MY WORD

How (not) to get a job...ach fall, at the beginning of thejob season, clusters of postdocs

nucleate in the lab, worrying aboutthe job market, which is probablyworse than ever before, and abouthow to maximize their chances ofsuccess. Each new rumor inducesmore anxiety. There ,are 400applicants for this job. No, 500.Only plant histopathologists have achance this year. Only those whocan teach gross anatomy need apply.After a few hours of this, everyone isupset and the most sensitive feel ill.

The irony of these discussions, ofcourse, is that the postdocs talk toother postdocs, who don't knowhow to get a job either. Surprisinglylittle information flows from thefaculty to those who need it. So hereis one view from the other side ofthe process. Searches are probablydifferent at other institutions, but anyfacts are better than none. Besides,this information might save valuablepostdoc time for experiments.

The application process in science isparticularly idiosyncratic. Not for usthe efficiency of medical residencymatching, where all the choices arefed into a computer March 1, andeverything important is settled tenseconds later. Not for us the intensityof the Modern Language Associationmeetings, where aspiring humanistsare interviewed by prospectivecolleagues in one action-packedweek. Instead, an advertisement goesout, somewhere, the applications flowin, and an achingly slow drama playsout over the succeeding months.

Most places want to hire the bestperson they possibly can. Threethings are important: productivity(measured in number and quality ofpublications), talent (the letters of

recommendation) and potential (theresearch statement). Many applicantsundermine themselves by ignoringone of these areas. Some apply beforetheir papers are written. Some letterssay "I don't know what he's up tothese days", a vaguery that could beavoided by giving the recommendera current CV Some research plansare too dense. The applicant thinks"I need to impress the world's expertin my field". But if the expert isdoing graduate admissions this year,she isn't reading any job applications.

The search committee consists ofeight overcommitted facultymembers from random fields. Eachapplication is read by two of them, soif there are 400 applications,everyone reads 100. Of these, 30-50applications will be read moreseriously, and only six to tenapplicants will be interviewed, so thebiggest cut is based on what that firstfile looks like. Each applicationdeserves hours of thought, but it'slucky to get thirty minutes. Nothinglowers the reader's spirits like a tenpage research statement. Two pages isideal. Everything needs to be at thereader's fingertips. If your papers arein the package, they'll be looked at;if not, the reader's unlikely to searchfor them. The strongest applicantsare those who are easy to appreciate.

One surprise to me was that the bestresearch statements were pitched at adifferent level from most scientificwriting. Unlike grants, which shouldbe conservative, statements allow theapplicant to show her creativity andlong-term goals. Again unlike grants,the applicant need not continue anongoing project. The most popularstatements are the ones that conveyexcitement about a new researchdirection. People who say they'll becompeting with their current lab orother well-established labs do poorly- the search committee thinksthey'll be toasted. A less visible, butmore serious, problem is overlap orperceived overlap with a faculty

member. A few places buildsupergroups with similar interests,but most prefer one of everything.Time spent positioning yourself as aBright New Thing is well spent.

No one - not the applicants, notthe search committee - realizes inadvance how slow the hiring processis. Academic campuses move at aglacial pace. In the first month, theapplications get organized on thefloor of a secretary's office. In thesecond, they go out for first reviews(by this time, the applicants arepanicking because they haven'theard anything). The third month isChristmas vacation. The fourth isspent trying to schedule eightfaculty for the first meeting, wherethey discuss only a fraction of theapplications. The interviewees maynot all be invited until May.

Is this the best process possible? Weall know political operators with fewideas and many papers, or brilliantstudents who mixed with theiradvisors like oil and water. Ideally, theletters, publications and statementsshould correct for each other. Mybte noire is the search committeelemming phenomenon: we must lookat X because Stanford wants her. Ialso have qualms about the weight theletters get. One applicant has lettersfrom Mendel, Darwin and Ramon yCajal, another from Newton,Copernicus and Fermat. Well, Iknow the first guys better... and whenI'm done, they're the only part of thefile I remember. It's human nature,but it encourages intellectual incest.Letting the applicant speak for him orherself, at least in the first reading,might soften the grip of old boy- andgirlism on the field.

And all this trouble only gets theapplicant to the interview... - butwe can deal with that another time.

Cori BargmannDepartment of Anatomy,University of California,San Francisco, USA.

© Current Biology 1995, Vol 5 No 5 451