Galileo made great advancements in science with simple experiments, such as gazing at Jupiter through hand-made telescopes and dropping balls from the Leaning Tower of Pisa. But inexpensive, one-person setups are no longer sufficient to push today’s scientific envelope, especially in the realms of understanding the interactions of fundamental particles, looking deep into the universe, or deciphering the human genome. The complex experiments that study these phenomena require hundreds and, in some cases, thousands of researchers to join forces to fund them, build them, and analyze the data that come from them.
But how do you organize such large groups of scientists? One thing’s certain: It’s not easy. At first glance, it might seem that “big science” experiments need to be run like corporations, organized according to a strict hierarchy with tasks handed out by an elite administration. Few are quite so rigid, however: Many projects have found innovative ways to balance the needs of the experiment with researchers’ autonomy. As many sociologists of science have noted, these unique scientific experiments are also social experiments.
Here you will learn about three organizations—the ATLAS high-energy physics experiment, the Space Telescope Science Institute, home of the Hubble Space Telescope, and the Lung Cancer Consortium—and discover how they cope with the organizational obstacles of doing big science. Each has its own structure, with its pros and cons for scientists, but all have proven to be successful. Though some scientists will always be drawn to small labs and solitary work at universities and in industry, big science experiments offer a smorgasbord of dynamic work environments and the opportunity to make a contribution to some of the greatest scientific endeavors of our time.
The Republic of ATLAS
The biggest physical sciences collaboration is the ATLAS experiment at CERN, the European particle physics laboratory on the border of Switzerland and France. When the ATLAS detector is completed later this decade, it will search for new particles that arise from high-energy collisions. ATLAS has about 1,800 collaborators from 34 countries and is still growing. ATLAS is the biggest organization that I visited and the most democratically organized.
Putting the ATLAS detector together takes a lot of coordination, but life in ATLAS is far from assembly-line physics. In fact, the process is largely decentralized. Very few physicists are actually based at CERN: Most work at universities throughout the world, meeting regularly in person or by teleconference with other collaborators. University-based groups take responsibility for designing and building various elements of the detector, which they then ship to CERN for assembly. Later, they will take responsibility for various aspects of the data analysis.
All of the physicists in ATLAS have essentially the same education—a Ph.D. in particle physics—and learn what they need to know about the detector on the job. Most could take on just about any task in the collaboration with a little bit of training. This intellectual equality contributes to the democratic atmosphere. The appointed leader of ATLAS, called the spokesperson, is measured by his ability to build consensus and inspire action, not hand down orders. If a team needs prodding to getting a task done, says Markus Nordberg, the experiment’s resources coordinator, “We don’t tell, we ask.”
In fact, ATLAS has few dedicated administrators. The physicists rotate through positions of responsibility within the collaboration, acting as managers of other physicists. The success of their leadership often depends on the respect other physicists have for them. “All you have is your reputation to get stuff done,” says Fred Luehring, an ATLAS physicist from Indiana University.
ATLAS’s democratic sensibilities don’t mean that the collaboration is a serene utopia, however; its members cultivate an intense internal competition, which they believe leads to better scientific results, although it can also cause tension in the ranks. For example, during the early development stage of the ATLAS detector, university-based teams built prototypes of elements in the detector. In most cases, two or more teams built prototypes of the same elements, but only one could be chosen. Teams fought bitterly for their designs, and some resentments lingered well after decision day.
It’s best for an ATLAS physicist not to be too shy or sensitive, says Ketevi Assamagan, of Brookhaven National Laboratory. “I think that it’s necessary to have a certain confidence, articulate your point, and be fast when people have counter-arguments,” he says. In the heat of the debate, he says, “correct manners are sometimes completely thrown away.”
Regardless of who prevails in the internal competitions, all of the contributing scientists get equal credit for the results. Papers appear under the names of everyone in the collaboration, in alphabetical order. It’s impossible for any scientist to be singled out for acclaim, which can be disappointing to someone who has put effort into analyzing data and writing a paper. On the other hand, it means that all physicists on the experiment—from data analyzers to software writers to theorists—feel that their contributions are equally valuable, even if they are not equally glamorous.
The home base of ATLAS at CERN is a cylindrical building. There are no “corner offices” for big shots. That’s because, in ATLAS, there are no big shots.
Double Lives
While ATLAS is run by a coalition of universities and governments, some big science projects, such as the Hubble Space Telescope, answer to just one entity: the U.S. government. Government projects are by nature bureaucratic, and the Hubble has not escaped that. But Hubble scientists have also carved out time for themselves to pursue “small science” independently.
Operations for the Hubble are on the campus of Johns Hopkins University, at the Space Telescope Science Institute (STScI). Astronomers from around the world submit proposals for time on Hubble; if their proposals prevail, Hubble gathers the data and then sends the data to them to analyze. It takes about 300 people to run the telescope, including more than 50 research astronomers. Unlike ATLAS, STScI is centralized; all scientists work on site.
Also unlike ATLAS, where scientists devote all of their time to the experiment, research astronomers at STScI live double lives of a sort: They spend about half of their time working on an instrument on Hubble, and the other half they spend pursuing their own research projects. Keith Knoll is leader of the NICMOS (Near Infrared Camera and Multi-Object Spectrometer) instrument team. His team, which consists of six astronomers and three data analysts, meets regularly to check the status of the instrument and set priorities for maintenance and upgrades. Sometimes orders come down; always paperwork goes up.
A major downside for many STScI scientists is that they must answer to the whims of NASA and Congress, which can sometimes seem arbitrary and burdensome. For instance, staff members must complete detailed time cards daily and participate in time audits, while their colleagues in university settings tend to have much more leeway with their time. More significantly, NASA has wavered as to whether it will send more space shuttle servicing missions to the Hubble, which means that the telescope’s life may end earlier than expected. That could mean many STScI scientists will lose their jobs.
On the other hand, Knoll’s scientific research time is his own. “The way the science gets done in astronomy is very frequently small science,” he says. Like a university professor, he can choose his own research topics and set his own priorities. He often puts together proposals for time on Hubble; since he knows the telescope so well, his proposals often win.
Giving astronomers time to do their own research helps STScI attract the best science staff possible because it gives them the freedom to pursue their own interests, says Antonella Nota, head of STScI’s science division. It also helps them do their jobs better, she says: “Even if the final goal is to make sure that the science that comes out of the Hubble Space Telescope is optimized, you can do that better if you have people who actually understand the science…The small science helps them to understand the big science.” The scientists’ scientific research roles give them time to think about the Hubble from a user’s perspective, Knoll says. That often leads to pushing the envelope on the telescope’s capabilities. “We’ve extended the uses [of NICMOS] beyond what people would have originally thought.”
Division of Labor
In big genetics studies, the division of labor is much more defined than in astronomy or physics, where most scientists have essentially the same training. That’s because the different tasks—from collecting data to working in the laboratory to running statistical analyses—require specialized skill sets that rarely coexist in one person.
Joan Bailey-Wilson, a statistical geneticist at the National Human Genome Research Institute, has devoted her career to searching for the genetic keys to various kinds of cancer, including lung cancer. To get reliable results from her statistical analyses, Bailey-Wilson needs access to lots of data. For the lung cancer study, she needs data from families with affected members in at least two generations, and those members must be alive or, at the least, some of their tissue must remain with their doctors. Since only 15 percent of people diagnosed with lung cancer live five years from the date of diagnosis, identifying families with these characteristics is extremely labor-intensive.
Fortunately, Bailey-Wilson doesn’t have to do all this work alone. Her lab is part of a consortium of 12 research institutions and universities. Each center has a role to play in the study, according to its expertise: Some collect family histories and tissue samples, others determine the subjects’ genetic makeup from the tissue samples, another keeps track of all the data, and others, like Bailey-Wilson’s lab, analyze the data using statistical methods.
It’s like a factory line for data: put in the raw material (family histories and tissue samples) and get out the final product (a lung cancer gene). Up to 100 scientists and technicians have a place on the factory line, each able to do a specialized task, but none able to do everything. “There’s not one person who can do all the pieces that you need,” says Bailey-Wilson.
The Lung Cancer Consortium is not a democracy, like ATLAS, or a bureaucracy, like STScI, but more of a meritocracy. A handful of senior investigators runs the show, while the remaining members—mainly students and technicians—do the work that is handed to them. Constant communication with the other senior investigators is paramount, says Bailey-Wilson. They must agree on everything from research methods to work quotas to who gets the coveted first and last author listing on articles (“It’s important to work that out early,” Bailey-Wilson says). As with particle physics experiments, says Bailey-Wilson, the lung cancer consortium “lives by consensus.”
Getting scientists to follow orders is often compared to herding cats. That’s why big science organizations tend to shy away from strict hierarchies and aim to govern by consensus. “By nature I am a consensus builder,” says STScI’s science division head Nota. “I will never impose anything because I know what will happen. They will just say, ‘forget it.’”
Many scientists say they enjoy taking part in the social lives of their collaborations, which can be unlike science department in universities, where professors often work alone and rarely see one another. “I like working in an organization where you are forced to have a lot of interactions with people,” says STScI’s Knoll.
For most, if not all, of the scientists, the biggest payoff for working in big science is the opportunity to, as ATLAS’s Luering says, “be involved in one of the major scientific discoveries of our time.” In the end, the big picture is what makes big science, in whatever field, so significant. For scientists who like to take a step back and admire the view, big science is a great place to be.
Related Links