Thursday, December 21, 2006

Google simply doesn’t give a shit about woman scientists’ publication records

Back in September, I wrote about how academic indices are incapable of realizing that a woman researcher who changes her name is still the same person. As a test name, I used my grad school colleague Helen (she has published under both names, and when I asked her, she said she didn’t mind me performing this experiment). In September’s typically bombastic, tongue-and-cheek post, I gave a couple of rather snide hypotheses which I could pretend to address with symbolic examples instead of rigorous testing.

I would like to revise those hypotheses, at least with respect to Google Scholar.

I now suggest that Google Scholar does not realize that Helen Degeling and Helen Tomkins is the same researcher because:

Hypothesis 1: The google search technology is not sufficiently savvy to determine that she is one person who changed her name.

Hypothesis 2: They have the technology, they just can’t be bothered to apply it to this particular issue (e.g. they don’t give a shit).

Hypothesis one is testable. All we need to do is to determine if any of the scientific literature accessible via Google Scholar turns up evidence connecting these two Helens. If no such evidence exists, then it would be fair to state that Google Scholar, as currently programmed, simply doesn’t know that Degeling became Tomkins.

The easiest way to do this is to search for “nee Degeling.” Here are the results.

Those of you with journal access can see that “Helen Tomkins (neé Degeling)” is listed in the acknowledgements, and that Google scholar is smart enough to have found it there (which is why Watson et al. turned up in this search).

Thus hypothesis 1 is refuted. Of the hundreds of articles published in Helen's field this year, Google Scholar unerringly picked out the one that connects her maiden and married names. It just doesn’t bother to use this information when asked to compile her publications. Thus hypothesis 2 becomes our best working hypothesis, given the available data.

Suppositions as to why Google doesn’t give a shit are left as an exercise to the reader.

3 comments:

Sara said...

you can add the arxiv (http://arxiv.org/, the biggest most well know location for papers in our field) to your list of "doesn't give a shit", to really quite a stupid level. i have a hypenated married last name (added The Husbands onto the end of mine) and searching for my old last name on its own won't bring up the new one, nor will searching for the hypenated version bring up the old unhypenated one.

EliRabett said...

A major issue like forever. The worst cases are those where someone marries, takes her husband's name, starts to publish, divorces, and is left with the choice of keeping the ISI version or changing her name back. And yes, I know of such cases. I guess we should all get DOI numbers (it would actually make searching a lot simpler)

Alex Villepique said...

Thanks for performing the experiment. Results look really depressing. :-(