The Safety of Work

Ep.26 Is good safety leadership just good leadership?

Episode Summary

Welcome back to the Safety of Work podcast. Today, we ask the question: Is good safety leadership just good leadership?

Episode Notes

We use the following papers to frame our discussion: Development and Test of a Model Linking Safety Specific Transformational Leadership and Occupational Safety and Contrasting the Nature and Effects of Environmentally Specific and General Transformational Leadership.





“How much do these things vary and how much do our explanations for these things explain why they vary? And the answer is, they don’t.”

“Don’t start measuring and tinkering with the statistical relationships between things until you’ve actually pinned down what those things are.”

“I strongly believe that we can’t easily change the values that people hold.”



Barling, J., Loughlin, C., & Kelloway, E. K. (2002). Development and Test of a Model Linking Safety Specific Transformational Leadership and Occupational Safety. Journal of applied psychology, 87(3), 488. DOI: 10.1037//0021-9010.87.3.488

Robertson, J. L., & Barling, J. (2017). Contrasting the Nature and Effects of Environmentally Specific and General Transformational Leadership. Leadership & Organization Development Journal.

Episode Transcription

Drew: You’re listening to the Safety of Work podcast episode 26. Today we’re asking the question, is good safety leadership just good leadership? Let's get started. Hey, everybody. My name's Drew Rae and I'm here with David Provan. We're from the Safety Science Innovation Lab at Griffith University. Welcome to the Safety of Work podcast. If this is your first time listening, then thanks for coming. If this is not your first time listening, then thanks for coming back.

The podcast is produced every week releasing on a Monday morning, and the show notes can be found at In each episode, we ask an important question in relation to the safety of work or the work of safety, and then we examine the evidence surrounding it. David, what's today's question?

David: Drew, the question for this episode is, is good safety leadership just good leadership? There are some common sayings in business about safety like good safety is good business, and safety and production are compatible goals, and then this question that we're exploring today: good safety leadership is just good leadership. The assumption here is that good leadership is, well, good leadership. If you're a good leader, then you'll automatically be a good safety leader. Some companies have a general leadership development program, and others have safety specific leadership development programs.

What we're really trying to explore in this episode is, is safety leadership something that's distinct from good leadership? Before we dive into that, Drew, do you want to give us an overview of transformational leadership as an idea?

Drew: Sure. I thought it might be worth—before we look at the particular papers—having an idea of how people study leadership. There are lots of different fields that study leadership. We do it in organizational psychology, we do it in psychology, we do it in human resources, we do it in management, pretty much every subfield, you have civil engineering safety, all have their own studies of leadership. When you talk about different leadership theories often we're not actually talking about different leaders or different ways of leading. We're just talking about different academic ways of studying it.

Some things you might have heard of—you might have heard of servant leadership, you might have heard of transformational leadership, you might have heard about leader-member exchange. All of these could be talking about the same leader. It's not like you make a choice to be one or the other. They're just different ways of studying the topic.

Transformational leadership comes mainly out of an approach to studying leadership, which is about using surveys to characterize different leadership styles. When you hear transformational leadership, usually that's in contrast to transactional leadership or laissez-faire leadership. The idea is we use a survey to say which style is someone tending to apply. The survey goes either to the leader to see what their preferred style is, or we spread the surveys out amongst the followers—usually the employees—and ask them about their boss, and to see what style they observe the leader doing. To see whether someone is following transformational leadership, there are four scales that we're giving the leader a score on.

The four things are, firstly, idealized influence. This is, is the leader acting as a role model that other people want to be like? The second one is inspirational motivation. Is the leader presenting a compelling vision that people want to follow? The third one is individualized consideration, which is, is the leader caring about the needs of each member of their team—as a separate individual compared to just you motivating the whole team? Then the fourth one is intellectual stimulation. Is the leader challenging people to grow themselves, and to innovate, and to move beyond the status quo?

When we talk about safety leadership, the heart of the question is how do you get from those four things to safety? Is it just you take each of those four things, and then you add a fifth thing which is how much you care about safety? You do each of the four good leadership things, but you do them by someone who cares about safety, or do you modify each of the four things separately so that you have idealized influence about safety, inspirational motivational about safety, individualized consideration about safety where it's actually a different thing because safety is part of it? You're not just adding the word safety.

That's really the question is, do we just train people to be good leaders and make them care about safety, or do we specially train people to be safety leaders because that's something different? That's what we're going to look at when it comes to the research. David has picked out two papers to talk about this relationship between general transformational leadership and specific or subject-oriented transformational leadership. We're doing it of course because we care about safety specific leadership.

David: Thanks, Drew. When I started looking—because I wanted to do a podcast on safety leadership, just looking back through the podcast that we've done, and I actually found that we talk a huge amount about safety leadership when it comes to safety management in our organizations. In safety science, there's not that much research really about safety leadership specifically, but I did find two papers. One about safety specific leadership, and one about actually environmental leadership, which we're going to draw some parallels back to safety.

The first paper is titled Development and Test of a Model Linking Safety Specific Transformational Leadership and Occupational Safety. This was published in the Journal of Applied Psychology in 2002. Drew, my understanding is that's a quite a reputable Journal for psychology publications.

Drew: Yeah. It's not one of the absolute leading edge, but it's reputable.

David: The authors are Julian Barling, Catherine Loughlin, and Kevin Kelloway. These were all from Canadian universities across different schools associated with business management and psychology. When transformational leadership was becoming popular or study of leadership more broadly was becoming popular in the early 2000s, as far as we can work out, as far as I could work out, this paper was really the first to expand that concept of transformational leaders—or not that expand the concept because it has been around for a while—but really try to develop this topic of specific transformational leadership to a particular priority within the organization.

They determined that safety specific transformational leadership might actually be a thing, and not only would it be important to understand if it is a thing, that they could link it to safety injuries within an organization. At some point, I'm expecting, Drew, you’re going to lob some kind of grenade into this podcast about trying to understand and use injuries to measure safety. If you want to have a go at that now before we jump into the method, feel free.

Drew: Let's put a pin in that one, but we will come back to it.

David: Okay, very good. There were two studies in this paper. For study one, what they did is they developed this model, which was based on the literature and particularly an earlier study by Hofmann and Morgeson in 1999 that really took 49 different dyads of variables—just two variables—and looked at 49 different relationships between two variables. They found in this study that accidents were mediated or influenced primarily by safety communication and safety commitment, and that these attributes were influenced by leader worker exchanges.

These authors in this study actually took the idea that, “Hang on a second. If literature says that there are these two things, we can draw some kind of model which links transformational leadership to workers’ safety consciousness and the perceived safety climate of the organization, and then that climate would link to safety-related events, which would then link to occupational injuries.” They drew this pretty linear flow model of what they believe the relationship might be between transformational leadership and occupational injuries.

The participants for the study were 174 workers, and they recruited them from the food and beverage industry. As an interesting side note, Drew, from my early psychology studies, when I read their recruitment method, they were putting posters all up around town trying to get people to participate. They went directly to hotels and restaurants. They sent emails throughout their university saying, “Hey, does anyone work in a restaurant, or a McDonald's, or something like that.” These are pretty old-school recruitment techniques from psychology studies. Do you see too much of those types of recruitment processes on campus today?

Drew: You certainly see still these flyers up on walls with the little tear-off things at the bottom. “Do you want to rent a flat chair? Do you want to be part of a fitness class? Do you want to participate in this study?” Yeah, it still does happen.

David: They found 174 workers. What they wanted to do, so they wanted to test this model, these five boxes in their diagram. They designed a battery of questionnaires, and there was a multifactor leadership questionnaire to understand transformational leadership. Questions like, “My supervisor talks about his or her values and beliefs about safety.” They had 10 items that’ll ask how to assess safety climate—remembering this was 2002 so Zohar’s work on safety climate was quite prominent at the time. These are questions like, “Upper management assigns a high priority for safety.” Then they generated these scales for safety consciousness, safety-related events, and occupational injuries.

I’ll mention your thoughts on the design of these, Drew, because what they said was for safety consciousness, they came up with seven items like, “I always wear my PPE. I'm well aware of the safety risks at work,” and those things. For safety-related events, they came up with 11 common events in the food and beverage industry like someone slipping on a slippery surface, or someone cutting themselves with a knife.

Then they came across eight common injuries like sprains and strains. Interestingly for this, Drew, one of the things is they didn't go to the organizations, and they didn't get the information out of a safety incident reporting databases. They just asked the participants how often these things had happened in the last 12 months. What are your thoughts on this group of this question set?

Drew: I was going to talk about this a little bit later on, but I think it's probably worth jumping in with it now. This is the big challenge of doing research where you are using one survey to try to measure the relationships between a bunch of different things. There are always at least two possible explanations. One explanation is that the apparent relationship between the things is a real relationship. You give someone a survey, you measure what they think of their leader, and the safety climate, and the safety outcomes. Those three things people give similar scores to. You say, “Okay. One causes the other causes the other.”

The second possibility is just that all of those things are blended together in people's minds. Someone who loves their organization says, “Yes, I like my leader. Yes, it's got a great safety climate. No, there aren't any injuries.” Someone who got hurt last week and fired says, “No, I hate my leader. My organization's got a terrible climate, and we have lots of injuries.” It's not that one is causing the other. It's just that the concepts aren't separate. There's no way within a survey, no matter how much you do sophisticated model analysis, there's no way to separate which of those two explanations it is. Whether it's cause or whether it's correlation because they're part of the same survey.

David: Thanks, Drew. That's a really helpful explanation of just when you think you might have these different constructive testing, but they may not actually be different. The results of this study—they performed a statistical analysis technique, which was predominately covariance analysis, which really just tests the dependency between two variables. There were seven variables that they were looking at. The five of them being the ones in the model. For people that remembered, there’s safety transformational leadership, safety consciousness, safety climate, safety events, and safety injuries. Then they added to demographic variables: age and tenure in role. That's important because I think they just want to look at how long people have been in the industry, in the organization, and things like that.

The conclusion was that each of the links in the proposed model was statistically significant. They found that from their data, the model accounted for a substantial amount of the variance in the raw data, which means that the model explained a fair amount of what might be going on between the different constructs and within the data. The authors concluded after study one that they felt that they had extended the work on the link between safety climate in injuries, and extended that out back towards transformational leadership, and its impact on safety climate. Drew, your thoughts on the findings?

Drew: For a single survey, they are overanalyzing the data that they have. The reason why you'd usually include something like role overload is not to include it as a variable in your model but just to test whether adding any random variable is just as good an explanation as any other. I almost suspect that that's what happened here is that someone's sort of misunderstood why some of those factors we're in their survey and just amalgamated them all into a model, which has got too many links and too many explanations for the amount of data that they had.

David: They rolled on from this first study, and they decided to do a second study. I suppose because they thought they're onto something with their model, and they thought, “Well, we were just studying the restaurant industry. How valid are these findings outside of the study population and so on?” They did a second study, and they got 254 participants in the second study. They specifically went after young workers. Drew, they said in the paper that they didn't want workers who had a lot of career experience to confirm their data, but I also feel like it was easy to get 250 for uni students into their second study. Nevertheless, they got 254 people, and they went back into the literature a bit more.

They were trying to now test their model against other things that they thought within the literature were linked to injuries. One of those things was goal conflict around productivity pressure and role overload. They added that into their model and said, “Well, we think we've got this relationship between transformational leadership, and safety climate, and safety consciousness, which then flows down to injuries.” But other parts of the literature then link productivity pressure to safety climate, and safety consciousness, like people cutting corners, and people not thinking safety is a priority.

In their second study, they wanted to introduce questions around that so that they could see if that would then account for a greater understanding of what was going on. They took the questions from the first study, they added some extra questions in. For example, “There's too much work in my job for it to all be done well,” or “I'm so busy that I can't get to take normal breaks.” When they ran this similar study with these extra questions across this younger and broader sample of people, they concluded that, like you said, they'd overanalyze it probably again a bit, Drew, where they concluded that the model provided a reasonable but not an outstanding fit.

My general interpretation, they probably identified some relationships that exist that we might all understand. “Okay, well, a safety transformational leader is going to have an impact on the safety consciousness of the workers, and that's probably going to have some impact on safety climate, and safety climate may have an impact on occupational injury.” They've found these relationships, which we might all think is probably okay, but we know that this analysis of this data does not give us some compelling case that this is all we need to worry about for safety.

Drew: Whenever you read a paper that says, “This is a good but not outstanding fit,” or “This is approaching significance,” always read that as, “We set some targets for ourselves that we were going to consider was a good outcome, and we failed to meet those targets, but we decided to publish anyway.” They were having a model, they hoped the model was going to fit, the model didn't fit, and the model is not a good explanation. There's a huge amount of variance that is not accounted for, which is what they're trying to do. They're trying to see how much do these things vary, and how much do our explanations for these things explain why they vary. The answer is they don't.

Safety leadership and safety climate don't explain the variability in what people perceive and how safe they think they are, which could mean that there's some other factor that's causing the differences, or could just mean that the constructs aren't very well-defined. I'm inclined to think it's the latter. I'd point people back to episode 20 when we talk about the manifesto paper that we published. One of the clear principles in that is don't start measuring and tinkering with the statistical relationships between things until you've actually pinned down what those things are.

In this case, your transformational leadership, it wasn't a well-understood concept. It was something they invented for the purpose of this study. They did it by taking—sorry, safety specific transformational leadership. They took the standard transformational leadership, and they just changed all of the words to say safety. That's a reasonable thing to do, but it's an experimental thing to do. It doesn't come from a clear understanding of what safety specific transformational leadership is. It just comes from this assumption that there is this thing.

That means that it's probably likely that there's a big overlap between what they're calling leadership and what they're calling safety climate. In fact, Zohar’s safety climate scale has been criticized for that exact thing. That a lot of the questions in it aren't actually about climate, they're about leadership behavior. What we've got is one way of measuring leadership behaviors sort of correlates with another way of measuring leadership behavior. Yeah, of course, it does, but what does that usefully tell us?

David: Yeah, that's a good example, Drew. The example question I go for safety climate is safety is a priority for upper management, so it really is specifically asking about leadership and leadership behavior. There's going to be a big overlap between those constructs of safety climate and safety transformational leadership. That is useful to say. I suppose there’s a saying that all models are wrong and some are useful, so I'd like to think that there's some usefulness in this model. We'll talk about that after we talk about the second study and the practical takeaways, but very clearly, the model is wrong like all models.

Drew: Yeah. I think as we get into the second study, we start to see how we can start exploring some of the questions in ways that do produce usable answers.

David: Should we dive in the second study, Drew?

Drew: Yes, let's.

David: Okay, great. The second study was actually the paper that I first started with when I was exploring this question before we went back to that original paper from 2002. The second study is titled Contrasting the Nature and Effects of Environmentally Specific and General Transformational Leadership. This was a study that really sought to understand if there are different outcomes associated with context-specific leadership or a general transformational leadership. The study was about environmental leadership as opposed to safety leadership, but we think we can translate some of the findings across into safety.

It was published in Leadership & Organizational Development Journal in 2017. It was submitted in 2015, and the authors here are Jennifer Robertson and Julian Barling. Barling was the first author in the first study that we just spoke about. From what I can work out, Robertson is an associate professor in Queens University in Canada. She got a Ph.D. in 2014. From what I can work out that this study was part of her Ph.D. that was done sometime around 2012-2013 because the conclusion of the paper refers back to the 2002 paper, and talks about it being about a decade earlier. This is one of the things in the research publication I suppose, Drew, is this study could have been done in 2012 and published in 2017. That's really not that uncommon.

Drew: No, that's fairly common. It's a very typical thing to happen that you have a professor publishes a speculative study but never has the time to follow up. Then new Ph.D. student, “Hey, here's my work from 10 years ago,” that really needs someone to look at a bit more rigorously. How about you do it.

David: Robertson’s research interest on a university profile includes corporate social responsibility and environmental sustainability. She was really in understanding of how leadership behaviors impact on pro-environmental behaviors of workers in the workplace. In this study, that's really what they wanted to understand is, is there a difference between environment-specific transformational leadership and general transformational leadership, or a control group on predicting worker engagement in pro-environmental behaviors. There were two studies in this paper again, and we'll run through them.

The first study really tried to explore if there was actually a difference between environment-specific transformational leadership and general transformational leadership. Drew, this is one of the things you called are they the same thing, or are they a different thing? So 185 participants completed an online survey with two questionnaires. One that measured this environment-specific transformational leadership, and one that measured general transformational leadership. These questionnaires showed that while these two constructs—if you like—are related, they are in fact distinct constructs based on the two question sets that people were given.

Drew: I was very disappointed when I looked at the detail of the statistical analysis here because I think they actually missed exactly what the question is. Essentially, they compared two models. One of which they said, “Okay, let's substitute everything and make it all about the environment, or let's do it generic about leadership, and let's see if people's answers are different.” Of course, their answers are going to be different. You ask people different questions, they're going to give you slightly different answers. The questions are very similar, your one bit has just got environment substituted in, so of course, your answers are different but correlate. They sort of point in the same direction.

The really interesting and important question here is to compare it to a third model where you've got an additional factor. Call it something like concern for the environment. The question is whether the data best fit a model where environment leadership is just general leadership plus scoring highly on concern for the environment, or is the best model one where you've got two actual different types of leadership?

That's really what we want to know. That's actually what's useful because that then tells you, “Do we train our leaders just in good leadership plus we tell them to care about things, or do we need special training for safety, special training for the environment, special training for financial leadership?

David: Drew, you would have given a third condition or a third set where you gave the general transformational leadership questions, and then one question on the end which is like, “My manager is concerned about the environment on a scale.”

Drew: No. I think this one is purely in the statistical analysis you do. They're trying to say, “Does the data best fit that there's one thing here or two things here?” Actually, the correct comparison isn't between one thing and two things, it's between two things and three things. Is there a third mystery factor that the variance in the data doesn't account for? If there is that third factor and that accounts for most of the variance, then that's just, do people care about the environment?

David: They've concluded in this research that they were distinct things, so now they wanted to go and see if these two distinct things—this environment-specific transformational leadership and general transformational leadership—created different outcomes amongst participants. They took—again, I suppose indicating it might be a Ph.D. project—a convenient sample of undergraduate students. They basically enrolled 196 participants, and like we've mentioned in other episodes, Drew, is that actually, in psychology studies, they actually didn't want the participants to know exactly what they were trying to measure.

When they recruited the participants, there was something like 40 participants or 41 that figured out what the study was about, so they got booted out of the study. They were left with 155 participants. They randomly assigned these 155 participants to three different conditions: either a controlled condition, a general transformational leadership condition, or an environment-specific transformational leadership.

Then what they did is they recruited a confederate, so that's basically someone who others think is a participant or some other real person, but they're really working for the research team. Drew, immediately think of things like the Milgram experiments with electric shock where they had participants give electric shocks to people who got answers wrong. Do we use this often in research?

Drew: Typically we do it less off. It used to be a real thing in the 1960s and 1970s where you'd have this person who's like pretending to be another participant who falls over, or bumps into people in the corridors, or encourages a participant to steal something, or to cheat on a test. In this case, the confederate isn't like a secret competitive veteran, they’re active presenting a video. There's no pretense that they're a participant. Everyone's just asked to watch a video of someone who is supposedly a CEO but is actually an actor pretending to be different types of CEO.

David: They've got this, let's call him a confederate, but let's start calling the CEO of this fictional company. They trained this person in environment-specific transformational leadership and general transformational leadership. Then they developed these really extensive scripts that the person would read out, and they tested those scripts with people who are familiar with general transformational leadership, and they practiced and rehearsed, and got the script right down.

They use one person to present across the whole lot. They presented the general transformational leadership condition as well as the environment-specific transformational leadership. They prepared these videos and those videos were then going to be watched by participants.

What happened is the participants—these 155 participants—they turned up in groups of 7-10. What they did is they arrived, and they were told that they're here because they're business students of the university, and they now have to pretend that they're working in a fast-food organization that was restructuring its budget to expand operations. They were told that the organization wanted the support from these business students to know how it could restructure their budget.

They needed to consider what they did about expenditure for marketing, training, product development, operations, environments, safety, and all the other areas of the business. Then they watched the video. This is from whom they thought was the CEO that we mentioned earlier. He was going to be giving them instructions and useful information about the company and how to complete the task.

After watching the video, the students filled out a survey that contained all the different budget questions and the things that they recommended that the organization do. These three conditions, like I said, there's a general transformational leadership where the CEO provided this big vision for the future. There was the environment-specific transformational leader, which provided a vision for the future and also recognized the organization's environmental priorities. Then there was the control where the CEO provided some transactional information about the tasks, the organization, and generic suggestions without trying to be visionary or transformational in any way.

Basically, they then gave a survey. These people watched this video. They did this survey that contained transformational leadership questions, perceptions of the leaders, environmental values, leaders in role behavioral priorities—so what people thought were the behaviors and priorities of the leader—and then what they thought about workplace pro-environmental behaviors. This is what they would then do in that type of organization. Interestingly, at the end, Drew, just how much they liked the leader.

Drew: Just to be clear here, even though it's presented as a survey, some of this is essentially doing the budget tasks in the survey. It's not just asking their opinions, it's also asking them to do things like assign amounts of money to different parts of the budget. We're seeing both what they think the leadership style is—how much they like the leader—and also how much that has influenced them in the performance of this particular task that they've been asked to do.

David: Yeah, it's a good point, Drew. They're like specifically how much money they would take out or add to the environmental budget in the restructured organization and so on. Fairly obviously, the environment-specific transformational leadership was a greater prediction of key environmental outcomes compared with general transformational leadership. For example, the environment-specific transformational leadership condition, the participants provided more budget to the environmental line-item within the new organization. They also responded that this condition provided a greater clarity of priorities for the leader.

Then they found, Drew, that I suppose there was no difference between the pro-environmental behaviors in the general transformational leadership and the control condition. When the leader didn't mention anything about the environment, there was no difference in whether they were a transformational leader or a transactional leader in the way that the survey was completed. That was probably one of the more interesting findings within this and maybe less obvious findings.

Drew: Something that you said at the start of the podcast, David, was that some people have this idea that, “Oh, you don't need to specifically focus on safety.” If you're a good leader and if you manage people well, then they will just want to do the right things. That's what this finding contradicts. It says that when someone—as the leader—doesn't just be a good leader but be as a good leader championing the environment that supports the environment. When someone is just a good transformational leader, “Oh, that makes really no difference in how much people,” because they don't see that as a priority for the leader. They don't follow along with that priority.

David: Maybe we’re into the practical takeaways, but we've got a little bit more to go. On this idea that you can just be an inspirational and motivational leader and make the jump to just assuming that people then be able to translate that into, “I need to do the right thing for safety,” is a jump that this research goes as people may not always make.

Drew: Yeah, absolutely. The rest of it though is obvious that if you have someone talking about the environment just before you fill out a budget, then you're going to focus on the environment a little bit more in that budget. The authors—to be fair to them—did recognize—in hindsight—that what they really needed to do was have the control condition mention the environment so that it was part of the leadership priorities—just not built into that transformational leadership.

David: There was another interesting takeaway that you thought from the paper as well, Drew. Do you want to talk about the training of the actor?

Drew: Yeah. I thought—and the authors obviously thought—this was interesting. If nothing else, what they proved is that you can actually train someone to be a different type of leader. At the very least, when you're getting the CEO to do a video, or an email, or communication, then with the right training and the right script, the participants not only saw the leader differently but actually did behave differently on the task.

A lot of people think that those vanilla communications from the CEO are just you're overly scripted and not useful. What it says here is that you know having an expert have input into how the communication is made does make a difference in the success of the communication.

David: Maybe this talks about safety for specific transformational leadership if you have a good leader, or just any leader, or CEO, and you, for some reason, want to increase your priority around safety, or it becomes a new priority for the organization, then you can actually train and help your CEO deliver safety specific transformational leadership messages into the organization. That's a useful finding. Drew, should we move on to practical takeaways now?

Drew: Sure. Although, I think this might be practical disagreements coming out of this one.

David: That's okay. I'd tend to get caught up in the author's hype around their own studies, whereas Drew tends to bring them back down to earth a little bit, so this is always good. My practical takeaway from looking at these couple of papers, and having a read, and to think about my experience in organizations, which is always a bit dangerous. I believe there is a difference between general transformational leadership and the safety of transformational leadership. That good safety leadership is not necessarily good safety leadership. Drew?

Drew: I agree with that, but I think that's actually a personal opinion rather than something that the evidence from these studies supports. The important question is whether safety leadership is just good leadership plus focusing on safety, or whether organizations should be investing in special training or education in safety leadership?

I'd strongly believe that actually, we should be investing specifically in safety leadership, but my reasons for that don't come from these sorts of studies. It's because I think safety presents a number of situations and challenges for leaders that make it really hard to consistently follow the same leadership styles and practices that make them good leaders in other areas.

Someone can be really good at showing individual concern and compassion for their workers, but it's not obvious what that means when you were responding to a safety incident you've had an investigation that says someone's at fault. That's where it actually helps to have dedicated training in what is a second victim, how do you treat second victims, how does the story of justice work?

It's not obvious how to be an ideal role model when you're working in an office and safety is about someone working at the side of a busy road. That's where your models of things like work as done versus work as imagined. You're understanding the relationships between procedures, rules, autonomy where it's not obvious, and we're actually getting someone to specifically think through how leadership applies to safety is going to be helpful.

Those are the reasons I agree absolutely that good leadership isn't necessarily good safety leadership, but I don't think the research is helping us here. I don't think it's giving us useful answers on how we could or should train people in safety leadership. Everyone's got an opinion, everyone's got something to say, but there's not a lot of evidence that we can apply.

David: If I throw that in my opinion on the table, Drew, as well. When we think about the second paper where you might have a hugely inspirational general transformational leader in your business, and in relation to things like company growth and market share. I think of people like Steve Jobs, and Elon Musk, and things like that, and they're hugely transformational leaders. Without knowing that safety is a priority—so even if it is just this plus concern for safety at the end of them being a great transformational leader—without knowing that, then people fill in the gaps from what they do see is important. Maybe things like production, or cost, or growth, or market share, or so on.

If safety is not something that's dealt with specifically in any transformational leadership, then workers are going to fill in the gaps in priorities, and they're going to make trade-offs in these conflicting goals. In my opinion, practically for organizations, if they assume that because they have a general leadership development program, they can't necessarily assume that they don't need to have a safety specifically development program as well.

Drew: Of course, that then just throws in the question, do you do that by having a general leadership program? And putting in a bit that tells your leaders to care about safety. I can see the skeptical thing on David's face, and I'm inclined to be skeptical as well. I don't think that would work, but I don't think we've got a really good evidence-based answer for why. It's more than just telling good leaders to care about safety.

David: The skeptical look for me is that, like you said about care for the safety and then how to translate that care and concern for safety into something that actually has the desired impact on the climate in the organization, if we go back to the model from the first study. I'll tell a quick story here, Drew. I was in an organization that did some safety climate work, and they had identified that they were concerned with responses about management commitment for safety about—like we go back to Zohar's climate questionnaire about upper management has a priority for safety. Management was like, “Well, of course, it's a priority, but why don't our workers think it is?”

They went to put in place some things to show their organization that safety is important to them. One of those things was, every time there's an incident, the executive general manager will drop everything, and go out to that site, and find out what happened, and how they can provide support. The reasoning of management, at that time, was that, “Well, if there's a safety problem or an incident, then we're going to show our people that the most important thing is that we drop everything, and we go out and provide help and support. That will have them see us being more committed.”

When I went and checked a little while later with the people, I said, “Well, how's that going in terms of management commitment?” They were like, “Well, management still doesn't care. All that happens now is as soon as something goes wrong. They fly out here, try to find answers, and try and cover their own butt.” I suppose the lesson then is just adding this concern for safety to general transformational leadership without specific instruction for how to translate that concern into actionable leadership behaviors that have the desired result when it comes to how it’s received in climate, you may actually be thinking you're doing something good and not doing something that good at all.

Drew: I strongly believe that we can't easily change the values that people hold. The only useful assumption to make is to assume that our current leaders care about safety, and that if they don't seem to care then, that is a shortfall in skills and knowledge about how to demonstrate that care and how to apply that care.

That's what we need to be fixing when we talk about safety leadership training is not telling people to care more. Either they do or they don't, we can't change that, but what we can do is help people who are generally good leaders who do care about safety. How do they translate that general leadership knowledge into the domain of safety successfully? I strongly suspect that that's a specific skill set that is teachable.

David: That's a great conclusion, Drew. I definitely couldn't have said it better myself, and that's why I do believe there needs to be safety specific leadership development training in all organizations. Drew, we usually ask the listeners what we'd like to know. I'm really curious to know—just following on from that topic, organizations that have safety specific leadership development programs in their organization, and whether they have any measures of safety climate to evaluate the effectiveness of this safety specific leadership development.

If anyone is out there doing safety leadership training and also measuring safety climate over time, we'd love to hear what you're finding in your organization in regard to that. Drew.

Drew: Also, given that, David, you make a living that includes coaching of organizations, how I make a living out of teaching safety leadership, so I'd love to know what people see the deficit in market gaps are. How are existing ways of helping safety leadership not filling what you think you need in your organizations? What skills would you like people to have? What opportunities would you like them to have that you in learning outside the organization to build up that skill set? What is good safety leadership for you that you need?

David: Drew, today we asked the question, is good safety leadership just good leadership? I saw the way we're doing the practical takeaways and disagreements. My answer was a flat-out no. Good safety leadership is not just good leadership. Drew?

Drew: Speaking for myself, I agree, no. Speaking for the evidence, maybe, we don't really know.

David: That's it for this week. We hope you found this episode thought-provoking and ultimately useful in shaping the safety of work in your own organization. Please leave us a comment or a recommendation on your podcast feed. It helps other people find the show. Send any comments, questions, or ideas for future episodes to us at