EDIT: Hmm, I wonder why people don't post preliminary results on their blogs all the time. Well, because sometimes they think they've proofread more carefully than they have. At least that was the case with me. When I reviewed my citation data again, I found that I'd accidentally categorized a set of IEEE journal articles as conference papers. Once I'd fixed the data, the split between articles and conference papers wasn't nearly so even. And IEEE moved well ahead of AIP in article citations. More details in the "future article" mentioned below.
I'm now a third of the way through my six week leave. And yes, I was happy to stay in my cozy home office today instead of waiting at the bus stop. Or on the four foot high mound of snow where the bus stop should be.
Leading up to the leave and during the first week, I read a fair amount of literature on engineering resources and reviewed many of the pages other librarians have built for their students. Thanks to others' thorough work, my "book knowledge" is pretty solid, and I've seen great examples of what others are doing.
But in order to get people to actually use the instruction modules I create, they'll need to know why they need them. The best way I can think of to do that is to educate them (and myself) about what other people just like them have found useful. And that means digging into the richest source of data available to me: Dissertations and Master's theses written by students in U of M engineering and computer science programs over the past few years.
I started with the Electrical and Computer Engineering department. Over the last ten years, ECE granted 27% of the Masters degrees in engineering and computer science, and 23% of the PhDs, so ECE grad students account for a healthy share of the target audience. To preserve my sanity and keep the time required reasonable, I put a few parameters around this experiment:
- I went back three calendar years, 2008-2010.
- I only used dissertations and theses that are available for download from either Proquest Digital Dissertations & Theses or from the University Digital Conservancy.
- If the bibliography didn't lend itself to parsing for any other reason, I didn't include the item in the sample.
I was able to collect all citations from 47 PhD dissertations and 10 Masters theses. For the record, the UDC turned out to be the better source for theses: Only one thesis showed up in the Proquest database. For dissertations, all were at least in the Proquest database and most--39 of 47--were in the UDC as well.
Oh, results? Well, I'll leave most of them for a future article. I'm anxious to compare the ECE results with those of other departments, and I think such an analysis may be useful for others as well. But here are a couple of interesting tidbits:
Citation sources were predominantly split between articles from academic journals (mostly peer reviewed) and conference papers.
8% of the citations were to books. An array of websites, magazines, patents, software, and theses made up the remaining 9%.
I also looked into who published the articles cited. I was surprised to find that IEEE was second on the list of publishers. American Institute of Physics edged out IEEE, 274 citations to 259. On the other hand, I'm confident that if I took the time to find the sponsors for all the conferences listed, the vast majority would be IEEE-sponsored.
My favorite citation? Has to be this awesome video of Boston Dynamics' BigDog.
I love my job.