a ph.d. is not enough a guide to survival in science

APh.D.

Is NotEnough

A Guide to Suroiualin Science

Peter ]. Feibelman

......

Addison-Wesley Publishing CompanyReading, Massachusetts Menlo Park, California New York

Don Mills, Ontario Wokingham, England Amsterdam BonnSydney Singapore Tokyo Madrid San Juan

Paris Seoul Milan Mexico City Taipei

Many of the designations used by manufacturers and sellers to distinguish theirproducts are claimed as trademarks. Where those designations appear in thisbook and Addison-Wesley was aware of a trademark claim, the designationshave been printed in initial capita1letters.

Library ofCongress Cataloging-ln-PubHcation Data

Feibelman, Peter J.A ph.d. is not enough: a guide to survival in science / Peter J.

Feibelman.p. em.

Includes biographical references.ISBN 0-201-62663-2 (paperback)ISBN 0-201-62717-5 (hardcover)1. Science-Vocational guidance-Handbooks, manuals, etc.

2. Scientists-Training of-Handbooks, manuals, etc. 3. Mentors inthe professions-Handbooks, manuals, etc. I. Title.Q147.F45 1993S02.3-dc20 93-26668

elP

Copyright 1993 by Peter J. Feibelman

All rights reserved. No part of this publication may be reproduced, stored in aretrieval system, or transmitted, in any form or by any means, electronic, me-chanical, photocopying, recording, or otherwise, without the prior written per-mission of the publisher. Printed in the United States of America. Printedsimultaneously in Canada.

Cover design by Lynne ReedText design by Diane LevySet in Ll-point Garamond by Carol Woolverton Studio

2 3 4 5 6-ARM-97969594Second printing, April 1994

Addison-Wesley books are available at special discounts for bulk purchases bycorporations, institutions, and other organizations. For more information, pleasecontact:

Special Markets DepartmentAddison-Wesley Publishing CompanyReading, MA 01867(617) 944-3700 x 2431

Contents

Preface: What This Book Is About ix

Acknowledgements xv

I Do You See Yourself in This Picture? 1A set of nonfiction vignettes illustrating some of the waysthat young scientists make their lives more unpleasant thannecessary or fail entirely to establish themselves in aresearch career.

2 Important Choices:A Thesis Adviser; a Postdoctoral Job 17A discussion of what to consider: young adviser versus anolder one, a superstar versus a journeyman, a small groupversus a "factory." Understanding and attending to yourinterests as a postdoc.

3 GivingTalks 27Preparing talks that will make people want to hire andkeep you, and that will make the information you presenteasy to assimilate.

Vll

viii

A Ph.D. IS NOT ENOUGH

4 Writing Papers:Publishing Without Perishing 39Why it is important to write good papers. When to writeup your work, how to draw the reader in, how to drawattention to your results.

5 From Here to Tenure:Choosing a Career Path 53An unsentimental comparison of the merits of jobs inacademia, industry, and in government laboratories.

6 Job Interviews 71What will happen on your interview trip, the questions youhad better be prepared to answer.

7 Getting Funded 83What goes into an effective grant proposal, how and whento start writing one.

8 Establishing a Research Program 95Tuning your research efforts to your own capabilities andyour situation in life, e.g., why not to start a five-yearproject when you have a two-year postdoctoralappointment.

Afterthoughts 107A behaviorist approach to professional success.

Preface:What This Book Is About

My scientific career almost never happened. I emergedfrom graduate school with a Ph.D. and excellent technicalskills, but with little understanding of how to survive inscience. In this I was not unusual. Survival skills are rarelypart of the graduate curriculum. Many professional scien-tists believe that "good" students find their way on theirown, while the remainder cannot be helped. This justifiesneglect, and perhaps not incidentally, reduces work load.There may be some sense to the Darwinian selection pro-cess implicit in "benign neglect," but on the whole, failingto teach science survival results in wasting a great deal ofstudent talent and time, and not infrequently makes a messof students' lives.

Since science survival skills are rarely taught in a directway, most young scientists need a mentor. Some will findone in graduate school, or as a postdoctoral researcher, orperhaps as an assistant professor. Those who do not, toparaphrase Mencken, have an excellent chance of movingfrom graduate study to scientific retirement without pass-ing through a career. The unmentored can only succeed

1X

x


by being considerably more astute than the naive, idealis-tic, and very bright young persons who generally choose ascience major.

These thoughts have been on my mind ever since I al-most had to tell mom and dad that their golden boy wasnot good enough to find a permanent (or any') job inphysics, a job for which his qualifications included eightyears of higher education and four more of postdoctoralwork. The agony of those days is not easily forgotten-theboy with the high I.Q., who had skipped a grade, gradu-ated from the Bronx High School of Science at 16 and fromColumbia, Summa Cum Laude at 20, found himself in amuddle at 28. How do you choose a research problem?How do you give a talk? What do you do to persuade auniversity or a national or an industrial lab to hire andkeep you? I hadn't a clue, until, midway through my sec-ond postdoctoral job, I had the good fortune to spendsome months collaborating with a young professor whocared whether I survived as a scientist. Although this men-toring relationship was brief, it helped me acquire a set ofskills that graduate education did not, skills without whichmy lengthy training in physics would have been wasted.

This book is meant for those who will not be luckyenough to find a mentor early, for those who naively sup-pose that getting through graduate school, doing a post-doc, etc., are enough to guarantee a scientific career. Iwant you to see what stands between you and a career, tohelp you prepare for the inevitable obstacles before theyoverwhelm you. In shaft, I hope to enable you to use your

Xl

Peter J. Feibelman

exceptional brainpower in the way that you and thosewho put you through school have dreamed about.

I begin with some brief "case histories." This may helpto put your own early career in better perspective. At leastI hope it will give you a feeling for how important men-taring can be. Succeeding chapters are arranged in parallelwith a career trajectory. Please skip ahead to whichevermay be relevant to your situation. Chapter 2 deals withchoosing a thesis or a postdoctoral adviser. My choice ofthesis adviser was based on two criteria: Who is the mosteminent professor in the department? And whose studentsfinish soonest? Was this intelligent, or did it represent a firstmistake? Chapter 3 concerns oral presentation of yourwork. However brilliant your insights, they will be of littleuse if you cannot make them appear interesting to others.If no one pays attention, what difference does it make ifyour results are clever! There are' of course Nobel prizewinners whose orations are Delphic, whose overhead foilslook as though they were scribbled during a particularlyturbulent flight, and so on. But you are not one of themyet, and if that is how your talks are prepared, you neverwill be either. There is more to Chapter 3 than "neatnesscounts," however. It contains several important ideas onmaking your oral presentations effective.

In Chapter 4 you will find a discussion of paper-writing.Through your scholarly articles, you can make yourselfknown nationally and internationally. This means thatyour reputation in science does not just depend on whatyour boss says about you, but on documentation that is

A ph.D. IS NOT ENOUGH

available in any science library. You should thereforeview publishing as a means to attaining job security, andtake the task of writing compelling journal articles veryseriously.

Chapter 5 is devoted to career choices, mainly the meritsand defects of positions in academia, and in governmentor industrial labs. The focus is on being reflective and ra-tional rather than naive or romantic about key decisions inyour scientific life. In Chapter 6 I discuss job interviews.There is more to an interview than wearing your Sundaybest, and having a firm handshake. Doing your homeworkand persuading your potential employers that you have asense of direction are the most important issues. Inciden-tally, this is not a matter of deception-knowing who yourcolleagues will be, and developing an idea of what youwant to know, scientifically, are keys to having a prod-uctive career. There are also a few choice words in thischapter about negotiations, once you do get an offer.Negotiating for what you will need, when your leverage ismaximal, can make a large difference to your happinessand to your success.

In Chapter 7, I discuss what to many is the bane of sci-entific life, namely getting money. This used to be the ex-clusive headache of academic scientists, but nowadays it isalso a significant part of the lives of government and in-dustrial scientists. I suggest that you view the preparationof a proposal as an important scientific exercise. Comingto see, and being able to articulate how your work fits into"the big picture" is not only essential to winning financialsupport, but also to being a first-class researcher. Learning

Peter J. Feibelman

to distinguish extravagant "pie-in-the-sky" from promisesthat you have a chance of fulfilling is also very valuable.

The most difficult problem in being a scientist is select-ing what to work on, and it is even more difficult whenyou are just launching your career. Therefore, in Chapter 8,I venture a few comments on establishing a research pro-gram. Jumping into the hottest research area may not bea very good idea, nor is taking on a project that you haveno realistic hope of completing before your short-term em-ployment comes to an end. The main idea is to establish aprogram that simultaneously maximizes your chances ofcontinuing employment and of scientific achievement.The focus is on strategic thinking.

As this book is written, economic times are tough world-wide and funding for scientific research is contracting. Ihardly need to emphasize that when resources becomescarce, competition intensifies for what remains available.To win a permanent position in scientific research, and thefunds to carry on serious work, you will have to be excep-tionally thoughtful about your career choices. My hope isthat this "pocket mentor" will help you to become more in-trospective about what it will take to succeed.

-Albuquerque, NMAugust, 1993

Acknowledgements

In order to make this handbook accessible to peoplewhose backgrounds, experiences, and scientific interestsdiffer from my own, I have prevailed on several friendsand colleagues for advice. I am very grateful to ProfessorsMichael J. Weber and Alison P. Weber of the University ofVirginia, for numerous constructive criticisms of the firstdraft. I would also like to thank Dr. Ellen Stechel, my col-league at Sandia National Laboratories, and ProfessorGeorge Luger of the University of New Mexico for theircritical readings of the manuscript. Finally, I thank mywife, Lori, for many editorial improvements.

IDo You See Yourself

in This Picture?

The brief stories in this chapter have a common theme:that understanding and dealing rationally with the realitiesof a life in science is as important to science survival as be-ing bright. Once you leave graduate school, the clock isticking. Unlike a fine wine, you do not have many yearsto mature. As a young professional, you must be able toselect appropriate research problems, you have to finishprojects in a timely manner, and you ought to be givingcompelling talks and publishing noteworthy papers. Whenjob opportunities present themselves, you should be ableto assess their value realistically. Romanticizing your pros-pects is a major mistake, and is likely to have serious con-sequences, not excluding dropping out of scientific lifeprematurely. The first story is an excerpt from my own sci-entific beginnings. The others are also nonfiction, though Ihave altered locations and personal characteristics to avoidinvading the privacy of the protagonists. I have deliber-ately identified the various characters with initials, ratherthan names, to avoid any ethnic implications.

1

2


What Do Scientists Do?Technique versus Problem Orientation

Virtually all classroom work and much of what happens ina typical thesis project is aimed at developing a student'stechnical skills. But although the success of your researchefforts may depend heavily on designing a piece of appa-ratus or a computer code, and on making it work properly,no technical skill is worth more than knowing how to selectexciting research projects. Regrettably, this vital ability isalmost never taught. When I signed on with a research ad-viser in my first year of graduate school, I was thrilled tobe given a problem to work in the physics of the upper at-mosphere. The fact that I had no idea what motivated theproblem did not prevent me from carrying out an analysis,on a supercomputer of the day, and publishing my first pa-per at the age of 22. For my thesis, I consciously switchedto a project that would require learning the tools of mod-ern quantum physics, but again I found myself assimilatingtechnical skills without ever grasping the significance ofthe problem, without understanding how or whether itwas at the "cutting edge" of science. This way of workingbecame a habit, one that seriously threatened my career.My first seven publications were in seven different areas ofphysics. In each case I relied on a senior scientist to tell mewhat would be an interesting problem to work on; then Iwould carry out the task. I assume that it was my ability tocomplete projects that impressed my superiors sufficientlyto keep me employed. It certainly wasn't my depth in anyfield.

3 Peter J. Feibelman

Four years and two postdoctoral positions after earninga Ph.D.-still having little sense of what I wanted to learnas a scientist-I was in the job market. More than anythingelse, I needed good recommendations from faculty at theuniversity where I was employed. I was asked to give theweekly solid-state physics seminar, and realized at bestdimly that my performance in this venue was either goingto make or break me as a scientist.

The talks I was giving at this point in my career reflectedmy approach to science. There was little in the way of in-troductory material. Much of the presentation was techni-cal. I would describe a few "interesting" problems I hadworked on and explain the methods I had used, but wouldgive little idea of context because I really didn't knowwhat it was. For the seminar at hand, I prepared my usualhodgepodge of this project and that, with no introduction,no theme, and ultimately, no meaning to anyone but an"expert." Fortunately, the professor supervising my re-search, C., understood what was about to happen to me,and asked for a preview of my seminar in his office. Thankgoodness I accepted this invitation. C. expressed surpriseat how poorly I had prepared my talk (though I don't thinkhe was surprised at all), how little grasp I seemed to haveof the reasons that the problems we had worked out weremeaningful, and consequently, how uninterestingly I wasgoing to present them to my audience. But, he told me, hethought I was too good technically to be allowed to fail inthe way I was about to, and he gave me the lesson Ineeded.

His most important advice was:

4 A Ph.D. IS NOT ENOUGH

1. There has to be a theme to your work-some objec-tive-something you want to know. There has to be astory line. (Do not start with: "I have been trying to explainthe interesting wavelength dependence of light scatteringfrom small particles," but rather "There is a widespreadneed to explain to one's kids why the sky is blue.")

2. If you know why you have chosen to work on aparticular problem, it is easy to present an absorbingseminar. Start out by telling your story, why the field youare working in is an important one, and what the mainproblems are. Give some "historical" material showingwhere the field is, the relative advantages of differentmethods, etc. Then outline' what you did, and describeyour results. Conclude with a statement of how your re-sults have advanced our understanding of nature, and per-haps give an inkling of the new directions that your workopens up.

Do not assume that your audience is comprised of "ex-perts." There may be a couple of them, but even expertslike to hear things that they understand, and particularly tohave their colleagues hear (from someone else) why theirfield is an important one.

3. Lastly, rehearse your talk in front of one or two ofyour peers and/or professional supporters. Choose listen-ers who will not be shy about asking questions and givingconstructive suggestions. Giving a seminar is serious busi-ness. Your future depends on the strong recommendationsof your senior colleagues. If your talk is a hodgepodge oftechniques or experiments or equations, if you seem to

5 Peter]. Feibelman

have no idea of where you are headed, if you reek of def-erence to the experts in the audience, you will not be per-ceived as a "rising star," a budding scientific leader. Youwill fail.

The wonderful result of C.'s mentoring was that I finallylearned what it means to be a scientist. By making mywork meaningful to others, it became interesting to me aswell. No longer was I just working on somebody else'sproblems. I was part of an intellectual enterprise with rela-tively well-defined goals, that might actually make a differ-ence to humanity. I scrapped most of the equations I hadplanned to show, and refocused my talk using thematicmaterial I had garnered from C. I gave an excellent semi-nar-people I scarcely knew complimented me afterwardon the interesting field I had chosen to work in, and re-marked on the clarity of my presentation. In science, thereinforcement doesn't get much more positive than that. Ihad learned a key lesson and was on my way.

Timing Is EverythingHaving completed a respectable thesis problem and hav-ing acquired a reputation in graduate school as an excel-lent sounding board and scientific consultant, T. accepteda postdoctoral position with a leading scientist at a first-rate government laboratory. There, he was offered andbegan to work on a computational research project thatfirst involved arriving at a numerically practical mathemati-cal formulation of a problem, and then a considerable


computer programming effort. As the months passed, andwith the necessity on the horizon of finding a permanentjob, T. absorbed himself totally in his very challengingwork. Whereas, in graduate school, under little time pres-sure, he would have spent a few hours each week visitinglabs and contributing to projects other than his own, as apostdoc, T. became utterly single-minded.

Working 12 hours a day and more, he managed to com-plete his computer program soon enough to be able to runtest calculations. The results were promising, but not espe-cially interesting, which meant that neither T. nor his audi-ences found his job seminar particularly exciting. Besides,since he had not spared time to meet and consult with sci-entists at his lab, his only strong recommendation wasfrom his postdoctoral adviser. The lab itself was unwillingto promote T. to a permanent position, which it sometimesdid, because he had not made himself useful, or evenknown, to a spectrum of its staff members.

On the outside, his job offers were a cut below what histhesis adviser had expected for him. In the competition forthe best positions, T. did not persuade potential employersthat he would ever derive useful results from his postdoc-toral project, even while T. believed that he would be ableto produce them within six months to a year. Other jobcandidates, whose postdoctoral work had been far lessambitious, but had resulted in two or three finished proj-ects, appeared much more impressive. Moreover they hadobtained excellent recommendations from the experimen-tal colleagues whose data they had analyzed.

7

Peter J. Feibelman

On the whole, it is hard to blame potential employersfor their view of T. To them he was "a pig in a poke," anunknown quantity. His thesis work might just have beendone by his thesis adviser, and his postdoctoral project-though in principle a worthy one-was incomplete. WouldT. be able to complete projects on his own? Is he a self-starter? The information simply was not there in the eyesof the interviewers.

To some extent T.'s fate was the "fault" of his adviser.Assigning a long-term project to a postdoctoral researcherwho has to go on the job market in 18 months is a clearrisk to the postdoc's future. But, had T. been as reflectiveabout his career as he was in carrying out his research, hehimself would have realized the dangerous path he wastaking. As exciting as his assigned project seemed, hewould have recognized that his postdoctoral years werethe wrong time for such a large effort. At the very leasthe would have reserved time each day or week toestablish contact with other researchers at the lab, andinvolved himself in one or two short-term projects witha clear chance for success. Many a graduate studentor postdoc spends time trying to understand what his ad-viser wants and getting it done. In fact, it is the young sci-entists who define and carry out what they want, wholearn to be scientific leaders, who find the best jobs andhave the most productive and satisfying careers. Makingyour thesis or postdoctoral adviser happy is sensible, andworth doing, but not more so than acting in your own bestinterests!

8


Know Thyself:A Sweet]ob Turns Sour

B. obtained a Ph.D. from a top-flight university in the Mid-west. He had two different thesis advisers during thecourse of his four years as a graduate student. The first wasa Nobel Prize winner, a theoretician whose name is ahousehold word to chemists. The second was an experi-mentalist, also a very widely respected scientist. Havingcompleted his degree, and cognizant of the scarcity of"real jobs," he accepted a "permanent" position at a majorlaboratory instead of a postdoctoral, temporary slot. It didnot take him long to realize that this apparently wonderfulopportunity was a trap. On arrival at his new location, B.was presented with two opportunities. A senior staff mem-ber, who was involved in a major experiment, suggestedthat B. begin his tenure by working in his lab. That way,B.'s knowledge of the experimental aspects of his fieldwould deepen, and after a couple of years he would bemuch better prepared to work on his own. Objectively,one would say that this was a wonderful opportunity, ef-fectivelya postdoctoral job, but at a regular staff salary andwith a reasonable approximation to regular staff job secu-rity. B.'s alternative option was to begin independent workimmediately. Talking to his younger colleagues, he heardthat, in the eyes of management, a full staff member wassupposed to run his own research program, and that at theannual performance review, if he was perceived to be

9 Peter J. Feibe1man

working as someone else's "assistant," his rating, salary,and job security would suffer, perhaps irretrievably.

One does not have to be a rocket scientist, as they say,to appreciate that B.'s two-year stint as a graduate studentin experimental physics was inadequate preparation forhim to perform at the level of his supposed "peers." Nev-ertheless, unmentored, B. was not willing to risk his all-too-sweet regular staff position by choosing the trainingthat he badly needed. This was a mistake. After three yearsof buying equipment and setting up a lab, B. had still notestablished a research program, and indeed had little ideaof what he wanted to accomplish as a scientist. Thus, de-spite its investment in his laboratory equipment, and de-spite his formally very impressive pedigree, B.'s employermoved him out of basic research. In an environmentwhere goals were clearly defined from above, he eventu-ally matured into a real contributor, and is reasonablyhappy. On the other hand, he is not doing basic researchany more, and he went through several very stressful yearsas a result of his bad start. Like many whose scientific ca-reers flounder, his failure at work coincided with thebreakup of his marriage.

The Ph.D. Technician1. spent two postdoctoral years at a prestigious lab, switch-ing into a new field. He had been hired as a postdoc therebecause of the technical know-how he had acquired viahis thesis work. As a postdoctoral scientist, his task was to


build a piece of equipment combining technology in hisnew area with that of his thesis work. The lab where hedid his stint as a postdoc was well-enough satisfied withhim. At the end of his two years, the desired instrumentwas in place, and L. had his name on a couple of publica-tions with his postdoctoral adviser. Of course it was recog-nized that 1. had not really learned the basics of his newfield, and so his postdoctoral employer did not offer him apermanent position.

A more aggressive or aware young man might havespent a significant fraction of his two years not simplybuilding the desired instrument, but rather asking ques-tions about the direction of his new field, reading aswidely as possible in its literature, and formulating a re-search direction of his own. 1. did not, however, and evenat the end of his postdoc, no one had told him, nor did herealize that becoming an expert in a field and having anexciting research program is an essential aspect of being ascientist. L. did manage to land a "permanent" job after hispostdoc. But as in B.'s case, permanency was an illusion.

In his new job, 1. again built an instrument. But henever participated as an equal member in the group thathired him. At seminars or in planning research proposalshe had little to contribute. When he went before his man-ager to explain what his research plans were, he could sayno more than that he planned to look around for "interest-ing" problems. L.'s employer was happy to possess thenew instrument that he had built and got running. But itwas not long before L. was moved from the research divi-sion of his company.

11

Peter J. Feibelman

Some will argue that 1. just wasn't suited for research,that his fate was predetermined by his personality. Thismay be the truth. On the other hand, I have the lingeringfeeling that if 1. had been appropriately mentored at somepoint during his decade of higher education and as a post-doctoral researcher, he would have succeeded in the ca-reer for which he had trained, or perhaps would haveswitched earlier to a more appropriate field of specializa-tion. It remains to be seen how well he will perform in hisnew job.

Institutionalized ConflictManagers make many mistakes. More often than not thesehurt the people they manage, rather than themselves. Con-sider F.'s experience as a postdoc in R.'s lab. R. had beenhired after a two-year postdoctoral position, but had thewit to appreciate that his "permanent" position would onlyreally be permanent if he proved himself a capable scien-tist in his first two or three years. He invested his first yearbuilding a lab around a major piece of equipment, andwas ready to begin to do science when F. appeared at histhreshold. F. had hired on to work on a project thatseemed rather exciting to its managerial proponents, buthad failed to get the hoped-for, and necessary, externalfunding. The result was that management had to findsomething else for F. to do, and had decided that since histraining was similar to R.'s, F. would be a postdoc in R.'slab. The results were inevitable. Being a clever youngman, F. realized that his future depended on gaining rec-

12


ognition for a significant piece of work, work that wouldhave to be done in short order. R., no less clever, under-stood that his probationary position required him to com-plete several projects and get credit for them. The resultwas not a happy collaboration, but months of bickeringover who would turn knobs on the machine and whowould get credit for the scientific progress. Despite its re-sponsibility for a bad situation, management did not like tohear the resultant whining, from either side. F. ultimatelywon credit for most of the work done in R.'s lab, with theresult that R., whose competence was felt to be more tech-nical than scientific, was moved out of research. But man-agement's distaste for F.'s complaining far exceeded itspleasure in his scientific achievements. F. was not consid-ered as a candidate to replace the hapless R. He did even-tually find another position in science, though, and I hopehe will succeed.

Post-rnortem: Successful collaboration is possible whenone or both contributors have established reputations, orwhen each researcher brings a different, identifiable skillto the collaborative project, for example when a theoristand an experimentalist work together. Collaboration doesnot work, as a rule, for two young competitors. Neither F.nor R. was mature enough to realize that F.'s postdoc wasa predictable nightmare, an arrangement that should havebeen refused by both of them.

If F. and R. had found or had been assigned appropriatementors, early on, they might have been able to deal withthe competitive relationship that had been imposed onthem. If management had explained to F. at the outset that


R. was to be "the boss," and had discussed with both howcredit for results was to be allocated, then F. could havemade an informed decision on whether to work in R.'s lab,and he would have had little reason to complain later on.However, on their own, F. and R. spent a miserable year anda half together, and R.'s scientific career is just a memory.

Impressing Mom and Dad:1Vhose Life Is It Anyway?

A common theme in the minds of young scientists is im-pressing Mom and Dad. This strong motivation is to becherished, of course, but only if it does not overwhelmone's ability to make rational decisions. H. is the eldestdaughter of a successful professor of microbiology. Havingobtained a Ph.D. in an area of limited interest to em-ployers, she decided to switch fields, hoping that hertechnical expertise would enable her to establish a niche.However, she decided to carry out this (wise) move as anassistant professor at a prestigious university (a question-able choice, at best).

A major factor in this decision was that she wanted toshow her father that she could succeed in the academicworld, just as he had. Had she thought her choice through,H. would have realized that when her dad was startingout, research funding was expanding dramatically, makingthe odds of success much better, and also that her next fiveyears were going to be a major struggle, a period whenany desires for a personal life would have to be sup-pressed. Between coming up to speed in her new field,

14

A r n.n. IS NOT ENOUGH

fulfilling her teaching assignments, writing proposals andbuilding equipment-all essential before any research re-sults could be produced-H. found herself spending 16-hour days in her office, the classroom, and her lab. Yes,she did receive tenure after five years. So in that sense shedid succeed. But in that time, she had no life outside of herwork, and by the time she was done, her marriage had dis-integrated. Did this impress Dad?

In a national or industrial lab, H.'s plan would havebeen much easier to realize. With no teaching assign-ments, no committee meetings, no insistent students at thedoor, wanting their grades explained, she could havemade her name working eight or maybe ten hours perday. After five years of building a lab and producing sci-ence, she would have had little difficulty landing a tenuredjob at an excellent university. In the meantime she wouldhave had time for her family-maybe even time to havethe child she wanted. She would have been earning 30 to60 percent more, she would have had better job security,and time for reading a novel or taking a vacation. Thingsare working out for H. now, but she paid what I see as ahigh price for the romantic notion that she needed tomove directly into academia to win her dad's approval.

Get a MentorI certainly hope that reading this book will help you recog-nize what is in your own interest. But no author can beexpected to foresee your own special pitfalls. The bestpreparation you can make toward the goal of having a sci-


entific career is to find yourself a "research aunt or uncle,"someone with little or no authority over you, who hasenough experience to act as a sounding board and giverof accurate advice. Do not be shy about getting to knowpeople outside your advisor's realm. The scientists at yourlab will very likely cherish the human contact. They spendlots of time behind the closed doors of lab and office, andeverybody likes to give advice.

2Important Choices:

AThesis Adviser;a Postdoctoral Job

As a young graduate student, I selected a thesis adviser onthe bases of his prominence in the world of physics, andhis reputation as one who would not require me to spendtoo much time in graduate school. As with other aspects ofmy early career, I now see these criteria as reasonable butinsufficient.

A Prominent Scientist as a Thesis AdviserChoosing a prominent thesis adviser makes a lot of sense,but not because brilliance is transferable. It is not, as Ihave witnessed more than once. Trying to be "another"Linus Pauling, Roald Hoffmann, James Watson, or P. W.Anderson is a common road to failure. What a prominentadviser can offer is: 1) being part of the "old-boy network"(he can help you survive if times are tough, sometimeseven if you don't deserve to), and 2) not competing withyou.

17

18


Point 1 is self-evident upon a moment's thought. Point 2is not so obvious to the naive. A young adviser, trying tomake his way in the world, does not want to be shown upby his student or postdoc. He has a lot to prove, and istherefore unlikely to be generous with credit for ideas orprogress. An adviser who has made his mark views the ac-complishments of his students with pride, even joy. Theyare his research "children." Thus, other things being equal,an established (tenured) professor is a superior choice foran adviser. This recommendation is a simple corollary ofthe way universities are organized. It is not an indictmentof young professors to recognize that they are likely toview their own scientific survival as more important thanthat of their students

A more senior adviser also offers you better prospects offinishing the thesis project that you start and of spendingyour entire graduate career at one university. Many assis-tant professors fail to win promotion to tenure. If this hap-pens to your adviser, he or she will either have to move toanother university or may drop out of academic scienceentirely. In either case you will face unwanted, difficultchoices: whether or not to move with your adviser, orwhom to choose as a new one, whether to select a newdissertation topic or to try to find another professor whois willing and able to help you proceed in your initialdirection.

Although a senior professor may also move to anotherjob while you are a student, the probability is lower, notonly because he or she holds a tenured position, but alsobecause the bother involved in moving an established,large group is substantial. If your senior professorial ad-


viser does decide to move, the consequences for your the-sis project are unlikely to be dire. A senior scientist relo-cates by choice, usually because the funding situation inthe new location, or perhaps other aspects of scientific life,are better. Moving with your adviser is thus likely to beboth financially possible and scientifically desirable. If youdo decide to move with him or her, the delay in your pro-gress toward a Ph.D. should be minimal.

Obviously, an older professor has a better chance of be-coming seriously ill or dying while you are a student. Oth-erwise, the chances of a senior scientist's dropping out ofresearch entirely are rather remote.

TENURE AND PROMINENCE ARE NOT ENOUGH: Althoughsigning on as the student of an established scientist hasmany clear advantages, choosing a good adviser is not aseasy as finding out who has won the most importantprizes, gives the most invited talks, or brings in the largestresearch grants. Does the established person you are con-sidering make himself available to his students? Does hegive real guidance? Is he comfortable talking to peoplewho are not his scientific peers, i.e., beginners like your-self? Does his group have a sense of purpose? Do they in-teract with each other? And does Professor Eminent teach"survival skills"? These are important questions. Making amistake in choosing your adviser can mean years of frus-tration. If you can learn the answers to the important ques-tions in advance, by talking to his current or formerstudents, you may save yourself a lot of grief.

Do HIS OR HER STUDENTS SEE THE BIG PICTURE? Prof. E.was obsessive. He was obnoxious. I have heard it said thathe didn't know quantum mechanics. But his contributions

20


to materials science were manifold-and his students havedone wonderfully well. They knew what they wanted tolearn, and they learned from each other. Thus, even if E.was often away consulting at industrial labs, his studentsthrived.

How do you find out in advance whether the group youare thinking of joining will be like E.'s? Visit the members.Ask them what they are doing. See if they can explain "thebig picture." If they cannot, find a different adviser.

Often a prominent scientist will lead a big group. Hemay have 15 or 20 experimental systems that permit him toemploy an equal number of graduate students, to study"trends." These students are guaranteed to finish their de-grees in a reasonable period of time. In total contrast to myown graduate student experience, they are assigned veryspecific problems. They take their data, report their results,and get their degrees. It all seems so easy. Should you bepart of this kind of group? Again the issue is whether thestudents have an inkling of "the big picture." Is it only theadviser who knows what trend is being studied, while stu-dent A is looking at rhodium, B has a sample of ruthe-nium, and C has some palladium? If the students cannottell a good story, move on!

Choosing a Postdoctoral PositionHow should you be rational about the choice of a postdoc-toral position? It is essential to understand what your inter-ests are, and how they differ from your employer's. Tobegin, you should realize that what you actually achieved

21

Peter]. Feibelman

in your thesis is not especially important to your postdoc-toral adviser. If you are one of the few whose thesis repre-sents a major breakthrough, you will probably be much indemand, and you will likely have few problems finding apermanent job. Probably you won't ever have a postdoc-toral position. Your problem may be that you will spendthe next several years trying to show that your initial tri-umph was not a fluke. This kind of thinking has paralyzedmore than a few young "geniuses," but is not an importantconsideration for the majority, for whom this chapter iswritten.

If your thesis, as is more likely, has not attracted muchinterest, despite your worries you will probably find apostdoctoral slot. Employers generally feel that a postdoc-toral employee is not a big risk. A postdoc will be gone intwo to four years. If he is helpful and productive, so muchthe better. If he launches a successful career as a result ofhis postdoctoral research, that would be wonderful. But ifnot, it would be viewed as unfortunate, but not unusual,and not disastrous. For the typical employer, a postdoc ischeap labor. Unlike a graduate student, who has to beshown the ropes and whose education may absorb somuch time that his net contribution to the progress of aproject may be slight, or negative, a postdoc is a trainedresearcher. He can be expected to be reasonably compe-tent, and not to be terribly demanding of his supervisor'stime. Simultaneously he is "cheap."

At the laboratory where I work, and this is not un-usual, a postdoctoral employee receives minimal bene-fits. The lab pays his medical insurance, but makes no

22


contributions to a pension plan. His paid vacation is onlytwo weeks per year and his salary is not "loaded" withsubstantial overhead or "indirect costs." Acquiring a post-doc is therefore much like buying a piece of laboratoryequipment. One assumes that it will work for a while,helping to produce results. Then it will be replaced with anewer model.

Thus, from the postdoctoral employer's point of view,the most important qualifications of the postdoc candidateare 1) that he should present his thesis research well- thisimplies that he will be a good spokesman for his supervi-sor's research program, 2) that he should not have taken 'overly. long to finish his Ph.D.-after all, he will be em-ployed as a postdoc for only a few years, after which timeone hopes he will have produced a paper or two per year,and 3) that he should be interactive-he should make theresearch group livelier.

If the postdoc candidate wants to change fields, that isnot a problem. Indeed this is a very common practice. Ifhe did not produce a major piece of new knowledge in histhesis, that is not either, because, in general, he is hiredfundamentally to further his supervisor's research program.If he breaks new ground, does something important dur-ing his postdoctoral period, he may be offered a perma-nent job. If not, he will go away, and not much will havebeen lost. This is the employer's perspective. What shouldyours be?

You have three important tasks in your postdoctoralyears: You must decide in what area of science to makeyour name. You must finish at least one significant project.


And you must establish your identity in the research com-munity sufficiently to land an assistant professorship or ajunior position in an industrial or government laboratory.You have little time to waste, because it will not be longafter you begin your postdoctoral work that you will beback on the job market.

These considerations imply that: 1) you do not want aposition where your field of research is undefined. Youwant to get to work on a significant research project on ar-rival or shortly thereafter. 2) You do not want a position inwhich a complex technique is being perfected (whichmeans that your chance of producing results in time foryour job-hunt is minimal). You want to be involved in oneor several short-term projects.

If you are changing fields, you want to start your read-ing and learning before you arrive at your postdoc site.The clock starts ticking when you get to your new loca-tion. Whatever you do before you leave the nest of gradu-ate school doesn't count, for all practical purposes.

Generally it would be wise to find a mature scientist fora postdoctoral supervisor rather than a relative novice. Themain reason is the same as for a thesis professor. You donot want to be in competition with your postdoctoral ad-viser. If there is only one apparatus in his laboratory, if hiscomputer budget is relatively thin, do you think he willlet you use it as much as you need to? If he has only sixyears to prove himself as an assistant professor, will he becapable of recognizing the importance of your achievingrecognition after only a year or so? Very likely not, logicdictates. Unless you can find an assistant professor or

24


junior industrial researcher who is a "superstar," or at thevery least, unless you can satisfy yourself that the youngscientist you want to work with understands, and agrees toaccommodate your needs, you would probably be betteroff working with someone established.

KEYS TO SUCCESS AS A POSTOOC: Once you do take apostdoctoral position, the keys to success are 1) finishsomething, and 2) make yourself known and useful. Yourfirst priority as a postdoc is to have something to talkabout when you go job hunting. No employer wants tohire a person who starts but cannot finish projects. Even ifyou have put a year and a half into developing a verypromising method, you will lose out in the job market toyour competitor whose methods may be less adventurous,but who has produced a kernel of new knowledge, whohas written it up and published it.

I do not recommend that you be careless in your re-search endeavors. Nevertheless, you should be aware thatit is possible and may be desirable to publish an interest-ing result before the last i's are dotted and t's are crossed.It is possible, and relatively risk-free, if you are honest inyour manuscript about the work that remains to be done.It may be desirable because someone who has an interest-ing story to tell, even if it is only supported by admittedlyplausible evidence, will win out in the job market oversomeone whose very thorough effort is not far enoughalong to allow conclusions to be drawn. Although atten-tion to detail is important, and publishing results that laterturn out to be incorrect is anything but desirable, finishingprojects and having a story to tell are essential. As a post-

25

Peter J. Feibe1man

doc, under time pressure, you may have to sacrifice yourdesire for perfection, you may have to live with the fearthat you haven't got everything just right, in order to de-velop a story that you can use to sell yourself. This is notcynicism but realism, and is worth remembering for yourentire career. The famous physicist Wolfgang Pauli is re-membered for complaining ironically that the work of ayoung colleague "isn't even wrong." Think about that!

Do NOT BE A SLAVE TO YOUR POSTDOCTORAL ADVISER: Ifyou just sit in your office working while you are a post-doc, your supervisor will know you, but no one else will.You will get one good recommendation letter, assumingyou have performed well, and that is all. If you chose athesis adviser with good connections, he or she may stillbe able to help you find a "real job" after your postdoc.But what you accomplished as a graduate student does notcount for much in later life, unless it is very exceptional. Ifyour thesis adviser helps you find a job via his or her con-nections, it may be looked on as being despite your per-formance as a postdoc, and the burden on you to proveyourself in a junior, "continuing" position may be greaterthan otherwise.

What you really want to achieve as a postdoctoral re-searcher is that three or four staff members where youwork will think highly of you and will write you good rec-ommendations. If you are a theorist, plan on spendingtwo or three hours weekly talking to experimentalists, andvice versa. Barge into peoples' labs, politely, and find outwhat kind of work is going on. Discover whether there areother research programs to which you can contribute. Get


copies of your lab's preprints. Read them, and if you havecriticisms, questions, or contributions, make them known.Every lab is eager to employ and to recommend interactivepeople. If you are congenitally shy, you have a real prob-lem, one that it would be helpful to overcome. In any case,try to focus on the idea that positive feedback from thepeople you help will help you psychologically, and therecognition that their positive comments to others will ad-vance your career.

Above all, during your postdoc years, work hard! Youhave only a short time to prove yourself. Do not slough offnow. There is no time to waste. Your postdoctoral yearsrepresent the most intensely important period in determin-ing whether you will have a career.

3GivingTalks

Tourist to New Yorkpasserby: "How do you get to Carnegie Hall?"Passerby to tourist: "Practice, practice, practice!"

On a job interview trip, your task is to persuade a signifi-cant fraction of the professionals who see you that theywould be excited to have you as a colleague. The seminarthat you present is your best opportunity to convey themessage that you are the person to hire. The same applieswhen you report on your progress after a year or two in anew position. The colleagues who know you best may al-ready think very highly of you. But they have only a few"votes." By giving a good seminar you can add to the baseof support that you will need to be kept on or promoted.Finally, remember that few professional scientists havemuch time for reading. The way that they learn of new andinteresting work, nowadays, is by going to meetings andlistening to seminars. If you present your work well inthese venues you will be much better able to attract a fol-lowing. Having a following is an excellent form of job se-curity.

Because oral presentations will play a vital role in yourcareer advancement, you must take their preparation very

27

28


seriously. Learning from scientists who present their talkswell is a good idea. In this chapter I hope to impart someof the basic concepts.

The Scientist as ShowmanAlthough a seminar is not a theater piece, there are com-mon elements. .As the speaker you are putting on a one-man show. Your listeners are investing an hour of theirvaluable time. Of course they want to learn somethingfrom you, but like theater-goers, they expect to hear agood story, with a beginning, a middle and an end. Theydon't want to squirm when you explain something poorlyor wrongly, when you put up an overhead containing anegregious misspelling or an inkblot, or when the end ofthe hour is approaching and you obviously have a lot leftto tell. Disappoint your listeners at your peril. They mightnot throw tomatoes or rotten eggs, but they might dismissyou, might be unwilling to find out how good a researcheryou really are-just because you put on a bad show.

The IntroductionA fundamental principle, in preparing a talk, is never over-estimate your audience. No matter how grey their beards,no matter how many papers a few might have publishedin your field, those frightening-looking people in the audi-ence want a complete performance. They want you to saywhat is important in the area of interest, particularly ifwhat is important happens to be their own work! Theydon't mind hearing things they already understand-itmakes folks feel good to understand something.

29

Peter J. Feibelman

The opening lines of a talk set the tone, make a "first im-pression." The main impressions that you want to makeare 1) that you know your field, 2) that you are possessedof the scientific curiosity that will make you a valuable col-league, 3) that you enjoy doing research, and 4) that youplan to convey some useful and interesting information.Tell the audience what the "theme" of your presentation is;or tell them that your work was undertaken to resolve aparticular controversy, and why it is an important one; ortell them that you have demonstrated a novel techniquethat permits access to new and useful information.

Do not simply launch into a discussion of the experi-ment or calculations that you did. Establish the context ofyour research to the degree that time will permit, give anoverview of the novel technique, ideas and/or shortcutsyou have employed, and possibly, intimate what the mostimportant conclusions are. C'These measurements, as youwill see, confirm the long-standing, but until now un-proven predictions in Feibelman's early, brilliant paper.")

This done, you can go on to discuss the specifics. If youare giving an hour's talk you will want to expand on yourintroductory remarks before launching into the details ofyour own work. In a ten-minute paper at a large meeting,a one- or two-slide introduction may be enough.

StagecraftBe aware of the importance of your demeanor, particularlyyour air of self-confidence. If you speak almost inaudibly,it will be assumed that you lack confidence in, or do notunderstand what you are saying. If your presentation is too


low-key, you may convey the idea that you are not enthu-siastic about your work, or perhaps about research in gen-eral. Scientists are like terriers, trained to chase down andpick apart reasoning that is not rigorous. If you appearconfident, your presentation is more likely to be acceptedat face value. If not, you can expect to be fielding insistentquestions early on, and may never get to finish your talk.Alternately, you may see people walking out of the semi-nar room. If you are interviewing for a job, that could berather disconcerting.

Time is of the essence when you are giving a talk. Youmust plan your presentations and rehearse them, to ensurethat you will be able to finish before your time is up, or atleast to be sure that you will have conveyed the main ideasby the time the bell rings. You can easily determine inpractice sessions how long it takes you to present an aver-age slide. This will make it easy to fix an upper limit on thenumber of slides to prepare for a given time slot. Person-ally, I can discuss six or at most seven slides in ten min-utes. If I prepare more than that, I know that my talk willbe breathless and that my audience will absorb little. Theymay well respond to a talk too crammed with informationas a "snow job," an attempt to disguise the flaws in yourwork by overwhelming your listeners with words andfigures. Designing a "modular" talk is a good idea. Afteryour introductory "module," you present several completeinformation packages in sequence. That way, if you seethat your time is running low, you can excuse yourselffor leaving out the last module and skip ahead to yoursummary,

31

Peter J. Feibelman

Don't Try Their PatienceOne of the first lessons students learn about giving a talk isto "prepare an outline." Many of them are also apparentlytaught to begin with a slide that gives "an outline of mytalk." I often find these slides a waste of time, if not down-right silly, and would like to dwell here on the structure ofa talk, not just to help you, but in the hope that I will haveto sit through fewer outline slides in the future.

Have you read a novel recently, or seen a play thatstarted with an "outline of the plot"? When a political can-didate gives a speech, does he put his outline on a chart?Of course not, and in general, neither should you. You cer-tainly should outline your presentation in the privacy ofyour office. But in giving your talk, you should just tell astory. Its structure should be "organic," invisible. Your lis-teners should be propelled from idea to idea with thesame sense of inevitability they feel on hearing a Bachfugue.

At meetings of the American Physical Society (largemeetings), contributed papers are allotted ten minutes plustwo for questions and discussion. Thus, I can present sixor at most seven slides in such a talk. What message do 1convey if slide 1 is "The title of my talk," and "these are thenames of my collaborators, and I want to thank the De-partment of Energy for my funding," and then slide 2 con-tinues with: "I will begin my talk with a brief introduction.Then I'll discuss our experimental apparatus. Followingthat, I'll present my results for system X, and finally, I'llend with some conclusions." All right, this is something of


an exaggeration, but it is not an enormous one. What itconveys is that "I don't have much to say, so I'll throwaway most of my time telling you how I planned my talkand who my friends are, leaving little time for any discus-sion of what I have learned." If you have nothing to say,you would be better off not giving a talk. If you do opt tospeak, you do yourself an injustice not using virtually allyour time to present your ideas and results.

One of the wonderful abilities people have is to takein different information with their eyes and ears, simul-taneously. If you have collaborators not announced ascoauthors, and a funding agency, do acknowledge themon your title slide (Fig. 1), but do not waste time readingtheir names. Someday, when you are a professor, and aretrying to place your students, then you can mention theirnames and good qualities (usually at the end of your semi-nar). Now, however, you are the person you are trying tosell. Acknowledging your co-workers is important, butshould not be overdone.

What you want to convey in your introduction, whileyour title slide is on the screen, is what got you interestedin the material you are about to present, or perhaps whyresearchers in your field are interested, or why the com-munity as a whole should pay attention. What you actuallysay should be geared not just to the subject of your work,but also to the nature of your audience. Clearly, if you aregiving a ten-minute presentation to experts in your field,you should dispense with remarks of too general and in-troductory a nature. On the other hand, if you are giving acolloquium, whose audience includes professionals who


Stimulative Properties ofElixir X

I. M. Balding

SUPERVISOR:

Prof. A. Barber

ADDITIONAL HELP FROM FELLOW POSTDOCS:

Sam Son (dendrite growth)D. Lila (cutting tools)

FUNDING:

Nat'l Hair Council

Figure 1

are expert in areas other than yours, and students, then along introduction is essential.

Attention to the technical aspects of talk preparationcan make the difference between a good seminar and anexcellent one.


Technical MattersExperimental solid state physicists always seem to show aslide featuring a schematic, or God help us, a photographof their apparatus. Occasionally there is good reason forsuch a slide. More often than not, it is a waste of time. HGetto the ideas!" I think in these cases. In putting together thebody of your talk, try to recognize digressions for whatthey are. If there is a good reason for showing an equip-ment slide, if it explains a novel technique, then do it. Ifthe measurement method is standard, if the slide onlyproves that your lab isn't empty, that you didn't make upyour "results," forget it. Nobody minds a short, informativetalk. Don't pad your presentation by design, or by inatten-tion to preparation.

Theoretical physicists, particularly inexperienced ones,often show slides covered with equations. (Molecular bi-ologists show DNA sequences.) Except in very specialcases, such as meetings of specialists devoted to technicaladvances, this is a bad idea. The audience cannot assimi-late more than a small amount of information in an hour-to say nothing of ten minutes. A talk comprised of detailedtechnical slides is likely to be received as a deliberate at-tempt to persuade the listeners that since the material be-ing presented is so complex as to be incomprehensible, itshould be looked on as important. Save this for after yourNobel prize. Then, most of your audience will be afraid toreveal that they have no clue as to what you have done,and/or that they despise your snow job. For now, youneed to please your audience, not to beat them into sub-

35

Peter J. Feibelman

rmssion. Put yourself in the place of an experimentalistamong your listeners. Why would he want to hire you?There is an outside chance he would act in your favor be-cause a colleague who actually understood your equationstold him that they are important. More likely he would pre-fer someone he thought he could talk to. To communicatewith him, you need to convey not the details of your mathbut the basic concepts, the approximations, the results,and the predictions. Think about that. Then throwawaythat slide full of superscripts and subscripts.

SLIDES: A few ideas on the preparation of slides arecertainly in order. When I see a beautifully computer-prepared, full-color slide, what first goes through mymind is "this guy obviously doesn't have enough to do."Granted, modern technology makes the preparation of"professional-looking" overhead foils and slides relativelyeasy. Nevertheless, you do not want to give the impressionthat thinking about how your slides look is more importantto you than what they say. If you are preparing a presenta-tion for a group of laymen--e.g., upper management or ageneral, or some such-by all means make your visualmaterial spiffy. If you are talking to your fellow scientists,go easy on the "professional" look. Remember that manyof them drive 25-year-old VW bugs, and the same reversesnobbism that keeps them in their clunkers probably alsoaffects their impression of your foils.

This, I hasten to add, does not mean that your overheadfoils should be illegible, smeared, or overcrowded. For themost part, they should contain a figure or two, a "cartoon"and simple text. Write big! This has two advantages. One is

36 A r n.r. IS NOT ENOUGH

that people in the back of the room, close enough to thedoor that they can escape inconspicuously, can read whatyou've written, and might be persuaded to stay. The otheris that it limits the amount of material that you can fit on apage. You don't want a lot-so writing big helps.

Using permanent ink in preparing foils is a good idea.Water-soluble ink will ball up if you sneeze or spill yourcoffee on it before your talk. "Permanent" ink can easily beerased from an overhead foil using vodka, or any similarsolvent.

SummaryBy now I hope you have realized that this chapter is or-ganized as a seminar on seminars, and I would like to reit-erate the main ideas:

1. Your seminar is a performance. It needs to be care-fully planned and thoroughly rehearsed.

2. Present yourself confidently. Act as though you haveenjoyed doing your research and that your results are ex-citing to you.

3. Respect your audience. They are spending an hour tohear you. They want to understand what you have to say,even if your specialty is not theirs. They do not want to be"snowed," nor do they want to be treated as IIexperts" in afield where they really are not.

4. Do not waste your time with filler. Make sure eachslide pushes your story forward. If your talk is a bit tooshort, no one will object.


5. Make your visual aids pleasing to the eye without toomuch of a Madison Avenue look.

Thanks for your attention!

ADDITIONAL READING

J. C. Garland, "Advice to Beginning Physics Speakers," PhysicsToday 44,42(991).

Vernon Booth, Communicating in Science: Writing andSpeaeing, Cambridge University Press, New York, 1985.

4Writing Papers:

Publishing WithoutPerishing

The negative connotation of the cliche, publish or perish,is seriously misplaced. Publication is a key component ofyour research efforts. It is widely accepted that a scientificendeavor is not complete until it has been written up. Theexercise of putting your reasoning down on paper will fre-quently lead you to refine your thoughts, to detect flaws inyour arguments, and perhaps to realize that your work haswider significance than you had originally imagined. Pub-lication also has strategic significance. As a beginning sci-entist, not only are your hours long and your pay low, butyour job security is anything but assured. To succeed, youmust make your talents well known and widely appreci-ated. Publishing provides you with an important way toaccomplish that. Your papers, available in libraries aroundthe world, represent not only your product but also yourresume. Compelling, thoughtful, well-written articles aretimeless advertisements for yourself. You can imagine thata sloppy resume is not worth preparing. A premature or

39

40


slapdash publication is far worse. It will remain in the li-brary indefinitely. These thoughts raise the two basic ques-tions that are addressed in the present chapter: whenshould one write a paper, and how should one write it?

TimingGenerally, papers that are written too soon are written inresponse to the fear that one's competitors will publishfirst, or as a result of intellectual laziness, i.e., inattention toimportant details. When they are written too late it is be-cause of the fear of publishing a blunder, or because of"writer's block." Overcoming these fears and frailties isnecessary for everyone in science. At the very least, theknowledge that they are not yours alone may help youdeal with them. (Read Carl Djerassi's novel Cantor's Di-lemma) Penguin Books, New York, 1991, for a poignantexposition of the problem of when and what to publish.)

Planning your research as a series of relatively short,complete projects (cf. Chapter 8) is the best way to achievea disciplined publication schedule, one that serves your in-terests in scientific priority, self-advertisement, and job se-curity. Even though you are working toward an importantlong-term goal, you report each project as an independentpiece of work that has produced a new kernel of knowl-edge (only half-jokingly a "publon," a quantum of publica-tion'). In the introduction to each paper of a series, you

1. The concept of the "publon" emerged from the graduate studentminds of M. J. Weber, now at the University of Virginia, and W.Eckhart, now at the Salk Institute.

41

Peter J. Feibelman

place the work reported in the context of the long-termgoal, to which you thereby lay claim, and you explain howthe present results take you a step closer. If your projectturns out to be as significant as you had hoped, after youhave published several papers in the series, no doubt youwill be asked to write a review. This will provide you withan appropriate forum for a long, definitive article, one thatwill be widely referred to and will help to make yourname in science.

There are many advantages to writing up your work asa series of short papers. Managers and funding agenciesneed concrete evidence that they have hired personneland spent money wisely. Nothing is more helpful in thisregard than the list of publications that their wisdom hasfostered. Of course they will be pleased if you eventuallyrealize a long-term research goal. However "funding cy-cles" are typically two or three years (cf. Chapter 7) and re-newal of junior scientific positions occurs on a similar timescale. Therefore deans, research directors, and contractmanagers cannot wait for your long-term dreams to cometrue. They need published evidence of your progress onan ongoing basis.

By writing numerous, relatively short articles, you cankeep your name in the spotlight. The title and authorshipof each of your papers is printed in Current Contents. Theabstract appears in abstract journals and computer databases. The number of citations of a long publication list in-creases more rapidly than that of a short list. You mustn'tbe overly cynical about these facts of scientific life, ofcourse. If you attempt to achieve name-recognition bypadding your publication list with repetitive papers, your

42


efforts will soon reap scorn rather than admiration. On theother hand, the little admiration you gain for publishing anawesome magnum opus in one paper is surely not worththe risk that this publication strategy poses to your jobsecurity.

If you publish frequently, you are less likely to be"scooped." The longer you hold back reporting your re-sults, particularly if they are important, the greater thechance that some other group will beat you into print. Youdo need to develop an appreciation for when a piece ofwork is complete enough to be written up. If the logic of amanuscript is clearly missing an important piece of confir-matory evidence, submitting it to a journal is likely tocause you endless, painful interactions with referees. Thisis the time to hold back. (Among other problems, the refer-ees may very well be your competitors. Their own publica-tion strategy is likely to be affected by their appreciation ofwhere your incomplete work stands.) On the other hand,if you have completed a project, the sooner you get it intothe hands of a journal, the better the chances are that youwill get credit for your accomplishment.

Writing a paper that presents one new idea or result ismuch easier than writing a long, complex article. This is areasonable way to address the problem of writer's block.Much of the introduction to a shorter paper can be pre-pared, at least mentally, when the long-term research proj-ect is originally proposed. The organization of a paper issimpler if there is not so much material to present, and it isalso relatively easy to explain the conclusions in that case.

Referees are generally busy people, and prefer to review

43

Peter J. Feibelman

short papers. You are likely to receive a more thoughtfuland positive report on a short manuscript than on a longone. Shorter papers are of course not only easier on refer-ees. They can be read and assimilated more easily by thescientific community at large.

Writing up individual kernels of new research alsoshould have some appeal for the perfectionist. It is easierto get everything right when one is dealing with a smallproject than in publishing the results of a major, complexeffort.

Eventually, of course, all the significant details of a re-search project need to be reported in an archival journal,so that others may repeat and confirm the validity of thenew science. Writing such technical papers is an importantexercise, and one that will win you credit from your peersif you do it well. On the other hand, in most cases the writ-ing of such papers can be carried out at leisure.

Writing Compelling PapersA journal article should present a careful and relativelycomplete account of your research. However, it is all tooeasy to write an accurate description of your work thatattracts no attention and that adds little to your scientificreputation, even when your results are significant. Learn-ing to write articles that people will read and rememberwill make you a more effective scientist. It will also en-hance your chances for survival as a researcher.

The structure of a newspaper article is a good mod-el to follow in preparing a scientific publication. Your


colleagues, like newspaper readers, rarely have muchtime for acquiring new information from the literature.Many do no more than skim some journals of interest andhope to come back to the interesting papers later. This isjust the reason that newspaper articles present a story sev-eral times, in increasing levels of detail. The headlines,equivalent to the titles of your articles, are what sell thenews. They are designed to attract attention by providing asuccinct description of what is noteworthy. Scientists at-tempting to deal with "information overload" often turn forhelp to Current Contents, a weekly report of the majorjournals' tables of contents. The titles of your papers, andnothing else of them, will appear there. This in itself is agood reason to spend time writing concise, accurate, andcompelling titles. ("Cute" should be avoided, as a rule.)

The abstract of a paper corresponds to the first para-graph of a news article. It summarizes the main informa-tion, what the important results are and what methods youused to obtain them. Numerous journals place a word limit(e.g., 75 words) on the abstract. It is a good idea to imposesuch a limit on yourself whether or not the journal does.An abstract that is brief and to the point has a betterchance of being read. A wordy one, that reads like the in-troduction to or the body of a paper will lose readers. Asin the case of titles, it is worth remembering that abstractsare circulated more widely, via abstract journals and incomputerized data bases, than the articles they summarize.A well-written abstract may thus attract additional readersto your article-readers who do not ordinarily skimthrough the journal where it is published.

45

Peter J. Feibelman

The introduction to a paper is where you tell your story,possibly illustrating the text with an important figure orsome key results, but without going into great detail. Hereis where you want to explain why your project was an im-portant one to undertake and how your results make a dif-ference to the way we understand the world. Many busyscientists read only the introduction and conclusion sec-tions of papers, leaving the technical details for anothertime. Therefore it is a good idea to highlight your results,for example by placing your most important figure in theintroduction. Even if your readers never take the time toplow through the complete description of your work inthe body of your paper, they may think enough of the in-formation in your introduction to make sure that theycatch your talk at the next scientific meeting.

Virtually everyone finds that writing the introduction toa paper is the most difficult task. It is easy to report theprocedures you followed and to describe the data you ob-tained. The hard part of paper-writing is drawing thereader in. My solution to this problem is to start thinkingabout the first paragraph of an article when I begin a proj-ect rather than when I complete it. I would not embark ona scientific effort if I didn't think that it was important andthat my work would answer a question of rather wide in-terest. The reasons that I found the project in question in-teresting enough to work on provide half the material Ineed for my introduction. The remainder is a summary ofmy key results. The decision to start writing a paper is gen-erally based on recognizing that a kernel of knowledgehas been produced. In my introduction I want to let my


reader know what this new information is, in a nutshell,and why it is worth reading about. Sitting at the wordprocessor, I imagine that I am on the phone with a scientistfriend whom I haven't spoken to in some time. He asks mewhat I have been doing recently. I write down my imag-ined response. If, when you try this, you feel an attack ofwriter's block coming on, tum on a tape recorder and ac-tually call a friend. It works.

Incidentally, if you know why you have carried out ascientific project and what makes your results interesting,there is no reason that your paper should start with an in-ane cliche like: "Recently there has been a resurgence ofinterest in ... (whatever the topic), H which bothers meevery time I see it. If you have been working on a projectfor several months or a year solely because other peopleare interested in it, you have a lot to learn about problemselection. (In this case, see Chapter 8 for some help. Donot pass go. Do not collect your next paycheck.) Beforeyou start on a research effort you must understand why itis important, and in the introduction to your publication onthe subject this is just what you need to explain.

In writing your introduction, as well as the body of yourpaper, it is extremely important that you place your workin context, not only by explaining what you did and why,but also by citing the relevant literature. This is important,not only to provide your readers with a way of under-standing your area of research, but also because your sci-entific colleagues are very eager to get credit for theirachievements. (This is not just vanity. Scientists' careers arebuilt on the perceived importance or usefulness of their re-

47

Peter J. Feibelman

search results.) You have much to gain and little to lose byscrupulously citing your competitors' work. I said abovethat many busy scientists read only the introduction andconclusion sections of papers. Even more turn directlyfrom the title and abstract to the references, to see if theirwork is cited. If someone's work is not but should be, yourisk losing potential friends or at least some respect.

In revising and editing your article before submitting it,you should constantly be asking yourself if you have dealtwith all the "loose ends" in your logic. Are there argumentsthat you have thought and used but not written into yourtext? Are you wishy-washy about inferences that you drawinstead of forceful because there are missing links in thelogic? If so, you either need to work a little longer beforewriting your paper or you need to be forthright aboutwhat are assumptions and what has actually been proven.Even if the referee does not catch the weak points of yourmanuscript, you must not forget that your paper will be onthe library shelves for a long time. Intellectual honesty isaccordingly a very good policy. This is not to say that youshould be such a perfectionist that you never feel comfort-able saying that a project is done and ready to be pub-lished, but rather that you should own up, in print, to whatyou think might be weak links in your reasoning. This is aservice to the community, in that it points to further re-search directions. It shows the world that you are athoughtful and forthright individual. Finally, and impor-tantly, it provides you an "out" if your reasoning is latershown to be incorrect.

The format of the body of a paper is often dictated by


the journal where it will be submitted. Within the journal'sconstraints, however, the key to organizing your work is tomake your text read like a story. Often it is a good idea torelegate detailed discussion of a technical aspect of thework to an appendix. That way experts or interested par-ties can try to understand your arguments in full detailwhile others do not have to guess how much of the text toskip to get back to the flow of the story.

Keep in mind that the function of a paper is to commu-nicate, not simply to indicate how wonderful your resultsare. In principle a paper should provide enough informa-tion that an interested reader would be able to reproduceyour work. It is your responsibility to ensure that the nec-essary information is made available, at the same time asyou try to make your paper as "snappy" and readable asyou can.

Snappy PapersIn archaic times, say 20 years ago, you generally had towrite your papers as though the work had actually beendone by someone else. You were discouraged from usingthe personal pronoun "I" in favor of "we" Of, even worse,"one." Journals seemed to require writing papers in thepassive mood, as in "the data were obtained using the fol-lowing novel method" rather than "I developed the follow-ing novel method to obtain the data." More recently, it hasbecome possible to drop the phoniness of this style and toreveal in your papers that you actually did the work thatyou are reporting. 1greatly prefer the more straightforwardstyle and recommend that you use it.

49

Peter J. Feibelman

People of a mathematical bent often connect the sen-tences in their papers with words like: now, then, thus,however, therefore, whence, hence, and so forth. If youwant your text to be readable to the nonpedantic, youshould be very sparing in using them. Go over your firstdraft, and challenge yourself to see how many of theseconnectives you can remove without undermining thelogic of your argument.

In this era of fast desktop computers and laser printersthere are few excuses for omitting evocative figures from apaper. A picture may be worth more than a thousandwords in a scientific article, if the thousand words are notread, while the thoughtfully prepared figure is examinedand the information it reports absorbed. This does meanthat it is important not to prepare figures that are too clut-tered. If they offend the eye, they may be ignored alongwith the thousands of words.

Some journals restrict the length of articles. This typi-cally forces one to go back through the first draft of amanuscript to rewrite more economically. In preparing afirst draft of a paper, it is a good idea to be as generous aspossible with words. You should write down everythingthat comes to mind as relevant. This may not be easy, buthelps get all the logic on paper. (Again, get out the taperecorder if you tend to be stingy with words.) If you havewritten a copious text, the exercise of cutting back may bemore difficult, but is less likely to lead to a paper whoseflow is compromised by something important's being miss-ing. I recommend the approach of writing generously andthen editing severely in all cases, Le., whether or not thejournal in question imposes restrictions on the length of


manuscripts. The exercise of rewriting as concisely as pos-sible leads to more readable text and thus, to text that isread more widely.

As in the preparation of a seminar, the last section of apaper should provide not just a summary of the results re-ported but also some idea of how they might affect thedirection of future research. The goal of the conclusionssection is to leave your reader thinking about how yourwork affects his own research plans. Good science opensnew doors.

RefereesFinally, since arguments with journal referees can takemany months to settle, and can be very frustrating, it is agood idea to forestall such arguments by having yourmanuscripts reviewed locally, by one or two of your col-leagues, before submission. If you have chosen your localreviewer well, you may discover the weak points in yourmanuscript in a matter of days rather than months. If En-glish is not your mother tongue (and if you are writing foran English-language journal) it is even more important tohave your paper reviewed and edited by a colleague, onewhose English is perfect. Your readers, including yourjournal's referees, are human, and thus impatient to somedegree. The easier you can make their task, the better willbe their response to your efforts.

Incidentally, as one who referees many papers, I muchprefer receiving a cogent, well-written manuscript that Ican learn from than the other kind. A paper that I enjoy

51 Peter]. Feibelrnan

reading disposes me favorably toward the author. Yourreferee may be your paper's most careful reader ever. Mak-ing a good impression on this anonymous potential em-ployer is not a bad idea!

If your referee does have serious complaints about yourarticle, getting angry is not a productive response. A betteridea is to consider why this thoughtful expert did not fol-low your argument and agree with it. If on reflection youbelieve that your results are correct, and that the refereehas simply misunderstood them, it is likely that spendingsome time revising your text will not only persuade thereferee to recommend that your paper be published butwill also ultimately make your ideas less confusing to yourjournal's general readership.

ADDITIONAL READING

Sylvester P. Carter, Writingfor Your Peers--The PrimaryJour-nal Paper, Praeger, New York, 1987.

Michael Alley, The Craft of Scientific Writing, Prentice-Hall,Englewood Cliffs, N], 1987.

Vernon Booth, Communicating in Science: Writing andSpeaking, Cambridge University Press, New York, 1985.

5From Here to Tenure:

Choosing a Career Path

As a scientist, your goals are to make exciting discoveries,to change the way your colleagues and maybe even thepublic-at-large view the world, and generally to improvepeople's lives. However, need I remind you, you will re-main a human being, with human needs, even while youare pushing back the frontiers of ignorance. No matterhow romantically you view your role in research, you willnot be happy without a secure, well-paid job. You willwant help in accomplishing your research goals and rec-ognition for your achievements. You will probably want tosee your family on a regular basis, and more generally, tohave enough free time to engage in activities outside yourprofessional life.

It is all too easy to lock yourself into a situation whereone or more of such basic desires will not be satisfied. Thismay adversely affect your productivity, your family life, andyour ability to enjoy yourself. Thus it is important to con-sider rationally, and in advance, not only the benefits anddisadvantages of the various kinds of scientific positions-

53


academic, industrial, and governmental-but also the mer-its of the different roads to permanent employment.

Economic conditions may limit your choices, but if youare fortunate enough to have more than one job possibil-ity, this exercise will save you considerable stress. It mayhave a significant effect on your financial well-being. Itmay save your marriage. I harbor a secret hope: If enoughof you start to act rationally, the system may eventually berationalized.

It is only natural to adopt as role models the people oneencounters in one's formative years. For this reason many,perhaps most of us, finish graduate school dreaming of anacademic career. For some the academic life may be ideal.For many it is not. Even if being a professor is the rightgoal, however, it is far from clear that rising up the aca-demic ladder is the most desirable way to get there. Myrecommendations, and the reasons for them, are the sub-ject of what follows.

The Pluses and Minuses ofa Job in AcademiaThe idea that a university is an "ivory tower" is common-place. The academic freedom embodied in the granting oftenure was originally supposed to protect the professoriatfrom political repercussions against expressions of minor-ity views of the world. However, tenure is in itself auniquely desirable and economically significant benefit.

Who wouldn't want the ultimate in job security? As aprofessor, if you fulfill minimal performance requirements(e.g., teaching a class every semester) and maintain at least

55

Peter J. Feibelman

minimal moral standards (love affairs with your studentsare sometimes frowned upon), and if your universitydoesn't shut down your department entirely in response tosevere economic stress, you have a guaranteed paycheck.In fact, universities have long since recognized the eco-nomic significance of tenure. University salaries would cer-tainly have to be higher if professors were subject to beinglaid off.

Tenure is a form of financial independence, and thusconveys corollary benefits. A university professor is free tochoose whatever research projects may interest him orher. There is no "boss" at a university to define researchprojects or to decide who will work with whom. In princi-ple, a professor can do research at his or her own pace.If energetic and ambitious, an established professor, to-gether with a group of students and postdocs, may pro-duce a dozen publications a year, or more. If "scholarly,"he or she may publish many fewer and may not havemuch of a group. The department chairman or the deanmay complain about a tenured "scholar's" lack of produc-tivity, but the scholarly professor will still receive a pay-check.

Although tenure and its corollaries are the unique bene-fits of a professorship, they are far from the only attractivefeatures of the job. A professor can anticipate the respectnot only of class after class of students, who pay a greatdeal of money to be exposed to what he or she has to say,but also of the community at large.

Typically professors are free to sell their services as aconsultant, perhaps one day per week, to supplement

56 A Ph. D. IS NOT ENOUGH

their salary. Many science professors found private compa-nies to develop the fruits of their research and sell them fortheir own profit. Others write textbooks on university timeand pay, and then are allowed to reap the royalties forthemselves.

Since classes are held only nine months of the year, theremaining three are in principle a very long annual vaca-tion or at worst unprogrammed time. Sabbaticals are typi-cally part of a university contract. Every several years,professors can look forward to six months or a year at adistant and often an exc

a ph.d. is not enough a guide to survival in science

Documents

addisonwesley books

science peter j

young adviser

waysthat young scientists

enough4 writing papers

special discounts

cover design

research efforts