Shifting access policies at "OA journals"

Hajar Sotudeh and Abbas Horri, Tracking open access journals evolution: Some considerations in open access data collection validation, Journal of the American Society for Information Science and Technology, July 11, 2007.  Only this abstract is free online, at least so far:

This article examines the evolution of a collection of open access journals (OAJs,) indexed by the Science Citation Index (SCI; Thomson Scientific Philadelphia, PA) against four validity criteria including a free, immediate, full and constant access policy for at least 5 years. Few journals are found to be wrongly identified as OAJ or to have a dubious access policy. Some delayed journals evolved into gold OA; however, these are scarce compared to the number of journals that withdrew from gold OA to be an embargoed or a partially OAJ. A majority of the journals meet three of the criteria as they provide free and immediate access to their entire contents. Although a lot are found to follow a constant policy, a large number has an OA lifetime shorter than 5 years, due to the high frequency of newly launched or newly converted journals. That is the major factor affecting the validity of the collection. Only half of the collection meets all the requirements.

Comment.  This result confirms my view that all journals --OA or not-- should post their access policies in some clear, up-to-date, and accessible form.  It won't keep policies from shifting, but it will keep shifting policies from confusing authors, readers, and libraries.