Regarding the announcement at: http://www.library.yale.edu/~llicense/ ListArchives/0611/msg00035.html This study is interesting, however as a librarian my comment is that the assumptions underlying the study illustrate a lack of understanding of the basic decision-making process of the academic librarian collections specialist. This study looks at 6 attributes and assesses librarian preferences, in an attempt to predict cancellations of subscriptions in favor of open access materials if articles are available in archives. Elements of the model examined: Version of Article Percentage of a Journal's Articles that are Available Reliability of Access How up-to-date is the content Quality of the content Cost The problem with this, is that the primary factors determining collections decisions are not taken into account: research and educational priorities of the university, and faculty assessment of the importance of journals. When we take these factors into account, we can see why it makes sense that librarians continue to subscribe to physics journals, even when prices are considered high and virtually all of the articles are available for free in arXiv. In other words, the answers this study have found really do not matter, because it did not ask the right questions. Research into librarians' collections decisions might be best led by librarians. Anna Creech hit the nail on the head when she said: Publishers could stop trying to squeeze blood from a stone * that might impact cancellations, too. For decades, libraries have been forced to cancel subscriptions due to prices rising far above inflation. No study of librarians' preferences and journal cancellations which not does consider this major factor can be considered even remotely objective. Heather Morrison http://poeticeconomics.blogspot.com