This post is standing in for a lecture and class discussion that would be happening today if I knew how to be in two places at once. (Welcome Phil. 133 students! Make yourselves at home in the comments, and feel free to use a pseudonym if you'd rather not comment under your real name.)
The topic at hand is the way relationships in research groups influence the kind of science that comes out of those groups, as well as the understanding the members of the group have of what it means to do good science. Our jumping off point is an article by Vivian Weil and Robert Arzbaecher titled "Relationships In Laboratories and Research Communities." 
Arguably, while one can learn quite a lot about scientific theory, problem-solving, and facts from textbooks and coursework, these don't teach you much about what's really involved in being a grown-up scientist*. That knowledge is generally transmitted within research groups, where the scientist-in-training is more or less apprenticed to an experienced scientist, from whom he or she will learn the necessary skills to go out into the world and be an actual scientist.
At least, that's the hope.
While a good number of undergraduates get some experience doing research in laboratories**, it's usually the case that the real apprenticeship starts in graduate school. And there, at the educational junction point where you're trying to make the transition from learning about science to doing science, you join a research group. In some ways, a research group is a microcosm of the larger scientific community. As you're figuring out how to use the instruments without breaking them, how to set up the experiments, how to collect and evaluate data, you're also learning things like how to engage cooperatively with other researchers or how to present a persuasive argument that a particular result is meaningful.
Weil and Arzbaecher think that these mini-communities of scientists have to cope with some of the same issues as do larger scientific communities. They write:
All research groups have an interest in fostering a climate of trust, an atmosphere that supports "responsible" conduct, conduct that meets justified standards of the scientific community. This sense of "responsible" takes in more than the avoidance of falsification, fabrication, and plagiarism, the usual examples of misconduct. It includes some ethical standards that are presumed to apply across research communities, as well as standards specific to the local research community. Each research group should make explicit the underlying assumptions and rationales for their standards, while recognizing that standards evolve over time, as circumstances change and new problems come to light. Ensuring that new researchers who join the group, including graduate students and postdocs, understand the prevailing standards is part of the group's business. 
In other words, one's instruction in the norms of science starts at home -- or at least, in your research group. (Good research groups will end up feeling like home, and in any case, many grad students end up spending many more hours in the lab than they do at home.) The research group is where you are supposed to learn, first hand, how to make good scientific decisions and how good scientists behave. The research group is where you are supposed to learn your scientific "manners" so you can interact with scientists from other research groups without embarrassing yourself or the research group that trained you.
Weil and Arzbaecher suggest that a good way to make sure the transmission of information is effective would be to be explicit in communicating the rules, the standards, the rationales for holding to one set of standards rather than another. Experience suggests that these explicit conversations don't always happen.
Another wrinkle is just how radically norms and practices can vary from research group to research group.
Different researchers have different personalities and different personal styles, both when it comes to attacking scientific problems and when it comes to interacting with other people. To the extent that research requires creativity of a certain sort, this is generally a good thing. But in many graduate programs, students have close contact with just one research group -- the one they're in. Some programs make students do "rotations" through multiple labs before settling down in one of them. In other programs, the "shopping around" for research groups can be rather more superficial (e.g., getting the signatures from N principal investigators to indicate that you spoke to them about their research groups before you're officially allowed to join a group).
Once you've joined a group, the usual move is to hunker down, learning the necessary techniques to get started on your first research project. Both your body and your mind suddenly spend a whole lot of time in this small territory and with this small group of people.
You don't necessarily get much of a view of how things are done in other lab groups. And, you don't always have a good basis for distinguishing which of the things that your advisor or your labmates do are done because they are essential for good science. How do you tell which of the things they do are their own quirky stylistic choices? How do you tell whether some of the things they do might actually be bad practices that their peers in the larger scientific community might reject?
Ideally, the scientist running the research group is making it clear what the rules of the game are, and is making an effort to model good practices rather than bad ones. As Weil and Arzbaecher note:
... the relative autonomy of research groups that permits them to establish their own procedures and standards, imposes on them the burden of figuring out how to incorporate any shared values and live up to accepted standards. 
Given that scientific research groups are interacting with each other on some level -- certainly at professional meetings and in the scientific literature (as well as through formal collaborations) -- they aren't free to completely make up their own rules and standards. They want their scientific work to be taken seriously. They want the scientists they train and send forth into the world to be taken seriously. The "accepted standards" around presenting scientific results (at least within a particular scientific discipline) are reasonably clear, if sometimes hard to articulate. Through various interactions around their own findings and the findings of others, researchers get a good sense of what will persuade their fellow scientist and of what is likely to be challenged. They figure out what sort of evidence they'll need to produce, what kinds of statistical tests will be demanded, what kind of theoretical and experimental worries will need to be addressed.
It's less clear what bounds there are on the "acceptable standards" of interactions between the members of a research group. This isn't the kind of issue repeatedly discussed in the literature or at professional meetings. In a sense, it would be almost like discussing parenting strategies or family dynamics. Some people are comfortable having these conversations, but some P.I.s would think it tremendously rude to stick their noses into how their colleagues run their labs and treat their underlings.
This makes conditions for grad students, postdocs, and lab technicians radically local. There are advisors who prioritize mentoring their students and there are advisors who mostly want their students to produce as much good data as is humanly possible in a given stretch of time. There are advisors who encourage cooperation and teamwork in their lab groups, and there are advisors who pit their students against each other in winner-take-all competitions. There are advisors who are constantly in the lab to see how things are going, and their are advisors locked into their offices writing grant proposals and manuscripts. There are advisors who show great empathy for their trainees, and their are advisors who yell at their trainees in public.
If you only ever see one way of doing things, you might conclude that all scientists do it that way. Even if you manage to discover that things are done differently in other lab groups, by the time you come to this realization, it may be too late to escape the situation in your own research group. It's a very good idea to talk to people at various stages of the pipeline in a research group before you join it! Listen to what they have to say about the experience, but trust your gut if you get a vibe about whether a particular group would be a happy or unhappy place for you to be. Even the best scientific training may not be worth feeling desperately unhappy for 3-7 years. On the other hand, generating data for an advisor who's nice and friendly to you may not be the best call if that advisor doesn't make a point of teaching you the skills you'll need to be a productive independent researcher at the end of your apprenticeship.
It's kind of a big decision.
Weil and Arzbaecher talk a bit about the main types of work research groups do:
Three major goals characterize and help to identify research groups: (1) to get research done; (2) to get students trained; and (3) to acquire the funding needed to achieve the first two goals. 
From group to group, there can be significant variation in how the goals are weighted relative to each other. (Shifts in how important they seem may also come about in response to external factors -- changes in the funding climate, new departmental priorities around effective mentoring, whether the advisor is pre- or post-tenure, etc.) And, given how much there is to do and the different abilities and talents lab members bring to the table, there will be some division of labor in accomplishing the goals -- something that will also vary in its particulars from lab to lab. (On this issue, you should read PhysioProf's discussion of why as a PI he shouldn't spend too much time at the bench.)
As in the larger community of science, research groups have to figure out how to function well together despite the fact that their members have different interests, abilities, personalities, and goals. Becoming a hive-mind (where all the variations between members have been removed somehow) is not the goal:
Participants in a cooperative enterprise must be able to accommodate their personal concerns to the objectives of the group. Sometimes observers write as if the cooperative, collaborative features were all that characterized these groups. In addition, however, research groups value independence in each of the members; it is a trait that the training aims to foster in students.
Moreover, competition pervades, in the broader structures and systems of science and within research groups. 
The research group is about the smallest meaningful unit of "scientific community". Even on this small scale, there's a lot to work out if the members of the group are to balance their individual interests and their obligations to each other -- while simultaneously thinking hard, getting the experiments to work, and so forth.
I agree with Weil and Arzbaecher that more explicit communication about the standards within the research group (and how these relate to more widely shared scientific standards) might make it easier to achieve this balance. And here, I think it's really important for trainees to get in touch with their own interests and communicate those clearly. A scientist-in-training may not fully understand what her interests as a grown-up scientist will be -- that's part of what she's learning from her advisor. However, she may have a better sense of what kind of life she wants to live than her advisor has. Not every trainees wants to grow up to be the spitting image (professionally) of his advisor. Good advisors know how to listen to their advisees rather than picking a career path for them with no consultation.
In their article, Weil and Arzbaecher include a hypothetical example of how a PI might communicate expectations to the research group. It comes in the form of a memo:
To: New & Used Graduate Students in the Laboratory of Last Resorts
From: Director Drake
Subject: General Rules
Welcome to our laboratory. As you know, research in this laboratory is funded by grants from NIH, NSF, and other agencies. The projects so funded have specific aims and a detailed research plan stated in the grant applications. Departure from these aims and plans requires re-application for the grant funds. We would only do this if the original ideas prove early to be without merit.
Therefore, students in the laboratory are not free to pursue ideas and activities of their own design, unless these fit the aims and research plans of the project that supports them. In accepting this fact you are surrendering a significant amount of intellectual freedom. It is important that you understand what you will gain here and what you will give up. Please be certain that the mutual agreement stated below is acceptable to you.
I agree to provide, as long as grant funds are avilable:
- Your tuition.
- A stipend to live on.
- Excellent laboratory facilities, including all necessary computers, instruments, equipment, tools, supplies, and desk space.
- Superior research training.
- Thesis idea and guidance.
- A long-term commitment to your career goals
You agree that, since the Laboratory's highest priority is continued funding, I may:
- Set your daily work schedule.
- Determine your research.
- Personally present your work whenever and wherever I deem it appropriate.
- Decide what and when to publish.
- Decide the authorship and order of names on all publications.
- Determine your readiness for PhD qualifying, preliminary, and final examinations.
- Approve your committee membership.
- Approve any communication you have with other laboratories.
- Have exclusiove ownership of your data -- before and after you leave the laboratory.
- Restrict your lunches to the usual Banana and an occasional Tuna Sandwich.
I'm going to throw it open for discussion whether this memo really embodies transparency in communicating expectations. Should there be a parallel memo from students to PI, and if so, what items might it include?
Are there any items in this memo that are troubling? If so, are there any good strategies for members of the group to renegotiate them with the PI -- or is it a "My lab, my rules" kind of situation?
I'll confess, given the experience I had with the research group I joined in grad school, I would not be clamoring to join the Drake lab. Still, there's something about having clearly articulated expectations that is really appealing...
*Yes, I know I'm way behind in delivering the next installation of this series. Once I get through some grading, it will happen.
**Of course, the set up of this discussion is going to bug the scientists who do fieldwork instead of labwork. To the extent that scientific disciplines centered on fieldwork also involve research groups and training relationships within them, this conversation isn't meant to leave you out. Please do use the comments to point out interesting differences in the interpersonal dynamics and training relationships that you think might be striking features of scientific fields that aren't laboratory based.
 Vivian Weil and Robert Arzbaecher, "Relationships In Laboratories and Research Communities," in Deni Elliott and Judy E. Stern (eds.), Research Ethics. Hanover, NH: University Press of New England (1997), 69-90.
 Weil and Arzbaecher, p. 70.
 Weil and Arzbaecher, p. 71.
 Weil and Arzbaecher, p. 73.
 Weil and Arzbaecher, p. 73.
 Weil and Arzbaecher, pp. 79-80.
I think a letter like that is a great idea. However... you should be getting it before you choose to work in a lab not once you've decided to work there. Me personally, I would never work in a lab that I couldn't do mostly my own work and develop my ideas.
The Weil and Arzbaecher memo is shockingly close to the truth in a lot of cases. But you can't blame a lab for following suite because it's sole existence often depends on the availability of grants. Although scientists are trained to be ethical, objective individuals, businessmen are not. Unless science funds itself without any third party involvement, it can be difficult to rid research of this "tampering". And even if science were able to fund itself, I'd be amazed if nobody grabbed the bull by the horns and operated on their own monetary agenda.
Research groups are a microcosm of a specific science. The lab I belong to is far different than any other in the field of Behavior Analysis. There is nothing else like it. However, no two researchers (or advisors) are alike, so it's difficult to maintain a strict "quality control" standard. But I think that maintaining a standard of research/lab groups could be devastating to a science because it does not allow for creative "outside of the box" thinking that can lead to something extraordinary. You have to find a happy medium, which may lie, for example, in the hands of the peer reviewed journals. But even then, journals can be biased... so I guess we'll never win :-)
I too think that the memo is a great idea and would have liked to see it (or memos sent earlier to the group) before I join the group.
When I joined grad school, I had no idea what is expected of me and what I can expect from the adviser. I figured it out slowly over the years. Adviser's great, but I would have liked better communication between the both of us.
As well as the lab versus field work type of science there's also, in my case, the clinic and the very large dataset crunching setting.
I'm not sure I'd like the Drake lab much either. Point 10 is pretty much just a bullying tactic. Who cares how long a grad student takes for lunch or what they eat as long as they are actually productive and do so in a timely manner? Point 9 seems highly autocratic as well- unfortunately I've seen that it's necessary when you have grad students who aren't productive in a timely manner.
Set your daily work schedule, Approve any communication you have with other laboratories, Have exclusiove ownership of your data -- before and after you leave the laboratory
-- bite me, bite me, and bite me entirely.
We'll decide experimental goals and directions together, and I'll decide how to get the necessary work done (though I will want your help there too); I'll talk to whomever I choose, and you'll trust me not to get you scooped; and we have an equal stake in any data I produce.
Don't like it? Find yourself another sucker. There are plenty of labs that don't treat students and staff as fungible data production units, obedient to the whim of the almighty PI. Yeesh.
I am extremely explicit about explaining both to my trainees and to anyone else who seems to give a flying fuck (1) exactly what my expectations are for participation as a trainee in my laboratory, (2) exactly how I mentor my trainees, and (3) exactly why I mentor that way.
That sample "memo to trainees" is fucking ridiculous, and bears little relationship to an effective lab management plan. If it's not satire, them it sounds like the wack-fantasy of a nutjob control-freak. A laboratory is supposed to be a collaborative scientific environment, not a prison license-plate factory.
Most of it is so absurd as to not require specific comment, but one edict is so hilarious I'd like to highlight it:
As you know, research in this laboratory is funded by grants from NIH, NSF, and other agencies. The projects so funded have specific aims and a detailed research plan stated in the grant applications. Departure from these aims and plans requires re-application for the grant funds. We would only do this if the original ideas prove early to be without merit.
(1) exactly what my expectations are for participation as a trainee in my laboratory, (2) exactly how I mentor my trainees, and (3) exactly why I mentor that way.
More info, please. I'd like a real-world example to counter the control freak bizarro list above. Would make a good front-page post if it turns out too long for a comment.
You might be interested in a book we publish that addresses many of these issues. It's called "Lab Dynamics" and was written by a PI and a psychologist. It addresses managing personnel in a lab both from the point of view of a boss managing students/employees, and from the point of view of a student/postdoc managing their boss and their peers. It uses real-world examples to make points (who gets to be the first author on a paper?) and it avoids the touchy-feely "let's all hug" type of psychobabble one would prefer to avoid.
I'm not sure I'd like the Drake lab much either. Point 10 is pretty much just a bullying tactic. Who cares how long a grad student takes for lunch or what they eat as long as they are actually productive and do so in a timely manner?
That sample "memo to trainees" is fucking ridiculous, and bears little relationship to an effective lab management plan. If it's not satire, them it sounds like the wack-fantasy of a nutjob control-freak.
I'm pretty sure that indeed much of it is tongue-in-cheek. Not the funniest thing ever, but "Restrict your lunches to the usual Banana and an occasional Tuna Sandwich" at least got a brief smile out of me, which is more than anything in the 95 comment discussion of humor accomplished. (Except for the Interrupting Sheep joke, which genuinely is funny.)
David: I might be interested, if you weren't charging $39 for a 176pp paperback. Is there a downloadable version I could actually afford?
JSinger--sorry, no, it's not online anywhere yet. And sorry for the price tag (that's actually $20 cheaper than the original hardcover). As with most things, you're always dealing with an economy of scale. When you have a book that is only going to sell limited numbers of copies (and there are only so many scientists out there, and only a small percentage of them buy books, and a much smaller percentage buy books on a given subject), you end up with higher prices to cover the cost of production. And by "production", I mean overhead, editorial and staff time as much, if not moreso than actual paper, print and binding costs, so putting something online only goes so far as to lower prices. We're certainly experimenting with online versions of our books and hope to do more in the future (our CSH Protocols journal is an attempt to move our manual publishing program online). As part of a not-for-profit research /teaching institute, we're very interested in exploring new avenues, but we don't have the funding of the big publishing corporations that will allow us to take big and expensive risks. So I'll apologize if we're not moving fast enough for some readers, but we'll get there eventually.
+1 to bill; PP please post an actual written document agreement you had for one of you trainees- if you are comfortable with that. It would be enormously useful.
Actually, I'd like to extend that request to any faculty in the audience (or grad students that want to share any letters they got).
As far as the norms of scientific culture, post-doc/PI agreement models are available at the national association website (I can't find the specific link right now, but I know it's there). There's also a helpful generic-type document in the appendix of Macrina's "Scientific Integrity" book.
I think that regulating science culutre via instituting rules is going to be met with extreme resistance (justifiably so)... but from a trainee's standpoint, I could do with a lot more regulation of the scientific culture at my instutition. I think it will be best if it occurs via the common adoption of good ideas- and this is why transparency in determining 'what others use' is so important.
It's pretty obvious to me this is a very important topic (I'd just like to see more details). Thanks, Dr. Free-Ride!!
@bill - many many libraries have this book (see a list here: http://www.worldcat.org/oclc/57366543) if yours doesn't, why not ask them to borrow it for you via interlibrary loan?
PP please post an actual written document agreement you had for one of you trainees- if you are comfortable with that.
We don't use written documents. It is informal, but very explicit, and based on detailed ongoing discussions.
I have to say I'm a little disappointed in you.
If you want to restore my faith- do an honest little experiment. *Without further discussion*, ask your trainees to write down the top 10 things they have agreed to, and you write out what you believe you have communicated about your expectations.
If they match up (mostly), I will be highly impressed. I'd even go so far as to say you have better communication with your trainees than a lot of marriages ;-)
If they don't match up... I guess that's something to think about.
Janet: Thank you very much for posting a lecture's worth of material on the dynamics of effective cultures in scientific research groups. One of my kids has survived 1st year grad school, been accepted into a research group and is about to have that possibly fateful "lets find a thesis topic that works for me and for this laboratory" discussion with the PI.
This post looks like valuable background .
The PI's commitments in that document should include:
1. I'll refrain from sexual harassment.
2. I won't tolerate anyone tampering with another's work, no matter how cute and flirtatious the perpetrator is, in bed or out.
3. I won't invite anyone with an established, multicontinental reputation for being an impossible asshole into the group (and not because there's only room for one of us in the lab).
4. I'll remember that you're here to learn to do good science and to do good science; making me look good is secondary.
It wasn't a research lab, but I worked in a commercial lab for 3 years - and for half of that time I had a boss who was exactly like this hypothetical one. No he didn't write any of this down as a lot of it comes down to harassment and creating a hostile work environment. He would actually stand in the upstairs window to see what time you arrived and berate you if you were a minute late (regardless if you had stayed overtime - for hours - the day before to finish rush work). This was my first job and I put up with it until I could move to another department. It was a great learning experience - as in what NOT to do when you are the boss, to do my homework before accepting another position, and to take out of line criticism with a smile!
My advice is shop for a boss (PI or otherwise) with the same thoroughness as you shop for a spouse. No you don't have to love them - but your personalities better mesh OK. The better you vet them ahead of time (what do the people working for them think?) and get expectations defined (what is expected of you and what you can expect from them) the better off you will be.