General Social Survey Panel Data (2016-2020)
DOI
10.17605/OSF.IO/HACZVCitation
Davern, M., Bautista, R., Freese, J., Morgan, S. L., & Smith, T. W. (2022, July 7). General Social Survey Panel Data (2016-2020).Summary
The General Social Surveys (GSS) have been conducted by the National Opinion Research Center (NORC) annually since 1972, except for the years 1979, 1981, and 1992 (a supplement was added in 1992), and biennially beginning in 1994. The GSS are designed to be part of a program of social indicator research, replicating questionnaire items and wording in order to facilitate time-trend studies. The 2016-2020 GSS consisted of re-interviews of respondents from the 2016 and 2018 Cross-Sectional GSS rounds. All respondents from 2018 were fielded, but a random subsample of the respondents from 2016 were released for the 2020 panel. Cross-sectional responses from 2016 and 2018 are labelled Waves 1A and 1B, respectively, while responses from the 2020 re-interviews are labelled Wave 2.The 2016-2020 GSS Wave 2 Panel also includes a collaboration between the General Social Survey (GSS) and the American National Election Studies (ANES). The 2016-2020 GSS Panel Wave 2 contained a module of items proposed by the ANES team, including attitudinal questions, feelings thermometers for presidential candidates, and plans for voting in the 2020 presidential election. These respondents appear in both the ANES post-election study and the 2016-2020 GSS panel, with their 2020 GSS responses serving as their equivalent pre-election data. Researchers can link the relevant GSS Panel Wave 2 data with ANES post-election data using either ANESID (in the GSS Panel Wave 2 datafile) or V200001 in the ANES 2020 post-election datafile.
Data File
Cases: 5215Variables: 2377
Weight Variable: WTSSALL_1A, WTSSALL_1B, WTSSALL_2, WTSSNR_1A, WTSSNR_1B, WTSSNR_2
Data Collection
2016, 2018, 2020Original Survey (Instrument)
2016-2020 GSS Panel CodebookFunded By
National Science FoundationCollection Procedures
Because there was no in-person contact with respondents and all outreach to panel respondents took place via mail, email, and telephone, having the most accurate contact information for all panel respondents was critically important. Before data collection, the GSS staff compared contact information from prior-round GSS interview data against existing public and commercial records in two phases of locating: 1) batch locating to confirm old information and find new contact information, and 2) manual locating to confirm unconfirmed contact information with up to two rounds of search and verification.GSS staff redesigned outreach to respondents for mostly mail-based outreach and tailored materials for each phase of data collection. Throughout the data collection period, panel respondents were mailed postcards and survey invitation packets and sent a series of emails urging them to complete the 2020 GSS survey.
An experienced interviewer team conducted phone outreach to respondents to complete interviews by phone and/or prompt them to complete the web survey. To the extent possible, long-time GSS interviewers were assigned to contact the same respondents with whom they had completed GSS surveys in prior rounds.
From the start of fielding, panel respondents who had completed their prior-round interview in Spanish were assigned to bilingual interviewers and were mailed and emailed the Spanish versions of materials.
Sampling Procedures
The 2016 GSS Cross-Sectional SampleTo support a nationally representative sample, the GSS uses an area-probability sample of census tracts in urban areas and block groups in rural areas (these are second-stage sample units), and the addresses they contain, called the NORC National Sampling Frame. The NORC National Sampling Frame provides the basis for an equal probability multistage cluster sample of housing units for the entire United States. Every decade, NORC reselects a National Sampling Frame based on the decennial census. The 2016 GSS was the third to use the 2010 NORC National Sampling Frame. First-stage units on the National Sampling Frame are called National Frame Areas, (NFAs), each of which comprises a Metropolitan Statistical Area (MSA) or one or more counties outside MSAs. The number of first-stage selections in the 2010 NORC National Sampling Frame is 126 NFAs, but GSS uses only 76 NFAs, which is half of the noncertainty first-stage selections. The 59 noncertainty NFAs in the GSS each represent one percent of the U.S. population, and the 17 certainty NFAs contain 41 percent of the U.S. population.
The second-stage selection yielded a sample of segments that are census tracts in urban areas (where U.S. Postal Service address lists are usually available), and segments that are block groups (smaller than census tracts) in rural areas where a physical listing is often required. The first-stage and second-stage units were selected with probabilities proportional to size (in housing units), and the sample housing units (third-stage units) were then selected to be an equal-probability sample, which resulted in approximately the same number of housing units selected per second-stage sampling unit. The number of second-stage selections in the 2010 NORC National Frame is 1,514, but GSS uses only 400, which is four per one percent of the U.S. population.
The 2016 GSS sample consisted of 6,200 housing units with an equal probability of selection (0.0000471825). For 67 of the 400 second-stage units, GSS staff located missed housing units and either subsampled or interviewed them using a half-open interval procedure. Only 18 missed housing units were added to the sample of 6,200 through this procedure.
The initial sample was larger than required so that 12 weeks into the field period, non-responding households could be subsampled, resulting in a smaller number of more difficult cases needing to be interviewed (allowing greater concentration of effort). Approximately 60 percent of the cases eligible for subsampling were kept (weights were inflated for these cases to account for the subsampled-out cases), whereas 40 percent were subsampled out. The final number of completed interviews was 2,867.
The 2018 GSS Cross-Sectional Sample
The 2018 GSS was the fourth to use the 2010 NORC National Sampling Frame, and it used the same 400 segments as the 2016 GSS. The 2018 GSS sample comprised 5,200 housing units with an equal probability of selection (0.0000388289). Since the missed-housing-unit procedure adds so few cases and increases costs, 2018 GSS eliminated it.
The initial sample was again larger than required so that 10 weeks into the field period, nonresponding households could be subsampled, resulting in a smaller number of more difficult cases needing to be interviewed (allowing greater concentration of effort). Approximately 39.5 percent of the cases eligible for subsampling were kept (weights were inflated for these cases to account for the subsampled-out cases), whereas 60.5 percent were subsampled out. However, to increase the number of completed interviews, 300 of the subsampled-out cases were rereleased. The final number of completed interviews was 2,348.
The 2016-2020 GSS Panel Wave 2 Sample
The 2016-2020 GSS Panel Wave 2 consisted of re-interviews of respondents from the 2016 and 2018 Cross-Sectional GSS rounds. All 2,348 respondents from 2018 were fielded, but a random subsample of 2,146 out of 2,867 (74.85 percent) of the respondents from 2016 were released for the 2020 panel.
