Wednesday, June 10, 2015

How to Collect Matches that Will Catch Fire


 
Today’s guest blogger is Abraham Loeb. Avi is the Frank B. Baird Jr. Professor of Science at Harvard University. He serves as chair of the Harvard Astronomy department and director of the Institute for Theory & Computation at the Harvard-Smithsonian Center for Astrophysics. Avi is also the father of two young daughters and is working to make the scientific world a better place by the time they enter university.
 
Science can only blossom if young researchers are rewarded for acquired skills and growth rather than inherited academic ancestry.
 
Will a match catch fire when it scratches against the rugged matchbox wall? Knowing the answer is of paramount importance if we want to collect useful matches in our box. One way to find out is to try them all. The only problem with this approach is that by the time we will know the answer, the burnt matches will be of no value. The challenge is how to select useful matches reliably in advance? Putting this challenge into an academic context, how can we select a cohort of promising scientists before they have made their discoveries? This is the fundamental challenge of academic planning. Prestigious universities are plagued by past hirings which led to ‘duds’ or ‘dead wood’, namely faculty who when hired were labeled as geniuses with great promise but in retrospect, decades later, had little impact on the progress of science. At the same time, some of their contemporaries who were not endorsed by prominent scientists and hence moved to faculty positions at lesser schools, carried the day. Without mentioning names, suffice it to say that this is a familiar occurrence. Why is this phenomenon so prevalent?
 
Senior scientists who serve on promotion, prize, or search committees are often asked to evaluate the promise of their younger colleagues. One would naively expect them to approach this challenge in the same way that they address a scientific problem, namely study all available data and construct a model that extrapolates into the future. In order to avoid biases, it would appear natural to adopt a dynamical model which takes into consideration the special initial conditions of an individual and allows for growth in forecasting the individual’s future. For example, a young researcher who did not benefit from the privilege of being nurtured by top quality mentors or had to transition from a different culture or an inferior socio-economic status, should be given more slack. This is common sense. But is it common practice?
 
My experience over the past three decades suggests otherwise. Young scientists are commonly assigned static labels without proper attention to their starting point or the time derivative of their career trajectory. Early-career evaluations reflect a frozen snapshot of achievements. It is common occurrence for a science department to under-appreciate a faculty position applicant who graduated many years ago from the same department, due to a frozen image of the qualifications of the applicant. These mistakes have serious consequences, as poor recruitments lead to drifts in the prestige of academic institutions. To make things worse, evaluators often resist updating their image of an individual later on, out of fear that admitting the need for this update would reflect badly as lack of foresight originally.
 
Insisting on a static image that is out of synch with the growth of a successful researcher often leads to persistent attempts to shape reality in such a way that it will justify the preconception. The inconvenient truth is that evaluators with pre-conceptions have the power to allocate resources so as to justify their original static images. On the positive side, when serving on prize committees they can favor those whom they originally supported. On the negative, when serving on a grant allocation committee, they could block support for others even in the face of evidence that contradicts their early impressions. Such actions lead to self-confirming prophecies, and can occasionally crash the rising career of brilliant individuals who were not recognized as such at an earlier stage of their career.
 
The above faults are sometimes driven by the misconception that scientific success is enabled primarily by raw talent that should be evident in any snapshot of an individual. After all, Albert Einstein showed brilliance at a very young age. But this presumes a static view of science itself, while in reality the landscape of science has evolved dramatically over the past century since the days of Einstein. Today much of the relevant scientific information is updating rapidly and there are many more scientists around. In this climate, success is often linked to acquired skills, such as the abilities to adjust to rapidly changing intellectual landscapes (e.g., Big Data) and to identify the right problem to work on while others are searching in the dark. Today’s science requires social skills, namely the ability to lead other scientists and to communicate results so that they promote progress. These acquired skills take time to develop and require any model that attempts to forecast success reliably to include evolution and refrain from static images.
 


Avi, his wife, Ofrit Liviatan, and daughters
Klil (left) and Lotem (front)
Yet, it often seems that the guiding principles are completely off target. A traditional obstacle to an honest evaluation process is that prominent scientists wish to promote their own research program in an effort to link it permanently to the mainstream. Frequently, this tendency takes the form of senior scientists promoting their own students or group members well beyond what may count as fair play. This tends to suppress independent thinking outside the boundaries of widely-held paradigms. Put simply, senior scientists tend to measure success by how much a younger colleague replicates their own research agenda or set of skills. For example, if they are fluent with mathematical subtleties, they will identify success with mathematical skills. In faculty recruitments, this tendency for self-replication is dangerous because it might not stop at academic qualifications, but could easily spill over to an unconscious bias based on the replication of one’s own gender, race or ethnicity.
 
There are multiple paths to success in today’s science. Some paths are mathematical and quantitative while others are qualitative and require conceptual vision. Rather than replicating ourselves and preserving a static past, we would do better to aim at diversity and promote scientists of all varieties to secure a vibrant future. When serving on committees, we should resist static images of our younger colleagues and replace them with dynamical models by paying special attention to initial conditions and embracing evolution in our assessments. To cultivate innovation, we should always encourage creativity beyond the comfort limits that we establish for ourselves. Keeping a wide variety of matches in our matchbox will guarantee that not all of them will be duds. Hopefully, a few will light up in the dark to guide us how to move forward.
 
For more, check out this podcast from Astronomy Magazine’s "Superstars of Astronomy" series, which covers a broad range of topics ranging from the promotion of diversity and innovation to some of the most exciting frontiers in present-day astrophysics.