Scientists fail every day. Failure is an essential and inescapable part of scientific research. It’s baked right into the scientific method: observe, measure, hypothesize, and then test. Of course, that hypothesis is often wrong. When it is, scientists go back, observe more, get new measurements, come up with a new hypothesis and test again. And again.
Despite this, scientific failure is rarely talked about openly, which was why when University of Arizona astrophysicist Erika Hamden used her TED 2019 talk last week to share how her work has been characterized by setback after setback, it felt like a radical act. As she spoke, she seemed at times near tears. And yet the talk, video of which is not currently available on TED’s channels, was not just brave; it was inspiring.
“The reality of my job is that I fail almost all the time and still keep going,” Hamden said on the very first day of the weeklong conference that usually celebrates triumph.
Hamden was onstage as part of the fellows cohort for this year, a group of promising change-makers who are on the cusp of reinventing the world. Most people were there to hype up their impressive work, to tell TED why it was so important and incredible that it demanded the world’s attention.
Hamden told the story of a balloon that popped.
The balloon was carrying a telescope that Hamden had already been working on for 10 years that fateful evening in September, 2018. The telescope is known as the Faint Intergalactic Redshifted Emission Balloon, or FIREBall, and its job is to measure giant hydrogen particles, which astronomers hypothesize pass between galaxies. Seeing them could help scientists understand why galaxies look the way they do, Hamden explained, and could help her to eventually measure every atom that exists. (You know, NBD)
“FIREBall is weird as far as telescopes go because it’s not in space and it’s not on the ground,” she said. “Instead it hangs on a cable from a giant balloon and observes for one night only from 130,000 feet in the stratosphere, at the very edge of space.”
One night only. Now you start to see why the balloon popping was such a colossal failure. And it came, Hamden explained, on top of failure after failure leading up to this night. Sensor failures. Mirror failures. Cooling system failures. Calibration failures.
Emily Dreyfuss covers the intersection of tech and culture for WIRED.
“Failures when you literally least expect them. We had an adorable, but super angry baby falcon that landed on our spectrograph tank one day,” she said, adding that despite the damage the bird did, it was still the greatest day in the history of the project, because adorable baby birb. “Falcon damage fixed, we got it built for an August 2017 launch attempt and then failed to launch due to six weeks of continuous rain in the New Mexico desert.”
And then the skies cleared, and the balloon took flight. “I have this picture that’s taken right around sunset on that day, of our balloon, FIREBall hanging from it, and the nearly full moon, and I love this picture,” she said. “God, I love it. But I look at it and it makes me want to cry. Because when fully inflated, these balloons are spherical. And this one isn’t. It’s shaped like a teardrop, and that’s because there was a hole in it.”
The balloon sank. FIREBall crash-landed in the desert. “We didn’t get the data that we wanted, and at the end of that day I thought to myself, Why am I doing this?” she said.
Not getting the data is just about the worst failure a scientist can experience. It’s also one that happens every day. Despite those failures being common, and even important, they’re rarely talked about openly.
As a journalist, I often try to talk about failure with scientists I’m reporting on, and though they mostly are quick to acknowledge it’s a huge part of their work, they are wary to discuss specifics. Very infrequently do you get a science process story, one where a scientist will lay out all the ways an experiment failed, and failed again, and failed again, until those failures taught them enough to get the right answer. If you’re reading a news story about science, it’s almost always about the successes—the breakthroughs, the cures, the mysteries solved.
On some level, this is understandable. No one likes to talk about their lowest moments. And in science, where work is mostly funded by grants, it can be tricky to have those conversations out in the open. Granting agencies want to see a proven track record of success before they take a risk on backing research; they certainly don’t want to hear about flops.
“Nearly everything that happens in the lab will never make it to print. The Journal for the Banal Failures and Self Doubt that Face Day-to-Day Life in the Lab does not exist. So a huge chunk of science goes unreported,” molecular biologist Maryam Zaringhalam wrote in Scientific American a few years ago. “Without failure, we lack a complete picture of science. And, a bigger shame, we lack a complete picture of the scientist beyond the brainy stereotype.”
I have witnessed the devastation of scientific failure firsthand, as the spouse of a scientist. I’ve watched my husband and his colleagues and friends across various disciplines lose sleep, lose hope, lose perspective when an experiment fails, a machine breaks and kills all the data, or a grant application is denied. Getting over those setbacks is excruciating.
And new scientists may be shocked to find how full of failure the research life is. “When I moved from medicine into research, the biggest shock to me was failure,” wrote oncology researcher Eileen Parkes in the journal Nature this year.
Part of what makes scientific failure so challenging is the timescale. Data is gathered over months, years, decades. When you’ve put in that much time chasing a theory and suddenly the data reveals that you’ve been wrong, or if the telescope you’ve been building crashes, it can feel like your whole life’s work crashed down.
“It took thousands of people and 44 years to get the Hubble telescope from an idea into orbit. It takes time, it takes a tolerance for failure, it takes individual people choosing every day not to give up,” Hamden said.
“It took thousands of people and 44 years to get the Hubble telescope from an idea into orbit. It takes time, it takes a tolerance for failure, it takes individual people choosing every day not to give up.”
Academic sciences have a huge turnover and dropout problem—a recent study found that around half of all people working toward an academic science career will drop out after five years. There are many factors at play, including a systemic failure to support parents and unequal pay and prestige along gender lines, but perhaps this lack of transparency about failure only makes things worse. Young scientists, facing their first failure, can be left feeling like they’re the exception, like their failure says something meaningful about their aptitude, or that their career is doomed. It can make scientists feel stressed and alone with their failures, rather than seeing it as a normal part of the process.
“Many students who began science degrees with me switched to other majors the first time a project failed. One failure and they were gone,” Sara Whitlock, a structural biology graduate student, wrote in STAT, about the importance of what she called “scientific resilience.”
Learning not to give up is one of the most important lessons for becoming a successful scientist. Studies have shown that resilience and higher tolerance for failure can keep people in science. But that learning doesn’t happen in formalized settings in graduate school. There aren’t usually classes that teach it, even though research shows that specific resilience training, when offered, can be effective. If it is learned, it happens privately, in conversation with helpful principal investigators, with labmates who’ve been there, at home or over drinks with empathetic ears. Rarely is it discussed on stages like TED, or in print, or in career counseling session with prospective new scientists.
“I have put so much of myself, my whole life, into this project,” Hamden said last week. And when she asked herself why she was doing this, after FIREBall crashed and the data was lost, she thought about Hubble. She thought about those atoms she wants to measure. “I’ve realized that discovery is mostly a process of finding things that don’t work, and failure is inevitable when you’re pushing the limits of knowledge, and that’s what I want to do, and so I’m choosing to keep going,” she said, demonstrating the resiliency science demands, while giving voice to a problem every scientist experiences but rarely has a chance to hear spoken about so candidly. It was refreshing to see this talk at TED. Maybe science conferences and graduate recruitment programs could follow suit, empowering young scientists to know failure is inevitable, and OK.
“It might feel like a failure today, and it really does, but it’s only going to stay a failure if I give up,” Hamden said.
She won’t. Hamden and her team will launch FIREBall again next year.