Listen to the podcast:
The traditional path to success has emphasized excelling in a single discipline or field rather than being a generalist. But one writer is challenging that wisdom, contending that it’s sometimes better to be a “jack of all trades, master of none,” as the old saying goes. In his book, Range: Why Generalists Triumph in a Specialized World, investigative journalist David Epstein looks at the strengths of generalists versus specialists, focusing on how keeping a broad range of interests, experimenting and changing course every now and then are essential to finding your true passions — and the success that comes with loving what you do. The idea has implications for how businesses recruit employees and define the skills required for each position. Epstein visited the Knowledge@Wharton radio show on SiriusXM to explain why intense focus on one thing isn’t always the best approach. (Listen to the podcast at the top of this page.)
An edited transcript of the conversation follows.
Knowledge@Wharton: You start your book comparing two of the greatest contemporary athletes, golfer Tiger Woods and tennis player Roger Federer, and how they reached success. Can you take us into that story?
David Epstein: I think most people have absorbed at least the gist of the Tiger Woods story. His father gave him a putter when he was six months old. He was physically precocious and dragged it around everywhere in his circular baby walker, started imitating a swing at 10 months. By 2 years old, he was on national TV showing off his swing in front of Bob Hope. By 3, his father started to media train him. Fast forward to 21, he’s the best golfer in the world. He’s very focused on golf — large amounts of deliberate practice where it’s like technical training.
Roger Federer, on the other hand, played a dozen different sports from skiing and skateboarding, rugby, badminton, basketball, soccer, all sorts of things. He delayed specializing. His mother was a tennis coach and refused to coach him because he wouldn’t return balls normally. When his coaches tried to kick him up a level, he declined because he just wanted to talk about pro wrestling with his friends. When he first got good enough to warrant an interview from the local paper and they asked what would he buy with his first check if he ever became a pro, [they thought] he said a Mercedes. His mother was appalled and asked if she could hear the interview recording. She did, and Roger had actually said “mair CDs” in Swiss-German, which just means he wanted more CDs, not a Mercedes, so she was OK with that.
He kept playing badminton, basketball and soccer years after his peers were focusing only on tennis, and obviously he turned out OK. So, which one of these is the norm? If you look at the science instead of just individual stories, which is a norm?
It turns out it is the Roger pattern. All around the world, sports scientists track the development of athletes and found they have a so-called sampling period, where they gain these broad general skills to scaffold later learning. They learn about their interests. They learn about their abilities. They systematically delay specializing until later than their peers, who plateau at lower levels.
Knowledge@Wharton: That’s noteworthy because sports is one area where you see more and more kids at a younger age only playing one sport year-round, instead of playing the sport of the season. You have kids playing chess year-round at a very young age and not having as much time on the playground.
Epstein: That’s true. Chess is a domain where I know that early specializing does work, so I’m not dogmatic about this issue. I wanted to say, “Well, in what domains should you be Roger, and when should you be Tiger?”
We certainly need specialists in some domains. Chess is one, and possibly golf, where specialization does work because [those domains are] what the psychologist Robin Hogarth calls learning environments, where all information is available, people wait for each other to take turns, the next steps are clear, and they are based on pattern repetition. If you’re in these kind learning environments, feedback is immediate and always fully accurate, so specialization does work quite well. The problem is, the more that kind of expertise is based on either pattern recognition or repetitive motions, and the more you’re in one of those domains, the more likely it’s getting automated.
“The more varied your training is, the better able you’ll be to apply your skills flexibly to situations you haven’t seen.”
Knowledge@Wharton: From the research you’ve done, it seems that success is more associated with having a variety of experiences in your life?
Epstein: Yes, the more dynamic sports are what Hogarth called more wicked learning environments, where patterns don’t repeat and you have to do things on the fly and at speed, and you have to solve problems you haven’t seen before. You have to take skills and knowledge and apply them to situations you haven’t seen before — do what psychologists call transfer. Whether you are a kid learning math or sports or a scientist working on an unusual problem, what you need to do is transfer knowledge because you’re trying to do something you haven’t done before. And the way you set that up is with this much broader-based learning.
The classic research finding goes like this: Breadth of training predicts breadth of transfer. The more varied your training is, the better able you’ll be to apply your skills flexibly to situations you haven’t seen. You’re trying to learn how to match a strategy to a type of problem instead of just learning how to do repetitive patterns.
Knowledge@Wharton: How much benefit do young adults get from having a variety of experiences in their formative years?
Epstein: You want to find out how important is specialization timing in education, whether kids are picking something pretty early, like when they’re in their mid-teens and studying that, or if they get to do this greater variety of stuff first before they settle in. [Hogarth] looked at different countries that have similar education systems, like England and Scotland, where they’re really similar except for specialization times. In England, the students by age 15, 16 have to pick what they’re going to study because they have to test into specific focus programs in college, versus Scotland, where they don’t. They experiment the first two years and can even continue taking some other courses beyond that, so they get this sampling period. His question is, who wins this trade-off?
The earlier specializers do jump out to an income lead, but they pick worse fits for themselves. They tend to pick things they already knew about, of course, because what else could they do? They pick something they already knew about when they were 16. The later specializers catch them and surpass them by six years out. Then the earlier specializers start quitting their careers in much higher numbers because they failed to optimize their match quality — this is the term economists use for the degree of fit between an individual’s ability and their interest in the work they do. It’s the same exact pattern as with the athletes: The late specializers get out behind, but then they fly past.
Knowledge@Wharton: This also plays into something popping up in business culture, which is that more companies want to assemble teams to complete projects. Instead of having one person who specializes in a task, they want employees with a diversity of experience.
Epstein: That’s a great point, and gets right at a study that I found to be fascinating about the comic book industry. A lot of research in the business world suffers from survivor bias because we’re only looking at the companies that did really well. The business professors who were doing the study picked the comic book industry so they could follow the commercial value of comic books up and down for decades. They were wondering whether individuals or teams could produce more valuable comics on average and be more likely to make a blockbuster breakthrough.
They made typical predictions that came out of industrial production research, that it would be the resources of the publisher, that it would be the number of years in the field that the creator had, or the number of repetitions they had done in the past. Those all turned out to be wrong. The most important thing turned out to be the number of different genres that the creator had worked across 22 different genres, from fantasy, crime, nonfiction. One of the really interesting findings was initially, they did not favor the individual who had worked in two genres. You were better off replacing that person with a team with three specialists who had only worked in one genre each. But after four genres, the individual flies past and can no longer be recreated by the team of specialists. What these researchers were saying was the individual is clearly the best unit of integration as this becomes much wider. They titled the paper “Superman or the Fantastic Four?” because if you can find a Superman who’s been across all these genres, get that person. If not, assemble this diverse team.
“Finding the right goal and finding your match quality in the first place is incredibly important, so you should be changing your interests and experimenting to maximize your match quality.”
There were exactly analogous findings in patent research starting in about the 1990s with the explosion of the knowledge economy, where the biggest impacts were not coming from people who had drilled down into the same technologies as classified by the patent office, but from people who had spread their work across a large number of technological domains and were often merging them. That trend wasn’t always the case. Prior to about 1990, the bigger contributions were coming from the specialists, and then it seems to have changed as the knowledge economy exploded.
Knowledge@Wharton: In the scope of our business culture, have we undervalued the generalist?
Epstein: I think so. I don’t think that means we don’t need specialists. But as eminent physicist and mathematician Freeman Dyson said, we need both frogs and birds. The frogs are down in the mud looking at the granular details of everything. The birds are up above and don’t see those details, but they can see multiple frogs and can integrate work. What he said is, our problem is that we’re telling everybody to be frogs and we’re telling nobody to be birds. That makes us inflexible, and all of our information is coming out of context. I think we need both, but we’re only telling everybody to be one, and I think that’s having perverse effects in some of the areas I wrote about. When I was doing investigative reporting at ProPublica, we were seeing some of the perverse effects of specialization in medicine that I think were well-intended, but just having effects that we don’t want because nobody’s integrating information and looking at it in context.
Knowledge@Wharton: What do you think that means for society, especially in this digital age where we have so much data coming at us?
Epstein: I read this really interesting research that I mentioned towards the end of Range about attempts to basically predict where [scientific] breakthroughs will come from. Basically, you can’t do it. A study of 10,000 researchers’ careers found that someone’s most impactful paper is as likely to be their first as their last as their 10th. The attempts to predict can cause what researchers called a dangerous purifying selection, where you get in a feedback loop where you’re selecting for the same kind of stuff.
These sorts of artificial intelligence where we’re using big data, they’re built for chess. They’re built to get what you got in the past, so if the patterns repeat, that’s OK. But for things like innovation — by very definition, the patterns do not repeat.
I end one of the chapters with the work of [University of Utah marketing professor] Abbie Griffin, who studied the so-called serial innovators. Her work is really interesting because she goes through her studies of what are the traits and experiences of these people who make creative contributions over and over again to their companies? She steps out of the more staid academic language at the end and says, “By the way, dear HR professionals, you’re defining your jobs way too narrowly and therefore accidentally selecting out a lot of these people because they have zig-zagged through their careers, they appear to flit among ideas, they need to talk to people outside of their domain, they use analogies from other domains.” She was saying that when you try to get the square peg for the square hole, you’re accidentally writing your job descriptions too narrowly. I was just at a conference where the head of a company that does machine learning to select employees was there. I think that’s a potentially dangerous idea if you’re in a realm where you’re looking for people to be creative and flexible and create new knowledge.
Knowledge@Wharton: You talk about having grit, which is something that Penn professor Angela Duckworth has researched and written about. What are your thoughts on grit?
Epstein: One of the things that professor Duckworth and her colleagues wrote in their studies was that people who are trying to get through orientation at West Point or to the finals at the National Spelling Bee … have been highly pre-selected for a specific goal and for other specific traits, you can’t really extrapolate the findings outside of those people. I think that’s very astute because life isn’t a six-week orientation or the finals of the National Spelling Bee.
I think that what a lot of other research shows is that finding the right goal and finding your match quality in the first place is incredibly important, so you should be changing your interests and experimenting to maximize your match quality. You don’t want to just blindly say, “Well, if I don’t stick with something, I don’t have as much grit.”
I was delighted because the day before my book came out, Angela’s most recent post was titled “Summer is for Sampling.” It was about making sure you had a sampling period over the summer. You don’t want to do grit just for the sake of grit before you’ve identified a good match and a goal. I was really happy to see her write that.
Knowledge@Wharton: Should sampling occur more within the structure of a company? It seems most companies want to stay in this pigeonhole mentality of what their employees can do.
Epstein: Some companies have it. But I think the pigeonhole mentality is like an oil tanker. You’ve got to start steering it from 40 miles out to shore to get it to go the right place. I think it’s a little bit of generational change.
“You can do all the strength finder quizzes you want, but your insight into yourself is constrained by your roster of previous experiences.”
I went on a podcast with Bill Simmons at The Ringer, which is probably the most listened-to sports podcast in the world. He was successful on his own, then got picked up by ESPN, then had a failure of a project at HBO and then started The Ringer. Some of my former colleagues from Sports Illustrated work there and were hired to do editing and fact-checking stuff. Some of them, like Natalie Rubin, for example, was hired just to edit online stories. Now she’s a famous podcast host because they’ll give people a shot. Some of these people who were hired to do one job have now become famous.
Knowledge@Wharton: Retention is a vital issue in business and one of the reasons why you see some change in the structure and hierarchy of companies. Human resources departments realize how much money they have to spend when they lose an employee. They would rather try to make them happy and incorporate them in more areas.
Epstein: Everything we know about everything is, keep your current employees, keep your current customers, etc., rather than having to develop new stuff, I think there’s a recognition of that. It reminds me of this famous essay called “The Mythical Man Month” by Fred Brooks, who went on to found the computer science department of the University of North Carolina. The theme of the essay is what’s called Brooks’ Law, which is that when you have a software project that’s late, adding more manpower to it will make it more late. The reason that happens is because managers always underestimate how difficult it is to assimilate new people onto that team. They would be much better off with the people on the team learning some new skills instead, but they continue to make the same mistake over and over and over. So, I think it would be good if there was some recognition that you’re way better off developing and diversifying the people you have to make those people Supermans instead of having the Fantastic Four, which probably becomes the Fantastic 100 as the job gets more complex.
Knowledge@Wharton: What is the 10,000-hour rule and its significance?
Epstein: Culturally, it’s extraordinarily significant and such a moving target at this point because so many different people have written about it. What you think it means probably depends on where you read about it. For some people, it’s this idea that there is no such thing as talent and only 10,000 hours of practice is what talent is. It’s just masquerading as talent, but actually it’s just practice. Practice is incredibly important; that is completely uncontroversial among people who study this.
In studies of developing chess skills, the number of hours it takes to get to international master status ranges from 3,000 hours to other people who are still being tracked at 25,000 hours and they still haven’t made it. If you average it, you get the 11,053-hour rule, but it tells you nothing about the real diversity of human skill acquisition. Actually knowing things about your strength and where you fit is really important.
Knowledge@Wharton: What would you like readers to take away most from your book?
Epstein: For one, the things that cause the most graphic progress often systematically undermine long term development. But also that researchers are giving us some good ammo to fight against the sunk cost fallacy in our professional lives — where we get information, we get signals, our insight into ourselves. You can do all the strength finder quizzes you want, but your insight into yourself is constrained by your roster of previous experiences. We learn stuff about ourselves, our interests and our strengths as we try things, so we should have a period of zig-zagging and experimentation like those athletes, like those comic book creators, like those technology inventors. And we shouldn’t just see it as a sunk cost, where you say well, I’ve started down this path, so now I don’t want to get off.
That’s not lost time. You haven’t wasted it. It doesn’t mean that transitioning is easy, but you can take what you learned in one domain and bring it to the other. Like when I was in grad school training to be a scientist, I was a completely ordinary scientist. When I was at Sports Illustrated, suddenly it’s extraordinary and my total advantage. I think we should think about it that way, and HR people should as well, to cultivate this diversity and not define jobs too narrowly, where they’re just getting whatever the LinkedIn perfect algorithm is sending right in their direction, because you’re going to screen out some of the most interesting employees.