Literature searching and maps
Strangely, many people have asked me for my tips and tricks on literature searching recently. So I thought, hey, I’ll write it down and then point people to the blog instead of describing it again.
This process is intended to produce a relatively comprehensive and systematic review of literature pertaining to your topic or question. It’s a good way to master a new area and locate key articles in a relatively short space of time. It replaces (for me) a lot of fumbling around and hoping I’ve got everything. With this, you won’t get everything but you can be pretty confident you haven’t left anything big out.
This is how it works.
Step One: Choose search strings.
Garbage in, garbage out, as they say. So the first thing to do is identify the area you are searching. Recently, I wrote part of a massive and evil literature review on media literacy, intended to cover all electronic media. I created the following initial search strings:
- “media literacy”
- “television literacy”
- “radio literacy”
- “telephone literacy”
- “film literacy”
- “computer literacy”
- “Internet literacy”
Some of them are stupid, of course (telephone literacy?) but then what do I know about literacy? It was a useful first step, as it turned out.
Step Two: Create initial annotated bibliographies.
This is where computers are really cool. The key tool you need here is EndNote, which LSE provides on its computers and at a rottenly expensive ?70 to its students (but the commercial version is hugely expensive, so what can you do?). Also you need a connection to various academic databases, again which LSE provides.
Then you search for your terms in this general order and export them to EndNote (I’ll explain this in a minute):
- Books and book chapters. A fabulous way to do this is to use OCLC’s WorldCat, which looks in academic libraries all over the world to see their book listings and brings them to you in exportable format. Lovely. The really great thing about WorldCat is that you get chapter headings for many of the books! So you can finally find that key article in a collected volume. Whoopee! Bad news: the exporting function seems to be really slow and horrible. I’ve found the best way is to export page by page.
- Articles. I use ISI’s Web of Science as my primary port of call. It indexes everything in the social sciences. But you should also check out the other databases available. Search for your terms, “Mark All” on the list (it’ll do up to 500), go to the “Marked List” page and tick the boxes that say “Abstract” and “Times Cited”. Then export to End Note.
Step Two-A: Exporting to EndNote.
This is cool. Sorry, I just had to say that, because it was a revelation to me when I started to do this. OK: The first thing to do is create a new EndNote file. Call it, for example, media_literacy_ISI. You’ll be creating a lot of EndNote files so a name that includes the search term and the source is helpful. (Also helpful to keep note of what you searched for when in a separate place).
To get the references into EndNote, you first look for the option in your database that says “Export to reference software”. This is on the “marked list” page in ISI and on the search results page in WorldCat. So, press the button. Depending on how your computer is set up, it will either automatically ask you for an EndNote library (point it to the one you just created), or you will have the option to save the results as a file.
If it goes to EndNote automatically, you will be faced with a long list of things called “filters” which makes sure that the bibliography arrives safely, with the author in the author field and not in the title. For WorldCat, choose the “WorldCat” filter from the list; for ISI, choose ISI. Simple.
NOTE: EndNote has lots of filters available, and you can install them (like, I installed a filter for the results from the Digital Dissertations archive). The best place, I’ve found, is the University of Queensland’s library. I love those guys. Also the EndNote site has some, but they are a bit harder to use. It’s pretty difficult to install new filters on the standard LSE configuration because the ‘filters’ folder is locked. If you’re using your own laptop, just save the filter you want to the filters folder and it will show up on the list.
If you saved your results as a file, just choose File/Import, and choose the right filter (as explained above)from the Import Option… list.
Step Three: Print your huge annotated bibliography.
You now probably have several hundred records in your EndNote file. Choose the “Annotated” bibliography format. I print them two pages to a page, double sided. You now have a giant book of abstracts of everything related to your field. It is a horrifying, insuperable amount of information. This is where Step Four comes in.
Step Four: Map your area
I discovered mind-mapping during my MBA, and I swear by it for this particular application. It was developed by Tony Buzan, who wrote a several books about it, and it’s really fairly simple. You basically take the key concepts that you see in your abstracts and link them together visually, one word or phrase per line. Kind of like an outline. So, with “literacy” as my central concept, I had a chain that said “media”, then “purpose”. “Democracy”, “pluralism” and “social justice” were all linked to “purpose”. Here’s an image of the map I made. I also add, after each final key concept (like “democracy”) the author’s name and year in capital letters. That way, when I want to read the article about democracy as the purpose for media literacy, I just look up the author’s name in my handy-dandy EndNote bibliography.
You can see I’ve used a computer to make my map, but I’ve also done them by hand, which Buzan advocates. I used MindManager, which isn’t free. FreeMind, an open-source application that was developed after I bought MindManager also gets great reviews. And it is free. Which is good.
You may want to highlight certain articles. I highlighted the most recent ones, because the review was meant to be a review of recent literature. Another good possibility is highlighting those cited above a certain number of times. If you ticked the “Times Cited” box on your ISI export, this is something you can find out. You can also search ISI to check the citation figures on the books you found in WorldCat, if you’re compulsive. Knock yourself out.
Step Five: Read your articles
You knew it would come to this. This is where the rest of your academic training comes to your aid, and you can put this nasty article away.
December 9th, 2004 at 8:17 am
I’m impressed. This is truly amazing, and hugely helpful. I wish I had this advice 5 years ago. Three minor comments.
First thing I don’t get is: you created the map form the printed abstracts, right? If not, why not use atlas-ti and do it electronically? (another 70 quid, I reckon)
Second thing is: this is exceptionally good for articles, far less for books as things are now. Firstly because have no abstract, and annotation takes far more time (though I don’t know about OCLC, I’ll take a look). Second because they are mostly printed at present. With e-books things may change.
Third, I seem to remember that the US Library of Congress does cited books as well, with titles of chapters. Far easier to access as the filter is great. But I should double-check again. I have made a few ENDNOTE filters myself, if you ned help with one just shout.
Thanks again for the great tips.
Wainer
December 9th, 2004 at 9:52 am
Hi Wainer,
Thanks for the comments!
Regarding comment 1. I know nothing about Atlas-TI - how would it work? But assuming that it would do some kind of electronic parsing, my impression from developing the abstract maps is that, like other academic work, there is quite a bit of judgement involved. For example, you could develop a map that looked only at empirical topics, one that looked at theoretical approaches, one that was a bit of a mix. So I don’t know how an electronic tool would handle that. Also I find it to be pretty quick work - the media literacy map was done in an afternoon once I had everything printed - so I worry whether the electronic setup, tinkering etc would be slower? I’d be very interested to hear the results if you try this.
Comment two: you are right, it’s much better for articles than books. But, if you abstracted the books or had a volume of notes, for example, you could still map them, or use an electronic tool to process them somehow. I think Sebastian used nVivo, for example, to work with his reading notes. I haven’t done this at all.
Comment three: cool, I will check out the library of congress.
December 10th, 2004 at 5:05 pm
Rather than searching Worldcat from the web then exporting you can search it directly from within Endnote using its “connect” feature. Of course the way that Endnote is set up on our machines internally doesn’t work too well but if you go to Edit/Preferences and change “folder locations” so it points to folders you create instead of the standard Endnote folders you can fill those folders with only the Endnote connection and formatting folders you have created. The OCLC WorldCat connection filter the LSE has is broken but has since been fixed and is downloadable here.
I should also mention that Scopus, the new Reed Elsevier search engine offering “access to 14,000 peer-reviewed titles” which we are trialling has some really good features and is worth checking out.
December 10th, 2004 at 5:41 pm
David, the SCOPUS thing is awesome. Excellent filtering, citations right there! Thanks for the tip. Haven’t tried exporting yet.