprevention procedures found within theNACS program may be traced directly tothat work.
In addition, they discuss a studyby James White which evaluated 72convenience stores within Gainesville,and he concluded that the number ofclerks working was the strongestpredictor of convenience store robbery,and the store environment was not foundto be a statistically significant factor.
16
They point out that Swanson'sresearch also analyzed the relationshipbetween store characteristics and numberof robberies, from which he concludedthat having two clerks on duty seemed tobe the primary element in deterringconvenience store robbery in Gainesville,as well as security cameras, time releasesafes, other 24-hour business and closingstores at midnight. The findings from the1987 Tallahassee study, Hunter andJeffery note, were utilized in justifying anordinance for convenience stores inGainesville, and later statewide. Hunterand Jeffery conclude from their casestudy (1992) that among the strategiesstudied, having two or more clerks onduty has received the most support. Theresearch cited supporting two clerks wasall from Florida research, however, andnone from outside of the state.
Table 4: Security Measures Related to Robbery
Florida Studies
Measures Swanson1986
Jeffreyet al.1987
Hunter1988
Hunter 1990
cash control + + + +
escape routes + + + +
lighting +
visibility + + +
training
# of clerks + + +
# of customers +
cameras\video +
alarms +
armed guards
unarmed guards
BRBs
location +
hours +
prior robberies
mirrors + +
selling gas + +
+=positive relationship
Swanson & Jeffery list good cashhandling as having support, followed byescape routes, so with these two factorsat least--money and escape route--thereis agreement with research done bothinside and outside of the state of Florida.Table 4 summarizes the results of theFlorida research.
Florida Implementation
The Gainesville, Florida PoliceDepartment analyzed convenience storerobbery data and implemented a cityordinance in July, 1986, which includedthe following components:
• an unobstructed view of the cashregister and sales area throughwindows
• conspicuous signs in the windowsindicating less than fifty dollars onhand
• signs that a drop-time release safe waspresent
17
• parking lots lit with approximately twofoot candles per square foot
• security cameras of a type and numberapproved by the city manager
• mandatory robbery prevention trainingfor all employees who work between 8P.M. and 4 A.M.
After these efforts failed toproduce the desired results of robberyreduction, a provision requiring two clerkson duty was added to the ordinance inApril, 1987 (Clifton and Callahan, 1987).By 1989, robberies had declined by 65%(Hunter and Jeffery, 1992). With theexception of cameras, this programcomports with the original WBSI studyand with the Southland and NACSimplementation. It is interesting to notethat, even without the cameras,Southland had also experienced a 65%reduction in robberies from 1975 to 1986(Erickson, 1996c).
Largely as a result of theGainesville ordinance, the state of Floridaadopted the Florida Convenience StoreSecurity Act of 1990 which required thatthe following security measures forconvenience stores be implemented:
tt silent alarms
tt security cameras
tt drop-safes and cash managementdevices
tt a minimum specific lighting standard
tt signs of limited cash availability
tt unobstructed views of cash registers
tt height markers
tt robbery training for employees
tt limited cash from 9 P.M. to 6 A.M.
tt prohibition of window-tinting.
As with Gainesville, thesemeasures mirrored the basic robbery andviolence prevention measures which had,by now, become accepted industryrecommendations, and all of which hadbeen tested in the original research, withthe exception of cameras and alarms. Because robbery and assaults did notdecrease in the Florida state experience,however, further measures were added tothe legislation. The 1992 Act addedrequirements for businesses openbetween 11 P.M. and 5 A.M. to have oneof the following:
n two or more employees on the premisesbetween 11 P.M. and 5 A.M.
n bullet-resistant safety enclosures
n security guard on the premises, or
n only conduct business through a pass-through window between 11 P.M. and 5A.M.
Evaluation of Gainesville
In 1993, The National Academy ofScience undertook, under the direction ofReiss and Roth, to determine theeffectiveness of the security measures inconvenience stores in Gainesville. Theyfound that evaluating the consequencesof the two-clerk rule has proven difficult.
The Gainesville Police Departmentconcluded that the rule caused a dramatic
18
reduction in convenience store robbery,basing its conclusion on the fact that theconvenience store robbery rate wassubstantially lower in the monthsfollowing the two-clerk rule than it hadbeen during the preceding year. Such asimple before-and-after approach, theysay, to measuring an effect is unreliableand, in the Gainesville case, "yielded anerroneous conclusion" (Reiss and Roth,1993). They go on to cite the work of
retired police chief Jerry Wilson (1990),with funding by NACS, in which hedemonstrated that there had been asharp spurt in the convenience storerobbery rate during the fall of 1986, andthat this spurt ended in December withthe arrest of three men suspected ofbeing responsible for a great many of therobberies.
The two-clerk rule had gone intoeffect the following April, with nodiscernible effect on monthly ratesaround that time. There were tworobberies each month from Januarythrough August. Thus, the fact that therewere many fewer convenience storerobberies in 1987 than 1986 can be morecredibly explained by the arrests inDecember than by the implementation ofthe two-clerk rule. This conclusion gainssupport from the fact that theconvenience store robbery rate in thesurrounding county followed the sametemporal patterns as in Gainesville, eventhough the stores in the county were notsubject to the two-clerk requirement.
Thus, Reiss and Roth (1993)contend that the conclusions that the two-clerk rule was effective fails when rivalhypotheses for the observed pattern areconsidered. They continue by saying thatalthough the two clerk rule did not havethe large impact that the GainesvillePolice Department claimed, there remainsa possibility that it had a small deterrenteffect, of say 5 or 10% on robberies,which they note is difficult to detect in acity as small as Gainesville. Reiss andRoth (1993) say further that the ultimateconclusion from the Gainesville story is
not that there was no effect, but ratherthat there was no miracle and that thesetting and circumstances of theintervention simply do not allow for amore definite conclusion.
The implication of the lack ofeffectiveness of two clerks, drawn byReiss and Roth (1993), however, gainssupport from the experience of the stateof Florida. Since the implementation ofthese measures in Florida, from 1992through 1995, robberies were down 45%,but homicides were up 50%, according tofigures from the Attorney General's office,as shown in Table 5 (Erickson, 1996c).
The Florida research and ensuinglegislation predated the Federalgovernment's involvement in research onconvenience store security and the issueof two clerks, as will be discussed in thenext section.
19
Table 5: Robbery and Homicides atConvenience Stores in Florida*
Offense 1991 1992** 1993 1994 1995 % Change
Robbery 3839 3099 2638 3214 2123 -45%
Homicide 12 15 15 12 18 +50%*Information from Florida Attorney General's Office**Law went into effect in 1992
Government Research
In the late 1980s, academic publichealth researchers began to address theissue of risk factors for employeehomicide, with the early work of Drs.Kraus and Davis, which identified not onlywhich workers were at risk, but whichfactors seemed to place them at risk(Davis, 1987; Davis, et al., 1987; Kraus,1987). These public health researchershad published individual state data onTexas and California, respectively, whichidentified the problem of workplacehomicide and especially the risk forfemales. Their publications appeared oneyear after the first Gainesville robberyprevention ordinance had already beenimplemented.
Later, the National Institute forOccupational Safety and Health (NIOSH),through the examination of state deathcertificates, identified high risk positionsand the factors contributing to that risk(Bell, et al., 1990). These studies and theCenter for Disease Control's (CDC)NIOSH Alert (1993) have addressed what
makes retail employees, specifically, athigh risk, and the results seem to centeron the following factors:
• an exchange of money with the public
• working alone or in small numbers
• working late night or early morninghours
• working in high-crime areas
NIOSH Surveillance and Conferences
The Federal government, it turnsout, was far behind industry, socialscience and local and state involvementwith identifying the problem of workplaceviolence and the issue of two clerks. Notuntil 1990, fully 15 years after the originalWBSI report was published, did NIOSHcome out with their first report on theextent of workplace homicide nationwide(Bell et al., 1990).
That report was followed the nextyear by a report on the risk of female
homicides at work nationwide (Bell,1991). The same year (1990) that NIOSH
20
published their first national data onworkplace homicide, they held aconference which was reported upon in1992. For that meeting, NIOSHassembled a group of experts to reviewthe NIOSH data on workplace homicideand make recommendations for thefuture. The report from that conferencecalled for evaluating the effectiveness ofvarious strategies alone, and incombination with other efforts, including:
nn lighting
nn drop-safes
nn visibility
nn increased staffing
nn training.
To accomplish this, theyrecommended that demonstrationprojects and evaluation research beconducted and\or funded to determine themost effective intervention and preventionstrategies. The next section discusses thefollow-up research undertaken by NIOSHon those recommendations.
The NIOSH data were based onthe National Traumatic OccupationalFatalities (NTOF) data base, which wastaken from death certificates from thestates from 1981 to 1985, revealing theinformation that homicide, at that time,was the third leading cause ofoccupational injury death in the U. S.,following motor vehicles and machinesand accounting for nearly 13% of thenation's total deaths from occupationalinjuries in the workplace.
In 1993, a new NIOSH reportextended the data from 1980-1989. Theresults were essentially the same asbefore. The report emphasized thathomicide was the leading cause ofoccupational death from injury for women,and it was still the third leading cause forall workers.
In the 1993 CDC Alert, it waspointed out that among workplaces, retailtrades had the highest number ofoccupational homicides, and services hadthe second highest--the two togetheraccounting for over half (54%) of allhomicide in the workplace. Industries withthe highest rates of occupationalhomicide were as follows, in this order,with the rates in parentheses showingdeaths per 100,000 workers:
rr taxicabs (26.9)
rr liquor stores (8.0)
rr gas stations (5.6)
rr detective\protective services (5.0)
rr justice\public order (3.4)
rr grocery stores (3.2)
rr jewelry stores (3.2)
rr hotels\motels (1.5)
rr eating\drinking places (1.5)
Taxicab companies, with a rate of26.9 deaths per 100,000 workers, hadalmost 40 times the national average of0.7 and three times the rate of liquorstores (8.0), which had the next highestrate. Taxicab drivers had the highest rateof occupational homicide at 21 times the
21
national average. High risk occupationalrates, from 1980-1989, were as follows,with the rates presented in parenthesis of
the number of deaths per 100,000workers:
rr taxicab drivers (15.1)
rr law enforcement (9.3)
rr hotel clerks (5.1)
rr gas station workers (4.5)
rr security guards (3.6)
rr stock handlers\baggers (3.1)
rr store owners\managers (2.8)
rr bartenders (2.1)
In addition to pointing out thevictimization rates, NIOSH posited thepossible risk factors, as noted earlier, andrecommended preventive measures.Recommendations made were based onresearch, which had been conducted bothinside and outside of government. Therecommended measures, in the 1993NIOSH CDC ALERT, were as follows:
• Make high-risk areas visible to morepeople.
• Install good external lighting.
• Use drop safes to minimize cash onhand.
• Carry small amounts of cash.
• Post signs stating that limited cash ison hand.
• Install silent alarms.
• Install surveillance cameras.
• Increase the number of staff on duty.
• Provide training in conflict resolutionand nonviolent response.
• Avoid resistance during a robbery.
• Provide bullet-proof barriers orenclosures.
• Have police check on workers routinely.
• Close establishments during high-riskhours (late at night and early in themorning).
A final recommendation was forposting a warning to employees aboutpreventing homicide in the workplace,with a one-page tear-out sheet found inthe back of the booklet, showing a gun, astore and a police car with a warningreading as follows: Workers in certainindustries and occupations are atincreased risk of homicide.
In 1996, NIOSH issued CurrentIntelligence Bulletin 57, which expandedupon the data, extending it from 1990 to1992, for the inclusive dates of 1980-1992. Homicide had moved up from thethird to the second leading cause ofdeath, exceeded only by motor vehicleaccidents, and 75% of all workplacehomicides were found to be robbery-related, compared to 9% of homicidesbeing robbery-related in the generalpopulation. The majority of homicides inthe workplace were stranger-to-strangerhomicides.
As in 1993, the majority ofhomicides were of males (80%) but 11%
of all occupational injury deaths amongmale workers were due to homicide, while
22
42% of deaths among female workerswere homicides. The majority of thosewere in the retail trade (46%), largelybecause men are found in occupationswith higher rates of injury and death(Crow and Erickson, 1989), whereaswomen are more predominantly employedin retail trades and not the occupationswith higher rates of injury and death.
In the homicide study ofconvenience stores (Erickson, 1991),67% of the homicide victims were male,and 33% were female. This somewhatparallels the research on robbery/murder,in which Cook (1987) reported that 85%of robbery murder victims were male, and15% were female. Zimring and Zuehl(1986) reported that 80% of robberymurder victims were male, and 20% werefemale. Since the ratio of male to femaleworkers is not known, it is impossible todetermine whether females are morevulnerable than males in conveniencestores.
The rankings of industries andoccupations, by NIOSH, were alsoconsistent with the Census of FatalOccupational Injuries (CFOI) data beingcollected by the Bureau of LaborStatistics. Unlike the NTOF data, whichrelies on death certificates, provided bythe states, the CFOI system uses multiplesources, including death certificates,medical examiner records, workers'compensation reports, and regulatoryagency reports. For both reportingagencies, the ranking of high-riskindustries and occupations wasconsistent with taxicab drivers, followedby law enforcement and securitypersonnel and retail trade with the largest
numbers of workplace homicides (NIOSH,1996; BLS, 1997).
The risk factors, advanced in 1996by NIOSH were essentially the same asthose advanced in 1993, which werelisted above, but with the addition ofdelivery of passengers, goods orservices; having a mobile workplace,such as taxicab or police cruiser; workingwith unstable people, such as health careand criminal justice. The preventionstrategies advanced in the 1996 NIOSHreport were also essentially the same asthose in 1993, although they added escape routes, guards, personalprotective devices, and they droppedreference to police patrols and closing atnight. These were the 1996 NIOSHrecommendations:
tt cash control
tt bullet-resistant barriers
tt visibility
tt lighting
tt limiting access and egress
tt security devices, such as cameras,alarms, mirrors, locking devices
tt personal protective equipment
tt increasing the number of staff on duty
tt guards or receptionists to screenentrance
tt training, and
tt administrative controls.
The final question in the 1996NIOSH report was "What are the mosteffective prevention strategies?" In
23
response, they recommend futureresearch and prevention efforts, notingthat 20 workers being murdered eachweek is unacceptable and should not be
considered the cost of doing business inour society.
This section has outlined the veryimportant reports issued by NIOSH,based on their surveillance studies andtheir review of the literature, while thenext section discusses the research thatthey have conducted.
NIOSH Nine-State Study
One of the studies done internallyby NIOSH, known variously as themultiple-city or multiple-state study, wasreported by Amandus and others in 1996. As with the other NIOSH studies, thisone again identified the nature and extentof the problem but was not itself a studyof the effectiveness of the presentprevention program in the conveniencestore (C-store) industry. The paper calledfor further research to determine thateffectiveness.
In a second paper resulting fromthe study (Amandus et al., 1997), thepurpose was to estimate the risk of injuryin a robbery situation for various riskfactors. Two findings have shownremarkable consistency over twentyyears. The original WBSI study found that5% of the robberies resulted in injury; thisstudy revealed that it is 12%. The originalWBSI study indicated that 65% of therobberies were from 9 P.M. to 3 A.M. This study indicates that 59% of therobberies were in that time frame.
For the surveillance, NIOSHselected State Statistical AnalysisCenters (SACS). Nine states were
chosen for the highest number of C-storerobberies. Seven of the nine SACS(Florida, Maryland, Massachusetts,Michigan, Pennsylvania, South Carolina,and Virginia) provided abstracted andcoded police report data on a sample of1,835 convenience store robberies,occurring during 1992 and 1993.
The robberies occurred in andaround the following selectedmetropolitan areas: Miami and Tampa,Florida; Baltimore, Maryland; Boston,Massachusetts; Detroit, Michigan;Pittsburgh and Philadelphia,Pennsylvania; Charleston, Columbia,Greenville, Spartanburg, South Carolina;and Arlington, Chesterfield and Henricocounties in Virginia. Injury data were onlyused from four states--Florida,Massachusetts, Michigan and Virginia fora sub-set analysis. These permutationsaccount for the variable study names ofmultiple-city, 9-state, 7-state, 4-state, etc.
The finding was that theemployee risk of injury was notsignificantly different between one-employee (0.106) and multiple-employee(0.111) stores. Similarly, the employeerisk of severe injury was not significantlydifferent between one-employee (0.029)and multiple-employee stores (0.022).Thus, the results of their study, overall,indicate that there is no evidence ofincreased safety with two clerks over oneclerk. In other words, the evidence isinsufficient to base a recommendation fortwo clerks on the results.
24
Because the risk of injury is notdifferent with one versus two clerks, thenwhen there are two clerks in the store,and there is an injury-producing event,the risk of injury is doubled because thepotential number of people who might gethurt has been increased. Even ifrobberies are reduced by having twoclerks, they would have to be reduced bymore than twice the amount to make upfor the number who would be injuredbecause of the increased exposure.
Florida's statistics are consistentwith this finding, with robberies beingreduced by 45% in convenience stores,but homicides increasing by 50% in thesame time period from 1991 to 1995(Erickson, 1996c). In sum, more peoplemay get hurt with two clerks on dutybecause it does not decrease theirchances of getting hurt to have two onduty, and it increases exposure; that is,the number out there to be hurt. In fact,the NIOSH statistics show that when theunit analysis of each robbery is used,rather than each clerk, the chances are infact double; that is .106 for one clerk and.206 for multiple clerks.
The article concluded the following(Amandus et al., 1997:447):
The literature thus remainsunclear on which ED[environmental design] measuresare most effective and specificallywhether the presence of two ormore employees at night willdeter robbery in C-stores. Thereis no information on the effect oftwo or more employees on overallinjury rates. The study presentedhere found that the employee riskof injury and severe injury in a
robbed store was essentially thesame in single- and multiple-employee stores. If there weresimilar injury rates given a robberyamong single- and multiple-employee stores, there wouldobviously be an increase in thenumber of injuries in multiple-employee robbed stores,compared with single-employeerobbed stores because therewould be additional employees atrisk in the former; this wasconfirmed in our data.
NIOSH Virginia Study
After the NIOSH epidemiologicresearchers identified industries with ahigh risk of work-related homicide,NIOSH undertook a collaborative projectwith the Virginia Department of CriminalJustice Services and local policedepartments to evaluate the deterrenteffect of crime prevention strategies inVirginia convenience stores, developed in1993. The original project officers--Dr.Harlan Amandus, Chief, Analysis andField Evaluations Branch, NIOSH, andDr. Jay Malcan, Crime PreventionSpecialist, Virginia Crime PreventionCenter, Virginia Department of CriminalJustice Services, state of Virginia--havesince left their positions and respectiveagencies.
Designed as a case-control studyof robbery-related injuries, NIOSH hasbeen studying the violent crimes whichoccur in selected cities within Virginia, butthe study has been extended in timebecause not enough robberies occurredin the one-year time period originallyplanned. Results are now being
25
analyzed and are expected to bepublished within a year. One of themeasures specifically addressed will bethe two-clerk issue and whether or nottwo clerks prevent either robbery orinjury.
This section completes the reviewof the research that has been conductedon the issue of two clerks up to this time.In a later section, the other researchcurrently underway will be summarized,but first an examination of the legislativeinitiatives being undertaken in statesother than Florida is presented, whichrelates primarily to two clerks.
26
LEGISLATIVE ANDGOVERNMENT INITIATIVES
Having dealt with the researchconducted by social scientists, theindustry and the government, let us turnnow to the legislative initiatives on thesubject of two clerks and other securitymeasures for convenience stores. Already discussed were the Gainesvillelegislation (1987) and the Floridalegislation (1990 & 1992), since thoselegislative initiatives grew directly out ofthe research in Florida.
Cities
An earlier effort to legislate twoclerks than Gainesville, however,occurred in Akron, Ohio in 1980. Akronwas the first city to try to legislate twoclerks, an attempt which was notsuccessful at that time. In 1990, the cityof Memphis, Tennessee, similarlyconsidered legislating two clerks butended up recommending the basicrobbery deterrence program instead.
States
In the 1980s, activity was alsogoing on at the state level. After a year ofstudy and testimony on the issue ofrobbery and violence in convenience
stores and gas stations, the MarylandGovernor's Task Force on Retail Securityrecommended the basic robberyprevention program, instead of two clerks,when they concluded the following (Crowand Erickson, 1989):
Proposals, such as limiting storehours, two employees on duty atcertain hours, store cameras, andelectronic alarms, have beenadvanced to control robbery ofretail businesses; but there is noempirical evidence or data toshow that such techniques willreduce retail robberies orminimize the violence toemployees. (Maryland, 1988:19).
Similarly, in the state of Washington, thetwo-clerk legislative effort failed, and thebasic robbery prevention program waslegislated, along with a video foremployee training, which the stateproduced. In the 1990s, the State ofVirginia entertained two-clerk legislationfor several years but they have now madeother recommendations instead. This brings us to the state of California, whichimplemented a model program in 1995discussed in detail below.
Cal-OSHA Model Program
Probably the most sophisticatedapproach on the state level was the effortby California OSHA. The effort was not,however, a form of legislation, but rathera model program recommended to
employers. Cal\OSHA was particularlyconcerned about workplace homicide when, in 1993, homicide became theleading cause of workplace death inCalifornia, joining Alabama, Connecticut,
27
Maryland, Michigan, South Carolina andthe District of Columbia (Howard, 1996). The most prevalent category ofworkplace homicide was retail storerobberies.
The typology that Cal\OSHAdeveloped guides much of the thinking onthe subject nationwide today. Outlined in1995 in the Cal\OSHA Model Injury andIllness Prevention Program for WorkplaceSecurity, the typology characterizes threetypes of workplace violence (California,1995).
For purposes of understandingwhich category or type workers fall into, itis important to remember that workplaceviolence is of both an internal andexternal nature. This distinction is oftenoverlooked by those studying or workingto prevent violence in the workplace orthe media's reporting of workplaceviolence. The three types, as developedby Cal-OSHA, are presented below, withexamples (California, 1995).
Type I. External--Assault or threat byoutside third parties, usually criminals. Nolegitimate relationship with the affected workplace,which is commonly a retail establishment.
Examples: robbers, rapists, murderers.
Type II. Service-related--Assault or threatby someone who is the recipient of a serviceprovided by the affected workplace, such as healthcare providers and the public sector, i.e., police,parole, welfare.
Examples: patients, clientele, customers.
Type III. Internal--Assault or threat by anindividual who has an employment-relatedinvolvement with the affected workplace.
Examples: disgruntled employees,troubled employees, management problems, co-worker problems, acquaintance problems,domestic problems.
The incidents involving Type I arethe most common type of workplaceviolence overall, and the most commontype of violence for retail workers, withestimates as high as three-fourths of thehomicides occurring in this category. Type I is characterized by the outsideactions of third parties, usually criminals,coming on the premises to commit acrime. Retail workers, however, are alsoat risk of Type II violence, such as attacksby customers.
While relatively uncommon, suchinstances do occur with angry or drunkcustomers. Even though clerks are oftentrained to defuse the incident, suchefforts are not always successful, and notall clerks are trained or willing to backdown to what they may consider to beunreasonable requests, comments orbehavior. Type III violence, based onanecdotal information, appears to be lessof a problem within the convenience storeindustry than it is in other industries, for anumber of possible reasons. Typically it isthe "loner" who is profiled as a potentialtroublemaker in internal workplaceviolence incidents.
First, the turnover rate inconvenience stores is high, as comparedto manufacturing, for example, so there isnot as much time for uneasy alliances toform; for high competition among co-workers to develop; for supervisor-subordinate problems to evolve; or forpeople to be passed over for promotion orlaid off, when their expectation was for
28
long-term employment. All of thesefactors are more common in large
hierarchically-based companies than inconvenience stores (Erickson, 1996a).
Also, the convenience storeemployee tends to work somewhatindependently. Unlike fast food workers,who work in close quarters with eachother, convenience store workers arerelatively autonomous. In sum, while retailworkers are at greater risk for homicidefrom the Type I category, they are at lessrisk from Type II, as compared to publicservants or hospital workers, for example.
Government workers make up18% of the workforce, but 30% of thevictims of violence are governmentemployees, who are likely to experiencemore Type II violence (U. S. Justice,1993). Retail workers are also atrelatively low risk for Type III violence. Fewer than 15% of convenience storeworkers who were killed on the job werekilled by someone who was known bythem, according to the study of homicidesin convenience stores (Erickson, 1991).
The recommendations provided byCal/OSHA for preventing Type I violenceinclude these physical changes (Howard,1996):
1) visibility
2) lighting
3) mirrors
4) cameras
5) cash handling and signs.
Other recommendations include workpractice changes of (Howard, 1996):
1) keeping an eye outside
2) staying away from register when not busy
3) limiting cash
4) having a time-access safe
5) using only one register after dark
6) leaving unused registers open and empty
7) training employees in not resisting.
The Cal\OSHA model programdoes not go beyond the measures whichhave been tested, with the exception ofcameras, which by 1995, had becomefairly standard in much of the industryanyway. The Cal\OSHA model program.does not make recommendations for theuse of bullet-resistant barriers, multiplestaffing or closing at night.
Dr. John Howard, Chief of theDivision of Occupational Safety andHealth, Department of IndustrialRelations, State of California, says thatalthough many of the causes ofworkplace violence have their origin inforces outside of the workplace, and thereare gaps in the knowledge about how toprevent the occurrence of some types ofworkplace assaults, there is enoughknown to develop effective preventionprograms in the workplace (Howard,1996). He goes on to say that theproblem cannot be solved by governmentalone.
Federal-OSHA Draft Guidelines
29
A year after the Cal\OSHA modelprogram was developed, Federal OSHAissued the draft of Guidelines forWorkplace Violence Prevention Programsfor Night Retail Establishments (Biles,1996). Patricia D. Biles, WorkplaceViolence Program Coordinator at OSHA,took the lead in preparing the documentand fielding the responses to it. TheFederal Guidelines, as with Cal\OSHA,were never framed as, or intended to be,regulations.
As stated in the noticeaccompanying the guidelines, they arenot a new standard or regulation and areadvisory in nature. OSHA, it was noted,would rely on Section 5(a) of theOccupational Safety and Health Act of1970 (OSH Act) "General Duty Clause"for enforcement authority. Further,employers could be cited for violating theGeneral Duty Cause if there is arecognized hazard of workplace violencein their establishments, and if they havedone nothing to prevent or abate it. Theguidelines stated that OSHA would notcite employers who have effectivelyimplemented these guidelines. OSHA hadpreviously issued guidelines for healthcare workers.
As a first step after the issuance ofthe draft guidelines, comments wereinvited from stakeholders with an originaldeadline of June 30, 1996, which wasextended to September 30, 1996. Theyreceived an overwhelming response fromindustry recommending changes. As aresult, a stakeholder's meeting was heldon November 4, 1996. Approximately tenpeople were invited from the industry tomake presentations, and OSHA invited
about the same number. This was asomewhat different arena for OSHA to beoperating in for a number of reasons (Erickson, 1996c).
The reasons included the fact thatthese workplace safety issues havenothing to do with asbestos, scaffoldingor accidents. They have rather to do withthe intentional actions of criminal thirdparties coming into the business to doharm. This is new ground for OSHA, andthis is the first time that OSHA hasadvanced prevention measures in thisarea. This is also the first time thatOSHA has been involved in trying toprevent such unpredictable behavior--behavior which others, in manydisciplines, have been studying for overtwenty years.
The issues before OSHA, itseemed, were those of both process andcontent, and dealt with both legal andresearch issues. The process deals withhow these guidelines are issued, and thecontent involves what is in them. Withinthe process and the content, there areboth legal and research issues, andwithin the legal and research issues,there are both process and contentissues. Some of the legal issues, relatedto both the process and content, arequestions of:
Legal Issues
• Liability• Definitions• Guidelines vs. Standards• Enforcement• Compliance• Burdensome Requirements
30
For example, some of thequestions are these:
_ Does the very presence of guidelines orstandards create potential liability for businessowners?
_ What are the definitions of night or latenight?
_ What type of businesses do theyinclude?
_ Are they in fact guidelines or are theystandards?
_ How can they be enforced?
_ How do such variant businessescomply? _ Is the recordkeeping, analysis,monitoring crime trends, and the training andretraining of employees too burdensome?
Some of the research issues,related to both process and content, arethese:
Research Issues
• Internal vs. External Violence
• Larger Social Context
• Validated vs. Non-Validated Research
Internal vs. External Violence
Examples here are understandingand distinguishing between internal andexternal violence--two very different typesof violence; one, as mentioned earlier,being the external violence perpetratedby outside criminals, and the other beingactions of employees within. The largersocial context in which the violent eventsoccur cannot be ignored either.
Larger Social Context
The problems of crime in retailsettings cannot be solved until the largersocial issues are addressed. The largerissues in society lead criminals to committhe crimes and become repeat offenders. Root causes of crime include these,among others, (Currie, 1985; Ericksonand Crow, 1980; Wilson and Herrnstein,1985):
n unemployment
n lack of job training
n poverty
n breakdown of the family
n child and spousal abuse
n racial inequalities
n inadequate education
n drug abuse
n youth gangs
n urban decay
n neighborhood decay.
The FBI suggests that the causalfactors related to the changes in thepattern of homicide are related to theillicit drug trade, the disintegration of thefamily unit and weapon proliferation (FBI,1993).
Certain problems which affectviolence occur within the criminal justice
31
system (Currie, 1985; Erickson and Crow,1980; Erickson, 1995b; Wilson andHerrnstein, 1985). Examples of thesefollow:
ll Crimes are committed with guns, but
there are few restrictions on them.
ll Police are out-numbered and out-
gunned on the streets.
ll Criminals recidivate and commit more
crimes, but they are released from prison
early anyway, because of prison
overcrowding.
ll Uneven and uncertain sentencing is
meted out by overwhelmed courts.
A study in Virginia, for example,determined that almost half (42%) of theconvenience store robbers were "legallyrestrained"--that is, they were in thecriminal justice system at the time of theirarrest (Virginia, 1993). Eighty-two percentof state prisoners have prior convictions,and nearly two-thirds of inmates havecurrent or prior convictions for violentoffenses (U. S. Justice, 1985). It isdifficult to run a retail business in aclimate of societal violence, and in asociety which is experiencing problemswithin the criminal justice system as well. In the meantime, efforts, such as thoseoutlined earlier, need to be continued inan attempt to harden the targets andmake the workplace safer for retailemployees.
The nature of retail crime hasbeen changing in a number of ways too: Convenience store robbers seem willingto rob for less; they may steal
merchandise, especially cigarettes, ratherthan money; they are more likely nowthan before to openly steal, without forceor threat of force; carjackings at gasislands are a reality, making some peopleafraid to go there at night.
Assaulting and robbing customersin the parking lots of stores is occurring,perhaps even increasing; gangs andgraffiti are a fact of life in many cities, andloitering and panhandling are majorproblems for stores located in centercities. Murders by strangers and unknownpersons represented 53% of the murdersin the nation during 1992, representing ahistorical high (FBI, 1993).
According to the FBI, everyAmerican now has a realistic chance ofbeing murdered by a stranger, rather thansomeone they know, because of therandom nature that homicide hasassumed in the recent past. This leads toa more generalized fear of violence. Criminals are also younger and moreviolent. The average age of murderarrestees has fallen significantly. It was32.5 years in 1965, and was 27.0 in 1992.The fastest growing murder circumstanceis juvenile gang killings, and the victimmay not necessarily be associated with a
32
gang. Such changes impact upon retailsettings.
Validated vs. Unvalidated Research
The third research issue,presented at the OSHA stakeholdersmeeting, has to do with promulgatingnon-validated versus validated research.
Erickson's paper centered on coveringthe research issues, while others at thatmeeting, and in written responses toOSHA, have dealt with the other issues ofcompliance etc. The validated measuresare those of the basic robbery andviolence program, recommended by theindustry and others which are as follows(Erickson, 1996c):
• Keeping low amounts of cash in theregister
• Ensuring good visibility
• Maintaining good lighting
• Using security cameras and videosystems
• Limiting access and escape routes
• Training employees in proper behavior.
The non-validated measures are these(Erickson, 1996c):
• employing multiple clerks at night
• using bulletproof shielding
• employing guards or off-duty policeofficers at night.
The primary concern expressed atthe November, 1996 meeting about thepreventive strategies was therecommendation for two clerks, whichcontinues to be the major point ofcontention between the industry andOSHA. At this writing, a revised set ofguidelines has not yet been issued.
Other Federal Initiatives
In the following paragraphs, someexamples of other federal involvementwith the problem of convenience storecrime and security are outlined.
FBI
The FBI has been keeping track ofrobberies for both convenience storesand service stations for over a decade,which has been a two-edged sword forthe industry. While it has enabled theindustry to have the information on theextent of its robbery problem and theeffects of ameliorative measures, it hasalso focused attention on this industryalone, and not others. There is, forexample, no FBI robbery data onhotels\motels, bars, liquor stores orrestaurants. There is no breakdown onhomicide for any categories of business,including convenience stores--informationwhich would be extremely useful.
Because of the recordkeeping,however, we do know that, according tothe latest FBI statistics (FBI, 1996),convenience store robberies decreasedby 33% from 1991 to 1995, and servicestation robberies decreased by 25%. TheFBI statistics also bring perspective to thepercentage of robberies that take place inconvenience stores. They account for
33
5.2% of all robberies, compared to streetsand highways, which account for over50% and residences at 10.8%. With theattention paid to convenience store crime,it is often believed that they are a fargreater percentage of the robbery
problem in this country than these figures indicate.
Congress
In an effort to aid small business, particularly convenience stores, the U. S.House Committee on Small Businessheld a hearing on Crime and SmallBusiness on July 21, 1994, chaired byRepresentative John J. LaFalce. Fourwitnesses testified at the hearing on"Strategies in the War Against SmallBusiness Crime" (1994): Dr. LindaRosenstock, Director of NIOSH; JereGlover, Chief Counsel for Advocacy forthe U. S. Small Business Administration;Teri Richman, Senior Vice President,NACS; and Rosemary J. Erickson,President, Athena Research Corporation.
Chairman LaFalce opened thehearing by commenting that smallbusinesses are the heart and soul ofmany urban neighborhoods and thatwhen they leave, city residents are not farbehind. He pled for broader policyinitiatives in the area and the need to findways to build on the research that isavailable.
Dr. Rosenstock spoke about theavailable research, testifying that there isgeneral consensus on measures whichreduce crime in convenience stores,which include visibility, lighting, and cashcontrol but that there is considerabledebate about the effectiveness of certainother measures, such as multiple clerksat night, surveillance cameras and bullet-proof enclosures because of study designlimitations and conflicting results across
studies. Thus, she testified, theeffectiveness of these security measuresin actually preventing violence againstemployees has not been assessed.
Erickson testified that governmentis often quick to suggest solutions, andbusiness is equally quick to reject thembecause of their fears of governmentintervention and increasing their cost ofdoing business. Further, in a plea forcontinued research, by both governmentand industry, it was noted people shouldnot go by common sense beliefs in anarea of study where lives are involved.Just as government would notrecommend a vaccine without testing it;neither should they recommend crimeprevention solutions without testing them.
Follow-up recommendations fromthe Committee included requests to theHouse-Senate conference on Commerce,Justice, State and Judiciaryappropriations for fiscal year 1995,asking for three things:
1) Provide funds for the Department ofJustice, Bureau of Justice Statistics, and theCensus Department to conduct a NationalBusiness Crime Victimization Survey.
2) Instruct the FBI to begin collecting morespecific data on business-related crime andviolence, by specifically breaking out homicideand other violent crimes by type of business.
3) Provide funds allowing the NationalInstitute of Justice to commission a study on
34
the effectiveness of specific robbery-deterrence measures.
Although the request was notgranted, interest was generated, and therequests were sent directly to NIJ, FBIand Justice (LaFalce, 1994).
Point number three above--to provide funds to NIJ--seemed a logicalextension of NIJ's early involvement withfunding the WBSI study in the early1970s to test the effectiveness of specificrobbery-deterrence measures. No large-scale experiments to test new measuresfor combatting crime in small businesseshave been funded by them, or any otheragency for that matter, for over 20 years.
35
THE STATE OF THE KNOWLEDGE
Before discussing the state of theknowledge, it is important to discuss threeissues related to understanding theresearch on the subject of robbery andviolence in convenience stores, which willbe explained in the following paragraph. The three issues are:
1) robbery vs. homicide;
2) injury to customers vs. employees; and
3) exposure data.
Robbery vs. Homicide
The question is whether we aretrying to prevent robbery or preventviolence with these deterrence efforts.Overall, it has been assumed either that1) they are one and the same, or 2) if youprevent robbery you prevent homicide.Based on Cook's (1987) research, hefound that if you reduce robbery, youreduce homicide, but the two-clerk issuemay be an exception to that premisebecause it may be a security measurethat, if introduced, may actually increasehomicide.
The NIOSH research revealed(though it was suspected before) that iftwo clerks were added, reducingrobberies may not reduce homicidesbecause you double the risk, i.e.,exposure. Using that paradigm, let usexamine each of the proposed security
measures and see whether the measuredoes one of the following:
1) increases or decreases robbery;
2) increases or decreases homicide;
3) decreases robbery, but increaseshomicide; or
4) increases robbery, but decreaseshomicide.
Based on a preponderance of theevidence of the research findings,discussed within this paper, andassuming proper training andimplementation for each of the measures,the results appear to be as follows:
Ø lighting (decreases robbery anddecreases homicide)
Ø visibility (decreases robbery anddecreases homicide)
Ø reducing cash (decreases robbery anddecreases homicide)
Ø training employees (decreases robberyand decreases homicide)
Ø escape routes (decrease robbery anddecrease homicide)
Ø cameras (may not decrease robberies;may not decrease homicides)
Ø alarms (do not decrease robberies; do notdecrease homicides)
36
Ø closing establishments at late nighthours (increases robberies at opening andclosing and may increase homicides atthose times, but not necessarily the totalnumber of robberies or homicides)
Ø guards (may decrease robberies, but mayincrease homicides if, as with banks, thereis more violence with guards)
Ø two clerks (may decrease robbery, butmay increase homicide because twopeople are exposed, rather than one)
Ø bullet-resistant barriers (may reducerobberies and injuries for the clerks, butmay increase robberies and homicidesamong customers).
By using the robbery versushomicide paradigm for the varioussecurity measures above, it is clear whythe basic robbery violence program is ineffect, with lighting, visibility, cash control,escape routes and training becausethose measures decrease both robberiesand homicides. Guards and two clerksare the only measures which mayincrease homicides, even though bothmay decrease robberies. In futureresearch, it needs to be clear whichoutcome is being measured--robbery orhomicide. Customers vs. Clerks
A second problem, in addition towhether we are measuring robbery orhomicide, is whether the outcomemeasure is injury to clerks or customers.Again, the result of a particular measuremay be effective in protecting one, but notthe other. The best example of that isbullet-resistant barriers, which on the faceof it, may seem the ideal solution to boththe robbery and homicide problem.
In fact, there are situations, and resulting civil litigation, in which the clerkhas been behind a bullet-resistant barrier,and the customer has been robbedand/or shot, or the customer has beenheld hostage, while the clerk is ordered togive up the money. One robber in theprison study said "They have to come outsometime, so I just wait for them"(Erickson, 1996b).
This disparity between reducingrobberies or reducing injuries may beresponsible for part of the schismbetween government and industry. Theother part may be due to whether injury isto employees or customers. If this is thecase, it may not be that industry andgovernment are at odds, but that theyhave been working unknowingly towardthese different ends: 1) robbery versushomicide and 2) employees versuscustomers. Both OSHA and NIOSH'smissions are employee safety andemployee injury--not customers and notrobbery, but the business owner has toconsider both employee and customersafety and robbery and injury.
Exposure Data
No research, to date, on crime inconvenience stores has had the luxury ofhaving exposure data. Without it, we willnever really know the answer to thequestion of whether two clerks are saferthan one. For exposure data, stores withtwo clerks need to be matched with storeswith one clerk to determine which aremore likely to have robberies and whichare more likely to have homicides.Without exposure data, it is meaningless,for comparison purposes, to state that the
37
majority of robberies or homicides arewith one clerk, rather than with two.
While it is not wrong to say thatmore homicides occur with one clerk thantwo, it is misleading because theassumption is that a clerk was robbed orkilled because there was only one clerkon duty, when in fact the number may justrepresent the reality of what was outthere. Since it has been estimated that85% of the stores may have just one clerkon duty on the late-night shift, thatexposure factor alone may be the reasonfor the majority of homicides occurringwhen only one clerk is on duty.
When NIOSH concludes that themost dangerous time for homicide at workis at night, working alone, dealing with thepublic, with the exchange of money, itmight be that homicides or robberies withone clerk on duty are a result ofcorrelation and not causation. Futureresearch needs to address exposure datato adequately answer the researchquestion of one versus two clerks as asafety factor.
Overview
Amandus (1993:3), in a proposalto study two clerks in Virginia, said thefollowing about the studies which hadbeen done to test the effects of crimedeterrence measures:
There is good consistency amongstudy results for the crime-deterrent effect of some designfeatures and preventionstrategies. However, studyshortcomings, including possibleconfounding, lack of specificity inthe measurement of individualdesign features, small sample
size, lack of standardization in thedefinition of environmentaldesigns and inability to extrapolateto store populations in a widegeographic area, suggested theneed for further research.
Amandus went on to say thatlighting, visibility, traffic, escape routesand training appear to effectively deterrobbery, but results were inconsistentwith respect to multiple clerks, securitysystems (cameras, videos, alarms) andcash-handling procedures. In addition,he noted the need for further research onthe relationship of previous robbery andneighborhood characteristics.
Another issue in need of study fortheir effectiveness as a security measurein convenience stores are guards. According to Baumer and Carrington(1986), guards have not proven to be adeterrent in banks, and banks hadreduced the number of guards at that timeto approximately 8% because they foundthat there was more violence if an eventoccurred and a guard was on duty(Baumer and Carrington, 1986; Crow andErickson, 1989).
Recently, however, with theincreasing use of guards and privatesecurity in general, retail settings appearto be increasingly using guards, eitherroving or stationary. Roving guards covera number of locations, rather thanremaining at a stationary post. The use ofpersonal alarms also seems to beincreasing, which allows the person tocarry an alarm with them. Though alarmsare not considered a deterrent for arobbery, they may have utility as an
38
assault or rape deterrent, though thateffect remains untested as well.
Other measures which need to betested in an experimental setting are theeffects of one clerk versus two clerks, asAmandus noted above. Also, testing isneeded on the effects of closing at nightbecause there is some evidence ofdisplacement in time where closing hasbeen tried; that is, that robbers are morelikely to rob at opening and at closing ifthe retail setting is closed at night. Fastfood outlets are an example. Robbersmay just rob at different times: "Just asbank robbers keep banking hours, soconvenience store robbers keepconvenience store hours" (Crow andErickson, 1989). That is, theirvulnerability to crime is a function of thetime at which they are open for business.
Table 6 summarizes the findings on the security measures which havebeen studied and their relationship to robbery. While location is listed as aseparate item on the table as a securitymeasure, it actually measures a numberof different variables, such as type ofneighborhood, commercial activitynearby, isolation, and so on. Hunter andJeffery (1992) had concluded that twoclerks have received the most support,followed by cash control and escaperoutes, which does not seem to followfrom their analysis. The results showrather that cash and escape routes havereceived the most support, in the studies,followed by visibility and lighting, andthen by two clerks and training, asfollows:
1. Cash
2. Escape routes
3. Visibility
4. Lighting
5. Two clerks
6. Training
The original WBSI study and theSouthland and NACS research andimplementation, as well as Cal\OSHA,include the first four measures in theirprograms, along with training. Thoseelements are considered by the industry,through both research and experience, tobe the most important variables known atthis time to reduce crime in conveniencestores.
In fact, all legislation, includingFlorida, Gainesville, and Federal OSHAincludes these elements, but add to itcertain other measures. For example,Florida and Cal\OSHA recommendcameras and alarms, even though theyhave not been validated throughresearch, and cameras have beennegatively associated in some researchas reviewed in the literature by Ericksonand Crow (1980) and in an experiment byCrow and Erickson (1984).
Though the industry's experienceis mixed on the use of cameras, theyhave essentially become a standardbecause of their widespread use andtheir increasing role in apprehension. Asmentioned earlier, they at least do notseem to increase robbery or homicide. The same may not be said for two clerks,however, as found in the recent prisonstudy of armed robbers, in which robberswho injure were compared to those who
39
do not, which is discussed in the followingsection (Erickson, 1997).
Table 6: Security Measures Related to Robbery: Overview
Measures Studies
Crow &Bull1975
Crowet al.,1987
Erickson1996
Schreiber1991
Figlio1991
Erickson1991
Swanson1986
Jefferyet al.,1987
Hunter1988
Hunter1990
cash control + + + + + + +
escape routes + + + + + + +
lighting + + + +
visibility + + + + + +
training + + +
# of clerks + + + +
# of customers +
cameras\video + +
alarms +
armed guards +
unarmed guards
BRBs
location +
hours +
prior robberies * * *
mirrors + +
selling gas + +
+=positive relationship*=relationship
Robbers Who Injure
Additional analysis has revealednew information about how violentrobbers differ from non-violent robbers,which can now be taken intoconsideration when planning programs todeter robberies and prevent injury. Money
was significantly more important to theviolent robbers than to the non-violentrobbers, so reducing the money not onlyturns away robbers in general, which wehave already known, but turns away themore violent robber.
40
A second finding is that the violentrobbers were much less likely to bedeterred by any measure than were thenon-violent robbers. In particular, theywere not deterred by people, whetherguards, customers or clerks, and weremore willing to take risks. The violentrobbers were willing to take on anaverage of 14 people with a gun and apartner, compared to 11 people for thenon-violent robbers. This agrees with theNIOSH finding that there is more injurywith two clerks because it may even bethe more violent robber that takes on twoclerks--a robber who is more willing toinjure or kill. So then the issue becomesnot a question of one or two clerks, butone or fourteen clerks.
A good summary of whichmeasures work comes from Amandus andothers (1996:717) at NIOSH, in whichthey said the following regarding thevalidated robbery and violenceprevention program:
Minimum elements of C-storerobbery prevention programs andsome state and local governmentregulations, include: keeping lowamounts of cash in the register;ensuring good visibility within, into,and out of the store; maintaininggood lighting within and outside ofthe store; using security camerasand video systems; limitingaccess and escape routes to andfrom the store and thesurrounding property; and trainingemployees in proper behavior,such as passive response, in arobbery situation. Althoughevidence is somewhat equivocalfor the effectiveness of theseminimum prevention elements,they are accepted industrystandards.
With the exception of cameras above, theother measures have proven effective inthe original field experiment, funded byNIJ, and in 20 years of experience in 7-Elevens, resulting in a 65% reduction inrobbery. Regarding the other additionaland non-validated robbery and violenceprevention measures, the Amandusarticle went on to say this (1996:717-718):
More controversial robberyprevention measures that havebeen recommended by somelocal and state governments andadopted by a small part of the C-store industry, include: employingmultiple clerks at night, usingbulletproof shielding around thecash-register station, andemploying guards or off-dutypolice officers at night. Althoughthese more controversialprevention measures haveintuitive appeal, they are notwidely used and theireffectiveness has not beenconfirmed.
Closing at night also has not been tested. The conclusion of the Amandus articlewas as follows (1996:718):
There is a need to evaluate thesespecific environmental designinterventions further to determinewhether they effectively reducerobbery risk and to estimate theircost benefit. Because the ED[environmental design] approachalso may be useful in other retailsettings, there is a need toevaluate the applicability of theseinterventions in other retailindustries.
Evaluating these measures andexpanding the test of them to other settings is precisely the point at which the
41
current research is, as discussed in thefollowing section.
42
RESEARCH UNDERWAY
There are at least five majorstudies underway by both governmentand industry to study the effectiveness ofconvenience store security measures.
n NIOSH 15-City Study
n NIOSH Virginia Study
n NACS Homicide Study
n NACS Camera and Two-ClerkStudy
n NIOSH-Funded UCLA Study
All but the NIOSH-funded UCLAStudy were discussed previously underresearch and implementation, as theyhave been ongoing for sometime. Thenewest and largest in scope is the UCLAstudy, described in the followingparagraphs.
UCLA Study
This four-year Workplace ViolencePrevention Project (WVPP), funded bythe National Institute for OccupationalSafety and Health (NIOSH), is acollaborative effort with the CaliforniaOccupational Safety and HealthAdministration (Cal/OSHA), the SouthernCalifornia Injury Prevention ResearchCenter (SCIPRC) at UCLA, and AthenaResearch Corporation. Drs. Kraus and
Peek-Asa at SCIPRC and Dr. Howard atCal/OSHA all have had extensiveexperience in work-related violenceresearch. Researchers from the SCIPRCdesigned the overall study and areresponsible for implementing theintervention program and also for thecomprehensive evaluation of theintervention.
Because workplace violence isresponsible for over 20% of work-relateddeaths nationwide, and because certainoccupations and employees areconsistently found to be at elevated riskof workplace homicide, these high-riskbusinesses were selected for study.Among the high risk occupations, factorssuch as hours of operation, cashhandling, and store visibility have beenidentified as increasing an individualbusiness's risk of workplace assault.
Businesses in Southern California were randomly allocated into interventionand comparison groups. The businessesincluded service stations, conveniencestores, bars and restaurants, liquorstores, grocery stores, hotels and motels,and taxicab drivers. The WVPP has twoobjectives:
1) to provide an intervention to reduceviolence in up to 1,000 Los Angeles Citybusinesses at high risk of crime-relatedworkplace violence.
2) to evaluate which measures orcombinations of measures are most effective in
43
reducing workplace violence in differentbusiness settings.
The intervention includesemployee training, cash controlmeasures, lighting, visibility, escaperoutes, entry\exit control, floor design,and the use of security equipment, suchas alarms, cameras and timed safes. Each participating business receives:
• a personalized action plan detailingrecommended steps for the individualbusiness
• a workplace safety manual whichdescribes the program and types ofsecurity equipment
• a video for employee training
• a safety poster
• brochures for employee training
• safety decals for the front entrance, thecash register, and the safe.
. The evaluation will determine theeffectiveness of the program overall aswell as the individual components of theprogram. The differential effects amongdifferent types of business settings, suchas type, size and location, will also beexamined. The overall goal is to identifywhich businesses can benefit most fromdifferent combinations of preventionstrategies, with consideration of theresources available to the business.
The UCLA study, along with thetwo NIOSH studies and NACS studies,should combine to bring the answers thatindustry and government have longsought on determining the most effectivesecurity measures.
44
CONCLUSIONS
The early identification of the riskto convenience store employees in the1970s has led to an overall reduction inrobberies since that time. There mayhave been a reduction, as well, inhomicides. Because the greatest risk toretail workers is homicide caused by TypeI violence, the concentration for theprevention of homicide for retail workershas been in the right place. That is, theemphasis needs to be on preventiontechniques which guard against thecriminal actions of third parties, includingrobbery, assault and homicide.
It is heartening, in a review of theresearch which has been done to date, tofind remarkable agreement on theeffectiveness of certain securitymeasures in convenience stores. Thatevaluation, presented herein, includesthe analysis by Kraus and his colleaguesat UCLA (1995 and 1996); a case studyby Jeffery and Hunter (1992); Reiss andRoth's evaluation at the NationalAcademy of Sciences (1993); andreviews by Amandus (1993) andAmandus and his associates at NIOSH(1996 and 1997).
Hopefully, with this increasedunderstanding of the research that hasbeen done and the research that iscurrently underway, a clearer vision willemerge of what the goals are, which canonly serve to help the efforts toaccomplish them. At the very least,
because of consensus on the basicrobbery and violence prevention program, work needs to be done toward a morewidespread implementation of thesevalidated measures:
tt cash control
tt escape routes
tt lighting
tt visibility, and
tt training.
Cameras and video systems havealso become more prevalent. At the sametime, we need to wait for the results of thefive studies underway to determine therelative effectiveness of the othermeasures often recommended. Thosenon-validated measures include:
n two clerks
n bullet-resistant barriers
n guards
n closing at night
Because the risk of injury is notreduced with one versus two clerks onduty, then even if robberies are reducedby having two clerks, it is not enough tooffset the number who would be injured,by having two clerks on duty.
45
There are times that multipleclerks are on duty or establishmentsclose at night for business reasons, suchas sales volume, but these should not beviewed as safety measures, until furtherstudy is done. Some of the currentstudies will address these issues andalso answer the research questions of therelationship between homicide and crimein the neighborhood and homicide and itsrelation to prior robberies.
46
REFERENCES
Amandus H. E. February, 1993. Protocol for a Case-Control Study of the DeterrentEffect of Environmental Designs on Violent Crime in Virginia ConvenienceStores. Morgantown, W VA: Division of Safety Research. National Institute forOccupational Safety and Health.
Amandus H. E. Zahm D. Friedmann R. Ruback R. B. Block C., Weiss J. Rogan D.Holmes W. Bynum T. Hoffman D. McManus R. Malcan J. Wellford C. Kessler D.1996. Employee Injuries and Convenience Store Robberies in SelectedMetropolitan Areas. Journal of Occupational and Environmental Medicine. 38:7.714-710.
Amandus H.E. Hendricks S. A. Zahm D. Friedmann R. Block C. Wellford C. BrensilberD. Bynum T. McManus R. Malcan J. Weiss J. C. Kessler D. May, 1997. Employee Injury in Convenience Store Robberies. Journal of Occupational andEnvironmental Medicine. 39:5.
Baumer T. L. and Carrington M. D. 1986. The Robbery of Financial Institutions: BriefSummary. Washington, D. C. National Institute of Justice, U. S. Department ofJustice.
Bell C. Stout N.A. Bender T.R. Conroy C.S. Crouse W.E. Myers J.R. June 13, 1990. Fatal Occupational Injuries in the United States, 1980-1985, Journal ofAmerican Medical Association. 263 :22:3047-3050.
Bell C. 1991. Female Homicides in United States Workplaces, 1980-1985. AmericanJournal of Public Health. 81:729-732.
Biles P. D. April, 1996. Guidelines for Workplace Violence Prevention Programs forNight Retail Establishments. WDC: Department of Labor\OSHA.
California Department of Industrial Relations, Division of Occupational Safety andHealth. 1995. Model Injury and Illness Prevention Program for WorkplaceSecurity.
Clifton W. Jr. and Callahan P. T. 1987. Convenience Store Robberies in Gainesville,Florida: An Intervention Strategy by the Gainesville Police Department. Gainesville: Gainesville Police Department.
47
Cook P. J. Robbery Violence. The Journal of Criminal Law and Criminology. 1987. 78, (2), 357-375.
Crow W.J. and Bull J.D. 1975. Robbery Deterrence: An Applied Behavioral ScienceDemonstration. La Jolla, CA: Western Behavioral Sciences Institute.
Crow W.J. Erickson R.J. and Scott L. 1987. Set Your Sights on Preventing RetailViolence. Security Management. 31:60-64.
Crow W.J. and Erickson R.J. 1984. Cameras and Silent Alarms: A Study of theirEffectiveness as a Robbery Deterrent. Jackson Hole, WY: Athena ResearchPress.
Crow W.J. and Erickson R.J. 1989. The Store Safety Issue. Alexandria, VA: National Association of Convenience Stores.
Currie E. 1985. Confronting Crime: An American Challenge. New York: PantheonBooks.
Davis H. 1987. Workplace Homicides of Texas Males. American Journal of PublicHealth. 77(10), 1290-1293.
Davis H. Honchar P.A. and Suarez L. 1987. Fatal Occupational Injuries of Women,Texas 1975-84. American Journal of Public Health. 77(12), 1524-1527.
Duffala D. C. 1976. Convenience Stores, Armed Robbery and Physical EnvironmentalFeatures. American Behavioral Scientist. 20:227-246.
Erickson R.J. 1991. Convenience Store Homicide and Rape. Convenience StoreSecurity: Report and Recommendations. Alexandria, VA: National Associationof Convenience Stores.
Erickson R. J. July, 1995a. Homicide in Convenience Stores. Proceedings of theThird Annual Spring Symposium of the Homicide Research Working Group. WDC: National Institute of Justice Research Report.
Erickson R. J. 1995b. Employer Liability for Workplace Violence. Tips on EmploymentLaw. Ed: Robert Fitzpatrick. Chicago, Il: American Bar Association.
Erickson R. J. April-June, 1996a. Retail Employees as a Group at Risk for Violence. Occupational Medicine: State of the Art Reviews, Workplace Violence. 11:2. Philadelphia: Hanley and Belfus, Inc.
48
Erickson R. J. 1996b. Armed Robbers and their Crimes. Seattle, WA: Athena Research Corporation.
Erickson R. J. 1996c. OSHA Guidelines--Meeting of Stakeholders. November 4, 1996. WDC: unpublished paper.
Erickson R. J. 1997. Robbers Who Injure and Those Who Do Not. Proceedings of the1996 Meeting of the Homicide Research Working Group. WDC: NationalInstitute of Justice.
Erickson R.J. and Crow W.J. 1980. Violence in Business Settings. AmericanBehavioral Scientist. 23 (5): 717-743.
Erickson R. J. and Stenseth, A. October, 1996. Crimes of Convenience. SecurityManagement. 60-64.
Figlio R.M. 1991. An Assessment of Robbery Deterrence Measures at ConvenienceStores. Convenience Store Security: Report and Recommendations.Alexandria, VA: National Association of Convenience Stores.
Howard J. D. April-June, 1996. State and Local Regulatory Approaches to PreventingWorkplace Violence. Occupational Medicine: State of the Art Reviews, Violencein the Workplace. Philadelphia: Hanley and Belfus.
Hunter R. D. 1988. The Effects of Environmental Factors upon Convenience StoreRobbery in Florida. Tallahassee, FL: Florida Department of Legal Affairs.
Hunter R. D. 1990. Convenience Store Robbery in Tallahassee: A Reassessment. Journal of Security Administration. 13:3-18.
Hunter R. D. and Jeffery C. R. 1992. Preventing Convenience Store Robbery throughEnvironmental Design. Situational Crime Prevention, Ed. Ronald V. Clarke. NY: Harrow and Heston.
Jeffery C. R. 1971. Crime Prevention Through Environmental Design. Beverly Hills:Sage.
Jeffery C. R. Hunter R. D. and Griswold J. 1987. Crime Prevention and ComputerAnalysis of Convenience Store Robberies in Tallahassee, FL. Security Systems,August, 1987 and Florida Police Journal, Spring, 1987.
Kraus J. F. 1987. Homicide While at Work: Persons, Industries, and Occupations atHigh Risk. Am.Journal of Public Health. 77:1285-1289.
49
Kraus J. F. 1995. Workplace Violence Prevention Project: UCLA Proposal Submittedto the National Institute for Occupational Safety and Health. Unpublished.
Kraus J. F., Blander B. and McArthur D. L. 1995. Incidence, Risk Factors andPrevention Strategies for Work-Related Assault Injuries. Annual Review ofPublic Health. 16:335-79.
LaFalce, J. L. September 6, 1994 (personal correspondence). Chairman, Committeeon Small Business, House of Representatives, Congress of the UnitedStates:WDC.
Maryland, State of. October, 1988. Governor's Task Force on Retail Security. State ofMaryland.
Reiss A.J. Roth JA. 1993. Understanding and Preventing Violence. Washington, DC. National Academy Press.
Schreiber B. 1991. National Survey of Convenience Store Crime and Security.Convenience Store Security: Report and Recommendations. Alexandria, VA:National Association of Convenience Stores.
Swanson R. 1986. Convenience Store Robbery Analysis: A Research Study ofRobbers, Victims and Environment. Gainesville, FL: Unpublished report to theGainesville Police Department.
U. S. Department of Health and Human Services, Public Health Service, Centers forDisease Control and Prevention, National Institute for Occupational Safety andHealth. 1993. NIOSH Alert. Request for Assistance in Preventing Homicide inthe Workplace. DHHS (NIOSH) Publ. No. 93-109.
U. S. Department of Health and Human Services, Public Health Service, Centers forDisease Control and Prevention, National Institute for Occupational Safety andHealth. (September, 1992). Homicide in U. S. Workplaces: A Strategy forPrevention and Research.
U. S. Department of Health and Human Services. Public Health Service, Centers forDisease Control and Prevention National Institute for Occupational Safety andHealth. June, 1996. Violence in the Workplace: Risk Factors and PreventionStrategies. Current Intelligence Bulletin 57.
50
U. S. Department of Justice. 1973. Grant #75NI-99-0002 awarded by the LawEnforcement Assistance Administration to the Western Behavioral SciencesInstitute.
U. S. Department of Justice, Bureau of Justice Statistics. February, 1985. SpecialReport. Examining Recidivism. Washington, D. C.
U. S. Department of Justice, Office of Justice Programs, Bureau of Justice Statistics. July, 1993. Violence and Theft in the Workplace. NCJ-148199.
U. S. Department of Justice, Federal Bureau of Investigation. 1993. Uniform CrimeReports.
U. S. Department of Justice, Federal Bureau of Investigation. 1996. Uniform CrimeReports.
U. S. Department of Labor, Bureau of Labor Statistics. April, 1997. Fatal WorkplaceInjuries in 1995: A Collection of Data and Analysis.
U. S. House of Representatives. Committee on Small Business. Hearing on Crime andSmall Business of "Strategies in the War Against Small Business Crime." WDC.
Virginia State of. Department of Criminal Justice Services. 1993. Violent crimes inConvenience Stores: Analysis of Crimes, Criminals and Costs. Richmond, VA.
Wilson J. V. 1990. Convenience Store Robberies: The Gainesville, Florida 2-ClerkLaw. WDC: Crime Control Research Corporation.
Wilson J. and Herrnstein R.J. 1985. Crime and Human Nature. New York. Simon andSchuster.
Zimring F.E. and Zuehl J. 1986. Victim Injury and Death in Urban Robbery: A Chicagostudy. The Journal of Legal Studies IV (1): 1-39.
51
INDEX
alarm................................................................................................................6, 7, 12, 34alarms.............................................7, 8, 10, 11, 13, 15, 16, 20, 21, 24, 32, 34-36, 39, 43Amandus .............................................................................. iii, 22, 23, 34, 35, 37, 40, 42Athena ...................................................................... 2, i, iv-6, 8, 9, 12, 31, 38, 43, 44, 50Biles.........................................................................................................................26, 42Bull ........................................................................................................3, 4, 8, 10, 36, 43bullet-resistant barriers............................................................... i, 7, 9, 11, 21, 26, 33, 40Cal\OSHA...........................................................................................................24-27, 35Cal-OSHA........................................................................................................... iv, 24, 25cameras..................................... 7, 8, 11, 13-16, 20, 21, 24, 26, 30-32, 34-37, 39, 40, 43cash........................................................................i, 3, 4, 8, 10-16, 20, 21, 26, 30-40, 50clerk.................................................. iii, 1, 2, 4, 6, 7, 12, 13, 16, 17, 22-24, 32-34, 38, 46clerks ...............................................i, iii, iv, 1-4, 6-9, 11-15, 17, 18, 20, 22-25, 30-37, 40Clifton ......................................................................................................................15, 42closing .................................................................................... i, 14, 21, 26, 32, 35, 37, 40Congress ............................................................................................................ iv, 30, 45Crow ................................................................... 3-6, 8, 10, 11, 21, 24, 28, 34-36, 43, 44Duffala .....................................................................................................................13, 43Erickson................................................ 1, 2, i, 4-11, 16, 17, 21, 23-31, 33-36, 43, 44, 50escape.................................................................. i, 4-8, 11, 13-15, 21, 30, 32-37, 39, 40FBI................................................................................................................... iv, 8, 28-31Figlio......................................................................................................... iii, 9, 11, 36, 44Florida ....................................................................... i, iv, v, 12-18, 22-24, 35, 42, 44, 46Gainesville................................................................... iv, 1, 12, 14-18, 24, 35, 42, 45, 46Griswold ..................................................................................................................13, 44guard .............................................................................................................7, 16, 34, 40guards .................................................... i, 7, 8, 10, 11, 15, 20, 21, 30, 33, 34, 36, 37, 40gun ......................................................................................................................7, 20, 36guns...........................................................................................................................7, 28homicide .................................i, iv, v, 1, 9-11, 18-21, 23-26, 28-35, 38, 40, 41, 43-45, 50homicides ................................................v, 7, 9, 10, 17-21, 23, 25, 26, 32-34, 40, 42, 43Howard ................................................................................................... iii, 24, 26, 38, 44Hunter.........................................................................................11, 13-15, 35, 36, 40, 44injury.................................................................. i, iii, 19-23, 25, 32, 33, 36-38, 40, 42, 46Jeffery.....................................................................................5, 11, 13-15, 35, 36, 40, 44Justice ........................................................... i, 3, 4, 19, 21, 23, 26, 28, 29, 31, 42-46, 50Kraus .........................................................................................12, 14, 18, 38, 40, 44, 45lighting..........................................i, 4, 8, 10, 11, 13, 15, 16, 19-21, 26, 30-37, 39, 40, 50
52
NACS........................................................................... i, iv, v, 9-12, 14-16, 31, 35, 38, 39NIJ .................................................................................................................3, 11, 31, 37NIOSH .................................................. i, iii, iv, 7, 9, 12, 18-23, 30, 32, 34, 37-40, 45, 50OSHA ............................................................... i, iv, 24-27, 29, 30, 33, 35, 38, 42, 44, 50police ........................................................... 5-7, 15-17, 20-23, 25, 28, 30, 37, 42, 44, 45prisoners........................................................................................................................29prisons.................................................................................................................... i, 6, 12Reiss .....................................................................................................11, 16, 17, 40, 45resist................................................................................................................................9resistance ............................................................................................................9, 10, 20Richman ........................................................................................................................31robber ................................................................................................3, 4, 7, 8, 33, 36, 37robberies .............................. i, 3, 5-8, 10, 11, 14-17, 22-25, 30, 32-34, 36, 40-42, 44, 46robbers .............................................................. i, iv, 3, 5-8, 12, 25, 29, 35-37, 44, 45, 50robbery .......................................................i, iv, v, 3-5, 7-18, 20-24, 29-37, 40, 42-46, 50Roth.............................................................................................................11, 16, 17, 45Schreiber ........................................................................................................9-11, 36, 45signs ................................................................................................3, 4, 9, 15, 16, 20, 26Southland ......................................................................... iv, 3, 5, 6, 9, 11, 12, 15, 16, 35Stenseth ....................................................................................................................6, 44Swanson................................................................................................12, 13, 15, 36, 45training.......................... i, 3, 4, 8, 10-12, 15, 16, 19-21, 24, 26, 28, 30, 32-37, 39, 40, 50UCLA......................................................................................................i, iii, iv, 38-40, 45video.......................................................................6-8, 10, 11, 15, 24, 30, 36, 37, 39, 40violence ...................... i, iii, iv, 3-5, 9, 11, 12, 16, 18, 24-29, 31-34, 37, 38, 40, 42-46, 50violent ..................................................................................23, 28, 29, 31, 36, 37, 42, 46Virginia ............................................................................. iii, iv, 22-24, 29, 34, 38, 42, 46WBSI ........................................................... iv, v, 1, 3-6, 8, 11, 12, 14, 15, 18, 22, 31, 35windows...................................................................................................4, 10, 12, 13, 15Zimring ................................................................................................................9, 21, 46Zuehl ...................................................................................................................9, 21, 46
53
54
ABOUT THE AUTHOR
Rosemary J. Erickson, Ph.D. is a nationally known expert on the subject of security.Since 1979, she has been president of Athena Research Corporation. As a sociologist,she has studied the problem of convenience store security for over twenty years,beginning with a study funded by the National Institute of Justice, to determine what wouldreduce robbery and violence in convenience stores. That work led to the measures seentoday in stores nationwide, which include cash control, visibility, lighting and employeetraining. Dr. Erickson conducts research on homicide, rape and robbery. She conducteda survey of robbers about what they look for when they consider robbing a place, whichresulted in the book, Armed Robbers and Their Crimes. She develops programs andtraining based on her research.
Dr. Erickson serves as an expert witness in premises liability cases and hastestified on the subject before city, state and federal legislators. In the summer of 1994,she testified before the U. S. House Committee on Crime Against Small Business. Sheserves as a liaison between business and government and consults with OSHA agenciesand the National Institute of Occupational Safety and Health (NIOSH) in planning modelprograms to protect workers. She has served on the local Seattle\King County Board ofDirectors for Crimestoppers. Dr. Erickson has received widespread media coverage andhas written numerous reports and articles on retail security and homicide in the workplace.