As of now, I will be collecting data from the articles that cite the Data Collection Article as collected using the Cited By function in ISI Web of Science. These citations were collected by Estephanie and placed in Dropbox as text files. Heather is tweaking them by adding a random number to the beginning of the title and converting them to BibTex files so they can be imported into Mendeley. I will then find the full text for a minimum of 100 citations for each data repository, search within the full text to find the sentences that were cited to determine if data was reused and apply tags to the Mendeley file. Our tagging structure is as follows and is likely to change/evolve as the project continues.
Tags for Use in Mendeley (the tag follows the letter under each heading)
- How was the article found?
- data cited – the article cites the data set
- article cited – the article cites the Data Collection Article
- Attribution Location
- attribution in reference
- attribution in caption
- attribution in text
- Data reuse classification
- dataset not reused
- dataset reused
- dataset reuse ambigious
- data perhaps from article – (not from dataset?)
- Confidence of data use classification
- high confidence
- medium confidence
- low confidence
- Other housekeeping tags
- feedback please
- no full-text access
- import error – incomplete citations
- nice example – articles that would make a good example of dataset reuse or citation of a dataset, etc.
also “in supplementary information” (though it is a limitation of our current method that we are unlikely to find any of these)
You may find cases in which both 1A and 1B are true, right? Since you may have different answers for 2 depending on which citation you are referring to, shouldn’t this really be two sets of questions, one for dataset citations, and one for data collection article citations?
Good point. Really the tags for #1 and #2 ought to be applied at a per-attribution level rather than a per-paper level. Since the schema for this study is article-based, maybe we go with this:
Attribution type (choose all that apply):
data cited in reference
data cited in caption
data cited in text
data collection article cited in reference
data collection article cited in caption
data collection article cited in text
or something like that?