Four years ago, I obtained my PhD which seems like a very long time by now. As scientists, we are trained to do a lot of different things, but the training how to run a lab is rather informal and largely based on trial-and-error learning. There have been a couple of nice write-ups (e.g., https://users.fmrib.ox.ac.uk/~behrens/Startingalab.htm) from far more influential and experienced researchers than me, but I thought it might still be helpful to publicly document my humble insights as they arise and evolve over time.
1. There is only so much you can do
And that’s okay. It is not great, but okay. The revelation that a typical day is not sufficient to get all necessary work done is probably the most difficult and important lesson that I had to learn recently. Frankly, I am still learning though I am getting a bit better every day (I hope). The longer you work in academia, the more responsibilities will pile up and there is an upper limit on how many hours per week you can work. Yes, even if you try harder, trust me. This has been nicely covered by the great Everything Hertz podcast (episodes 37 & 38). One day, you will have to adapt by saying “no” more often if you are asked to do stuff that is not worth the time. You will have to accept a certain share of sunk cost if you give up on things, but a fragmented schedule will take its toll and kill productivity all too quickly. Letting go of some old “side projects” is easy and reviving once you get the hang of it.
It is much more difficult to realize that you will also have to say “no” to stuff you would like to do. Stuff that you would have been more than eager to take on just a few years ago. Stuff you know how to do well, but can’t do on top of the other things. The silver lining is you may not have to turn the more promising projects down entirely. Instead, for example, you can try to empower other people to do analyses, to set up models or simply code a few steps that can make a big difference in everyday workflow. This way, it doesn’t have to feel like you are losing out on the fun, which is hands-on data analysis to me, and it is a good way to forge collaborations.
The key is not to feel swamped and overwhelmed for too long before making the necessary changes in daily routines because it can easily drag you down. The past few months, I have been super busy writing grants and getting experiments up and running so I had to pause several of my pet projects (for example, this blog in case you have wondered already). Sure, I wish I could have finished more of my old work in parallel, but that was just not feasible in the end. I have been lucky enough to secure basic support for my junior group’s research for the next 2-3 years so I can gladly turn back and finally finish the old work that is still haunting me. I have learned to accept such delays as part of academic life without becoming paralyzed by the thought that it feels impossible to keep up with everything.
2. When there is luck, there is also bad luck
Not every bit of your track record is under your control. Don’t feel disheartened by incomprehensible outcomes. I slowly had to get used to the fact that many decisions in science follow a semi-transparent highly probabilistic diffusion process that hits one or the other decision boundary based on accumulating too little signal embedded in a lot of noise. Hence, it might be difficult to figure out why things turned out right or wrong without becoming superstitious. Of course, you can try to figure out what works and what does not and wonder why reviewer number 2 did not see the substantial additional value in your >4-year research project. Who needs bigger studies with better estimates if it can’t replicate the previous surprising results from a small sample? Or you can try to let go and keep pushing what you believe is right.
I can’t say the random element does not bother me anymore. But once you sample more decisions in the tight race for high-impact publications or prestigious funding schemes, the more obvious the element of luck in a single outcome becomes. More importantly, what you might be able to control is your average rate of success if you push through the instant disappointment. Indeed, the reliability of most decision-making processes in science is inadequately low (Bornmann et al., 2010; Marsh et al., 2008) given their importance in particular for early career researchers (ECR) and it is not clear yet if open reviews will help to fix it anytime soon. Accepting luck and bad luck helped me not to take things too personal. Not every failure will be your fault. Don’t focus too much on learning from one single outcome that could have gone the other way on a different day.
Learning to accept this decision noise as part of the job does not mean that we should not continue in demanding changes. Scientific decision-making must be improved upon and everyone can try to contribute towards an increased transparency and fairness. However, be prepared to face opposition from faculty members who benefited from the past system and are afraid of losing out in a fair race. Not everyone will embrace the change.
3. Don’t count on instant gratification
When I look back at what I have accomplished after one week of long working hours, I regularly have trouble in identifying key achievements. Many daily obligations will not instantly lead to obvious effects making it hard to infer if you are on the right track. Also, if you think that peer review of papers takes way too long, you have not been in for grant reviews. Thus, external gratification will commonly take ages to materialize so it is good to revert to other means of reinforcement to keep going.
One of the reasons why I am still so passionate about doing science is that I have been more than privileged to work with many talented and principled students, who knew the right way is not necessarily the shortest path to success and were willing to fight for what they believed in. Even it made things harder for them. If you want to find it out how researchers feel about the truth, try this: point out an error and observe how they either legitimate and marginalize it or literally display the pain at being (slightly) off in their facial expression. While working with students does not always yield prompt consequences, it is probably the most gratifying experience in the long run. Keep this is mind when your schedule is packed again and emails keep piling up.
4. Trade off quantity, not quality
When I was still a PhD student, I did not understand why some researchers would not read the entire paper or check the numbers multiple times to make sure that everything is correct and consistent. I couldn’t imagine how fast the transition of responsibilities and expectations is. Now I clearly see the pressure to speed up. By now, I have greatly optimized the way I read a paper to keep up, but there is always the risk that you will miss some subtle details hidden somewhere in the narrative. Focused attention has become a limited and precious resource that you will need to spend wisely.
I have tried to keep many projects going at the same until I realized that I was not happy with touching them only at a level that I felt was too superficial. Conducting a thorough analysis involving fMRI data is not a trivial one-hour task that I can do well on the side. Taking a different perspective here helped me a lot. I am reviewing a lot of papers that basically say: We did A, but then A did not really work out and then we eventually found B, which is close enough to A to be of interest, and perhaps C or D if you are a bit more forgiving in terms of statistics. I don’t mind endorsing null effects at all, but it sometimes seems as if the pace of publishing results after results is inviting just the fastest solution that you can think of: Run the obvious stats, and if it leads nowhere, run some more, count the stars, move on. Don’t deliberate for too long or you might be losing out against your competitors.
Of course, there is a lot of good work out there. Yet, it feels as if it is swamped by the many papers trying to me a quick BOLD statement about the brain. How these “novel” brain mechanisms truly relate to behavior often remains a mystery. Mapping brain response dynamics onto complex behavior is not an easy task, for sure. Still, we have amassed enough candidate regions for aberrant brain function across numerous mental disorders without necessarily gaining a deeper understanding of what is going on. I think it is about time that we sit down, take a second look at what is out there and re-evaluate what we know indeed. Our time is too limited to chase the scant shadows of what might look like an effect if you apply the spotlight in a very specific angle.
5. Science must be more than just business
As a junior faculty member, I happen to receive a lot of advice about the way things work and I am truly grateful for that. However, a great deal of advice also speaks volumes concerning the state of academia. Yes, science is to some extent business. But doing science must be more than simply doing business. When career advice clearly conflicts with good scientific practice, this crosses a red line for me. Just because you can game a system, does that mean that you must exploit it to your advantage? Science is a zero-sum game. Essentially, it is like playing cards and cheating on your friends: your benefit will have to be paid out of another person’s pocket. And this person might have deserved to win the last round more than you.
Sometimes, such an objection is frowned upon and labeled as naïve by older faculty members who surely know the ins and outs. But is it naïve these days to insist that we are not just here to further our personal careers? Tax payers have invested a lot of money in my education and scientific work not for me to build the finest house of cards. James Heather just wrote an excellent manifesto-ish blog post about why we have to expose bad science. I have also blogged about the Wansink pizza gate before, but it is evidently not only about that one rotten apple in the fruit crate of food research. The truth is that there are many Wansinks out there. They build their careers on overhyped conclusions drawn from weak evidence after extensive p-massaging combined with vague storytelling why such observations do make sense (at times to the extent that a formerly surprising trend would become significant once a one-sided test was applied).
Heeding for such well-intended advice makes the scientist of tomorrow go down the same rabbit hole again that handed us the current mess. But then again, we are all competing with researchers who deliberately cut some corners to make a career and if you refuse to play by the rules of business, you might have to work even harder to stand out of the crowd. I think this is a rather worrisome lesson to learn, especially for students and ECRs that we would like to take over some day. Turning science into straight business is not a viable option if we are primarily committed to building knowledge for the sake of humanity. For me, the lesson has been crystal clear though: if I can’t succeed with the way I think science should be done, I would rather do something else than compromise on what I believe in. I am just a little afraid that this is not what is currently being selected for.
Bornmann L, Mutz R, Daniel H-D (2010) A reliability-generalization study of journal peer reviews: A multilevel meta-analysis of inter-rater reliability and its determinants. PloS one 5:e14331.
Marsh HW, Jayasinghe UW, Bond NW (2008) Improving the peer-review process for grant applications: reliability, validity, bias, and generalizability. American psychologist 63:160.