Supplementary Materials Related to the NEA's 2012 Survey of Public Participation in the Arts

June 30, 2014

Dear fellow researchers:

On the week of June 22, 2014, the U.S. Census Bureau confirmed that it had used a different technique for identifying completed interviews for the 2012 SPPA than had been used in prior years of the survey.

As this methodological choice has serious effects on how the survey data are weighted, the National Endowment for the Arts has requested a fully revised data file for the 2012 SPPA.

The NEA has removed, therefore, all related research tools from its website until the changes are received and can be incorporated into these resources. Please check here during the summer/fall of 2014, and/or email if you have any questions or would like to receive notification when the new data file and related resources are ready. The NEA also will post a comprehensive report of the survey results. Meanwhile, the Census Bureau has provided a user's note to assist researchers working with the current SPPA data file for 2012 (please see below).

On behalf of the NEA's Office of Research & Analysis, I apologize for the inconvenience.

Yours sincerely,

Sunil Iyengar
Director, Research & Analysis

National Endowment for the Arts


There was a methodological change made to the weighting procedures for this survey between 2008 and 2012.  In 2008, cases with answers of "yes," "no," and "don't know" to the first survey item were treated as interviews.  In 2012, cases with a response of "refused" were also treated as a interviews. The net effect of this revision was to shift some of the control population used for weighting away from the yes, no and don't know responses in 2008 into the refused answers in 2012.  It is our estimate that overall, all statistics in 2012 were decreased by a factor of approximately 1.043 for any yes, no, or don't know response.  If one were to attempt to break this down by core, the factor would be 1.0412 for Core 1 and 1.0448 for Core 2.  Statistics for various demographic cohorts will differ somewhat (+ or -) from the factors provided above.