John Antonakis is Professor of Organizational Behavior and Director of the PhD program in Management at the University of Lausanne, Editor in Chief of The Leadership Quarterly, and a prominent researcher on leadership and power, individual differences, social cognition, and applied psychometrics and econometrics. John is best known for his research on charisma as well as for his work and advocacy to advance the field of social science through robust causal analysis. Find John on Twitter @JohnAntonakis Interview by Iulia Cioca and Pietro Marenco John, you have a multitude of roles: you’re a professor, researcher, Editor-In-Chief of The Leadership Quarterly, and a consultant. How did you get into Organizational Behavior and what drives you in all these activities? I started out in pursuing Bachelor of Economic Sciences degree. I guess that’s what also helped to see things in more formal quantitative ways, which characterizes much of what I do today. I switched to hospitality and tourism management and then got an MBA. I even managed a hotel in Greece for a while. In the winter the hotel was shut for renovation and I wanted to do something, so I applied to teach at a private business school. I realized that I loved teaching, so I quit my job in the hotel industry and focused on teaching. At one point we were getting audited by some professors from a UK university that validated the degree we gave. The professor that went through my course said something like this to me: “what the dickens are you teaching here in your class about managing organizations? It’s all about managerial accounting, strategic management, cost accounting and economics. Where’s the psychology? Where’s the leadership”. I was lost. The professor said “You should be teaching about the psychology of motivation, how to lead people, how to inspire them”. I thought this was a bunch of hogwash; my motto in the class was “Trust is good, control is better”. So that I could continue teaching it I started reading everything I could on psychology. I loved it! I had to learn more. So I decided to go get a PhD (at Walden University by distance learning) where I focused on leadership. Then, after a postdoc in cognitive psychology at Yale University I became an Assistant Professor at University of Lausanne in 2002, where I’m now a Professor of Organizational Behavior and the director of our PhD program in Management. Lots has happened since then. I’m now Editor in Chief of The Leadership Quarterly, which is the premier research journal in leadership, and considered a top outlet in many of the rankings. For me, the key issue is that we need to do better science and accurately report the causal forces which are operating in explaining a phenomenon. Unfortunately in taking stock of the literature I typically find that about one out of ten published articles gets it right. So, in my journal, I’m paying very close attention to making sure that the causal claims that are made by articles are justified. What is the current state of leadership practice in business – what are the fads and fashions? How do organizations select leaders? Very often, I notice that selection procedures have systematic bias against certain populations, especially women and certain ethnicities, so invariably these selection methods are not necessarily taking the best people for positions. In a master’s thesis I supervised, my student showed that good looking managers earn a lot more money that less good-looking managers, controlling for their intelligence, personality, seniority, education, and tenure; this research follows from a study I co-authored in Science, showing kids can predict election outcomes from the photos of the candidates. So, just simple things like your face can affect whether you’re hired and how much you’re going to earn. Selectors don’t realize to what extent they’re biased by very little things. Psychometric tests, when properly used, ignore your age, your sex, what you look like and everything that’s irrelevant and they try to capture your trait: ability to learn, personality, preferences, what have you. The companies that try to do it better have properly trained I-O psychologists and in-house metrics, and they mostly get it right. But most of the companies don’t. For example, many use the MBTI, which is a good as a horoscope. There’s so much fake science that makes its way into practice; overall, unfortunately I think there’s more fake science that’s impacting practice than real science. What about the practice of leadership development – is it any different? Typically companies may invest in some programs that are sold by consultants, you know, “make your leaders more emotionally intelligent”, “make them more charismatic”; whatever. That’s very nice but unfortunately most of this stuff has never been tested and so we don’t know what the policy implications are. They advertise on the basis of “research” that has never been scrutinized by scientists and cite a long list of happy clients. Imagine if people could sell medicines on the basis of clients’ testimonials: “Company X used this program and they loved it”; “Here’s a manager who said the program was fantastic”. That’s not evidence and many consulting companies have ZERO evidence that they have tested their products scientifically. Popular books on leadership are similar. Daniel Goleman on Emotional Intelligence, for example, said “we studied the top performers and we found that all had something in common: they are all emotionally intelligent”. Good to Great, Built to Last, 7 Habits of Highly Effective People, all these books studied “the best” and they found what they have in common. It’s like studying the top CEOs in Switzerland and saying their names are Hans, Ulrich, Juergen, Joerg, and what have you. What do they have in common? They’re male. If the top 10 performing companies have a male with a Germanic name as CEO, does it mean that you need to have a Swiss German male to run a top performing company? No. Because the bottom performing companies probably also have Swiss German male CEOs running them. The problem is where do you start looking for commonalities and where do you stop? They all wear clothes, probably have a nose, and went to school. Maybe it’s that what’s causing good performance? We don’t know from “research” that claims to have studied top performance and unlocked it secrets. Much more needs to be done to understand what drives top performance by also studying bad performance. At ScienceForWork, we go through a lot of research on leadership and how it’s associated with various outcomes, from followers’ satisfaction to bottom-line results like turnover. We often see that transformational leadership in particular is associated with positive outcomes. What are the perils and the benefits of the transformational leadership framework? In my paper on what I call “Instrumental Leadership” we showed that it’s a bit naive to think that the Alpha and Omega of leadership is transformational leadership. First of all, the effects of transformational leadership are vastly exaggerated if you don’t account for the leader’s task oriented expertise. If you only ask a person “is the leader transformational?”, they will rate the leader higher or lower on transformational leadership; this rating will especially be correlated with things you don’t ask or observe like if the person is an expert in what they are doing. If you don’t measure that expertise, you can’t take out that variance that’s due to the expertise. There are many omitted variables that can account for the observed correlation between transformational leadership and outcomes. Alas, most studies do not consider such statistical issues and make very extravagant claims. The same goes with calling for leaders to be more authentic, servant leaders, or whatever. A lot of the research behind this is based on ratings of people’s perception of the leader: “Is the leader authentic?” and then it correlates those perceptions with some outcome. But my impression of someone being authentic can be driven by many things: by how the leader behaves, by how the leader looks, by my evaluation of how appropriate I think the leader’s behavior is. For example, in a paper written in 1978, Lord, Binnings, Rush and Thomas showed that if you ask someone to rate the leadership style of a person, that they will be very significantly biased by their knowledge of how performing the person has been; raters “fill in the blanks” in a stereotypical way that makes sense to them. So for example, if I ask you to rate your boss and you know that your organization is doing well, whether or not this is due to your boss, whether or not your boss has any causal influence in that, the simple fact is that knowing the outcome will make you attribute the performance cue to the boss and you will rate the boss higher on any intuitively prototypical characteristic of leadership. So you will say the person is more charismatic, you will see them as more authentic, more servant, whatever. By asking people’s perception of leaders, we cannot identify what causes leaders behavior, we simply observe an outcome that we correlate with another outcome; both those outcomes are probably caused by something else. We need to model this “something else” or to use some other appropriate statistical correction. We need to look for the distal predictors, distal meaning ”what caused the outcomes”? What causes the leader to behave and be seen in a particular way? It might be intelligence, it might be personality, or something else. We need to find these causes, in this case traits—and then we can select leaders on the basis of these traits. At this time, the research on leadership has been spinning its wheels a lot by finding correlations between outcomes and not looking for causes. What is the alternative? Which good leadership models have been established in research and could be used in practice? We know from personality research that certain aspects of personality matter a great deal to make leaders more effective; traits that matter include being more extroverted, conscientious, open minded, emotionally stable. We also know that we can measure it very well, with good personality tests. Intelligence also matters for both objective and subjective performance. However, the outcome makes a difference: observations of effectiveness are very different from actual effectiveness. So for example, a leader must be smarter than the average person in his or her group, in order to be effective. But, if you are too smart, people might not see you as effective, because they might not understand what you’re talking about, they might not identify with you, and so forth. So, intelligence actually has a curvilinear relationship to subjective performance. In other words, if leaders are too smart, sometimes they might not be perceived as being effective by others, as we have shown in some recent work. Another thing that’s very important is good old task-oriented leadership. As a leader, you need to understand the system in which you are operating, the strengths and weaknesses of the organization, the strategy to choose, to show your followers the path to the goal, to provide them feedback on how to improve their task-oriented expertise. Only then can you communicate an inspiring vision and be charismatic and so on. But being charismatic is pretty much useless if you don’t know what you are doing. It’s like getting on a plane: I don’t care if the pilot is charismatic. I want to know that the pilot knows how to fly the plane. If the pilot is charismatic and convinces me to get on the plane all the better, but I’m still not going to get on that plane if the pilot doesn’t know how to fly it. It’s the same thing with running a company or being a doctor: knowing the task is very important. I am not saying managing humans is not important. Most of my work specializes on charisma and how leaders can inspire others, so yes, motivating and inspiring is very important. But I must say it over and over again: it’s naïve to think that we can just run companies by being charismatic; you need to know the nuts and bolts of the organization, you need to know the operating environment, you need to know how to choose the strategy. Once you know that, then managing the people aspect makes complete sense and is really important. Thus, we should go back to the basics a bit with leadership and figure out how to walk first before we can run. We see more companies pilot and test ideas before implementing them in the whole organization. Is this a good way to become more evidence-based and improve practice? Well depends on what we test and how we test it; at the outset, “yes” experimentation is good but it needs to be well done. I once said to an HR Director “we could randomize your managers to a control group and a training group and I bet I could make the managers in the control group improve their leadership style without doing much with them”. So we ran a test, which actually resulted in a published article (see Study 1). We took some managers, we randomized them, they assessed themselves on leadership, they got assessed by their coworkers, then we trained one group, and then three months later we re-measured them, and both groups improved! But we did nothing to the control group. Why did they improve? They knew they are on a training program and they are being observed and would be reassessed. So they may have just changed their leadership styles in a way that they believed was making them appear more effective. They also might have reflected on what we were measuring. It’s like in a weight-training program where you know your weight will be measured in a month from now, so you might start taking the stairs more often, you might not have that ice-cream at the end of the meal; also, instead of having a siesta on the weekend, you might go on a walk. There are many things that can happen. This is why we need a control group, we need to measure the effects of interventions compared to a placebo effect or some kind of bogus training. Gary Loveman, who was a Professor at Harvard Business School and former CEO of Caesars Entertainment, a giant gambling company, liked to say there are three things that can get you fired from Caesars: theft, sexual harassment, and running an experiment without a proper control group. Now, that’s the kind of thinking I want to see more in companies, which is the only way to find whether policy changes actually worked. It is not hard to run an experiment or use some sort of quasi-experimental design. Behavioral economists and experimental psychologists can help. Just ask an expert! Related to the above, a new trend in business and HR is people analytics, where basically practitioners attempt to show a measurable impact of interventions on hard outcomes that are relevant for organizations. Does this make you optimistic and what are your thoughts around this trend? There are many interesting developments in this area and there is much to learn still. From what I’ve seen, however, a lot of the stuff is very much predictive and correlational. It’s not really causal; it doesn’t estimate the treatment effect of a policy change that’s given randomly or using some causal identification mechanism. What people analytics is really trying to find is what x variables correlate with y outcome. Oftentimes, these analytics processes use machine learning or deep neural networks, or exploratory regression algorithms. I mean, that’s very interesting, it’s very nice to know; however, it all depends on how things are done. Oftentimes, machine learning algorithms become a “black box” and not even the scientist who developed the algorithm knows exactly how that decision has been taken. Still there is use for these things and with time more progress will be made so I am optimistic about how analytics will help in the future. At this time, I think these methods are interesting, and very useful especially if they can make predictions ex-ante about what might happen. What do you see as a good way to go forward? How should collaborations be like between academics and businesses in order to bring benefits for both? More should be done to get researchers and companies to work together. There is really not enough done; it seems to me that most of the top researchers are in the ivory tower, write publications for other academics, and do not interact with the business world. Businesses have problems that they need to solve and they need to solve them quickly. I think behavioral economics is an example, which shows that a lot of collaboration can happen. There are more and more field experiments that are done by behavioral economists in companies where they show that leadership matters. There’s an example of a great field experiment to show that management works, which was published in the Quarterly Journal of Economics. Here they randomized plants and focused on improving management practices – they showed it paid off – good management works! Things like this can be done too on a small scale, as has been documented in this publication. I think managers should also read a couple of basic books – beginning with Leadership BS by Stanford Professor Jeffrey Pfeffer and The Halo Effect by Phil Rosenzweig – about how they should select training programs, how they should speak to consultants, on what basis should they invest their money in training programs. Managers should also seek to buddy up with academics who are very much driven by understanding causal processes. With time, we will find solutions to the world’s problems and how we can make better use of resources. Not only is it the economical thing to do. It is the ethical thing to do. John, thank you for your work to advance science, policy and practice through rigorous and robust research, and for offering our audience a dose of healthy skepticism on leadership in organizations. ScienceForWork is an independent, non-profit foundation of evidence-based practitioners who want to #MakeWorkBetter. Our mission is to provide leaders and decision-makers with trustworthy and actionable insights from behavioural science. Did you like this interview? Share it with your network by clicking on the buttons below! Follow us on LinkedIn and Twitter to receive the most trustworthy scientific research summarized in less than 1000 words! Resources for researchers Antonakis, J., Bendahan, S., Jacquart, P., & Lalive, R. (2010). On making causal claims: A review and recommendations. The Leadership Quarterly, 21(6), 1086-1120. Hamilton, B. H., & Nickerson, J. A. (2003). Correcting for endogeneity in strategic management research. Strategic organization, 1(1), 51-78. Ketokivi, M., & McIntosh, C. N. (2017). Addressing the endogeneity dilemma in operations management research: Theoretical, empirical, and pragmatic considerations. Journal of Operations Management, 52, 1-14. References Antonakis, J., & House, R. J. (2014). Instrumental leadership: Measurement and extension of transformational–transactional leadership theory. The Leadership Quarterly, 25, 746-771. Bloom, N., Eifert, B., Mahajan, A., McKenzie, D., & Roberts, J. (2013). Does management matter? Evidence from India. The Quarterly Journal of Economics, 128(1), 1-51. Blumberg, A., & Goldstein, J. (Producers). (2011, November 15). From Harvard Economist To Casino CEO [Audio podcast]. Lord, R. G., Binning, J. F., Rush, M. C., & Thomas, J. C. (1978). The effect of performance cues and leader behavior on questionnaire ratings of leadership behavior. Organizational Behavior and Human Performance, 21(1), 27-39. Pfeffer, J. (2015). Leadership BS: Fixing workplaces and careers one truth at a time. New York: Harper Business. Rosenzweig, P. (2014). The halo effect:… and the eight other business delusions that deceive managers. Simon and Schuster.