Dan Petersen: Why Safety is a People Problem

Jan. 1, 1997
If you want to improve your safety performance, says consultant, author and educator Dan Petersen, put away your accident statistics and start listening to the real experts on safety -- your employees.

He is arguably the best known safety professional in the United States. His consulting clients are the cream of American industry. He has written 17 books and appeared in 10 videotapes on a variety of safety-related topics. He is a past president of the National Safety Management Society and a former vice president of the American Society of Safety Engineers. But despite 42 years of success in safety, Dan Petersen"s prescription for improved safety management still seems less than enshrined in America"s workplaces.

Petersen attributes some of that to timing. The Occupational Safety and Health Act went into effect in 1971, the same year his first book, Techniques of Safety Management, was published. "My book said to go a certain direction and OSHA said to go the exact opposite, and OSHA was a little bit more powerful in terms of influence."

Petersen has long maintained, as he wrote in the 1988 edition of Safety Management, that safety is a "people problem." In his view, the two keys to a "world-class" safety program are to make the line organization accountable for safety performance and to have employee involvement. Yet, he complains, the "fundamental principle" that undergirds OSHA is that "accidents are caused by things, not by people."

Petersen has also taken safety management to task, arguing that the field has been slow to adopt new management theories and react to changes in the workplace. Instead, the safety field has developed its "own little mystique," insisting on using measures of performance such as injury frequency and severity rates that "weren"t very meaningful" to company executives. As a result, he says, executives have concluded that the same techniques that work for improving quality and productivity are "not usable in safety because it"s different."

Management Function

In new editions of his books Safety By Objectives, Analyzing System Safety Effectiveness, and Human Error Reduction and Safety Management, published by Van Nostrand Reinhold, as well as in a new series of video training programs produced by CoreMedia, Petersen emphasizes that "safety is just another management function and should be managed in the same way." In fact, in Analyzing System Safety Effectiveness, Petersen argues that five widely held beliefs about safety management are simply wrong:

1) Accidents are caused by unsafe acts and conditions. That idea, Petersen noted, is derived from W. H. Heinrich"s Industrial Accident Prevention, first published in 1931. Heinrich said unsafe acts account for 88 percent of accidents, while unsafe conditions account for 10 percent. "That was his first axiom and was something that he just made up. There was no research behind it whatsoever," Petersen said.

By focusing on unsafe acts and conditions, he said, safety managers end up dealing with accidents on a symptomatic rather than a causal level.

"Behind every unsafe condition, there is a management system that could have allowed that hazard to exist," said Petersen. "Behind every unsafe behavior or unsafe act, there is a reason that those people engage in those behaviors. A lot of times, that has to do with the management system -- the way people are measured and rewarded, the culture of the organization that leads unsafe behavior to be O.K."

Instead, he has stressed, managers have to look at the causes of accidents as being "a combination of a management system and a culture or environment that leads to human error."

"We have to deal at the culture level, not at the unsafe act level. We have to understand why people do these things. What are we in management doing that is allowing or even encouraging those behaviors? Until we get to that causal level, we"re not going to get very far," he warned, adding: "Unfortunately, that unsafe act or condition or domino theory of accident causation undergirds most safety programs and all legislation. It"s a difficult problem."

2) There are certain essential elements to a safety program. Take a common element of safety programs, such as safety meetings, safety committees or job safety analyses, says Petersen, and compare its use at different organizations or even at different locations within the same company. Thes ideas will be "highly effective" in one place and "a total waste of time" at another.

"There are no essential pieces or elements to the safety program. Rather, it"s the environment and the culture in which those things are placed that determine whether or not they"re going to work," said Petersen.

Yet in terms of organizational culture, he added, some elements do seem to recur among the best companies. They include accountability, employee involvement, flexibility to make choices about how results will be obtained, and whether or not employees perceive that the safety system is good for them.

In the end, he says, it is "how people feel about their organization that dictates how they behave on the job every day." If companies consistently place a high value on safety and make decisions daily that support safety, employees will respond and operate in a safe manner.

3) Accident statistics tell us something. For most organizations, particularly smaller ones, recordable injury rates "have no statistical validity and very little meaning whatsoever," Petersen says. Moreover, he points out, they neither "direct you very much" nor "diagnose and tell you what"s wrong." He added: "You go from an occupational injury rate of 6 to 4 in one year, does that mean that you are x percent better? If so, what is it that you did to get better? Organizations usually say: "I have no idea. Just lucky, I guess.""

Moreover, injury statistics are downstream measures. "What you need are upstream measures that tell you whether or not your system is getting any better as you go."

Instead, he offers three suggestions:

  • An activities-measuring approach where you identify for people in each level of the organization the kinds of things that they are supposed to do that you think will get results and measure whether or not they have done them.
  • Use behavior sampling, in which you systematically look at the behavior of people on a regular basis, identify safe versus unsafe behaviors, and compile behavior statistics. Over time in a large company, those behaviors will predict what your accident rate is going to be.
  • Use a validated perception survey, in which you ask people in the organization at the hourly employee level what they think about the effectiveness of the organization. "You ask the questions of people at different levels of the organization and then compare the responses. You find a tremendous difference of opinion among what management thinks, what supervisors think, and what the employees think. We believe what the employees say is reality. It tells you not only exactly where you are but what you need to fix to get better tomorrow. That"s because you can identify those various areas within the safety system that they feel are working and those areas that are a waste of time."

4) Audits predict results in safety. Petersen once backed the use of audits, but he said correlational studies between audit results and injury records in large companies show very little predictive value. That is because audits "tend to measure regulatory compliance as much as they do the effectiveness of the safety program."

For example, said Petersen, most audits start with paperwork. "Paperwork doesn"t necessarily dictate what the actions have been. There"s a little slippage there," he noted. "Also, the more paperwork you require people to do, the less they"re out there talking with their people and working on improving their behavior." Nor do the parts of audits that focus on physical conditions prove very helpful, he added, except to ensure regulatory compliance.

Can audits be a good measure of safety effectiveness? Yes, says Petersen, if the audit interviews employees about their perceptions of the safety program. This leads back, he said, to asking employees not only "what is wrong, but how to fix it. That gets back to involvement."

5) Regulatory compliance ensures safety results. Being in compliance with OSHA and ensuring you have a safe workplace, Petersen stresses, are "totally different things." Compliance with OSHA, he said, "demonstrates to the hourly workers that their safety is important to management, but it certainly doesn"t ensure" that a company will have a safe workplace.

Petersen said injury statistics bear him out. Over the last 25 years since OSHA was founded, he said, the physical layouts of factories and other workplaces are "safer than we"ve ever had before." Fatalities have been reduced. But there is no "significant change" in either minor or lost-time injuries and lost days are "50 percent worse than they were 25 years ago."

He said Heinrich"s maxim that the vast majority of accidents are caused by unsafe acts provides a clue to the problem with OSHA. "If we concentrate on the 10 percent [of accidents that are due to] unsafe conditions and we have a law that fixes the 10 percent, how much improvement do you think we can get?"

Little wonder, then, that Petersen opposes the movement toward standardizing safety programs embodied in California"s SB 198 and OSHA"s guidelines for safety management, now being examined as a possible standard. Petersen said he finds it "difficult, if not upsetting" to find the country going in that direction "when the research clearly says "That ain"t it.""

Asked whether we need OSHA, Petersen said the issue is debatable. "I think OSHA has been, and probably is, good in making the worst companies better. I don"t think it"s been good in making the good companies better."

Management and Leadership

If concentrating on physical conditions produces only limited benefits and there are no "essential elements" of a safety program, what does work? Petersen says the answer lies with industry "working toward the concept that achieving results in safety is a management function."

For their part, managers must learn that "they have to do things within their line management structure on a regular basis that develop the culture" that produces safe behaviors.

Petersen professes not to know what those "things" are. Instead, he makes the case that organizations must "unleash" their management skills on safety problems. "We"ve got brilliant people out there that can do anything in terms of productivity, quality, design of products and other things," he notes, but when it comes to safety, they throw their hands up and say they don"t know how to do it.

"That"s baloney," he continued. "All they have to do is think about it and they could do it. But, we"ve successfully taken it away from them and said this is a safety guy"s job, this is a government function, this is not yours. The more we do that, then all they have to say is "Tell me what to do and I"ll do it.""

Most safety experts believe top management support of safety is crucial, but Petersen points out that its role in safety is rarely defined. For many CEOs who come up the ranks from legal, financial or other nonoperations routes, says Petersen, their role has been to "sign a policy and that will take care of it." That won"t do, he adds, unless the policy is "backed up by action, by decisions that show that safety is important." Systems are needed to measure safety actions and reward structures are needed to demonstrate that safety performance is valued as highly as other areas. Yet, he notes, "All those things are missing in most organizations."

"We need to be able to say to a CEO, "It is crucial that you not just support but lead this function. You lead it and we in staff and in the line organization will try to assist you in your safety efforts,"" he said. ""Now, let"s sit down together and figure out what are the activities and actions that you feel comfortable doing that are going to demonstrate to people all the way down to the bottom that it"s a core value to you." I"ve very seldom come across a CEO who wouldn"t be happy to do something, if he just knew what to do."

The Measure of Employees

Employee attitudes, Petersen says, provide a true barometer of an organization"s values, its culture. Asked to define culture, he recalled talking several years ago to some hourly employees in a company. "They defined culture as being "that"s the way it is around here,"" he said. "Everybody knows what"s important and not important. Everybody knows the core values of the company because they live it every day. They live it in what comes down from management, the way they"re pressured on the job or pushed for production, and on the way they"re measured. Those are the dictates and inputs to culture."

To change a culture, says Petersen, you have to start by carefully assessing what things are like at the time. That means asking employees how management actions and decisions define the work environment. Then, management needs to decide where it wants to go and the steps that will take it there. Management needs to set up "accountability systems for activities and actions at each level of management," he said. When managers have demonstrated to employees that they are serious about safety, they can then go to the employees and ask "Will you help us?" He notes: "It"s not them versus us anymore. It"s a team. We"re doing this together."

Voice your opinion!

To join the conversation, and become an exclusive member of EHS Today, create an account today!