Page 1 of 1

Automated text analysis: your reactions to these transparency guidelines?

Posted: Mon Dec 19, 2016 3:58 am
by AlanJacobs
Dear Colleagues,

As a way of prompting some conversation about transparency in automated text analysis, we're inviting responses to a recent, brief essay by David Romney, Brandon Stewart, and Dustin Tingley that lays out some general transparency principles for this set of methods. The essay can be downloaded here:
Romney, Stewart, Tingley.pdf
(415.23 KiB) Downloaded 1347 times

Please read the piece -- it's about 4 pages long -- and let us know your thoughts. Are there additional ways, beyond those discussed in the essay, in which scholars can or should be transparent about text-analytic procedures? Are there additional challenges to transparency in this area that we should be wrestling with? Do newly developed methods/technologies in this field have implications for what transparency ought to look like? Any disagreements with the principles articulated here? Can you point us to specific text-analytic studies that you consider exemplary in their handling of transparency issues or that demonstrate useful ways of operationalizing transparency?

All comments -- short, long, identified, anonymous -- are most welcome!

Many thanks,

Alan