The principle of no surprises for Ph.D. students

The principle of no surprises applies to user interface and software design — stating that a UI should be something that users can predict (rather than be surprised at). However, I find this principle extremely useful in a professor-Ph.D. student relationship. This idea of applying the rule to advising context comes from my own father, who is a professor as well. At the time that I heard, I did not fully understand. As usual, his advice resonates in me in retrospect.

An advisor is like a driving instructor.

Professors' goal in advising Ph.D. students is to turn them into independent researchers. And getting a Ph.D. means getting a license for doing independent research; it is exactly like getting a driver’s license. Likewise, a Ph.D. advisor is like an instructor in driving lessons

Let’s keep using the metaphor. Imagine how annoying it will be for an instructor if a student tries to impress the instructor in driving lessons. The student surprises the instructor only if the student makes their own move, and does something inappropriate(e.g. speeding, driving in one hand, parking with 360-degree spin), typically from what they believe what driving is. What do instructors want then? One thing for sure, they do not want to be surprised at their students. They want students to follow their instructions because novice drivers do not have certain skills for driving— understanding of vehicles, laws, signs, norms.

Ph.D. advising = endless disapproval.

It is not too much different when it comes to an advisor-advisee relationship. They want you (student) to be awesome in the end but they do not expect you to be so from the beginning. In fact, professors’ expectation of a junior Ph.D. student’s work is pretty low. It is like an instructor’s perception of a student driver who has never driven a car regardless of their background, life experience, and intelligence. It is because doing research needs a set of skills like driving and they cannot assume that students without Ph.D. (a.k.a driver’s license) will have these skills already. There may be some traits that are good for research but nobody is born with these research skills equipped. You just need to get trained to learn these skills through iterative practice.

Often times, professors do not explicitly tell you what to do but rather want you to learn by doing. Therefore, the activity of professors advising students are, by definition, full of feedback, questions, criticism, and disapproval. A part of the reason is that they do not want you to fail externally (e.g., submitting a poor paper). That is why they need to, unfortunately, make you fail internally over and over again until you are ready.

Academics are skilled and trained to criticize anything. They would even feel bad about themselves if they cannot give any criticism of a work. In other words, probably, it is safe to think that you are not ever going to satisfy your advisor. Even if they are satisfied, they will not likely to say it aloud. Unfortunately, professors are trained to be tight-fisted. Of course, they may like your works and progress made but they tend to save positive feedback for external approvals (getting a paper accepted). The best positive feedback you can get from an advisor is an okay sign to submit a paper.

Don’t try to surprise (impress) your advisor. Be predictable.

One mistake that lots of students make in this endless negative feedback loop is to seek for their advisor’s recognition, to satisfy them, to impress them, and to surprise them. If you surprise them by doing something “unexpected”, it is highly likely that you did not do something that they expected — following their instructions, addressing the feedback, and practicing the skills. What an adviser wants is boring as much as it sounds.

The even worse thing to do is to hide their work to surprise their advisors with larger progress. Many professors would want to see every inch that their students make because the research on a wrong track means much more efforts needed later or simply a waste of time. Typically, they will wait until you bring it back voluntarily because they want to believe in you, because that’s the definition of independence, or because they simply don’t have time for that. The longer you keep your work to yourself, the more things there are for them to disapprove by the definition of Ph.D. advising. For example, spending 2 weeks to produce a progress that you think you will surprise them is much worse than presenting not-so-satisfying progress every other day. Surprises necessitate closedness. You can get rid of surprises by opening up and sharing the progress frequently.

Being predictable is a better strategy for a Ph.D. student. Your advisor expects you to follow their instructions. If you do as instructed, your advisor can predict what you will bring to the table. In a sense, they get bored with you if you are predictable. They will let you graduate. You adviser’s recognition needs to happen only one time in the end. Of course, having external approvals is necessary for graduation (dissertation committee, publications), which is now a matter of time. And then you are ready to be an independent researcher. Getting a driver’s license can be boring but, once you get a driver’s license, you can go anywhere you want.

Exploit the resource.

The better approach as a Ph.D. student is to exploit your adviser — to collect feedback as much as possible and to pick up necessary research skills. Bring the work and report the progress as often as your advisor can handle. Address their concerns and come back shortly. If they disapprove your work and do not tell exactly what to do, keep bugging them until they say what they want. It may be the case that they want you to figure out by yourself. On the other hand, it may be the case their research skill is internalized so it may be difficult for them to explicitly state what it is — like we, as a driver, cannot explicitly explain when to slow down and how much we should slow down in front of road humps; we just know. It is easy to make them tell what they want; do not be afraid of being ignorant and asking questions. Keep iterating until you get less feedback. The shorter the feedback gets, the closer you are to getting the license.

If you have an awesome idea, keep sharing it, discussing it, arguing for it and polishing the idea with their feedback, until they feel like “I knew what you were going to say and it is not surprising at all to me”. Make them feel like they come up with the idea at some point.

There is one problem here with this approach. You have to do lots of work. I did not find a way to exploit a Ph.D. advisor without working hard yet. They can tell easily whether you put some efforts or not, and they don’t give feedback if you don’t do your homework. If you do work hard, they are happy to be exploited. You cannot learn driving outside the car. Get in the car and grab the wheel, the instructor has nothing to do, otherwise.

You are on the same team as your professor.

Driving lessons may not be the most pleasant experience. It is the part of the deal. They will try their best to keep you safe, to help you learn, and to give you feedback when necessary. Therefore they have to sometimes, give commands, yell at you, stop you, and point out your mistakes. If they don’t do that, it means that either you become a good driver, or they are not doing their jobs!

I know this does not apply to all the Ph.D. students. Some professors take different approaches — being hands-off, delegating the duty to senior graduate students. And the problem gets more complicated. But remember, they hired you, even though they are hard to please, your advisor is on your side. Their goal is to help you conduct research independently and it is beneficial for them if you can conduct research independently. Until then, guess what, you will feel like you are disapproved.

Don’t try to surprise them with your awesomeness. Your adviser is not the one you need to surprise. They are the one who should surprise the world with you as a team.