It’s hard to talk about specific examples of searches that “don’t work” in Summon, isn’t it? Firstly you’ve got the problem that a search in one institution’s implementation won’t match another’s. Secondly, and more worryingly, the examples we talk about are generally ones that we come up with, rather than ones our students bring to us. Specifically, those one word searches that don’t bring up what one expects.
For example, you could, if you wanted to, just search for Dementia in Summon, and for our implementation the first 5 results don’t look particularly useful. The first is from 1987.
Alternatively, you could compare this to Google – first three results are paid for ads, then wikipedia, then useful stuff aimed at general public from NHS.
Or let’s look at Google Scholar – first five results seem to look better, but again they are old! 1968, 1982, 1950.
I think what’s missing from this conversation is the context that students place on their search terms when they start using Summon. Students are rarely, if ever, given one word assignments. They will be told to concentrate on the effects of dementia, or the treatment of dementia.
Additionally, now that we have Summon, we’ve more time for concentrating on identifying search terms and alternative words in our information literacy teaching.
It’s not that scientific, but I asked a couple of students what exactly they would search for it they needed to look for dementia.They both replied that it would depend on what the assignment was, but would start with dementia, and then add stuff like treatment, diagnosis, symptoms etc.
We also know that the number of students that actually only use one search term is low. Stats from Huddersfield pin this to around 8%.
One word searches are not useful to demonstrate positives or negatives of Summon or its relevancy ranking or its ability to link (or not link) to other material. But we need more conversation about the context within which students search in Summon and web-scale discovery.