How do you avoid getting buried by all the references' references' references' references?
October 17, 2010 10:18 AM Subscribe
I have a question about optimum levels of reference and avoiding getting snowed under by an avalanche of references. My question is clumsy because I am unsure what this is even called.
I call it source explosion, but I do not know what anyone else calls it. You are studying a topic. Let's say for illustration it's the Mandelbrot set. You have three or four papers to begin with. They all have references. You get a bunch of the references, which all have different references. Iterate two or three times more. You soon have a stack of a thousand or more pages to plumb and there is no way to read every word. There is a trade-off between being complete and not re-inventing a wheel.
Is there an expert wisdom on how to manage this problem? I have a very good HOW TO on "how to write a research paper" by one of the top guys in my (narrow) field. He does not cover this specific problem which I am having a particularly hard time with right now.
I call it source explosion, but I do not know what anyone else calls it. You are studying a topic. Let's say for illustration it's the Mandelbrot set. You have three or four papers to begin with. They all have references. You get a bunch of the references, which all have different references. Iterate two or three times more. You soon have a stack of a thousand or more pages to plumb and there is no way to read every word. There is a trade-off between being complete and not re-inventing a wheel.
Is there an expert wisdom on how to manage this problem? I have a very good HOW TO on "how to write a research paper" by one of the top guys in my (narrow) field. He does not cover this specific problem which I am having a particularly hard time with right now.
That is an abstract question, so I will give you an abstract answer.
Generally, you need to keep in mind what you're writing about. You're not going to write a treatise; you're not going to cover the entire field. Rather, you're going to write about a relatively small topic that is only a blade of grass in the entire field.
A lot of the references won't be about that blade of grass. You should be able to scan the title, the abstract, or the text of an article and tell whether it has a significant relation to what you're writing about.
If you can't do this, it is likely that you don't have a well-formed idea of what you're writing about. This is a sign that you need to go back to the first step, which is to get a good idea of what you're writing about.
In short, the difficulty you are having is likely a sign that you want to start researching and writing before you have given the topic sufficient thought. Your paper will be much better if you sit down for an hour or two, read a basic text or two on the topic, and start mentally drawing lines that separates what's in your topic from what's out.
posted by Mr. Justice at 10:38 AM on October 17, 2010 [1 favorite]
Generally, you need to keep in mind what you're writing about. You're not going to write a treatise; you're not going to cover the entire field. Rather, you're going to write about a relatively small topic that is only a blade of grass in the entire field.
A lot of the references won't be about that blade of grass. You should be able to scan the title, the abstract, or the text of an article and tell whether it has a significant relation to what you're writing about.
If you can't do this, it is likely that you don't have a well-formed idea of what you're writing about. This is a sign that you need to go back to the first step, which is to get a good idea of what you're writing about.
In short, the difficulty you are having is likely a sign that you want to start researching and writing before you have given the topic sufficient thought. Your paper will be much better if you sit down for an hour or two, read a basic text or two on the topic, and start mentally drawing lines that separates what's in your topic from what's out.
posted by Mr. Justice at 10:38 AM on October 17, 2010 [1 favorite]
When I'm starting out on something like this in a field with which I am unfamiliar, I tend to read a few review articles to get an idea of where to go with it. At this point, I have a an idea who the main players are, their opinions and what the latest concepts are and I can start to form an opinion. At that point, I compose a vague outline use that to focus my reference search. I then start writing - I've been criticized for not having all my references in line before writing, but this works for me - and I fill in the needed references after a first draft of the paper.
When you're reading those first papers, there are only a few relevant references in them to begin with. Let's say out of 20, three are relevant to your topic, so you get them. Some of them are dead ends - they show something specific that you need but their background isn't what you're writing about. Some of them are good and have a lot of stuff - maybe just print out abstracts or save a list of abstracts to a file. Write a one or two sentence summary of what these papers are telling you that is relevant to your topic.
Use a good abstract search tool. I use pubmed, which has things like the ability to only search reviews, or search related papers, or search based on a single gene or disease. In the area where I worked, I wasn't hugely interested in papers older than 5-10 years - this work had usually been already been reviewed somewhere.
posted by sciencegeek at 10:55 AM on October 17, 2010
When you're reading those first papers, there are only a few relevant references in them to begin with. Let's say out of 20, three are relevant to your topic, so you get them. Some of them are dead ends - they show something specific that you need but their background isn't what you're writing about. Some of them are good and have a lot of stuff - maybe just print out abstracts or save a list of abstracts to a file. Write a one or two sentence summary of what these papers are telling you that is relevant to your topic.
Use a good abstract search tool. I use pubmed, which has things like the ability to only search reviews, or search related papers, or search based on a single gene or disease. In the area where I worked, I wasn't hugely interested in papers older than 5-10 years - this work had usually been already been reviewed somewhere.
posted by sciencegeek at 10:55 AM on October 17, 2010
My experience is that even though you start out feeling like you're entering a reference fractal, if your topic is sufficiently refined, you'll find that all of the papers are actually referencing eachother. The set of 30 references in 10th paper you read will have very few unique citations, and those that are will probably be the least relevant to the topic you're trying to write about. I don't know what kind of paper you're trying to write, but I would hazard that if you are not starting to experience the reference mobius strip in the first dozen papers you read, you are either writing about something too broad, or having trouble distinguishing which references are the most important.
On preview, it sounds like I'm saying you only need to read 10 papers. I'm not. What I mean is that in the first 10 papers, you'll start to generate a consensus set of important references.
posted by juliapangolin at 10:55 AM on October 17, 2010 [1 favorite]
On preview, it sounds like I'm saying you only need to read 10 papers. I'm not. What I mean is that in the first 10 papers, you'll start to generate a consensus set of important references.
posted by juliapangolin at 10:55 AM on October 17, 2010 [1 favorite]
If you are a grad student, your advisor plays several roles. One of these is to be a filter for the enormous body of work on your topic that is out there (and which, as you have correctly surmised, you would never ever finish reading all of). There are also literature reviews, survey papers, and other resources that your advisor could point you to, thereby serving as (to coin a phrase) a meta-filter.
But seriously: ask your advisor, ask other researchers in your field that your advisor might point you to as knowing a lot about prior work on your specific question, and ask whatever librarian your institution your institution has designated as the point person for your field.
posted by AkzidenzGrotesk at 10:59 AM on October 17, 2010
But seriously: ask your advisor, ask other researchers in your field that your advisor might point you to as knowing a lot about prior work on your specific question, and ask whatever librarian your institution your institution has designated as the point person for your field.
posted by AkzidenzGrotesk at 10:59 AM on October 17, 2010
This is kind of difficult to answer because I don't know your level of experience or field of study....
Research is time-consuming and the papers add up. Try first to target the papers and books that seem the most relevant and provide the basic info you need to understand the topic. In your stack of 100s of papers, there will many that are only tangentially related to your topic. Read through the abstracts and do a quick sort. As you go along you will be able to recognize the ones that seem relevant and which ones don't more quickly. Are there any papers that are cited frequently? Put these at the top of your list. Have secondary piles of ones that look interesting and keep them on hand to read more thoroughly if you need them.
Then use what you have found in your literature search to craft your argument, and cite the papers that back up your argument. Think about the points that you want to communicate in your research paper. Then use the papers that illustrate your points. Once you've made your main points, you may need to go through your other 2ndary papers to find more support for any finer details within your argument.
As far as having too many references, you should use as many are needed to back up your arguments. Here's one rule of thumb I find useful to weed out sources, as stated by Applied Vegetation Science's Author Guidelines: "A reference that is cited only once, and that time along with several others, may be redundant."
I don't know if this is helpful for you or not! I don't know of any expert wisdom, I've just come to my own methods based on experience, and many late nights of reading and shuffling through papers. Good luck!
posted by feidr2 at 11:15 AM on October 17, 2010
Research is time-consuming and the papers add up. Try first to target the papers and books that seem the most relevant and provide the basic info you need to understand the topic. In your stack of 100s of papers, there will many that are only tangentially related to your topic. Read through the abstracts and do a quick sort. As you go along you will be able to recognize the ones that seem relevant and which ones don't more quickly. Are there any papers that are cited frequently? Put these at the top of your list. Have secondary piles of ones that look interesting and keep them on hand to read more thoroughly if you need them.
Then use what you have found in your literature search to craft your argument, and cite the papers that back up your argument. Think about the points that you want to communicate in your research paper. Then use the papers that illustrate your points. Once you've made your main points, you may need to go through your other 2ndary papers to find more support for any finer details within your argument.
As far as having too many references, you should use as many are needed to back up your arguments. Here's one rule of thumb I find useful to weed out sources, as stated by Applied Vegetation Science's Author Guidelines: "A reference that is cited only once, and that time along with several others, may be redundant."
I don't know if this is helpful for you or not! I don't know of any expert wisdom, I've just come to my own methods based on experience, and many late nights of reading and shuffling through papers. Good luck!
posted by feidr2 at 11:15 AM on October 17, 2010
Glancing at the title is usually enough to assign a priority code (2A-5D) to every paper I plan to read. I then read them in order. I make sure all the 2s and 3s are read. The 4s and 5s I read when I have time.
posted by grouse at 11:20 AM on October 17, 2010
posted by grouse at 11:20 AM on October 17, 2010
You might want to look into citation analysis and citation impact, and/or approach your university's science library with a few articles in hand and ask for help on conducting a citation analysis on them and/or deriving & making sense of a study of citation impact in your chosen area.
posted by johnofjack at 4:22 PM on October 17, 2010
posted by johnofjack at 4:22 PM on October 17, 2010
« Older How to export Outlook PST mail archives to a text... | How to entertain younger visiting relatives? Newer »
This thread is closed to new comments.
posted by proj at 10:33 AM on October 17, 2010