Against academic “Up or Out”

Academic science runs on an “up or out” model: you cannot stay in the same position without getting promoted for too long or you will be forced out (fired). This is sometimes imposed by law, other times it is a cultural phenomenon.

Whether legally or culturally imposed, it is a fixture of the system and a negative one. As more and more alternative structures (Arcadia, Arc Institute, CZI,…) are being tried, maybe it is time to rethink up or out.

Here are a few thoughts on why:

1. It structures everything! There are very few features and pathologies of the academic world that do not trace their roots to up or out.

   
   As I now run my own group, I am often trying to balance both the scientific success of the projects I run (which is the mandate I receive from the funders and what I will personally be most evaluated on) with the needs of trainees (students and postdocs).

   
   In the life sciences, the first author on each paper is generally the person who is driving the project and generally the one who gets most of the career benefits. There can be more than one first-author, but there is an order within first-authors, so that the first first-author is more important.¹ This is a big constraint on how projects can be structured and pursued.

   
   Science currently combines a ridiculous emphasis on field hyper-specialization with a ridiculous lack of job specialization. To be successful, besides the actual sciencing, one must be decent at writing and data visualization for manuscripts, one should be able to give a convincing presentation at a conference and write a good grant proposal. And even doing the science often involves stitching together multiple skills (write code, and do wet lab work). Some of us may actually enjoy most of these activities and achieve proficiency at them, but I also often see people who do good scientific work but write it up poorly (particularly, but not exclusively, non-native English speakers).

   
   Is the current structure of academic journals optimal? Very few people would say it is. This has led to initiatives like DORA Declaration which intend to draw the focus away from where something is published are well-meaning. However, without addressing the career structure, I frankly feel that they miss the point.

2. Projects take too long for the evaluation cycle. People are evaluated every couple of years and each project takes at least as long. Particularly ambitious projects take very long. This means that postdocs (and pre-tenure faculty) do not have that much time for executing a project and seeing if it gets the expected results and is accepted at a high-profile publication before recalibrating or moving on to the next project.

   
   This is like how the Olympics are the only competition that matters in some sports and, if you happen to be injured those two weeks in a 4-year cycle, you might miss your once-in-a-lifetime chance.²
   

   Besides, the system is incredibly noisy and biased. A very small number of people have to make a judgement call on whether to accept your paper and this necessarily leads to noise (see François, 2015). If it takes 3-7 years to get feedback on whether someone is talented and there is a very large amount of noise in the system, then the outputs are going to be horrible: talented people will lose out because they got unlucky. The result is less risk taking.³

3. Not everyone who works at SpaceX is Elon Musk. In fact, only one person who works at SpaceX is Elon Musk.

   
   The Silicon Valley cliché is that academic science should be more like a start-up: identify (young) scientists and give them the conditions to pursue their ideas. But surely “the conditions” must include access to support and staff? What makes Silicon Valley successful is not just that it empowers individual founders, but access to talent. Imagine a VC firm which gave out capital on the condition that you had to fire people after 3-5 years if they hadn’t started their own company.

4. It is (almost) the only model. Despite the existence of 100s of countries, 1000s of universities and other research institutions, 99% of the jobs available are up or down. Across the world, across a breadth of institutions, the same system everywhere….

   
   Even if you think it’s the best system (and I don’t), does it need to be the only one?
   

5. It is a huge source of stress for the people involved. I left this argument for last, because while I think it is an important argument, I also believe it backfires as concerns about the academic career path can easily be interpreted (and dismissed) as nothing more than self-interested plea for more career security. It may be good coalition politics to link this cause to other vaguely-similar ones (“precarization”, the gig economy,…). However, coalition-building is a double-edged source and by becoming aligned with a larger movement, not only are the specifics of your issues lost, but you also inherit the larger movements enemies.⁴

How would I fix it if I was starting my own research institution? I would define non-PI career tracks where researchers would not be expected to have first-author publications and that it would not be part of their evaluation. Instead, middle-authorships (evidence of collaboration) would be explicitly valued for career advancement in this non-PI track. This is not new, but researchers that are not on up-or-out tracks should be the vast majority of the institution, >80% of researchers.

Acknowledgment. I thank Clarice D. Aiello for helpful comments on a previous draft

1. Some people propose more complex credit allocation schemes, but if the system remains up or out, then the problem will not go away.

2. Over a 4-year cycle, the Olympics Committee actually gives out more medals than Nature publishes papers: about 3,000 in the Summer Olympics and 300 in the Winter ones, while Nature publishes ~800 papers per year (2400 total). Nature papers do often have many authors but there is only one first author.

3. I have also seen the opposite happening: a mediocre researcher who hits a home run early in their career and gets a lot of funding and responsibility. Then, they ruin their trainees’ careers for 10-15 years by being incompetent before being removed from the system. In some systems, it takes even longer than 15 years.

4. Aligning with the “anti-gig-economy movement” is not exactly joining a winning coalition, either.