Abstract
The number of data collections produced in the social sciences prohibits the archiving of every scientific study. It is therefore necessary to make decisions regarding what can be preserved and why it should be preserved. This paper reviews the processes used by two data archives, one from the United States and one from the United Kingdom, to illustrate how data are selected for archiving, how they are appraised, and what steps are required to retain the usefulness of the data for future use. It also presents new initiatives that seek to encourage an increase in the long-term preservation of digital resources.