Today I fooled around with Google Scholar a bit. During this, I searched for papers matching the exact phrase “climate change” for year 1903. For this search, Google Scholar returned 59 “papers”. But when I looked the results more closely, I was rather disappointed. Let me tell you why.
There’s this website, kbd.kew.org, that has a link named as “Kew & Climate Change” in their bottom link bar which appears in every page of the site. Google Scholar catches this as a “citation”. This website of Kew Royal Botanic Gardens has in it “Kew Bibliographic Database” which contains entries of scientific papers, and rather lots of them and many are old ones, too. This results in Google Scholar finding that all the papers in the database match the search criteria. For more modern searches this doesn’t matter much as there are plenty of search results for “climate change” so this kbd.kew.org thing just adds a bit noise to the results. However, when searching for historical papers this flaw can make the search results meaningless as almost all the results are due to the Google Scholar’s flaw.
Fortunately, Google Scholar has an option to exclude the citations from the search (citation results can be useful, but I wonder if they should be included by default). If we exclude citations from the search above, we only get 5 hits.
Of these, 3 are still from kbd.kew.org.
One result is the document:
At first glance it is obvious that this is not a peer-reviewed publication and second glance tells that it’s not from 1903 either. Apparently the document is from 2013. It also is not about climate but seems to contain all kinds of stuff, and one section does mention “a documentary about the effects of climate change on the oceans, as told by a grandfather to his 5-year-old grandson”.
One result is from The British Medical Journal:
The Róle Of Eyestrain In Civilization
George M. Gould, Sept. 19, 1903
Climate is mentioned 2 times in this article. Both of these times the exact phrase used is “change of climate”, so Google Scholar’s exact phrase search doesn’t actually search only for the exact phrase. (CORRECTION: Kevin ONeill notes in the comment section that the text actually contains “climate change”. Text has this sequence of words: ““change of climate,” “change of scene,””, where word climate is followed by word change as long as we forget the special characters, like this: “change of climate change of scene”. So, I was wrong about this, thanks to Kevin for this correction. I’ll also correct the text below for this.)
So, using default Google Scholar exact phrase search for “climate change” returned 59 results but none of them actually matched the intended search. One can argue that the article from The British Medical Journal is a quite good hit but beyond that Google Scholar returned just rubbish – 58 bad hits for one almost good hit.
Furthermore, this simple little search revealed four flaws in the Google Scholar’s search process; 1) the method for searching the citations is rubbish, 2) returned results are not always scientific documents, 3) the year assigned to document is sometimes wrong, and 4)
the exact phrase search returns papers not having the exact phrase the exact phrase search catches extra phrases because it ignores special characters. Also, while not actually a flaw, the choice to include citations (and patents) by default is a poor one, in my opinion.
It might be kind of nice to have an access to real scientific search engines.