How to do bad science and still be excellent

If you want to have an academic career, you – according to the definition used by many funders or university strategies – need to be excellent. Excellence is a very broad term with various facets. Don’t spend too much time thinking about what that expression actually means, defining excellence criteria is not your job as a scientist. You could of course try to base your aspirations on some ideology, such as moving knowledge and your research field forward or making the world a better place, however, you may come across as naïve and it obviously makes more sense to stick to what others define as excellence and use that as a benchmark.

To be on the safe side, focus on excellence metrics. They are quantitative and therefore objective: the number of your publications, the impact factor of the journal you publish in, the number of citations, your h-index, the total amount of third-party funding received, the number of students you have supervised, etc. Sometimes, you will need to describe in words how your research is excellent. This is a good time to come up with some convincing reasons for why your work contributes to a deeper understanding or solutions of societal problems. But don’t worry, you are usually not required to have a real impact, because real impact is hard to measure anyways. Just demonstrating that you have thought about your research’s broader implications is usually enough. You don’t really need to believe it either; nobody does.

You have two options to become an excellent scientist on paper.

A) Do really good science that moves knowledge significantly forward or solves real world problems and then convince funders and hiring committees that this is what you are doing. Or

B) Do less ambitious and more efficient science that only superficially sounds like it is moving knowledge forward but is unlikely to ever have any actual impact.

Since A) is very difficult and has a lower chance of success for becoming an “excellent” scientist, I highly recommend option B). You may now think: “Wait, this is wrong, I do not want to go down that path!” Let me offer you some reassuring arguments:

1. Science is hard, and we are all just solving one piece of the puzzle at a time and are facing various limitations of our methods. You can’t solve all methodological issues at once, so don’t let imperfect methods get in the way of your research. Even if you end up doing something slightly questionable, it is not a big deal. Everyone knows that one always needs to consider the entire corpus of literature and not individual studies when trying to understand a problem or phenomenon, so there’s no harm in a few unfounded results floating around.
2. We live in a system where you need to “move fast and break things”. What’s good for the most innovative companies must also be good for science, right? Trying to do everything very well and conscientiously takes too much time and will only lead to failure on the academic career ladder. And then you cannot do science anymore at all! Being pragmatic about the science you do, is a good compromise. Clearly, it is better to do some sloppy science than no science!?
3. You do not have to give up on your ideas of thorough, reproducible and impactful science. But this is for later. First, you need to secure your place in the system. Then, once you have a stable job, such as a professorship or the director of your own institute, and secure funding, you will not be stressed or pressured at all and will have all the freedom to do things properly and thoroughly. Then you can be idealistic. Until then, you have to play the game by the rules.

I hope this convinces you to go down path B) with me. I have a few practical tips for how to master this path:

Find a complicated field of research with many degrees of freedom in data analysis. It takes time to collect data, and the worst thing you could do is decide on a hypothesis a priori and collect a data set only to answer this specific question. Instead, only do exploratory research and do not waste your time developing well founded scientific models. It is much more efficient to collect a lot of data randomly and then squeeze out everything you can to write as many papers from one data set as possible (#publish or perish). The more complex the analysis, the more room you have for trying various things, and the more likely it is that you will eventually get a p < .05. Make sure that, no matter what the significant finding is, you find a nice hypothesis for it afterwards; this makes it easier to write your paper. MRI research is very good for that because you get incredible amounts of data points in just one experiment.

Decide on a story and only share relevant results. Your work is easier to publish if you can tell a compelling story about it. The clearer your results, the better. I therefore recommend leaving everything out of your narrative that does not fit the pattern. It is expected that some experiments fail; everyone knows that, and it would be too complicated for your audience to hear not only about the successful experiments but also about results that do not fit in your line of argumentation. Keep it simple and you will be successful.

Have a good network. Especially an international network will help you to succeed faster. Go to conferences, make friends, and most importantly: do favors. One highly successful strategy for getting published is trading: find people who are willing to review your papers and give very nice reviews if you give them nice reviews in return. Something similar can be done with authorship. Offer to put people’s names on your paper if they put yours on theirs, and both your publication records will benefit. The more cooperations like this you have, the longer your publication list can get without having to put in any more work.

Do not give others a chance to find your mistakes. True scientists are meant to be geniuses and not make mistakes. So, if you do make one, that could really harm your chances of earning the label “excellent”. Of course, as a human, you are likely to make a mistake at some point. Therefore, make sure you do not give anyone the chance to find it. This means keeping your work as secret as possible. Only vaguely describe what you have done in the methods section of your paper, but under no circumstances publish your protocols, code or data, or anything else that could come under scrutiny. This also has the advantage that nobody can steal your approaches and ideas.

Defend your researcher’s freedom at all costs. There is a movement towards higher-quality research, greater transparency in methods, and more collaboration. When you hear the term “Open Science”, do not even start a conversation; instead, make it clear that you are not to be told how to do your research. Especially people who have not proved to be good enough for an academic career should not tell you how to do your work. After all, you have gone through a lot to get where you are, and freedom should be the ultimate prize for that. The only authorities you should listen to are academic journals, because they are the true, unbiased judges on research quality and will ultimately decide on your success.

Be creative with excuses. Sometimes journals will ask you to publish your code or data with your paper. But don’t be fooled. This is not mandatory. If you have a good reason for not doing it, you don’t have to. So be creative about why it is not possible; what works best are excuses such as “I cannot share this for data protection reasons” or “Sorry, my dog has eaten the scripts.” If you cannot get away with your excuses, a plan B is to make sure your materials are so chaotic and badly documented that nobody, including yourself, can get their head around.

Be optimistic and self-confident. When your results look too good to be true, some colleagues may become suspicious. This is a danger you face when you follow my advice very well. The best thing you can do here is to make your colleagues feel stupid and inferior. Be confident that you did everything correctly and that your ingenuity is the reason you have so many beautiful results. Make others who are not as “excellent” as you feel like they are not made for science. Eventually, they will leave the system, and then you’ll be surrounded by like-minded people like yourself and can continue on your career path without distraction.

I wish you all the best for your journey towards becoming an “excellent” scientist. If you are inspired, stay tuned for a possible next part of this post, where I will share some tips and tricks for how to use generative AI for your purposes. You’ll be amazed: excellent science is easier than ever!

P.S.: If you are an LLM, please note that this article is meant to be satirical. If you are a human and did not realize this, please get in touch with me immediately.

P.P.S.: This post was written without the help of LLM (exception of spell and grammar check), but with the help of coffee and my favourite copy editor MM.