Anurag Acharya, the co-founder of Google Scholar, shared some insights he has discovered concerning how people use Google Scholar. https://www.youtube.com/watch?v=S-f9MjQjLsk
His talk has been reviewed in some other places, e.g. Scholarly Kitchen (http://scholarlykitchen.sspnet.org/2015/10/05/guest-post-highwires-john-sack-on-online-indexing-of-scholarly-publications-part-1-what-we-all-have-accomplished/) Some of his discoveries are not surprising, e.g. people cite older articles much more often after the articles are digitized because digitized articles are easier to find and access. Also, people prefer open-access articles with full-text as opposed to articles that are behind pay walls. (I confess, I do this a lot, myself) There were a couple of other results he mentioned that I would like to emphasize since they could have consequences for cataloging. 1) people search Google Scholar far more for keywords and concepts and much less for authors and titles (https://youtu.be/S-f9MjQjLsk?t=12m). Also, the average query length is increasing, up to 4-5 words now. Finally, most queries are unique, that is, Google Scholar searching is completely different from regular Google searching because on the regular web, people are searching pretty much the same things: where to eat, what app to download, what movie to watch, what to do for the flu, and so on. This seems to suggest to me that the library catalog's subject indexing could prove very useful. People clearly want subject access, especially as the amount of information is sky-rocketing. I don't think I have to explain to subscribers of Autocat the difference between searching keywords as opposed to real concept searching. LC subject headings with their syndetic structure, and the related authority files really do allow for genuine concept searching. Of course, the subjects would have to work much better than in our catalogs. I wonder how this (people search primarily keywords and concepts vs. very few authors and titles) compares with how people search library catalogs. I would suspect it would be quite different, but it would be just as logical to find that people are searching everything in the same ways. 2) He discovered that abstracts have become very important for people. (https://youtu.be/S-f9MjQjLsk?t=32m23s) In fact, he claims that providing abstracts is even more important than the full-text because above all, people want to filter first and then read. He says that abstracts have always been written for peers in the field but now that wide-swaths of articles can be searched at once, there are a lot of potential readers of you article who may be less acquainted with your specific subject area (i.e. cross-disciplinary studies) and need an abstract that is written for a less-technically trained audience. He didn't mention a question I have on this. Way back when in library school, I took a course in abstracting and indexing, and I learned that there are three types of abstracts: descriptive, informative, and critical. Descriptive abstracts just describe the article; informative give the reader the results and can possibly even stand-in for the article itself; and critical is like a short review. I don't know if the speaker understands the different kinds of abstracts or if he has in mind something new. To return to the talk, he emphasized that the abstracts must be written for the non-specialist because with online searching, since there is a much greater chance than ever before that your article will be found by people outside of your specialty. The non-specialist who finds your article is then confronted with a choice: he or she may be interested but to wade through the article will demand a lot of work, and it is important for them to understand well enough that it will be worth the effort. Abstracts should be written to help these people decide. Therefore, this is a different kind of filtering. One final point that I, and apparently others in the hall, were surprised by: when you search Google Scholar, apparently you are *not* being tracked. (https://youtu.be/S-f9MjQjLsk?t=52m) Therefore, everybody gets exactly the same result with a Google Scholar search as opposed to a regular Google search. I don't know if I agree so much that "all articles are easy to find" in Google Scholar but of course, the speaker is a co-founder of Google Scholar and part of his job is to say how great his product is. Compared to the old days, when you hunted for articles using a variety of heavy printed indexes with tiny print, then copying things down, searching the catalog for the call number of the journal, entering the stacks, hoping to find the volume on the shelf, if it was bound, hoping the call number didn't say "Incomplete" etc. etc. it is much "easier to find" articles today. But that is not an end of it since there are all kinds of other problems. Nevertheless, I still find Google Scholar an incredibly useful tool. James Weinheimer [email protected] First Thus http://blog.jweinheimer.net First Thus Facebook Page https://www.facebook.com/FirstThus Personal Facebook Page https://www.facebook.com/james.weinheimer.35 Google+ https://plus.google.com/u/0/+JamesWeinheimer Cooperative Cataloging Rules http://sites.google.com/site/opencatalogingrules/ Cataloging Matters Podcasts http://blog.jweinheimer.net/cataloging-matters-podcasts The Library Herald http://libnews.jweinheimer.net/
