Revision as of 14:21, 1 June 2014

Tracking relations between scientific works, such as disputes

Prof. Yang tells Lane and me that students have trouble interpreting a dynamic immunology literature which includes published papers whose findings are wrong or irrelevant.
We are imagining a wiki platform that made the state of research in a science more clear.
Here's a feature it could have.

Examples:

note: the platform doesn't generally resolve claims or disputes ; it's finding a shallower kind of Truth -- that a dispute exists

Similarly could incorporate other relations between scientific works: A cites B ; A is an important predecessor to B ; A and B use the same data set or the same clinical trial information ; A reproduced or could not reproduce result from B

Could scale this up with big lists of relevant papers and works from many places including PubMed, SSRN, and Wikidata

The list is not an innovation but the relations between the works can be useful if they are formalized a bit

Trees of relationships can then be made visible ; could give a picture of a dynamic literature across fields, journals, and languages
Could be useful for:

students of this research or
researchers not at a "central" place that is well-connected to the latest news (global South?)
somebody with some other expertise (or technology or skill) that helps to address a particular issue

Opportunities are rare if subfield has knowledge & institiutional barriers around it.

===> A good site could save time for scientists and bring in more scientists ===> speed science along

Report on disputes between papers on this site
who is qualified to make commentary & identify disputes on such a wiki? (to be worked out)

We need more people and ideas of what a wiki with scientific lit should have. It needs to spark critical mass (fun, addictive?)

Sources:

developed with Lane Rasberry (WM NYC), Otto Yang (UCLA medicine), others
drawing from Yaron Koren's discoursedb.org ; Retraction Watch blog; many other sites

@@ Line 13: / Line 13: @@
 * Substantive comment is formalized:  A disputes B     (using Semantic MediaWiki)
 :* note: the platform doesn't generally resolve claims or disputes ; it's finding a shallower kind of Truth -- that a dispute exists
-* Similarly could incorporate other '''relations between scientific works''':  A '''cites''' B ; A is an '''important predecessor to''' B ; A and B '''use the same data set''' or the same clinical trial information
+* Similarly could incorporate other '''relations between scientific works''':  A '''cites''' B ; A is an '''important predecessor to''' B ; A and B '''use the same data set''' or the same clinical trial information ; A '''reproduced''' or '''could not reproduce''' result from B
 * Could '''scale this up''' with big lists of relevant papers and works from many places including PubMed, SSRN, and Wikidata