I've *realized that a greatly overlooked topic around these parts is the panic felt by the **N-th year graduate student who is finally sighting down the barrel on the thesis defense date. This poor individual is in high gear, finishing experiments, making papers out of data, thinking s/he'd better finally write a dissertation and, finally, facing the biggest hurdle of graduate education- getting all of the committee members in the same room at the same time.
And then somebody throws out a casual mentoring thought: "You know, most postdoctoral training slots are arranged a year in advance".
Ok, Ok. First of all, calm down. While it may be true that many postdoctoral slots are arranged with significant (many months to a year) lead time, there are plenty that emerge within a month or two of the defense. Sometimes the newly minted PhD will actually stay in the same lab for a few months to a year (as a postdoc) while seeking the perfect opportunity. So there is no reason to think that you have missed your opportunity to find a good postdoc slot, no matter where you are relative to defense date. So stop panicking already!
Do not decline a job which has not been offered to you yet! This is no doubt an old and very popular bit of advice but in certain of the drug-abuse research superfamilies this is one of "Joe Brady's Rules" (see p. 12). For grad students searching for postdoctoral opportunities, this means if you have an inkling you would like to train with someone, apply. Do not hesitate because you think you are not good enough, do not come from a famous enough lab or any other reason. Apply.
Selecting Labs to Approach:
This takes a little bit of thought, but hopefully not much.
The Science is teh Hawt: Perhaps obviously, the bottom line is to think about what really interests you scientifically. What papers have you been reading in your past N years of graduate school that really excite you? What thing have you wanted to do as a grad student that just had to be put on the back burner because your current lab couldn't do those experiments? Take some time to go back to PubMed and run all your usual searches, this time thinking about the people and labs that are currently involved. Perhaps even venture over to CRISP and perform some of the same keyword searches. If you haven't bothered to retain the information about where all these labs are located, do so now. Check the most recent papers carefully because scientists do move about, at every level of seniority.
Geography: I advise people to fully credit geographic location in their choices. You may be there for five or six years, find a spouse and/or start a family. These are better bets than is the one that you will spend two-three years as a science acolyte and then get an assistant professor job in Geographical Paradise. It had better be a damn unique scientific opportunity to suffer in a place that you hate. Nevertheless, make sure it is not just reflexive bigotry driving your choices...Think about the differences between the red state/blue state distinctions (in the US) versus the county-by-county maps. Many University towns offer distinctive cultures.
Connections: You will also want to approach the PIs that you know in your department, University or even elsewhere. Simply say you are looking and would appreciate any thoughts or advice. The reason you do not do this first, is so that when they ask the inevitable questions about the above issues, you have thought of them already and can look reasonably with it. People like with it. The reason to do this is that older figures in the field know a lot of information that may not be readily apparent to you from the public record. Who has a great training record for particular job types or subfields. Who has just recovered from a disaster of one kind or another and who is facing one. Who is on the job market. Who is about to be hired in a new position. Etc. It adds to your data set. Also, you never know when the person you are asking will say "Um, do you want to come work in my lab"? This happened, very successfully as it turned out, to YHN.
Applying: This is a simple process. Send your CV (with the contact info for three references) and a cover letter which outlines why you want to get training from that lab to the PI. In my view, the most important part of this is not to show your intellectual depth and list off a bunch of novel experiments because you have no idea if the PI is really interested in that area. What you are trying to do is show that you understand what the lab is working on, is capable of working on and what they have worked on in the past. In short, no spam mail letters.
Notice that I got to this before discussing any strategy about funding, competition, publication hawtness, reputation and the like? This goes back to Brady's Rule about jobs- there is no point making strategic contrasts until you have received an offer.
Improving Your Chances of Receiving an Offer:
In rare cases, I suppose, there are super star graduate students in super star labs for whom their own reputation is sufficient to get any and all PIs to pay attention to their application.
Postdoctoral position listings: Wondering why I left this so far down the list? I'm not saying you shouldn't pay attention to job postings on your society email list or web site. These are good things to read over because it may give you some ideas. But do not assume that just because a PI doesn't have a postdoctoral solicitation that you can't approach her. Nevertheless if one of your target PIs does happen to have a job listed, it may give you some clues as to how to get her to pay attention to your cover letter. It can tell you which of the lab's many scientific directions is most urgently in need of a postdoc and has grant money ready to spend on that postdoc's salary.
Networking: This is the biggie and one of the main reasons you need to be talking to your local PIs and postdocs about your job search. It is unlikely, (even if you bother to search out pedigrees on Neurotree.org or similar) that you will have a full appreciation of the personal networks within your subfields. Some of the strongest may be between like-minded folks who hang together at meetings, study section, editorial boards, on blog (!), etc.... but may not have obvious connections otherwise. Some may be distant branches of an academic tree who still have tight interactions with the core scientist(s). The point of this is not because it guarantees an offer in your PI's BestBud's lab (although this can happen). It is to get you and your CV past the usual fog in that PI's brain so that your talents, accomplishments and qualities get considered. This is no small thing, believe me.
This is running way long and I am going to have to continue with how you evaluate postdoctoral offers in another post.
*despite what any specific readers may think, this is motivated by a pastiche of issues that have recently arisen IRL and as DM.
**it is not polite to discuss this number for values greater than 4