Monday, September 05, 2011

Advice to a new graduate student

A friend who is starting graduate school asked for some advice.

1. There is often a tradeoff between the advisor from whom you will learn the most vs the one who will help your career the most. Letters of recommendation are the most important factor in obtaining a postdoc/faculty job, and some professors are 10x as influential as others. However, the influential prof might be a jerk and not good at training students. The kind mentor with deep knowledge or the approachable junior faculty member might not be a mover and shaker.

2. Most grad students fail to grasp the big picture in their field and get too caught up in their narrowly defined dissertation project.

3. Benchmark yourself against senior scholars at a similar stage in their (earlier) careers. What should you have accomplished / mastered as a grad student or postdoc in order to keep pace with your benchmark?

4. Take the opportunity to interact with visitors and speakers. Don't assume that because you are a student they'll be uninterested in intellectual exchange with you. Even established scholars are pleased to be asked interesting questions by intelligent grad students. If you get to the stage where the local professors think you are really good, i.e., they sort of think of you as a peer intellect or colleague, you might get invited along to dinner with the speaker!

5. Understand the trends and bandwagons in your field. Most people cannot survive on the job market without chasing trends at least a little bit. But always save some brainpower for thinking about the big questions that most interest you.

6. Work your ass off. If you outwork the other guy by 10%, the compound effect over time could accumulate into a qualitative difference in capability or depth of knowledge.

7. Don't be afraid to seek out professors with questions. Occasionally you will get a gem of an explanation. Most things, even the most conceptually challenging, can be explained in a very clear and concise way after enough thought. A real expert in the field will have accumulated many such explanations, which are priceless.

Please feel free to contribute your own advice!

23 comments:

sykes.1 said...

Truly excellent advice, all of it. I would add, Find a niche that is unoccupied, and fill it.

MtMoru said...

All advice on competition. Beat the other guy at a game. Really bad.

The right advice:

1. Make sure there is a very healthy non academic job market for what you do or make sure your school is top 5.

2. Don't go unless you've got nothing else to do.

3. It shouldn't and it doesn't need to be this way, but in Les Etats-Unis Merdeux it is.

LondonYoung said...

IMHO, if the new student intends to measure his/her life's success in meaningful contributions to humanity's scientific knowledge MtMoru's (2) won't apply.  If cash is the metric, (2) dominates all other advice.

MtMoru said...

Agreed if you can make "meaningful contributions to humanity's scientific knowledge". Unfortunately, 1. natural science is still too academic and 2. you're unlikely to get a job as a PI. Wasn't intended as a comment about the money.

If you're chances are good for a professorship or one of the few serious non-academic science groups and your field is such that you can make "meaningful contributions to humanity's scientific knowledge" then go, absolutely.

rz said...

On 1., the influential prof wins out.  Pick the most influential adviser that is willing to take you on as a student.  Seeking mentorship in the form of doing research projects with other professors, post docs, and more senior students is not too hard.

LondonYoung said...

Hmm, well I'll say this.  Markets are mainly creations of govts - no regulations and contracts, then no modern mkt - so I don't really know how to pick apart govt failures from market failures.  However "Les Etats-Unis Merdeux" sure seems to produce a lot of science per capita relative to the world as a whole!

Dima Klenchin said...

Steve's points are all very good. Go work for big shot professor that has good track of caring about his mentees careers. And schmoozing is very important. Strive for your name and face recognition. Ultimately though, an advice that I give to every new graduate student is this: Quit now. The odds are against you and chances are high that you will regret not quitting now.

steve hsu said...

I often say the same thing: Quit now unless you are very sure you can't be happy doing something else.

Note the person who originally asked for advice has already had a career in software and is going back in (to biology, not physics) with his eyes open.

silkop said...

Advice from a C.S. PhD dropout: Assume that your advisor isn't all that interested in research and doesn't care much about what you are doing or whether or not you succeed. Assume he's got enough of other grad students to choose from and is interested in getting funding grants, not science. Assume that nobody else in your environment is going to understand what you're trying to do either. So unless you already have or find very early an excellent topic that truly fascinates you and which you find deeply meaningful, quit! Save yourself the torture of having to pretend to be working on something that you later realize is pointless, irrelevant and wrong, the pity and misunderstanding from non-academics ("you know, that cousin's wife's friend who is dumb got his PhD in social sciences now, and you?"), and the embarrassment of having to admit "failure" later on.

Sanjay said...

This is all good advice, though in my field (psychology) I'd amend #1. Letters matter, but publication record matters a lot more - number of pubs and prestige of the journals yes, but also the innovativeness of the research. Many of my colleagues whose work I most respect worked with junior faculty as grad students, or with faculty of any level who supported them develop their own line of work, rather than just minting clones.

Andrew Gelman said...

I like point #6 because it is framed as zero-sum advice but is actually positive-sum:  If all these students work extra hard, they'll contribute more in total.

MtMoru said...

It's a game that shouldn't be, so unless you're sure of victory don't play.

Steve's yet to post his own ideas about how science should be done. Maybe he doesn't have any.

My own:

1. Scrap academic science, that is, science as part of the education non-industrious complex, and science as the usually useless curiosity of scientists.
2. Replace professor groups with Manhattan project like mega groups with very clear goals, like "land a man on the moon by the end of the decade". Today science is measured by the subjective and inscrutable esteem of the totally artificial society of scientists.

All naivete to the Inner Party members I'm sure, status quo bias being human nature.

Albert Magnus said...

I was the first student of my advisor, but he came from a "pedigree" research group and he made sure I got introduce to all the right people and got a good post-doc. 

Also, I worked at an accelerator and took lots of shifts which is a great way to network and get experience. You don't really know anyone till you've sat with them for 8 hours on a graveyard shift.

However, though I had a lot of useful technical skills (programming, electronics, plumbing), my IQ was more in the 120-130 range and physicists can tell if you are super smart or not. Also, I wasn't in a trendy field. (Had I picked something to do with neutrinos I would have done a lot better), so I was stuck competing with a lot of hard working and super smart people for very few positions and decided to do something else. 

The education was first rate, but its just a hard way to make a living doing fundamental research.

Albert Magnus said...

I was in software before going to grad school, too and I'm glad I did. I didn't make a career in academic physics, but what I'm doing now is more interesting (at least to me) than an software job I might have had. 

Ricky Smith said...

Steve, what would you say to someone who, after getting a physics PhD in a non-computer-related area of his/her field, is considering going back to school to get a BS in computer science?

MtMoru said...

That's like the hen asking a fox what she should do with the remainder of her eggs.

steve hsu said...

I don't think you need the extra degree as a credential. Probably enough just to take whatever CS courses you think you need or are interested in.

Ricky Smith said...

What if such a person is interested in cryptography and computer security?  Needed courses only, no degree?

Mark said...

1) The problem-selection problem is where I see many otherwise brilliant people trip up. Only work on interesting problems*. Where necessary, maneuver around your supervisor, your grant funding, any other constraints. Whatever people say in the short term, ultimately you will be thanked for doing interesting work.

2) Have fun.

*: Personally I would say "problems that can have large impact" rather than just "interesting problems". However, this I think is more a personal bias, doesn't necessarily apply to all disciplines.

steve hsu said...

Yes. Take it from someone who has worked in information security.

Robert Rota said...

do something no one has done before or better than anyone else has done it previously or likely to do in your lifetime. if lazy or unintelligent collect more data about any chosen object that has ever been collected before, find a pattern or lack of pattern and label it with your name. 

christain said...

Further, as the college and university administrators continue the decimate the ranks of adjunct professor at every opportunity through mass layoffs, the unavoidable reality is that the remaining adjunct faculty has to teach the overflow of students, thus further reducing the actual income from college teaching. It is possible, however, for an adjunct academic to gain control over the situation by engaging the online adjunct instructor jobs generated by distance education technology

Odoacer said...

I've been told to watch out so I don't end up like this guy:

http://www.ebaumsworld.com/video/watch/822298/

Blog Archive

Labels