XYZ, subject of more than #### peer-reviewed articles

April 10, 2016 2:13 PM

What's the best way to estimate the number of peer-reviewed articles written about a topic?

I want to know how many peer-reviewed articles have been written about a specific topic within a certain date range.

I have access to a university library system. I can search there, but it's a bit tricky because it will give me articles on the topic as well as show results for articles which have cited other articles on the topic in their references, so I think I would end up double counting a bunch of articles.

Ideas?

Thanks!!

posted by Lutoslawski to Technology (10 answers total)

Best answer: Google Scholar will give you a better estimate.
posted by k8t at 2:30 PM on April 10, 2016 [1 favorite]

Best answer: Limit your search to the abstracts, or even titles if that's still yielding overlap.
posted by teremala at 2:35 PM on April 10, 2016

If it's health-related, try PubMed.
posted by mynameisluka at 2:43 PM on April 10, 2016

Best answer: Your university probably has Web of Science, which would be the standard academic librarian's answer.
posted by aspersioncast at 2:43 PM on April 10, 2016 [5 favorites]

Best answer: Web of Science is the usual answer for these kinds of studies - since it has consistent authority control.

Google Scholar catches a lot of things outside the purview of Web of Science, but there is a lot of non-peer-reviewed material in there, and that seems like an important distinction for you.

You will want to explain how you defined the articles being "about" a topic - WoS has an official taxonomy of subject terms you can refer to which will help you do that.

Depending on the topic - it might be worth it to include couple of sources. Google Scholar and WoS, for example.
posted by pantarei70 at 2:48 PM on April 10, 2016 [1 favorite]

If you are in the biomedical sciences area, a pubmed search based on the MESH subject headings, narrowed by primary subject, would give you a very solid and curable number. A science librarian should be able to help you out quickly.
posted by rockindata at 3:25 PM on April 10, 2016

If you have some programming skills and your search domain is focused on biosciences, you can convert MEDLINE data to a queryable form that makes it easier to eliminate duplicates.
posted by a lungful of dragon at 3:59 PM on April 10, 2016 [1 favorite]

This is really going to depend on the topic. Ask a librarian at your university which databases would be best for your topic, and what search strategies will be most effective at identifying peer-reviewed articles *about* the topic (usually by identifying/searching subject headings), rather than articles that simply have the words that you search somewhere in them.
posted by unknowncommand at 4:10 PM on April 10, 2016 [1 favorite]

Google scholar is not limited to peer reviewed sources. Do not use google scholar if you only want peer-reviewed articles (also, in addition to having non-peer-reviewed articles, google scholar will have books and book chapters, both peer reviewed and not).
posted by If only I had a penguin... at 5:03 PM on April 10, 2016

I am honestly kind of baffled by the Web of Science answers. Maybe this is field-dependent? In the qualitative-friendly social sciences, Web of Science misses so much stuff that you simply can't use it for much meta-analysis or quantitative research about research.

By contrast, yes, Google Scholar picks up extra material (M.S. and PhD theses and book chapters are mostly what I see). But that's actually pretty easy to pick out of results.

So in my field, Google Scholar would be your answer. This is because you can sample down from too much information, whereas I have no idea how you would get a good methodology trying to estimate what Web of Science is missing. In other words, you can take a subset of GS results, go through (say) 1000 or so - which is 100 pages, so doable - and figure out the number of non-peer-reviewed or questionable docs. And then compute your confidence intervals and ranges from that.
posted by migrantology at 4:44 PM on April 11, 2016