Feb04NII.zip

Sample Design: • Dual-frame RDD sample includes landline and cell-only households. • Weighting: • : For full sample analysis. • : Special weight for Q54a to correct underrepresentation of TX, CA, PA, NY, MI, GA, IL, FL, AZ. • Q54a Correction: • Due to a fieldhouse error, only one-third of respondents in nine states were asked to rate their own state. • : Combines responses from Q54b–k into Q54a. • : Flags affected states. • : Multiplies weights by 3 for affected respondents to balance topline estimates. • Constructed Variable – Q38tot: • Aggregates Q37a–d into a single measure of perceived deficit causes. • Categories include domestic spending, homeland defense, Iraq war costs, tax cuts, and “none.” • This variable is archived and does not require re-creation. • Privacy Protocols: ZIP codes, counties, and phone numbers removed from public dataset. • Geographic Variables: Based on self-reported ZIP codes or sample geography when unavailable.