Home

Advice for Graduate Students

Copyright (c) 2007 by Frank Vahid
(A synopsis of the book: "How to Be a Good Graduate Student -- Tips for succeeding in graduate school". Available as free pdf. See bottom of page for info.)

Introduction

Graduate school is the best kept secret in America.

Many students, exhausted from their 4-6 years of undergraduate courses, can't imagine why anyone would subject themselves to 2-6 more years of that torture to get a master's or Ph.D. degree, when a bachelor's degree is sufficient to get good jobs. However, contrary to popular misconception, graduate school is not just "more of the same."

Good reasons for going to grad school:

  • You've already learned how to learn
  • Professors run grad classes in a more enjoyable manner, because students are mostly A and B students
  • After finishing courses, you spend time (months for a Masters, a couple years for a Ph.D.) doing research, which is more like a fun job than "school."
  • You'll work closely with great fellow grad students who often become lifelong friends
  • After all this, you get an advanced degree, higher pay, more job satisfaction

    Bad reasons for NOT going:

  • You plan to work now and return to grad school later. Good luck ever coming back -- it's hard to wean yourself from the salary.
  • You intend to get a masters while working. Even taking one extra class can be tough on an already long work day (plus family life), and stretches on for 6-8 years. Not particularly fun.
  • You heard there are fewer available jobs for advanced degrees. If quantity is your metric, McDonalds is hiring on every corner.
  • You don't want to delay starting a family. But grad school can be a great place to start a family -- campus family housing is inexpensive and convenient, fellow grad students are supportive close friends. I got married before grad school, and had two kids while in grad school -- my wife and I loved it, wouldn't change a thing.

    Choosing a Graduate School

    Choosing a school is hard, but these are the good choices in life (better than choosing which leg to amputate). There's no "right" choice -- every school has pros and cons. So relax, don't stress over this decision.

    Many students look for highly-ranked schools. To be frank, ranking does make a difference. Fellow students will be sharp, making good friends to have in the future. If you plan to look for a job as a professor, the ranking of your Ph.D. school plays a key role. But, you can do great research at other schools too. Ranking is an important factor, but not the only one.

    If you have a preferred research area, you might give more weight to schools with strengths in those areas. Check department web pages, or ask professors you know.

    Likewise, if you happen to have heard good things about professors in your preferred area, that could help. But really, most students don't know professors beforehand.

    Apply to many schools; the cost of applying is trivial in the big scheme of things. Admission offers often include financial support that pay tuition/fees, and offer free money (fellowship, aka scholarship), a teaching assistantship, and/or a research assistantship, for some number of years. If you get multiple admission offers, don't just consider the number of years of financial support -- most students switch to research support from a particular professor after a year or two, so the difference between a four-year offer and a three or even two year offer may never materialize.

    Some schools expect every student they admit to succeed. Others fully intend to fail some percentage of their admitted students. If that matters to you, ask around to find out which schools are which.

    It is unethical to accept a support offer from one school with the plan to switch schools after a year or two. However, if you are stressing out over the decision, realize that if you are truly unhappy at one school, you aren't forced to stay -- you could switch schools later if necessary.

    The First Years

    You should plan to work hard in grad school. But, take the easier classes first. That gives you time to adjust to your new housing, new campus, new professors, even new country. It also gets you some higher grades in the "bank" before you take tougher courses that may lower your GPA. And, it can help with your motivation. Ask senior students if there are professors with overly-harsh grading standards; consider avoiding them altogether if possible.

    Working as a teaching assistant (TAing) can be fun and rewarding. Don't count on there being much training. Partnering with an experienced TA, at least for the first few weeks, helps a lot. Read the textbook before the semester; you'll already be way ahead of the students (who may *never* read it). Learn to communicate heavily with fellow TAs and professors, such as restating verbally agreed upon tasks in a followup email; miscommunication (or undercommunication) is the source of many problems. IT'S BETTER TO OVERCOMMUNICATE THAN UNDERCOMMUNICATE.

    Young TAs (and young college teachers in general) tend to grade too harshly in their first years, expecting students to know as much as the teacher does. This not only is unfair, but demotivates students, causing even worse performance. Good TAing is a balance between giving students answers versus teaching them how to figure out the answers themselves ("Feed a man a fish and he eats for a day; teach him to fish and he eats for a lifetime"). Finding the balance is the art of teaching. Motivation is also critical -- if you can motivate students to WANT to learn, they'll learn much more. Speaking to a class requires speaking loudly -- practice this (it feels funny at first). Don't be put off by American students' "coolness." High-school in the U.S. is socially brutal; a tough facade helps the kids survive. Pay attention to grooming and clothing; you may find students have more respect (why do you think airline pilots wear uniforms?).

    Help students help each other -- having them work in pairs or groups, in class or out, can yield great results. Keep "lectures" to 10 or 15 minutes at a time, then have students do something active. Otherwise, their attention wanes. Remember -- lecture is generally an ineffective mode of teaching. A bit of lecture, combined with students actively doing something, seems to lead to more substantial learning.

    Talk to fellow students and faculty as much as you can. Never underestimate the importance of casual conversation; a huge amount of information is learned just by chatting with people in the hallway, in the mailroom, and at lunch.

    Establish a habit of reading everyday, even if just for 15 minutes, perhaps when arriving to lab in the morning, just after lunch, etc. Read more when necessary of course, but stick to this minimum habit. Have a stack of items that you want to read (papers, textbook chapters, magazines, websites, etc.); reorganize the stack as priorities change. When you read, first skim the item to get the main idea, and to decide whether a second deeper reading is even necessary. On a second read, decide approximately how much time you are willing to invest in this item (e.g., 10 minutes versus 1 hour), and read at the appropriate level of depth for that allocated time.

    Choosing an Advisor

    Your advisor oversees your graduate research and approves your thesis. He/she is really more of a "boss" than an advisor. Obviously, your advisor should be in your desired research area, but other factors are important too.

    Talking to a professor's current and former students helps one learn if a professor is a good advisor. Does he treat the students well, and do his students graduate and get good jobs? A good student/advisor match may also require matching of personality, level of guidance, work schedule, ethics, etc. It's more like dating than it is making a rational choice like buying a car. Taking a course from a professor can help, but doing a small project for a professor, perhaps an extended course project or a volunteer project outside of any class, can help the student and professor decide if there's a good match. In doing projects with professors, be aware that professors are competing for students, so be careful not to unintentionally suggest that you have affiliated with a professor before you really have.

    Time Management

    Time management is the job of balancing multiple tasks that you must do. The key to balancing multiple tasks in graduate school is: DON'T. Graduate school is a time to study one problem very deeply, an opportunity you may never again have. Let yourself have this enlightening and molding experience -- avoid taking on lots of tasks. Maintain your personal life of course, but avoid overcommitting to other things at school.

    You will of course still have multiple tasks competing for your time. Give priority to tasks on the path to your main goal, which might be perhaps to get your Ph.D. and become a professor or get a great industry job. Most students instead give priority to tasks with closer deadlines. Be disciplined in allocating time. For example, given an on-path task of writing a paper draft due next week, and an urgent task of grading papers due in two days, allocate time to the former and do it FIRST. This might mean you have less time to grade -- everything is a tradeoff. You can't do the best job possible on every task, but rather the best job given the available time, which you decide. Keeping a "to-do" list of tasks, either electronic or in a notebook, can be a great help. Keep a notebook and write in it often, taking notes at meetings, writing down new tasks, recording ideas, etc. The trusty paper notebook is still essential; the only difference from 100 years ago is that today we might scan in the pages once in a while.

    Avoid distraction tasks having little importance, like frequent email checking, instant messaging, reading news sites, etc. Create blocks of uninterupted time; 3 hours uninterrupted can be far more productive than 6 30-minute work periods.

    Get sufficient sleep -- it's critical to clear thinking and to learning. Furthermore, strive to stay on a reasonable schedule -- coming in late and working late may mean less time interacting with your advisor and fellow students.

    Research

    Your advisor has the main job of teaching you how to do research. A few things can help. Take a baby step approach -- start with a small clearly defined project, write a paper, submit to a non-stellar conference, and present the paper if accepted. Gradually tackle more-complex less-well-defined problems. Tackling too hard a project initially is a common mistake, leading to little progress and to frustration.

    Don't just work on a project; think specifically about the contribution of your project to the research community. A contribution should be: (1) new, (2) useful, and (3) hard. For a given project, ask yourself what's new, why it's useful, and if carrying out the project requires advanced knowledge and cleverness. If not, your later paper might not get accepted, and even if it does, it may not have any impact. Coming up with research projects that make a real contribution is hard -- your advisor has the duty to help in your early years.

    Meet regularly with your advisor. Don't wait for him/her to set meetings; if he/she doesn't, request to meet. Meet at least weekly. Come to meetings prepared, with a pen, your notebook, printouts, materials from previous meetings, etc. -- don't just show up empty handed. Ask questions -- pick your advisor's brain. Go to lunch or for coffee sometimes, as those informal conversations are often just as important as formal ones.

    Writing a Paper

    Writing a good paper has very little to do with writing, and much more to do with two other things: (1) Having a clearly defined contribution, and (2) Organizing the subject for the reader. The goal is to make a point, not to write good English. The latter would be like saying that the goal of driving is to hold the steering wheel correctly.

    Regarding (1), first think of the one main point that your paper should make. What should the reader learn? It should be summarizable in perhaps one sentence, e.g., "Smoking increases lung cancer rates" or "My new cache memory design reduces power". Many paper writers just summarize the work they've done ("Here's what I did for the summer"), but the single main point isn't clear.

    Regarding (2), after determining a paper's main point, but before starting to write, you should organize the information you will provide. The information should clearly support your main point. Too many students throw in extra information because they spent time learning it or creating it, but including it dillutes the main point. Focus, focus, focus. Create an outline of the paper's sections and the subject of each paragraph.

    Create pictures for main ideas -- "A picture is worth a thousand words." Pictures also help convey the main point to readers skimming the paper. Holding off on writing sentences requires discipline, but the resulting papers are far more comprehensible. It's like painting a house -- disciplined painters don't just start slapping paint on the walls. Instead, they do a lot of preparation work first -- washing/sanding walls, covering furniture, taping, repairing holes, etc. The resulting paint job is far better.

    When learning to write, avoid potentially vague references like "It" or "This means." Replace such references by the explicit object being referred to, like "The computer" or "Increasing transistor densities means" even if such explicit reference means more words. An experienced writer learns to only use references when those references are unambiguous -- a tough skill to master. Also when learning to write, use simple sentences consisting of subject, verb, object/prepositional phrase. Don't use contorted sentences with lots of commas. The writing may not be artistic, but the point will be more clear. Be sure to have your paper read and proofread by others before submitting it for review -- a second set of eyes can find many ambiguities and mistakes, which can otherwise kill a paper in the review process.

    Be aware that reviewers rarely read a paper thoroughly; many just read the abstract and most of the introduction and conclusion, look at the pictures, and skim the content. Such skimming is not because reviewers are lazy, but rather because they are busy. Try to make your main point in the abstract, early in the intro, and through pictures. The first page of the paper should be a "masterpiece" (source: E. Keogh) -- concise, no mistakes, ideally a descriptive figure. Reviewers make their decision early and then look at the rest of the paper to support their decision. Think explicitly about potential criticisms, and address them directly (briefly in the intro, and in more detail in the paper).

    Only after all of the above is done should you start writing English sentences.

    Giving a Presentation

    Presentations are common in grad school, in classes, for Ph.D. exams, at conferences, etc.

    Being very nervous is normal. Nervousness decreases with experience, so don't be concerned about it -- it will get better.

    Think explicitly about your audience, and cater your presentation to them. Undergraduate students typically want several different explanations of the same subject. They like examples. They also want to be motivated. And they'll appreciate a few jokes. Technical conference attendees typically want to know the main idea and the results, and only want detail enough to understand the main idea, not to re-implement your project. Don't assume audience members are experts in your specific research area; think about what most audience members would already know, and start from there working towards your specific research contribution.

    Having a clear main point is even more important in a presentation than it is in a paper. Don't try to include everything you did -- that's a presenter-centric view of the presentation ("I did this work and I want you to know it"). Instead, focus on what the audience can learn from your talk, and emphasize that in the talk. Tell them the main point up front, rather than unveiling it slowly -- you are presenting technical information, not telling a TV story with a surprise ending.

    Don't compete with your own slides. If you create text slides and then speak the same things, the audience must decide whether to read or listen. You set yourself up to compete with your slides. Instead, use pictures and plots, and explain them to your audience -- now you are the good guy who is helping the audience understand. Keep text on slides short, supplementing pictures/plots just enough to organize information, remind yourself what to say, and make a key point explicit.

    For a conference/industry presentation, put your name at the bottom of slides -- you are selling yourself as much as you are selling your ideas. Number your slides, ideally like "5 of 12," so the audience knows how many more slides exist.

    Start a presentation by engaging your audience. Don't start by making "necessary" acknowledgements of colleagues and funding sources or by giving a dry outline -- audience members will decide to tune out. Start with an interesting fact, surprising data, a story, even a small joke in an informal setting -- audience members will perk up and decide to focus.

    Practice talks extensively. A conference talk should be practiced 5-10 times, ideally at least once in front of fellow students. Note time throughout the talk and know times at milestones, so during the real talk you know if you are short or long. More experience means less practice is needed, but even experienced presenters should practice a couple times.

    Respect allotted time -- avoid going long, which can destroy a schedule or keep people from their next appointment. If questions may be asked during a talk, meaning you don't know how long slides will take to cover, then know beforehand which slides you can skip. Experienced presenters actually prepare a shorter talk than the allotted time to allow for questions, and then include additional content to account for fewer questions, and which the speaker optionally will skip. Don't be attached to your material; your audience won't digest material that you plow quickly through while running over time. Better to skip things, focus on the main point.

    Conclusions

    Following the above advice (even assuming it was good advice) doesn't guarantee success -- there are too many other factors. A key ability to success in graduate school and in your career is to tolerate failure . Do your best, and hope to succeed, but don't feel bad if you don't. Most graduate students WILL experience failures along the way -- low grades in a course, bad TA evaluations, rejected papers (LOTS of those), rejected job applications, etc. The key is to tolerate those failures, learn what can be learned from them, and move on. Eventually, successes may come, and they may outweigh the failures. Even long-established professors and industry leaders experience failure after failure -- it's normal.

    A final note: all the success in the world may mean little if you don't have good people in your life. Seek relationships, nurture those you currently have, and you may find success to be much more meaningful.

    For More Information

    A book on the above subject is available: How to Be a Good Graduate Student. The book covers the above subjects in more detail, covers more aspects of each subject, and includes additional chapters on traveling and on preparing for an academic career. The book is available as a free pdf file. 

    For undergraduate students considering whether to attend graduate school, we at UCR CS&E have been using the following slides for several years at our annual graduate school information meeting: Is Graduate School for You?.