Hannah’s reflections

Friday 2022-04-11

aRxiv:

“Open research”” produced many irrelevant results
Forgot to apply date filter - Can re-run once when we have confirmed that the search strategy is suitable

rScopus:

scopus_search doesn’t attach abstracts, and I couldn’t get abstract_retrieval to work (seemed to produce all metadata other than the abstract)
Many irrelevant journal articles due to abstracts mentioning “Data/material being available on the Open Science Framework” – Should we filter this out now or later (manually)?
Do we want to exclude “corrections” and “corrigendum” results?
Next time: Find out how to attach abstracts

Still unsure what we are expecting from the data, though I guess that is the point!
I am super keen on Luke’s idea RE: using a broad search strategy and then using filtering afterwards, once we know what sort of search results we get
Struggling to get up and running with rscopus, but just need more time

I’m wondering how “systematic” we should be in conducting the database searches, without turning this into a systematic review
Updated protocol
Trialed search of MEDLINE (via Ovid)
Need to get a better sense of the data analysis approach (e.g., come up with step-by-step guide to share with SC)

Today, I worked mostly on building a list of terms we will automatically exclude
This list is provided via https://github.com/science-collective/scoping-review/issues/56#issuecomment-1576677260
I wrote some preliminary code on this and created a Pull Request to upload this to R/
Note: These terms will be particularly helpful with the PubMed search (i.e., 947 to 583 results) but less so with the Zenodo search (i.e., only removed a handful of records)
It appears that most of the Zenodo search results include “open science” and “open source” in the title
I didn’t get a chance to work on writing but I will start with this next time