"A GSS needs government funding to reach a scale allowing in-depth study.  For a simple national reading, the standard sample is only 1,000 respondents, as used, for instance in surveys of the International Social Survey Program and Eurobarometer.  But suppose one wants to focus on social attitudes of women (half of the people), of age 18-35 (one-third of all adults), who are college graduates (at most 15 percent of adults)?  Then 1,000 would be inadequate, 2,000 to 3,000 would be fair enough, and 10,000 would be extravagant."