When you need to find papers on a specific topic, you search the literature. Maybe a Google Scholar search, maybe a search of databases like Web of Science or PubMed, whatever. The point is, you have some idea of what you’re looking for, so you can search for it.
But what if you don’t know what you’re looking for? At least not with sufficient precision to do a search? What if you’re just looking for “papers I’ll find interesting, even if they’re on topics I didn’t realize I’d find interesting”? Or “papers that I should probably read in order to remain broadly-informed about what’s going on in ecology”? Reading for these purposes is very important to me, not just because I enjoy it, but because I’m sure I’d come up with fewer and poorer scientific research ideas if I didn’t do it.*
I’m old, and so I mostly skim the tables of content of a bunch of journals to identify papers I might find interesting. Brian does the same thing. I also periodically check what Google Scholar recommends for me (basically, it’s an automated recommendation engine that recommends papers for you to read, by loosely matching to the topics that you’ve published on yourself). And occasionally I’ll run across a citation that sounds interesting, and go read it. That’s pretty much it. I’m not on social media, so don’t find any papers that way. I don’t read Faculty of 1000 Prime. And while the blogs I read sometimes discuss published papers, they’re almost always either papers I have no desire to read or papers I already knew about, so I don’t find any papers that way (I do find other sorts of reading material via blogs). And conversations and correspondence with colleagues very rarely involves them saying to me “Have you seen this new paper?”, so I don’t find papers that way. My approach works for me. But presumably it wouldn’t work for everyone, which is what this post is about. (Just as an aside, when I need to find papers on a specific topic, I search for them, just like everyone else. But that’s not what this post is about.)
I’m sure my approach misses some stuff I’d be interested in if I found it. But so would any approach. Plus, I lack the time to read as much as I’d like in as much detail as I’d like, so the rate at which I read interesting-looking papers isn’t limited by my ability to find such papers. So I’m not too bothered by the fact that my approach is likely to miss some papers I might find interesting. For instance, I’m only likely to find papers in Plos One or other unselective, high-volume journals if Google Scholar recommends them to me, or if they concern a topic on which I’m searching.
I have the impression that a small but growing number of ecologists not only filter the literature using different methods than me, but with different goals. That increasingly, people are focusing their reading on papers they already knew they wanted to read, thereby enabling them to search for such papers (or set up automated filters to alert them to such papers, which amounts to the same thing as searching). And mostly don’t read anything else. That is, they only read stuff that can be searched for. I’ve even heard people say that the whole idea of “keeping up with the literature” is a quaint, outdated notion, because the literature these days is far too large and growing far too fast for that.
But those are totally anecdotal impressions. In an attempt to get some data more anecdotes, here’s a little two-question survey. Tell us, how do you find papers to read when you’re not looking for papers on a specific topic?
I’ll share the results in a future post.
*Throughout the post, by “read” I don’t necessarily mean “read in great detail”. The closeness with which you read a given paper is orthogonal to whether or not you found the paper by searching for it.
Dynamic Ecology 
Love this post 🙂
I just did your survey and chose ‘Other’ as an option, but there was no way for me to tell you about what ‘Other’ means to me so…
I often find interesting papers to read when I’m searching for something else. For example, when I’m searching topic X and in the results for that I find something not relevant to what I’m looking for (but Google thinks so). Nevertheless I may still find it interesting for reading another time so I add it to my short list (which is actually long) of things to come back to another time.
So I do find interesting things to reading by searching, but not by searching for them directly.
Thanks!
I’m curious about the extent to which your “method” is analogous to me following up citations in papers I’ve found by searching. Obviously the citations in the papers I read are somehow related to the citing paper. Presumably the irrelevant search results you follow up are somehow related to your search terms. But how? Can you give an example?
I guess what I’m really asking is, how is the original search helpful? Presumably the fact that you’ve started by searching on a term of interest somehow constrains or filters the “irrelevant” search results in some way. Does that make any sense?
Yes, the search results that I deem irrelevant for the current search purpose are definately related to the search terms, just not in the way I was expecting or desiring when I originally searched.
So a specific example: by searching ‘climate change insect Australia’ I’m looking for information on the impact of climate change on Australian insects, but there are also results for forest management under climate change scenarios that are not directly about Australian insects, or even Australian forests for that matter (but they do appear in the search as maybe the possibility of insect outbreak is mentioned in passing as a threat to forests or a paucity of studies in Australia might be identified… I could go on… and on, and on!)
So I guess the ‘irrelevant’ results are tangentially related to what I was originally searching but potentially less directly than if I was looking at reference lists or citations (or maybe not depending on the specific citation).
Hey Jeremy,
One targeted (and yet simultaneously general) approach I use is Web of Science Alerts. I enter all the keywords and combinations of key words that I’m interested in and every week or so I get a list of new articles that touch on those subjects in their title, abstract, or key phrases. I get a lot of junk that I don’t read, but I also see tons of stuff that I’d otherwise miss on my areas of interest.
I also use the other methods (table of contents, following citation threads, random chance, etc) but I like the alerts method above because it keeps be up to date on the breadth of literature in particular areas of interest. I get recommendations for articles across all the subdisciplines of ecology, in more specialized journals that I’d normally never read.
Chris
Thanks for the tip Chris.
Your comment illustrates something I perhaps should’ve noted in the post: with sufficiently many, sufficiently broad search terms, searching starts to shade into non-search-based methods.
I’m curious how common it is for folks to use your approach or something close to it.
Rather than keywords, I use WoS alerts for new papers by authors who I want to keep up with, or new citations of some key papers in my sub-field. It’s a bit of a pain to set up and maintain these, though, so my topics are a bit out of date.
Hi Jeremy,
I also chose “other”, but specifically, I use F1000 recommendations. Wash-U has a free subscription to F1000, so I get at least an email a week. I search through those, and choose the most interesting papers.
Yeah, in retrospect I wish I’d included F1000 as a separate option rather than lumping it in with “other”. Although so far, very few people are choosing “other”, which would seem to imply that few respondents use F1000.
I assume you can request F1000 alerts only on certain topics? How narrow are the topics you use? This gets back to an earlier comment on how sufficiently-broad searches shade into non-search-based filtering methods.
I have general search terms like “community ecology”, “genetics” and “thermal ecology”. 90% of the papers I receive are genetics, molecular evolution or adaptation focused, but I still enjoy reading them.
When I find a paper on something I am particularly interested in, I often check out which papers cite them in Web of Science, and those I then sort by number of citations. This makes me find the most influential papers on a general topic that is kind of new to me without using particular keywords.
I typically find semi-related, or just plain interesting, papers through twitter. It’s a lot like perusing through the table of contents through Science or Nature e-mail alerts. I like it because it alerts me to journals I typically wouldn’t read and its a great time management tool; I can scroll through on the train, over lunch, or, as I often prefer, on the toilet.
Presumably your choice of who to follow is pretty key with Twitter? I’d imagine that, if one chose poorly, one could find oneself scrolling through an awfully big haystack looking for needles. Or alternatively that you’d end up with lots of people alerting you to the same few high profile papers, or the same few papers on some topic near and dear to the twitterverse. How do you decide who to follow? Do you find you need to put some effort into optimizing who to follow in order to use Twitter as a filter on the literature? Or does “follow my friends and whoever’s working in my field” seem to do the job? (I’m assuming here that you are mostly following people, not following particular hashtags, right?)
I find that a small set of people commonly tweet papers within their specialty. So I get updates on a few sub-topics (mostly statistical) on Twitter from these people, but they don’t cover the spectrum of what I’m interested. It’s rare to find someone who matches your area of interest AND is an active tweeter of literature.
I enjoy finding interesting papers from the ” by association” method. For example, If I look up Jeremy Fox’s user profile under Google scholar, I see who his co-authors are and then look up one of the co-author’s Google scholar user profile or personal web page. This generally will end up with me staying in a specific field but can sometimes go down a nice rabbit hole of topics I may have never considered.
I pick out potentially interesting new papers (often not directly related to my research subject, so I would not have searched for them) by setting aside 10 min each day to scan through my RSS feed (Feedly) of 200 journals (roughly 50-100 new papers each day). Google Scholar only picks out a few of the ones I find in the RSS, and very few from Google are papers I did not already find in the RSS. Papers of lower potential interest I only bookmark in the feed and return to later to read the abstract; papers of higher potential interest I save directly to my reference manager (Mendeley) in which I later read them and makes notes.
Older papers that I did not know I wanted to read before I found them, I find at blogs, ResearchGate or from the “by association” method described by Kevin Chase above.
Books! Books when I’m thinking about something out of my normal and I’m not sure quite what it is that I’m thinking about and probably someone’s thought about this before if I just knew what it’s called.
That grammar was on purpose, as it describes the mindset when books works best for me. I’ll pull 5 to 10 books off the shelf that sound broadly related to what I think I’m looking for, and then flip through. A good book covers a lot of territory and you can search it without knowing any of the keywords before you start. Publication date is pretty irrelevant. Most of the books get dismissed within pages, but often one or two will feel promising. Those get flipped through with more vigor. Often, I just learn about something I wasn’t looking for but was interesting and became part of my “I’ve heard about that” quiver. Twice in the last 7 years (I’m 32), I came across a technique from another field that was very closely related to what I was looking for.
Hi Jeremy,
Very interesting post!
One of my prefered way is to read on subjects in general wide audience journals (in economy, sociology, philosophy, wired, seed, the Times, etc) and then ‘translate’ the interesting aspects into possible scientific subjects. It takes a bit of gambling to get the keyword right, but it usually does a very good job of leading you to novel topics, definitely connected with my general interests but isolated from what I read as a scientist. I think it also helps think more broadly.
A related one would be how to balance breadth with depth (I am often struggling with my curiosity).
Thanks!
Your way of staying broad is interesting, that never occurred to me. Although now that I think about it, I basically do the same thing, even if I don’t think of it as keeping up with the literature (and even though for me it very rarely leads to concrete ideas for research projects–but reading broadly is useful even you can never trace any specific research idea of your own back to something specific that you read outside your own area of expertise).
Reading broadly is definitely useful for being a blogger, though. 🙂 Lots of my posts are basically me taking stuff I’ve read on economics blogs or in popular science books or whatever and turning it into blog fodder.
Reblogged this on A Knowledgearian's Hat and commented:
I never really thought about how I find interesting reading materials outside a specific search, so it’s great that Jeremy Fox has not only thought about it in detail but also put it out there and has gotten such fabulous detailed responses from others in the science community.
Pingback: Survey results: how do you find papers to read when you can’t do a search? | Dynamic Ecology