From 1972 until 1993, the GSS was administered almost annually. The target sample size for the annual surveys was 1500; actual sample sizes ranged between 1372 (1990) and 1613 (1972). Additionally, there were oversamples of black respondents in 1982 (oversample of 354) and 1987 (oversample of 353). There were no GSSs in 1979, 1981, or 1992.
Since 1994 the GSS has been administered to two samples in even-numbered years, each with a target sample size of 1500, rather than to a single 1500-person sample each year. Total sample sizes for these biennial GSSs ranged between 2765 (2002) and 2992 (1994). In 2006 a third sample was added and the sample size was 4510. There have been no oversamples during this period.
The GSS has transitioned from a replicating cross-sectional design to a design that uses rotating panels. Under the new three-wave, rotating panel design in each round of the GSS, a new panel is started, the first reinterview from the previous GSS is conducted, and the second and final reinterview from the GSS panel started four years ago is conducted. In 2008 there were two components: a new 2008 cross-section with 2,023 cases and the first re-interviews (panel) with 1,536 respondents from the 2006 GSS. The 2,023 cases in the cross-section have been previously released as a part of the 1972-2008 cumulative data. This new release includes those 1,536 re-interviewed panel cases along with the 2,023 cases. In 2010 the new rolling panel design was fully implemented for the first time.
For more information on the GSS sampling design, and completion/response rates, see
Appendix A: Sampling Design & Weighting.