Meg’s recent post on #365papers inspired lots of questions and comments (and other blog posts). It led into questions about what kind of papers, how to read them (skim vs in detail), how to choose them, etc. But it led me to wonder if there was a consensus opinion on the even more basic question of how much time we should be spending on reading papers (and scholarly books such as monographs or others aimed at graduate students and above)?
This is in many ways a core question. It gets at how important is our role in tying our work to the literature and work of others vs just doing new science.
First a few caveats (feel free to skip straight to the poll if you’re not the type who wants all your terms carefully defined first!).
I think this question could easily vary by career stage. And also by which industry one is involved in and perhaps by which subdiscipline of ecology (do theoreticians read more or less) so I’ve asked those questions in the survey. There are a lot of flavors of “established researcher” so I’ve tried to capture a few key dimensions. Just answer the question assuming the established researcher is like your current job (or what you hope your future job will be like).
I wanted to keep this real world (in a perfect world I would spend 30 hours a week reading literature!), so I didn’t want to ask the question as “would like to spend reading”. Probably the best phrasing is something like “how much time would you like to spend reading assuming you have to balance it with all your other responsibilities and time commitments and assuming you are able to allocate your given time optimally?”. But I have left the word “should” in the questions for brevity and simplicity. Should in this case is definitely not a moral judgement or a form of peer pressure. Its intended to get back to the core question of where you say reading literature ranks in your universe of priorities.
I have also left out the questions of what and how you read. Some of you may spend your X hours mostly skimming and some spend the same X hours deeply reading one paper. I’m asking this mostly to get a sense of priority to reading literature vs other tasks so count all forms of reading literature.
So what do you think?
I’m not going to express my own opinions yet to avoid contaminating the survey. I will add my thoughts in a day or two in the comments. In the meantime, here are a few of Jeremy’s thoughts. But I would be really curious to hear other peoples thoughts, elaborations and explanations in the comments.
UPDATE – poll answers are in the comments below
Dynamic Ecology 
I’m really interested in the answers to this, partly as a result of my own realisation last year that I had inadvertantly stopped reading the literature (http://treesinspace.com/2015/09/25/why-i-stopped-reading-the-literature/). I’m making a more concerted effort now but it’s still hard to carve out time among all the other things we’re supposed to make room for.
“Hours per week” is hard for me. I just don’t read that way. I did a bit of math to come up with a mean value for your survey. But I think it’s important to note variance (high) and not just mean!
While I answered I’m not sure how relevant the questions are to the type of reading I do. I probably read 20-30 hours per week. This is almost all scanning or reading sections of papers. The only papers that I now read entirely are the papers that I review (about 10/year maybe). I read a huge amount for my teaching. I teach mostly introductory classes (because quantitative modeling classes aren’t too popular at the upper level!) and I love casting my net broadly to learn more background for these classes. But my research also chases lots of methods, so this requires me to read very broadly in many fields and even into econometrics, epidemiology, machine learning, etc. etc. I’m a serial dilettante. But it’s amazing how many research come up with ad hoc computational solutions to problems that are well known and well studied in other (and even their own) fields. They just don’t read enough.
Hurrah for serial dilattantism! And as a fellow s.d.er, I totally agree about people constantly reinventing the wheel. Drives me a little nuts — it doesn’t take that much effort to put your technique into a search engine and see if someone else has done something similar. But I don’t think it’s laziness, I think it’s lack of awareness about what’s going on in other fields (or even that those fields exist).
Just a quick update on the polls (with 192 respondents, I don’t think the results will change much but please feel free to continue to take the poll – I will post an udpdate) and comment.
For students the answer seems to be about 10 hours/week:
https://docs.google.com/spreadsheets/d/1Dq-8iNhR-aPmHmUQboMIYW1vJoZM2YzelmO8Kz49Oh0/pubchart?oid=105430793&format=image
For permanent positions, the answer seems to be about 5 hours/week:
https://docs.google.com/spreadsheets/d/1Dq-8iNhR-aPmHmUQboMIYW1vJoZM2YzelmO8Kz49Oh0/pubchart?oid=1719319426&format=image
My own thoughts and some cross tabs will come later
OK some cross-tab analysis.
First in terms of expectations*/desires of established researchers there is very little difference by industry they are in (to my surprise). Academic, NGO and Government were all extremely close.
In terms of research vs teaching appointment, no surprise that generally people expect/desire* more reading if their job gives them more time for research (with >50% appointments expecting a full 1.2 hours more than the split research/teaching R1 appointment – i.e. 6.7 vs 5.4 hours). The <40% research (i.e. teaching appointments) had a very fanatical hard core of 10.8% who expected 15-20 hours (compared to only 3% of R1), but other than that group expectations for teaching positions were a tad less than R1.
Where it gets interesting is expectations/desires by career stage. Sticking first with expectations/desires for established researchers … people who are PHD students, postdocs or early career established (0-5 years) all expected 5.6 hours from an established researcher (although the variance in expectation of an established researcher gets higher as you move up from PhD to postdoc to early career). A general theme is that early career people don't know what to expect of themselves or students (highest variance in each case). People whose current career stage is senior researcher (6+ years) however expect a full half hour more of established researcher (6.2 hours/week).
And for expectations for graduate students, people who are currently students, postdocs and early career all expect about 9.8 hours of a graduate student (postdocs slightly less). Senior researchers however expect 2 hours more (20% more) of students at 11.8 hours.
So the net net is nearly everybody thinks graduate students “should”* do about 9.8 hours/week of reading and established researchers should do about 5.5 hours/week (with some adjustments for job allocation to research but not industry). However senior researchers (6+ years permanent position) expect more of both themselves and students. Are they more delusional, manage time better, or have come to value reading more?
Do you think people are actually do this much or are they wishing they could do a bit more than they are? My guess is people are not that far off of what they think they should do. So I’m not sure why there is so much angst around how much people are reading – maybe the same phenomenon as thinking everybody else is working 80 hours a week?
* “expect” (and should and desire) should be taken as the shorthand for how I defined the poll – not some peer pressure expectation. Its just too awkward to write “thinks X should do about Y hours” over and over.
Very interesting!
This makes me wish I’d tracked my time as a grad student, especially first year when I recall reading a *lot*. I recall spending most of my time most working days doing nothing but reading the literature and making notes on the ideas it gave me. So I’m sure I was way over 9.8 hours/week, at least early on as a grad student.
I’m way less than that now and have dropped off a lot in the past year. Again, I’m not tracking my time, but I’m well under the 5-7 hours/week average that the crowd suggests for senior researchers. I need to get back to reading more, I’m starting to get out of touch with where the field is going and I can feel my flow of good new ideas starting to slow. My plan is that working on the book I’ll be writing will force me to get back to reading the literature in a big way.
If the goal is staying in touch or mining for ideas, do you include reading blogs or searching google scholar as reading?
No. I do spend more time reading blogs than I used to, but not science blogs.
One issue I had with the grad student part of the poll is that I think it depends a lot on what year you are in grad school. I think those in their first couple years should do a lot more than those in their last couple years.
But I thinking about this from a yearly perspective. If you’re arguing for 10 hrs/wk, that’s essentially three straight months of reading if you were to bunch it all together — or reading solid for one week out of every month. I think that’s a bit high for anyone. So I think people are being pretty aspirational in their votes. (“I’m answering what I think people expect I should answer.”)
And (no offense intended), I think the idea of thinking about how much time to devote to reading is a bit useless. One should read as much as one needs to to get the job done. If you’re a faster reader, you may spend less time reading than a slower reader. Ditto if papers in your field tend to be shorter than those in other fields. If you only need to read narrowly, you may not need to read for as many hours as someone who needs to read broadly.
Which raises the question of how broadly or narrowly one “needs” to read. And maybe “needs” isn’t even quite the right word here. I have a post gestating on this…
This is interesting. I really want to go back to starting my day with a couple of hours of reading and writing – I think it worked really well, not only for keeping on track with the literature, but also writing regularly. It would fit with the ~10hrs/week as well + it generally fits with the 5 papers/week too (which is another thing that I keep seeing as recommended).