The third body of data is the most broadly defined, since it includes data collected by many different organizations, some of which no longer exist. Until the mid-1960s, there were relatively few survey firms. Most of those that existed regularly sent their data to the Roper Center, but even during this time some surveys were missed. In the days of punched cards, for example, if the survey sponsor wanted the original data, there was not necessarily a backup copy to send to the Roper Center. Beginning in the 1970s, the number of firms increased dramatically, and many did not regularly archive their data. As a result, a large fraction of the surveys done in the last few decades remains outside of any archives.

We have used the Roper Center's iPOLL database, which includes records for about five hundred thousand survey questions, to identify surveys of particular interest. The list includes over two hundred surveys and continues to grow. Examples of surveys on our current list are "How Nations See Each Other," from 1948; "World Poverty And Development" (1972); “Knowledge Of Business And The Economy” (1984); and the “American Family Values Survey” (1991). Unlike the USIA and NORC surveys, these data are highly dispersed—there is no central source to which we can go. Our primary approach will be to contact the survey organizations and survey sponsors. When possible, we will also contact individual researchers who obtained the data for their own work and may still have copies.

We know that we will be able to obtain and archive much of the data that is currently held by USIA and NORC. Whether we will be able to obtain much from the third group is uncertain. It is possible that most have already been discarded, and that the rest is too scattered to be recovered. If we do not try to find them, however, they will certainly be lost, so the attempt is worth making.

The pioneering figures of survey research were aware of the value their data would have to future generations. The Roper Center was founded in 1947, only about ten years after modern surveys began, and a great deal of survey data has been preserved there and in other archives. It has been difficult, however, to keep up with the growing volume of data produced by the expansion of academic social science and the survey industry. The Data-PASS project is the first systematic effort to go back and identify data that have been missed. We hope that in addition to recovering older data, the project will establish standards and procedures that will help archives effectively manage data produced in the future.

David Weakliem is a professor of sociology and acting director of the Roper Center for Public Opinion Research at the University of Connecticut. The Roper Center urges readers of Public Opinion Pros and others in the field to help in the effort to recover data for the Data-PASS project. Anyone who has or knows of data that might be of interest should contact the project through the Data-PASS website, or pass along word of the project to others who might be able to help. The Roper Center will, of course, obtain any necessary permission from the collectors and sponsors before archiving any data, but knowing that they exist is an important first step.