What Do You Mean When You Say Workplace Violence?

Security professionals must proceed with caution when wielding powerful terms and important statistics.

NOTE: This is an annotated pre-publication draft of an article submitted for publication in the March 2012 issue of Security Management magazine. After a flurry of active consideration it was rejected. Since then the charts here have been obsoleted by a 2012 NCCI Workplace Violence study due to the fine grain analysis completed by its authors, Tanya Restrepo and Harry Shuford.


Early on Wednesday morning, October 5, 2011, Shareef Allman brought an AK47 and a 9mm pistol to work at Lehigh Hanson's Permanente Cement Plant, a landmark in the hills above the South Bay city of Cupertino, California. During a morning safety meeting he opened fire on his co-workers and then fled. Three persons were killed, seven were wounded, and the police engaged in a community-wide manhunt while schools and high-tech businesses operated under lockdown conditions. The search ended the next morning when Allman was shot and killed as he threatened law enforcement officers nearly a day after the killings. Since then we have learned that Allman displayed many different warning signs before the fatal day. There was a history of domestic violence (which does not always occur in these cases). Allman made comments to family members describing violent use of military-style weapons. There were long standing grievances at work. Ominously, Allman made arrangements to block exit routes before he began shooting. As with other acts like it this incident illustrates the complexity of connecting the dots in advance of violence, something which is all too easy to do after the incident. Yet another disgruntled employee committing mass murder at work. It happens all the time. Or does it?

Open the newspaper, read the news online, peruse your security magazines and safety blogs. Find an article about workplace violence. With very few exceptions we are treated to lurid headlines exclaiming that workplace violence is the second leading cause of death at work, that employees succumb to violence at work at the rate of two per day, that it’s an epidemic, and it’s getting worse. Says one security magazine article, “Distraught employees are injuring or killing co-workers and supervisors at an alarming rate. Customer rage is at an all-time high. Domestic violence has spilled over into the workplace. And it’s not that these incidents are getting more media coverage – the Center for Disease Control has officially classified workplace violence as a national epidemic."

Security professionals are at their best when offering well reasoned business advice in calm discourse, but something about workplace violence seems to activate an innate tendency to engage in hyperbole and fear-mongering. The truth is that, despite the media drumbeat to the contrary, workplace homicide has been declining steadily over the past 18 years and is only 50% what it was when the Bureau of Labor Statistics began tracking the category in 1992.

Coming to terms

Some terms and concepts are so powerful that sometimes even security professionals have trouble responding to them thoughtfully. While conducting research for my Master’s I asked private security professionals at several LinkedIn groups, the management staff of a contract guarding company, line officers at a guarding contract client site, and three security-related blogs were asked to answer a brief survey related to workplace violence. Among others two questions assessed the knowledge of the magnitude and nature of the problem.

When asked the number of workplace violence murders committed each year in the U.S. 57% of respondents under-estimated. Only one in ten (9.6%) private security professionals correctly answered the question. 32% over-estimated the number. 15% of respondents did not know the answer. When asked the percentage of workplace violence murders in the U.S. committed by co-workers, former co-workers, clients, family, or friends of the victim 10% of respondents underestimated. Four out of five (83%) private security professionals over-estimated the percentage of workplace violence murders in the U.S. committed by co-workers, former co-workers, clients, family, or friends of the victim. Only one in 12 (8.0%) private security professional correctly answered the question.

This sort of disconnect has been explained by those who study the psychology of risk. We tend to overestimate the risk of encountering exotic hazards (air crashes) and downplay the frequency of more mundane hazards (car crashes). This specific case may be the result of lack of access to or awareness of the available statistics. Or it may be the consequence of the very human tendency to get it wrong when we conflate our feelings about risk with actually thinking about the real numbers. The survey represented very small, self-selected sample. The opinions expressed were just that, opinions. But two fact-based questions were asked. The idea that self-selected private security professionals with strong opinions about the risk of workplace violence were unable to accurately describe the magnitude or nature of lethal workplace violence is striking.

These are the terms used by the Federal Bureau of Investigation, the National Institute of Occupational Safety and Health, the Occupational Safety and Health Administration, and the Bureau of Labor Statistics use to describe workplace violence and its perpetrators:

Type I: The offender is engaged in criminal activity, frequently robbery

Type II: These offenders have or had some sort of business relationship with the business where the violence was perpetrated by clients, customers, or patients

Type III: Includes coworkers and former coworkers

Type IV: Family, friends, or acquaintances of the victim; frequently an intimate partner

I’ve been told by a wise peer "I don't think employees are really concerned about the workplace violence perpetrator's classification." I couldn't agree more, especially when the violence is in progress. I know it seems clinical to poke and prod at our understanding of these issues. The value of understanding the nature of these offenders, their needs in some cases, their motivations in others, and their methods, is to help us prepare to detect, deter, prevent, or defeat them.

Rethinking the statistics

What if we break down the data into less alarming yet more useful chunks? Breaking apart the categories would help us put the different elements of the problem in better perspective, allow our profession to develop a variety of effective preventatives, and encourage firms to tailor their responses to their risks. Let's walk through the numbers while I explain.  The Bureau of Labor Statistics (BLS) keeps the records for workplace fatalities and injuries. The most recent year for which they have published final numbers is 2009. The details are found in their 2011 publication, Homicide: Occupational homicides by selected characteristics, 1997-2010.  [It shows that over 13 years there were 8,654 homicides in the workplace or on the job.  Of these, 6,512 (75%) were Type I offenses.  Type II offenses numbered 616 (7%).  There were 894 (%) Type III events.  Type IV accounted for 632 (8%) deaths.  Please note, as I lacked access to a precise breakdown of  workplace homicides by type on an annual basis I applied these ratios to the 2009 workplace homicide statistics to approximate their relative impact. See Tanya Restrepo and Harry Shuford's Violence in the Workplace for revised statistics.]

First, how many people die at work or on the job. According the BLS in 2009 the total number of employee fatalities was 4,551. Far and away, year after year, the leading cause of death on the job has been transportation accidents. In 2009 there were 1,795 fatalities in this category. Other categories include “contact with objects and equipment,” “falls,” “exposure to harmful substances or environments,” “fires and explosions,” and “other events or exposures.”

Using the BLS numbers we see that “assaults and violent acts” was the second leading cause of death in the workplace in 2009 but, unbeknownst to many with strong opinions on this topic, this category includes both 542 homicides and 263 suicides.

[click on chart to enlarge for easier reading]

If we break homicides and suicides out of assaults and violent acts category in 2009 then homicide falls to fourth place and suicide becomes the sixth place hazard.

[click on chart to enlarge for easier reading]

According to the BLS between 1997 and 2010, on average 75% of all workplace homicides are perpetrated by criminals, many while engaged in armed robbery. These are called Type I offenders. The remaining 25% are divided between clients, customers, and patients (7%); coworkers and former coworkers (10%); and family, friends, and associates (8%). Offenders in these categories are Types II, III, and IV, respectively. If we regroup the 2009 homicides according offender types then the 124 cases of what most people think of when the phrase “workplace violence” is used (or misused) assume their proper place in relation to other workplace hazards.  [See Tanya Restrepo and Harry Shuford's Violence in the Workplace for revised statistics.]

[click on chart to enlarge for easier reading]

When we break down the Types II, III, and IV homicide into their own columns we run the risk of being asked why these cases attract more attention and funding than other safety hazards. 

[click on chart to enlarge for easier reading]

At this level of granularity I propose the solutions our profession needs to develop and deploy are four in number.

Focused solutions for complex problems

Robbery/homicide is a risk to cab drivers and retail personnel, especially at night. Robbery prevention calls for facilities improvements, physical security measures, changes to business practices, and employee training. A reduction in homicides and injuries during robberies represent a reduction in workplace violence. Likewise reducing the number of robberies can reduce the number of persons harmed even if the rate of homicide and injury during the crime do not change.

The prevention of line-of-duty killings of law enforcement and security personnel calls for specialized safety training and personal protective equipment unlike that provided to employees engaged in non-enforcement work. This has become a well developed science in this segment of our field.

Preventing suicide in the workplace suicide is an extremely complex issue that calls for attention from management, our peers in Human Resources, the Employee Assistance Program, insurers, labor unions – where applicable, and the family members of employees. Suicide challenges our ability to deal with taboo subjects and to avoid stigmatizing those suffering from depression and other illnesses that put a person at risk for suicide.

Workplace homicides perpetrated by coworkers and former coworkers; clients, customers, and patients; or family, friends, and other acquaintances is what most people (and the news media) think of when they hear the phrase “workplace violence.” Yet, even combined, these three categories account for the smallest fraction of workplace deaths and murders. If we focus on solutions for this issue to the exclusion of others we will ignore the great majority of workplace deaths due to assaults and violent acts. Gratefully, there is new help on the horizon.

ANSI Workplace Violence Prevention and Intervention Standard

ASIS and SHRM have just completed a collaboration to create the ANSI Workplace Violence Prevention and Intervention ASIS/SHRM WVPI.1-2011 American National Standard. I’ve been looking forward to the results of this collaboration so I read the standard with enthusiasm. Considering the size and scope of the project, the involvement of three associations, strict standards criteria, hundreds of participants, and four years of development it turned out pretty well, but it is a mixed bag.

The new ANSI national standard was put together by many of the top names in the business and does a solid job outlining a corporate response to threats and violence by coworkers and former coworkers (Type III), and family members, friends, and associates (Type IV). It emphasizes the value of a cross-disciplinary team and a broad-spectrum response when assessing and responding to threats within the business. To the degree that this sort of violence is what most people, employers, and the news media think of when they hear the term "workplace violence" the standard is a strong move in the right direction.

Unfortunately, the new ANSI national standard pays but the briefest of lip service to violence perpetrated by criminals (Type I) which represents 75% of all homicides at work. Likewise violence by clients, customers, and patients (Type II) is given scant attention. I corresponded with a friend and peer who served on one of the standard's development committees. He made a good case that many Type II offenders can be addressed using this model. On reflection I agree. The behaviors exhibited by many clients and customers presents many of the same challenges and opportunities as Type III violence. Still, attacks and injuries by patients in the health care setting - especially in custodial units - remains more like Type I violence. Violence by patients in health care and social services, especially in mental health setting, account for a large fraction of injuries to employees [61% according to Restrepo and Shuford] (and until recently has been the only category of workplace violence subject to specific OSHA regulation). The new standard does not even contain the word suicide, let alone examine the idea that troubled employees sometimes choose to end their lives at work. The 263 suicides in 2009 represent roughly a third of all workplace deaths due to assault and violent acts. There are twice as many suicides at work as there are Type II, III, and IV deaths combined. By not addressing the prevention of, or response to, these issues in any way the new standard ignores the majority of the problem of workplace violence.

OSHA has never been more engaged in the issue

In September 2011 the Occupational Safety and Health Administration (OSHA) released a guidance document titled Enforcement Procedures for Investigating or Inspecting Workplace Violence Incidents CPL 02-01-052 which constitutes what OSHA regards as a "significant change," "the first instruction on the enforcement procedures for investigations and inspections that occur as a result of workplace violence incident(s)." It can be a useful document to those building or maintaining WPV prevention and response programs as one can reverse engineer OSHA instructions to refine the structure of the record keeping elements of a program. This directive specifically and correctly focuses on late night retail and health care & social service settings.

The document outlines OSHA’s "criteria for initiating inspections."

a. Known risk factors to consider, listed by NIOSH in its report NIOSH Current Intelligence Bulletin #57: Violence in the Workplace: Risk Factors and Prevention Strategies (1996).

b. Evidence of employer and/or industry recognition of the potential for workplace violence in OSHA-identified high risk industries, such as health care and social service settings and late night retail (See Section X, C, 1 and 2.).

c. Feasible abatement methods exist to address the hazard(s). [Appendix B]

Enforcement Procedures for Investigating or Inspecting Workplace Violence Incidents offers several scenarios where investigation and enforcement are appropriate, notably in health care and late night retail settings. Strikingly, it describes a case of acquaintance on employee violence (Type IV) as not meeting any of the three criteria for enforcement. It also describes a shooting of employees at a financial services company as meeting some but not necessarily all of the enforcement criteria.

Security practitioners need to be conversant with CPL 02-01-052 because it is the guidance OSHA inspectors will use to determine whether a company that experiences workplace violence injury or death should be subject to enforcement action. It provides some very interesting clarifications under which OSHA investigators might invoke the General Duty Clause, Section 5(a)(1). It describes a specific obligation for employers in late night retail and health care & social service settings. It does not seem to create specific obligations for other industries, but it is an interesting change of which all security professionals should be aware.

Does Anyone Have Any Numbers?

As we strive to implement effective workplace violence prevention and intervention programs another question looms. Is the apparent decline in workplace homicides the result of education, prevention, and mitigation plans or is it part of the overall decline in violent crime? Is it possible that organizations have been so busy implementing their workplace violence prevention programs that they have not had time to measure the contribution they make to the security, safety, and well-being of their employees? Those of us who have managed workplace violence cases to their conclusion are happy that our anecdotes are usually about potential harms successfully deflected, but these are not statistics. Will workplace violence prevention programs become a pro forma best practice without regard to their actual effectiveness?

We are all quick to agree that a workplace violence prevention and response plan is a reasonable element of a company’s employee relations policy, but can we prove it? Are there any data that demonstrate that having an organized WPV program creates a detectable reduction in deaths, injuries, threats, or cases? Of course, a very successful program might easily increase the number of reported cases while reducing their severity. While we all strive to reduce fatalities, with the exception of homicides during robberies, they are relatively rare and a single severe incident can throw off the numbers. Tracking injuries, assaults, and threats may prove a more useful measure.

There is some good news in this area. In 2009 James T. Wassell published a paper titled Workplace violence intervention effectiveness: A systematic literature review, one of the few studies that considers the question. The bad news, he didn't find much high quality evidence either. Of 100 studies, half dealt with Type II offenders in the health care setting. Ten percent discussed homicide prevention in retail establishments (Type I offenders). He recommended more research.

When it comes to Type II, III, and IV offenders - clients and patients, coworkers and ex-coworkers, and family and friends, respectively – an effective workplace violence prevention and response program should be able to track successful interventions, cases of which the company/client was aware but did not result in violence, changes in corrective actions and terminations for behaviors that are workplace violence risk factors, a reduction in assaults and threats reported by employees, a reduction in harassment and toxic manager cases, a reduction in some sorts of worker's comp claims, and increased employee job satisfaction and retention metrics. We do not have these numbers yet, but we should be careful to collect them.

Why not scare them if it works?

As a citizen, neighbor, and parent there are days when I regret that there is no shortage of security needs in our society. There is plenty of work to be done and I submit that most of it can be accomplished without resorting to alarmism and fear mongering. Still, there are security businesses who seek to leverage fear and create a sense of urgency. I’ve been told by some security professionals – especially in sales – that sometimes fear is required to get executives to act, so alarmist language is used to sell security services, products, and programs. The temptation to leverage fear, uncertainty, and doubt is a hazard every security professional must be careful to avoid. In the long run working to create informed consumers who correctly take confidence in our responses and preparations will be more rewarding and more ethical than fanning the flames of fear.

I remain concerned that we distract employees, employers, and our communities from the larger – and perhaps more tractable – problems of robbery/homicide and workplace suicide when we let the media reinforce the faulty notion that deadly violence at the hands of disgruntled coworkers is common. I have seen too many security professionals - especially those of us selling products, services, or books - misuse statistics like these to promote a response based on fear rather than sober analysis.

The security industry needs professionals who are committed to advancing the interests of his or her enterprise rather than simply reacting to fear, acting as private police, or closing the sale. Do not scare people; not deliberately, not accidentally. Thanks to world events people are already scared; thanks to sensational news reporting people tend to remain scared. When people are scared they frequently act without thinking. Our families, communities, and employers look to those in security profession for confidence in the face of troubling disorder.

Guiding thoughtful responses

Here are a couple more facts to consider. First, despite the media drumbeat to the contrary, workplace homicide has been declining steadily over the past 18 years and is only 50% what it was when the Bureau of Labor Statistics began tracking in 1992. Second, when we focus only on fatalities we risk losing track of the impact of 41,000 lost time injuries resulting from nonfatal assaults and violent acts by persons. Third, workplace suicides appear to be on the rise.

Referring to the aggregate number of all workplace deaths due to “assaults and violent acts” while discussing only homicides perpetrated by "disgruntled employees” significantly overstates the frequency of what most Americans think of when they hear the term "workplace violence,” “going postal,” or “active shooter.” This sort of overstatement is not unusual in news reporting, but I propose that we in the security profession will have an easier time addressing these important issues if the debate is not contaminated with inflammatory rhetoric.

There is still plenty of work for security professionals – along with our peers in human resources; environmental, health, and safety; and legal – to do in this area. All forms of workplace violence - threats, assaults, injuries, homicides, and suicides - remain important issues to address. Unless the businesses in question are enterprises with all the known high risk factors we do our profession no service by using the aggregated statistics to motivate our peers, employers, or clients. Sooner or later the decision makers will catch on those who rely on hyperbole to move their programs forward.

Imagine the impact security and peer professionals might have if we insist this problem be understood in its true complexity and approached as a set of issues requiring a variety of solutions applied across disciplines. We have much work to do. Let’s be certain we’re using our finite resources – time, money, and good will – where they will do the most good.

Photo Credit: http://www.photos8.com/bullet_against_pen_on_book-wallpapers.html