Cohen's take is valuable, especially for already existing and publicly available, often quant, datasets. As someone who uses non-English language archival sources that I am legally not allowed to publish, I have been wondering how to take on this good intuition -- make research accessible