Hopkins Biology Schleif

Some Comments for Graduate Students

Doing good science requires much training, learning, experience, and hard work. Many of the skills required must be self-taught, but some can be learned from others. Here I offer a few comments on some of the more practical aspects of graduate school and doing science. There is no need for everyone to have to relearn all of these for themselves, and perhaps another's experiences in these areas will help you.

A simple summary of the objective of graduate school is that here you learn that you can think and how to think. You will acquire more facts and you will learn how to carry out different experimental or theoretical procedures. The much more important objective however, is to learn how you can learn about nature by thinking about your results and others' results and then by designing and executing experiments, to test your ideas. Good science is much more than the mere collection of data. Finally, you may acquire a taste and a sense of style of doing science, although this often develops over a decade or two of doing science.

The training one obtains in graduate school in Biology and Biophysics prepares one for a rather wide variety of professions. Teaching and/or doing research are common, but it is also possible and common to pursue careers in law or business. Frequently one follows graduate school with several years of postdoctoral training.

Once one has learned something about a particular field, it is tempting to want to learn more in this specialty. Thus, in coming to graduate school after undergraduate training that includes some research, students often want to continue in the same field. Similarly, having done a Ph.D. thesis on a topic, students sometimes want to continue in the very same field. Often this is not a good idea. Most employers, whether they are in academia or industry and whether it is for a teaching, research, or administrative position, are more favorably impressed by people who have demonstrated that they can learn and produce in several different areas. Of course, you want to choose areas that complement one another, but if you have determined that you want to spend the major part of your career in one area, it is often a good idea to do your thesis in something complementary and perhaps more fundamental than your ultimate objective.

Doing anything well requires more than 40 hours a week. On the other hand, very few people can or should work 12 hours a day seven days a week for extended periods. When enduring the rigors of graduate school, and of work after graduate school, it is at times tempting to focus too intently on the task at hand. Remaining mentally, physically, and emotionally well requires eating well, plenty of sleep, physical exercise, and maintaining a reasonable balance between work and play. Ignoring the need for this balance for periods longer than a couple weeks or months seriously jeopradizes one's abilities to function well mentally, and then one's physical health. In this same regard, it is necessary to try to smooth out the highs and lows that inevitably come in work that is unpredictable. A steady and thoughtful approach is much more productive over the long term than periods of inactivity and very intense activity.

Theory of Exams

By now in your education, you are familiar with taking written exams, but now you will be subjected to oral exams, and it may be helpful to consider how they may be given. Before discussing oral exams, it is helpful to consider written exams. Generally the objective of giving exams is to induce students to study and organize the material and also to reveal to the student and to the instructor just how much learning and understanding has taken place. In the absence of trying to flatter or cow the students, the most efficient design for an exam is to aim to spread the scores widely around an average of 50. It provides little information to ask a question that no one can answer, or to ask one that everyone answers. Nor does it make sense to make the more difficult questions worth more than the easy ones. This merely skews the scores, making it harder to evaluate the understanding of the average and below average students.

Oral exams provide the opportunity for a much better determination of a student's boundaries of understanding, and if properly given, can do so in a very short length of time. The problem is that giving an efficient oral exam requires considerable preparation beforehand, and alas, many faculty members don't do this. Hence the Socratic exam (more about this later). The objective of an efficient oral exam, as in a written exam, is to learn the boundaries of a student's knowledge. To do this, you ask a simple question, say, and the student begins the answer. The moment you see that the student knows the answer, you stop him or her and ask a more difficult question in the same area, and if you have jumped beyond the student's boundary of understanding, he or she will fumble. Again, the moment you see that the student does not readily know the answer, you stop the student, and either ask more questions to narrow in on the student's level of understanding on this topic, or jump to a new topic. With such a strategy, the examiner and the student quickly learn the boundaries of understanding of multiple topics without a complete answer ever having been given.

Unfortunately, sometimes faculty members use Socratic examination methods. In this approach, a one topic is exhaustively covered by asking a long series of leading questions and quite frequently giving strong hints so that the questioning can proceed to some major "understanding" that the student reaches. An exam given along these lines takes a long time and reveals rather little about the student's depth of knowledge. Such an exam does, however, demonstrate to everyone present that the examiner knows the subject well.

A Few Considerations in Choosing a Graduate Advisor

One of the considerations in choosing a graduate advisor is the area of his or her work, but it is certainly not the only reason, and frequently it should not even be the most important reason. The particular facts and experimental techniques that you can learn from an advisor and in a lab are the least important consideration. Very soon the facts and techniques will be replaced by newer and more encompassing facts and better techniques will replace those currently in use. No one can tell you how to weigh the various other factors of importance in choosing an advisor, but it is a good idea to learn as much as you can about the issues raised by the following questions.

One of the most reliable measures of a potential advisor is his or her track record. It is quite astonishing when one considers the scientific output of a laboratory and the human output of a laboratory. A disappointing number of prominent laboratories, ones that frequently publish in the foremost journals produce few future scientists of note. Other equally prominent laboratories, and frequently many that are not so prominent, produce handfuls of active, prominent, productive scientists in academic and/or industrial positions. Quite why students from some labs tend to be highly successful after they leave the lab and why students from other labs rarely are successful is unclear. The records, however speak for themselves. If you are considering a younger faculty member, you will have to make a decision without the benefit of a track record. In this case, one thing to look at is the track record of the Ph.D. advisor and post doctoral advisor of the faculty member you are considering. Perhaps in this case, you might want to pay closer attention to some of the factors listed below.

How interested is the advisor in working with graduate students?

Be aware that scuttlebutt and rumor are highly unreliable and are often propagated by vocal students who are trying to justify their own decisions or rationalize something. Be aware too, that an advisor who may be unsuitable for 90% of the students may be the ideal choice for a few. Know yourself too.

Is the advisor available to each student an adequate amount of time?

Is time with the advisor scheduled, or does it occur randomly, whenever needed, or is it limited by the advisor's patience, interest in his or her own work, or activities outside the lab?

How well does the advisor tolerate opposing positions and scientific disagreement?

Is the advisor sufficiently organized and responsible that you will be confident that important letters of recommendation will be promptly sent, even after you are no longer on the scene?

Is the advisor in touch with a large number of scientists elsewhere so that the lab is aware of the very latest findings and techniques and will your advisor know whether your project is or is not being pursued in other laboratories?

What is the practice of the advisor with regard to competition, both between different laboratory members and with scientists elsewhere?

How much freedom does the advisor allow in pursuing a question? If the interesting questions lead in unexpected directions, can you pursue them? If you run into difficulties, will the advisor help out, redirect your research, or expect you to work things out yourself? Any of these possibilities can be just fine, but it is good for you to know beforehand what to expect.

How much does the advisor work with students in teaching them how to write? That is, how do the papers usually get written? Similarly, what about learning to speak about one's work?

Does the advisor encourage attending scientific meetings? What about meetings somewhat outside his or her own area of interest?

How hard does the advisor work to help students find their next position?

How willing is the advisor to let a student learn about other areas and other labs in deciding what to do after graduate school?

Does the advisor encourage or discourage interaction with other labs in the department and elsewhere in the world? Does the advisor attend a wide variety of seminars and does he or she expect laboratory members to attend a variety or seminars, only very closely related seminars, or none at all?

Giving Scientific Talks

Not only is it necessary to do good work, but it is necessary to communicate this fact effectively by writing good papers and giving good talks. It is surprisingly difficult to give a good scientific talk, and much practice, over many years, normally is required for most people to become good speakers. Avoiding a few of the common pitfalls can accelerate the learning process however.

One way to become aware of the quality of talks is to try to understand them. Pay attention to Tuesday research progress talks and seminars. Analyze what makes a talk clear and understandable or unclear. Figure out what the speaker could have said that would have allowed you to have understood it more easily. Do not conclude that someone who gave a talk that was largely incomprehensible must be very smart, doing very advanced work, or has important conclusions. Likely the presenter of such a talk is a bit insecure and is afraid to state the question and the findings in a simple and comprehensible way. The very best scientists can communicate well and make their work seem clear and simple.

It is very difficult for a listener to keep everything in mind and assemble it at the end of the talk. In contrast to reading a paper, the listener of a scientific talk cannot go back and reread difficult sections. Therefore, important or complex points have to be repeated. It is crucial that right at the beginning you explain what the major question is and what the major conclusion is. Next, for any talk longer than 20 minutes, it is most helpful to describe the major points you will make or provide an outline of the argument you will make. This permits the listener to focus on the issues relevant to the conclusion of your talk. At all costs, avoid the murder mystery approach of giving a lot of data and at the end wrap it all up and say "thus, the conclusion is ..."

After the initial statements laying out the talk and its conclusions, it is necessary to put the work in scientific context. Although most of us can understand the experimental data and conclusions to be drawn from them, few of us possess the background knowledge necessary to know why the experiments were worth doing and the importance of the findings. Most poor talks lack adequate explanations of the scientific reason for performing the work.

Beware of PowerPoint pitfalls. Many people are annoyed to have slides read to them. Don't do it. Slides can present figures and some data, but need not contain the text of your talk. Similarly, it may seem cute the first three times you see new data dance in from the right or left or materialize in some unusual way, but soon these contrivances become trite. Focus on clear presentation of ideas, results, and conclusions, and skip the glitz.

Keeping the audience in mind while lecturing is essential for maximum communication. To a general audience one should not mention specific strain numbers or models of equipment. No one knows what they are and they only complicate matters. The general rule is to avoid providing information that is not essential for an understanding of the points you are making. This is not to say that you shouldn't be ready to provide this information if you are asked, but the point of the talk is to inform others of some general scientific conclusion, not the details of your buffers.

Overpreparation for a talk is harmful also. People process the spoken word not only by the actual words, but also through inflections, intonations, speed, and loudness. If you are actually processing the information contained in your talk as you are giving it, your voice will speed up and slow down and change in a way that greatly helps the listeners understand. Overpractice and memorization of the words you will speak wipes out important vocal clues.

If you have worked on a subject for months or years, your ideas, experiments, and conclusions will naturally seem very simple to you. It is only natural to want to present your work to others in a way that does not make it seem trivial. It isn't, and what seems simple to you is far from simple to people hearing it for the first time. All too often a desire to keep one's work from appearing trivial results in making a talk far more complex than it needs to be. The real conclusion may be simply that any two of three proteins can bind to a fourth protein or DNA site, but a good many speakers will likely disguise this conclusion in some grand statements about the importance of the system and the insights their results will provide and then overwhelm you with a recreation of the torturous path by which they learned the relatively simple conclusion. In the end, nature is usually simple, and our conclusions can reflect this simplicity.

Reading the Literature

To be able to do important and original work it is necessary to know what has been done and what needs to be learned. To do creative work it is necessary also to be exposed to many ideas from outside your specialty! Everyone within your specialty will know the same facts and be aware of the obvious experiments. To produce work distinguished from that of everyone else requires having been exposed to additional ideas and facts, and then using the information creatively. Learning almost everything in your specialty and also being exposed to new and fresh ideas requires spending significant amounts of time reading, talking with people inside and outside your specialty, attending talks, seminars, and scientific meetings.

Reading the literature is a daunting task. A variety of approaches can help maximize your efficiency in this important task. As a graduate student beginning on a research problem it is most logical to begin by reading half-dozen papers most closely related to your work. Then read a lot more related papers and soon you will be able to read the new papers which are related to your work as they come out in the current journals. This might mean following five to ten journals. Initially, the reading will be slow because you will have to read much background information to understand a paper. Eventually your rate of reading will increase and you can expand the range of papers you read. Probably it is useful also to develop a generic search that will reveal new relevant papers and to run this search once a week or so.

Wholesale storing of papers "to read later at a more convenient time" is a trap to avoid. Very quickly a large number of unread papers accumulates, but not much reading gets done. Even spending fifteen minutes scanning a paper on line is better than storing it in your computer or printing it and never reading it. I now only read papers online, and I add about one fifth of the papers I read to my references database. Fortunately, the university subscribes to the RefWorks database, and I use this extensively. By all means begin to use this program or something similar. Such use will save you enormous amounts of time over your graduate career and afterwards.

It is very easy to overlook issues of journals. To overcome this problem, I list the journals which I follow and check off the issues after I have checked the contents, read a good many abstracts, and perhaps read a paper or two.

With the availability of complete bibliographic databases like Medline, Google, or Google Scholar, finding old papers is greatly speeded. It is also possible to scan sectors of the literature, reading just the abstracts of many papers. Because many papers actually demonstrate findings quite different from what is claimed in the abstract, this approach cannot substitute for reading complete papers. It is a pity that the literature database programs do not return some totally erroneous hits, essentially randomly chosen. In the past I have stumbled upon important papers because they were next to papers I thought I should read.

Research Notes

Because our memories are extraordinarily imperfect, we need to record for future reference our thoughts and experiments. These notes must tell us why we did an experiment, what was done and found, and what it means. They must be sufficiently complete that we, and others, may repeat the experiment and obtain the same findings.

It requires great discipline to force oneself to write a clear explanation of why an experiment is being done, and then after the data has been collected and analyzed, to write another section describing the conclusions of the experiment. This material is highly important, however, for it permits understanding the experiment and its findings at a later time. Often the reasons for its execution will have been forgotten, and occasionally the hypotheses used in its interpretation will also have been forgotten, been altered, or found to be incorrect. The additional paragraph placing the experiment, its findings, and conclusions in context greatly helps in extracting useful information from "cold" notes at a later time.

Simply keeping track of experiments is no easy task. The notes can be dated, but this doesn't overcome ambiguities when a number of experiments are being done at the same time. Sometimes it is better to identify experiments by number rather than by date.

Once upon a time a lot was written about the need to keep notes in ink in bound notebooks. (I wrote most of mine in pencil and kept them in manila folders.) Now, for many people the best approach seems to be to keep laboratory notes on a computer. Not only do computers solve the storage problem, but search programs allow finding most anything in a few seconds. At the same time, the storage capacity of computers is so high that one can keep a pdf file of every paper you have carefully read and have virtually instantaneous access.

No matter how the notes are being kept, the notes for each experiment or roughly each day's work need to contain the following:

I. A date and/or an experiment number. The need for this is obvious, merely to order things.

II. A prose paragraph that describes the context of the experiment.

a. It or the corresponding paragraph in the notes for an earlier experiment must clearly describe what is known and what is to be learned. Often the theory on which the experiment is being based needs to be mentioned. (All this information is exceptionally important later because science will have moved on, and without the description, it will be hard to recollect or reconstruct what the question was and therefore what the experiment was about.)

b. The principle of the experiment must be mentioned or described.

c. What the theory predicts and what the experimental data may look like.

III. The actual notes of the experiment. These can contain calculations and preparations, instrument settings, and the data from the experiment or links to the data.

Notes are most helpful when they contain sufficient information that mistakes can be caught. My notes often describe how I make the reagents. For example, in making a 2 M solution of KCl, you can write down the mass of KCl to weigh out and the volume of water it is dissolved in and even date the resulting bottle of 2 M KCl. Later if there is question about the reagent, sometimes you can look back in my notes and sometimes find at least how you intended to make up the solution. From time to time mistakes are made, and this relatively painless procedure seems to catch a few of them.

IV. Results and conclusions written out clearly and completely.

V. What comes next.

It is amazing how useful this final part is to one's experimental notes. Often times a line of experiments gets dropped, and when you look back over your notes, it is unclear why you stopped. Was it because the idea was wrong, adequate data could not be obtained, or something more interesting came along? Parts I and V prove to be exceptionally important in interpreting older research notes. Of course you know what you are doing day to day, and usually week to week, but properly interpreting data that may be several years old and after an idea you were pursuing has undergone several modifications and improvements is quite another story. Parts I and V greatly aid in understanding old notes and in extracting valuable information from them.

Computer Programs

Computers are invaluable tools for handling verbal and graphic information as well as for performing calculations. One should therefore, use and become expert on the following types of programs: word processor, spreadsheet, drawing, reference manager bibliographic, and image editing. Use the top programs, Microsoft Word, and Excel and RefWorks. I used to recommend Adobe Illustrator and Adobe Photoshop, but due to their present avaricious pricing structure, I recommend other drawing and image editing programs. When you are bored with a project, rest a little by exploring features of the program that are new to you. The variety of uses of a spreadsheet program is truly amazing. For three dimensional drawing the Google program SketchUp seems quite good.

I strongly recommend obtaining a powerful file indexer-finder. As an educated person, you will be judged by the economy and precision of your communication. For your writing and speaking to improve, you need a access to the equivalent of a dictionary, and fortunately looking up words and their appropriate useage is simple via Google, as you don't even need to know how to spell them correctly.

You will be able to keep your entire life's work on your computer plus all the email you receive, as well as every book and paper you carefully read. As cloud computing and information storage grow, it may soon be possible not to put much of anything on your own computer. For example, the bibliographic programs now allow rapid access to papers via the internet, and thus there is little need to download them into your own computer. The real problem is not storing the information, it is finding it again. Of course you will want to be organized and systematic about where you put things, but this still won't help when two weeks or 20 years later you are looking for something. File indexing programs are a good solution if the information is stored on your own computer. I expect it will not be long however, until you can conveniently list online papers or sources to be indexed by your own personal indexer. At present these programs prepare an index of essentially every word in every file in the directories or folders and web sites that you specify, and then when you are looking for something, you can do a Google-style search or a more sophisticated and precise search that instantly finds all the files that meet the search specifications, open a window into the most relevant and show you a half a page of text surrounding the first occurrence of your search item. I am currently using an indexer-finder called dtSearch (PC). Alas, it costs $200, but it is well worth it.

Filing Papers and Other Stuff

At one point I found my desk and office to be cluttered and almost unusable. On looking at the stuff I realized that it hadn't been put away because what was left out didn't fit within the categories for which I had file drawers or folders. Subconsciously, I must have been afraid that if I put the stuff away, I'd never find it again. The solution is simple and has been most helpful to me. Now, as such random stuff accumulates, I put it in a pile until the pile is an inch or two thick. Then, I number each item starting with the next number from where the indexing last time left off. In addition, in a word processor file that lists the items that have been numbered, filed, and described in the past, I add the new numbers and descriptions using as many unique words as possible, for example, " 458 photograph and list of attendees to the FASEB transcription meeting 2013 in Vermont". Usually I need to do this indexing every few months. Once labeled and indexed, I can safely file the items away by number. When I am looking for something, the first place I look is the word processor file listing my filed items. A simple word search usually yields the description, and from the date and number, I can then quickly find the item in my file cabinet.

Music in the Lab

BiP, that is, before iPhone and iPad, it was relatively easy to make the case that music should not be played in the lab simply because some people don't like some types of music, and it is very difficult to ask one's peer or superior to turn it off. Now, with smart phones, this argument cannot be made. It must be pointed out, however, that being plugged into music is the equivalent of wearing a "Do not disturb" sign. Why bother to be in a laboratory then? Who is going to ask for help or even share a spur-of-the-moment idea with you?

A more important reason for questioning the utility of music involves creativity. The argument goes as follows. While it is very nice to be hearing one's favorite type of music while performing boring tasks, this background distraction is precisely what one doesn't want. The major reason for doing science and research is to have good ideas and come to deeper understandings of things. When one is horribly bored doing some menial task, there is a chance that your annoyance with the task will spur you into devising a better way to do the job, or a way to avoid the task, or provoke you into questioning your whole experiment or even coming up with a creative idea on somehinge else. If many of your neurons are occupied with the background music you are using to dull your senses, you are less likely to come up with a good idea.