Year after year, the machine of scientific enterprise becomes ever more vast—but is it more productive? If the measure of innovation is the rate of economic growth that results from new discoveries, the answer is no.
If we are throwing more resources than ever at the problem of innovation, but are only maintaining the status quo, it would seem that at some point our ability to invent and discover new things might decline. Before we reach that point, is there a “science of science” that can tell us how to accelerate the production of knowledge, or at least arrest its tendency to deliver diminishing returns?
Ben Jones, an associate professor of management and strategy at the Kellogg School of Management, has suggested an answer to these questions. His hypothesis adopts the perspective of an individual researcher, and derives from his or her plight an entirely plausible explanation of the overarching challenge to research productivity. It is a way of looking at scientific progress that has implications for everything from the sustainability of economic growth to the career choices of our civilization’s best and brightest.
To understand Jones’ work, it helps to use one of his visualizations: All scientists are born knowing nothing. If we imagine that knowledge is a circle and learning is a journey to its periphery—the frontier of knowledge—everyone starts at the center, including scientists. For a scientist born 100 years ago, the distance from knowing nothing to knowing all there is to know about a particular field would have been significantly shorter than it is in 2011. That is, the knowledge frontier that a scientist must reach before he or she can add to that body of knowledge has expanded. The circle of all that is known has grown larger.
If acquiring a certain amount of knowledge is a prerequisite for a major breakthrough, and it takes longer to acquire that knowledge simply because there is more of it, we might imagine that across the twentieth century there would be a trend for young scientists to make their first major contribution later in life.
Jones discovered that this is true. The age at which a researcher achieves “great achievement”—such as a Nobel Prize-worthy discovery—trended up by between five and six years across the twentieth century. The age at which scientists obtain their PhDs has also trended up proportionally. “As our understanding gets deeper, it takes people longer to get to the knowledge frontier from which they can step forward,” Jones says.
Delayed by Education
One possible explanation is that the need for more education is simply delaying the productive peak of scientists’ careers, but this turns out not to be the case. Rather, the years spent learning are in a sense replacing the earliest, and some of the most productive, years of discovery. Education, it turns out, has an opportunity cost even higher than we might have imagined. Jones’ explanation for the observed drop in average productivity per researcher is, primarily, that they are not making breakthroughs at a young age like they used to.
One way to adapt to the problem of ever-expanding knowledge is for young scientists to narrow the range of their studies. Anecdotally, at least, this seems true—as journals proliferate and subfields beget subfields, it appears that specialization has become the norm in science. Jones found that this is more than just an intuition—scientists really do seem to be tackling an ever smaller “arc” of the circle of knowledge. The narrower a student’s focus, the less information there is in the pie wedge between him or her and the knowledge frontier, allowing the student to reach it in a more reasonable amount of time.
Alternate explanations for both the delay of great discoveries and the specialization of the sciences abound, and in his paper Jones tackles the most obvious. For example, in the past 100 years, lifespans have increased considerably. It is possible that the sample population includes more people who are older, so that it only appears that discoveries are on average happening later in life. This turns out not to be the case. Neither is the phenomenon accounted for by increases in experience or health later in life.
Per-capita research productivity is declining, but economic growth, which is dependent on innovation, remains steady, because we continue to throw more resources at the problem.
Another explanation could be that over the past century, the apparatus of science has increased in scope. Experiments require more equipment and grants than ever, so research has become more capital-intensive. Yet the aging phenomenon also occurs in economics, which requires little or no equipment at all.
“If I reach for Occam’s Razor, I think the simplest explanation for all this is that there’s more to know in this field; there’s more to know in general. This is also prima facie true if you look at, for example, the number of papers published in a given year—the volume is now in the millions,” Jones remarks.
From the perspective of the entire economy, Jones’ work has a certain explanatory power: Per-capita research productivity is declining, but economic growth, which is dependent on innovation, remains steady, because we continue to throw more resources at the problem. The system works as long as the population can continue to grow. When population peaks, we could be in trouble. “By the time that happens maybe we’ll have artificial intelligence and other ways forward,” Jones says.
A second implication of Jones’ work is more subtle: As scientists try to cope with the ever-growing volume of what is known, they engage in an activity that is not without costs—specialization. “By being narrow, your innovative capacity may decline, and you can only see the value of ideas through a narrow lens,” Jones points out. Without sufficient understanding of other fields, how is anyone to understand the potential applications, in those fields, of discoveries within their own?
Jones’ work subsequent to his award-winning paper outlines his thoughts on how to deal with the growing mismatch between the institutions of science and the needs of scientists. Improvements in K-12 education, reduction of the busywork that distracts professional scientists, and new incentive structures that recognize the work of individuals within ever-larger and more collaborative teams are all possibilities.
But until these solutions are realized, one pernicious effect of the delay in scientific productivity worries Jones: other fields could already be attracting the individuals who might make significant contributions to research.
“This is a well-known issue in life science, where the average age for getting your first grant from the National Institutes of Health is 41,” says Jones. From the perspective of a bright undergrad, he notes, “I can go into molecular biology and get my first grant at 41, or I can go work at an investment bank and retire by then. That worries me.”
Jones is currently studying this problem so that he can empirically test his intuitions on the subject—just as he did with his notions about the consequences of our surfeit of scientific knowledge. Taken as a whole, his work on the “science of science” is an impressive body of work. And it all began with his first paper on the subject, which despite being published in February 2010, he completed in 2007. When he was 34 years old.
Editor’s note: Jones’ paper “Age and Great Invention” was awarded the 2011 Stanley Reiter Best Paper Award, an honor bestowed by Kellogg School faculty upon a research paper judged to be the “best” of those published between 2007 and 2010.
Related reading on Kellogg Insight