Is there an easy way to find free, full-length articles that contain biological contingency analysis?

September 27, 2009 4:52 PM

Looking for a free resource for articles in biology that contain contingency analysis.

I'm a college student and I need to find an article that contains statistical contingency analysis in it to write up a summary on. I have access to JSTOR, EBSCOHost, LexisNexis, and Academic Search Elite, but searching through them is a huge pain. I need *full* articles that contain all of the necessary data of a contingency analysis, but searches result in very general abstracts with little statistical information and that won't cut it for my assignment. I may just be using poor search terms (specifically "contingency analysis" in these various resources), but otherwise I'd have to blindly go through way too many articles, and I'm convinced that I'm not just being lazy. I really need to see the chi-square test statistic, degrees freedom, specific test used, enough information to create a RxC contingency analysis table, etc. I was referred to the free resource PubMed as well, but I had the same experience with that. Any suggestions???

posted by steampowered to Education (5 answers total)

For biology or biomedicine, Pubmed is the way to go. Most articles published in the last 15 years should be online and accessible via the full-text link that appears at the upper right of the listing; this will have a different appearance depending on the publisher. If that's not there, sometimes you can find the article in question by going directly to the journal's web site. If you are accessing the internet from on campus, or if the paper is in a special depository called PubmedCentral, you should have access to the journals' full text (I am assuming you are at a college big enough to purchase a wide range of journals.) Otherwise, hie yourself over to your local medical or science library and buckle down.
posted by monocyte at 5:03 PM on September 27, 2009

Try searching PubMed Central for "contingency table" instead. Don't attempt to search the biological literature with any of the commercial services you mention. PubMed is usually a much better resource. Google Scholar can be helpful as well sometimes.
posted by grouse at 5:10 PM on September 27, 2009

Here's what I'd do. Search on google scholar for 'contingency table' as grouse suggested. You'll get a bunch of articles about how to build one, statistics articles. Choose one or two that give good descriptions or instructions and look at the articles that cite them (there's a little link right under the listing). Amongst them should be biology articles where the researchers have used that technique and are using the original one as a cite for what they've done. Often there's one or two seminal articles describing a technique which are ref'd by lots of people, if you find one that's specific for the type of analysis you're looking for you should find something. If you have access to science citations (I think it's part of web of science now) then it's built for this kind of research, so try that too.

Then to track down the full text, if you don't find it straight from google scholar and can't find it at the journal homepage then use pubmed (as described) and your commercial databases to see if you can access it from there. Also look into your college's library interloan service, often they can get you a pdf in less than a day because they just buy a copy.

Oh and a meta-analysis can also be a good lead since it will talk about the strengths and weaknesses of the trials it analyses, including sometimes a discussion of how much information the authors give in the papers. Traditionally full details of statistical tests are glossed over in journal papers but the advent of online supplementary data has increased what can be added, so supplementary data may be another thing to look out for (you'll get the feel for which journals encourage this kind of thing once you'll looked at a few).

I suggest google scholar because it indexes the full text so I've personally found it the most useful for finding technique-based articles. The other approach is to come up with a list of things you must find mentioned in the text (chi squared table, etc) and search for those. You can often tell by how they come up in the snippet if it's really going to be a worthwhile paper.

However you do this you'll probably have to go through quite a few to find what you need. You just want to narrow it down as much as possible so that you're trawling through a bunch of strong possibles rather than a random selection. Using citation networks can be a powerful way to manage this.
posted by shelleycat at 5:56 PM on September 27, 2009

Response by poster: Thanks, everyone! I feel so silly for overlooking the "full text" links available on pubmed. I've been going to that site for a couple weeks now and have failed to noticed it since then until now!
posted by steampowered at 7:40 PM on September 27, 2009

Pubmed also has some useful tools for limiting your search (see the 'limits' tab under the search box). For this assignment you probably want to exclude reviews since they just summarise findings instead of give statistics. Even more useful though, you can limit it to just results that have links to free full text. Then you're only looking at article summaries where you *know* you can see the pdf right away.
posted by shelleycat at 7:49 PM on September 27, 2009