In science, we are taught that failure is part of discovery. Yet academic careers often leave little room for it. Graduate students work toward milestone exams and fellowship deadlines. Postdoctoral researchers depend on publications, funding, and future job opportunities. Scientific training encourages uncertainty, but academic careers often reward predictability.
Early in my training, I remember a day when an experiment fell apart in a matter of minutes. A delicate surgical tube bent during a procedure, undoing hours of careful preparation in a matter of seconds. It was frustrating. When I told my advisor, he responded calmly. “This happens,” he said. “It’s part of the process.” We learn by trying, failing, and trying again.
I left that conversation reassured. But the relief did not last. Later that same day, I found myself thinking about upcoming project and fellowship deadlines and the expectation to produce results. I began to wonder how many failed experiments I could afford before they started to matter for my future.
This tension sits at the center of scientific training. We are encouraged to embrace uncertainty, yet we are evaluated as if uncertainty should be minimized.
Scientific discovery depends on the willingness to explore the unknown. Experiments rarely unfold as planned. Projects often begin with one question and end somewhere entirely different. In many cases, this is where the most interesting insights emerge. Unexpected results can reveal assumptions that researchers did not realize they were making. Some experiments become interesting precisely because they fail to behave as expected. Learning what does not work helps refine what might. Over time, this iterative process builds understanding.
As trainees, we are taught to tolerate this ambiguity. We are told to ask difficult questions, to pursue ideas whose outcomes are unclear, and to accept that progress may be slow or uneven. Failure is not seen as a setback but as a necessary step in the process of discovery.
At the same time, academic careers are structured around measurable outcomes. Progress is often defined through papers, results, and other tangible outputs. Graduate students and postdoctoral researchers work within limited time frames, with milestones that shape how their work is evaluated. These expectations do not always align with the unpredictable nature of research.
Navigating these two systems requires constant adjustment. In my own work, I have noticed how decisions about projects are shaped by these pressures. Questions that once felt exciting can begin to feel risky if they do not lead to results within a reasonable time. Over time, there is a tendency to gravitate toward projects that are more likely to produce clear outcomes.
This shift does not happen all at once. Instead, it builds gradually through small decisions about how to spend time and effort. Each choice makes sense in isolation. Together, they can change the direction of a research program and even entire scientific careers.
For researchers working outside their home country, these constraints can feel even more immediate. In addition to the usual expectations of academic training, these researchers often depend on maintaining continuous employment and steady progress to remain in the country. Missing a fellowship deadline, a grant cycle, or a period of funding can carry consequences beyond research progress alone. It can mean relocation, financial strain, or major disruptions to daily life. These conditions can make timelines feel tighter and reduce the room to explore uncertain ideas. While this does not create the underlying tension between uncertainty and evaluation, it can make it more difficult to remain in that space.
The result is not simply a change in individual behavior. It may also influence the kinds of questions that are pursued. When early-career researchers learn to prioritize predictable outcomes, the space for more exploratory work can narrow. This may steer researchers toward smaller or safer discoveries at the expense of more ambitious or transformative work. This may not only shape individual careers but also influence the novelty and ambition of the research itself.
At the same time, this tension carries an emotional dimension. Curiosity and caution begin to coexist in uncomfortable ways. The desire to follow an interesting question can be tempered by the need to show progress. Over time, this can shape how researchers think about failure itself. Instead of being seen purely as a source of learning, failure can begin to feel like something to avoid or something more consequential.
Scientific training depends on the idea that uncertainty is valuable. But if the structures that evaluate researchers reward certainty, then the ability to remain in uncertain, exploratory work becomes uneven. For example, researchers with greater access to funding may have more room to take risks, while others feel pressure to stay within safer boundaries to stay afloat. If only some researchers are able to tolerate uncertainty safely, then opportunities to pursue novel or unconventional work may become concentrated among those with greater resources and security. Over time, this could reinforce existing inequalities in who is most likely to produce high-impact findings and attract future support.
This raises a broader question about how we train and evaluate scientists. If we want researchers to pursue ambitious and open-ended questions, then the systems that shape their careers need to allow space for the uncertainty that such work requires. If we train scientists to embrace uncertainty and learn through failure, then our evaluation systems should leave room for those experiences as well.
Science advances by exploring what is not yet known. Maintaining that spirit requires more than encouraging curiosity in training. It also depends on creating conditions in which early-career researchers can afford to learn through failure, without feeling that a setback could jeopardize their future.
Jung Uk Kang, PhD, is a postdoctoral fellow in neurobiology at Harvard Medical School. His research focuses on how neural circuits support perception, decision-making, and flexible behavior across species.
Photo by Isaac Davis on Unsplash
