Kept all my notebooks; what good are notebooks?

In the discussion on the earlier post about what policies should govern lab notebooks kept by graduate researchers, the commentariat identified a number of important considerations. At least a few of the commenters were sure that a one-size-fits-all policy wouldn’t work, and collectively the comments identified some central questions that go to the heart of how, precisely, lab notebooks are supposed to function:

1. The permanent record of what was done and what was observed.
   To the extent that research is a matter of setting up particular conditions and recording careful observations of what happens in those conditions, the observations are the empirical data from which scientific conclusions will be drawn. Having a precise accounting of these data, and of the experimental details (or in the case of fieldwork, the conditions of observation), is supposed to make it possible for other scientists inspecting this permanent record to get to the best position short of actually being there for the experiment or observation. Such complete records save each individual scientist from having to collect all the data she cares about herself.
   The permanent record of what you observed and how you set up the conditions to observe it is not only useful in attempts to reproduce particular findings, but is also, as some commenters noted, a legal document that can be requested (or subpoenaed) by funding agencies that, among other things, have an interest in ensuring that the research they fund is actually conducted and that important results are communicated to the community of science (via journal articles, for example) and possibly even the broader public.
   Who maintains control of the permanent record?
   If the graduate advisor is the person who secured the funding from granters and who is identified on publications resulting from the research as the author to whom correspondence should be addressed, it seems most likely that the graduate advisor would also be the one pressed by granters, journal editors, or other scientists working in the same area for access to that permanent record. (Graduate students sometimes move around a lot, and by the time a piece of research is communicated in a journal, they may have moved on to a different lab.)
   While it is prudent for the PI to keep the notebooks when they function as this sort of permanent record, this does not obviously rule out the graduate researchers who generated the notebooks keeping photocopies of their contents.
2. The log of how an idea was hatched, pursued, and refined.
   Some lab notebooks give more than just the facts of how the experiment was set up and run and what was measured. Rather, they document more of the thinking that motivates a particular kind of experimental approach and its refinement. What were your hunches? How did you test them? Which hunches did you decide to discard in the face of the data (or the insurmountable experimental challenges) and which ones actually held up? Which interpretations of the data seemed most reasonable and how did you rule the other interpretations out?
   In many (most?) cases, a notebook of this sort will also include the nuts and bolts information about procedures and data that you’d get from #1. The extra component is a record of your thought process at various points in the research.
   Who maintains control of the record of the thought process?
   It’s conceivable that having a record of the thought process driving different stages of the research might be useful in communicating with journal editors or other scientists trying to reproduce a result. To the extent that the graduate advisor is the one most likely to be contacted for such communications, there’s a good argument that the graduate advisor should maintain this record. As well, to the extent that there may be patentable intellectual property that comes out of the research, a documentation of the thought process (dated, of course) can help establish the IP claim. (There is, of course, the question of whether the graduate researcher should get to share in the patent rights; different universities have policies on this.)
   An important issue here may be who came up with the initial idea — the student or the PI. If the PI has handed the student an idea to start with (especially one for which the PI has secured research funds), the PI probably has a reasonable claim to keep “control” of that idea until it’s played out. This wouldn’t mean that the graduate researcher couldn’t also pursue this idea, so long as she properly credited the source of it. In many contexts, there wouldn’t necessarily be a reason the graduate researcher couldn’t keep photocopies of the notebooks (which might make it easier for her to remember which ideas in the research originated in someone else’s head). In the case where the research is funded by industry and aimed at generating proprietary knowledge, however, leaving the lab and the industry funding may mean kissing your rights to the idea goodbye.
   If the student hatched the initial idea, I can imagine the PI might be able to mount a claim to partial ownership of that idea on the basis of providing the facilities and guidance necessary to pursue it. And again, if the PI is the corresponding author on the publications that result from the research, access to the original notebooks may be a practical necessity.
3. The place to capture ideas for future pursuit and refinement.
   This is a variant of #2 where the notebook serves as a memory extension to record ideas that occur to you as you’re doing research. They may be ideas related to your current project, but sometimes they are ideas for completely different projects that you may or may not get to before you leave the graduate lab.
   Who maintains control of the log of ideas to pursue in the future?
   If these are the brainstorms of the graduate researcher, there’s a good argument that she should at least have access to them and permission to pursue them after she has left the graduate lab. If they are brainstorms of the PI (perhaps scrawled in the notebook by the graduate researcher during group meetings or conferences with the PI), there’s good reason to identify them as the PI’s to pursue. If, as is often the case, the brainstorms are the result of the intellectual labor of both graduate researcher and PI, both parties should be credited in further development of these ideas, and ideally both parties would be involved in such development.
   Some graduate researchers maintain a separate “random ideas” notebook that is distinct from the permanent research record so they can take this notebook with them with no hassles. If you do this, it’s worth being excruciatingly clear in the “random ideas” notebook about the contributions of others (including your PI) in sparking and shaping those ideas — so you can give credit where credit is due.
4. The complete set of recipes for getting the experiment to work.
   In my view, every lab notebook (at least in experimental sciences) ought to serve this function. I came to this view having inherited an experimental system from a recent graduate who did not give the full run down on how to make the experiments run in his notebooks (nor in his thesis); this cost me quite a bit of time trying just to get things started in the lab.
   Since research necessarily involves measurements or observations whose outcomes cannot be known ahead of time, there’s no way to know up-front which experimental parameters in your set-up are going to make a difference to the outcome. Recording them all precisely makes it possible for someone else to pick up your notebook and (once his technical skills are sufficient) have a reasonable expectation of being able to reproduce your results.
   This kind of notebook is usually a fuller version of #1 (and of #2, especially to the extent that the ideas whose refinement you’re documenting concern getting the system to behave).
   Who maintains control of the detailed experimental protocols and recipes?
   Once again, if the PI is the one likely to be fielding the calls and emails from people trying to reproduce a particular experiment, it makes sense for the PI to keep these notebooks on file. As well, the PI may well be training new graduate researchers to do the same experiments (and since “training” has come to mean in many cases “passing on the notebooks of the last graduate researcher who worked on that system or with that technique). Again, though, this doesn’t necessarily rule out the graduate researcher keeping photocopies of the recipes.
5. The master index of raw data, analyses, crucial literature, and protocols.
   A bunch of commenters noted that nowadays very little data is actually “collected” by means of ink notations made by a graduate researcher in a bound notebook — you’ve got strip-chart recorders, printouts from scintillation counters, photos of gels, and measurements collected electronically and stored directly to computer files. However, to be of any use in a scientific conversation, each of these sets of data needs to be matched to the conditions used to produce them (and the date on which they were produced, etc.). This means that something like a notebook is vital in keeping track of which data correspond to which measurements. Notebooks do further organizational work in matching up raw data sets with products of data analysis. They can also identify a particular article (or set of pages in someone else’s notebook) as the definitive description of how to set up the experiment or perform the data analysis.
   Who maintains control of the master index?
   My pragmatic answer to this is “the person who maintains all the raw data (strip-charts, scintillation counter printouts, photos of gels, etc.)”. The raw data aren’t of much use without the information about where they came from, and the master index is fairly superfluous without the raw data. If the raw data end up being part of a continuing conversation with (or audit of) the corresponding author of publications that resulted from this research, the corresponding author (usually the graduate advisor) had better have access to the guide by which she can make sense of the boxes and boxes of raw data.
6. The repository of ground already covered.
   This is another version of the lab notebook as memory extension, and is especially valuable in terms of documenting the things that didn’t work, as these are less likely to be described in a journal article.
   Who maintains control of the record of ground already covered?
   Commenters mentioned that unnecessary repetition costs time and money. Certainly it’s useful for a lab group — especially one with a rotating set of members whose efforts have extended over many years — to have the notebook version of an institutional memory, so that its cumulative efforts can be most efficient; this is an argument for the keeping the notebooks in the lab where the research was performed. However, given that it would also be wasteful for the graduate researcher who performed the research to repeat experiments that she can’t remember having conducted, that might be a reason to give her access to (or photocopies of) her graduate notebooks.
   Of course, if we were really serious about avoiding unnecessary duplication of work, that might mean that everyone in the community of science should have access to the notebooks.

That last point gets to the crux of what I think really underlies some of the disagreements about who has a claim to lab notebooks and what sort of claim that is: Are lab notebooks primarily a way of recording the empirical evidence (and the steps involved in getting that evidence) that underwrites scientists’ efforts to understand a piece of the world, or are lab notebooks primarily a way to establish one’s intellectual property?
If scientific research aimed primarily to turn the efforts of the scientific community to figuring out how various bits of the world work, free and full communication among the members of that community would be a top priority (since the community builds the knowledge faster and better if the efforts of its individual members are coordinated) and open access to laboratory notebooks would be widely viewed as a good thing.
However, in the contexts where science is practiced, there seems an inescapable element of “keeping score”. Because positions and funding are dependent on establishing your individual contribution, and because the contributions that count the most are generally arriving at an important or surprising result before anyone else, scientists can be quite protective of their “turf”, hesitant to share all the crucial details of their experimental protocols or the full enumeration of promising experimental directions that ended up being dead ends. When you’re trying to look out for your individual interests, sometimes it ends up being convenient if your competitors waste their time covering old (and fruitless) ground or thrashing about trying to get the experimental system to behave.
And where this makes the control of lab notebooks a sticky subject is when the trainee (and collaborator) completes her training and leaves the lab to become a competitor.
If this post were not already so long, I could launch into an argument that the community of science would be better off ratcheting down the competition and embracing more collaboration. Instead, let me just wonder whether a PI’s awareness of a graduate trainee as a future competitor might not subtly undermine the potential for real collaboration between PI and trainee during her graduate research. Then, a tug-of-war over notebooks might be an indication of much larger issues to work out.