Supporting Statement B - 2012 GSS Arts Supplement

Supporting Statement B: Arts supplement to the 2012 General Social Survey

B.

COLLECTION OF INFORMATION EMPLOYING STATISTICAL
METHODS

B.1. Universe and Respondent Selection
GSS data are collected every two years, and made available to the research community
and the public as soon as possible after data collection is complete. Until 2004, the GSS
was designed as a repeated cross-sectional survey. Beginning in 2006, a panel component
was added to the GSS design. In 2012, there will be three panels: a new cross-section,
the first reinterview wave of cases from the 2010 GSS (1,530 respondents), and the
second reinterview wave of the 2008 GSS panel (1,300 respondents). The arts
supplement will be asked of the 2010 and 2008 panels, a total of 2,830 respondents.
Through the use of appropriate sampling weights, each biennial GSS will provide
nationally representative estimates of distributions of survey items measuring a wide
variety of social and political attitudes, opinions, and behaviors of U.S. adults.

B.2. Procedures for Collecting Information
The majority of GSS data is obtained via face-to-face interviews. Since the 2002 GSS,
Computer-assisted personal interviewing (CAPI) has been used. In limited cases, when it
is difficult to arrange an in-person interview, GSS interviews may be conducted by
telephone. The 2012 GSS will commence collection in March 2012.

B.3 Methods for Maximizing Response Rate and Dealing with Nonresponse
The GSS has high responses rates, exceeding 70%. Field interviewers are trained in data
collection methods and given specifications for issues that may arise (e.g., what to do if a
respondent does not understand a question).
Nonrespondents are subsampled and weighted to represent all nonrespondents as of a
cutoff date. This approach has been used in many other surveys, including the Census
Bureau’s American Community Survey. In 2010, the last time the survey was
conducted, 1695 temporary nonrespondents were identified, subsampled at a 47% rate
and pursued for interview. After about 7 weeks, 137 cases that had been removed for the
subsample were included back in the sample. Over an additional 4-week period, these
and other remaining cases were pursued for interview. At the end of collection, n=2,044
cases were obtained. The data include weights adjusted to account for nonresponse. For
more detailed information on sampling design and weighting, refer to the documentation
prepared by NORC (included in Attachment D).
1

B.4 Test of Procedures and Methods
Prior to the fielding of the arts supplement, the questions were pre-tested by NORC in the
summer of 2011, with a quasi-national sample of individuals. The average time to
respond to the questions was 3.5 minutes. Following the pretest, some minor revisions
were made to the instrument. These are incorporated in the copy of the instrument
provided in Attachment C). After the questions are programmed for CAPI, the system
will be tested to ensure that the program functions correctly.

B.5- Consultations on Statistical Aspects of the Design
The questions were reviewed and revised based on feedback from the Director of the
General Social Survey, Dr. Tom Smith, and following a pretest. The NORC Board of
Overseers reviewed and approved the supplement in 2012.
In addition, as mentioned in Section A.8, the NEA’s Office of Research & Analysis
consulted the survey questionnaire and methodology employed by Francie Ostrower,
formerly of the Urban Institute, in a pioneering study of motivational and socialization
factors affecting arts attendance; reviewed extensive literature of arts participation
surveys conducted in the U.S., U.K. and Australia; and convened sociologists,
statisticians, and survey methodologists, both for a November 2010 event and for a series
of post-event teleconferences.

Attachments
A.

Federal Register Notice 1, 06.07.11

B.

Federal Register Notice 2, 11.15.11

C.

2012 GSS Arts Supplement Questionnaire

D.

GSS Codebook – Documentation on Sampling Design and Weighting

2