Survey Explorer
Descriptive views of the released data
What were likely voters thinking in November 1963? Browse the raw survey responses — see how people answered each question, which questions moved together, and how one group compared to another.
Frequencies, cross-tabs, and correlations. Use this to orient before turning to the explanatory tools — what looks prominent descriptively may add little once other measured factors are considered together.
Only Presidential Approval has a published model. The Model Builder lets you build your own for Vote Intention and Tax Cut Support, the simulators run scenarios on all three outcomes (using a model Auto-Build selects where none is published), and the Nonresponse Test covers all three released scores. The Explorer works across every outcome — including the three-way leaned vote (Kennedy / Goldwater / Undecided).
How to Use This Tool
Pick what you want to know, then which respondents and variables, and click Run Analysis. Expand for details.
How to Use This Tool
Pick what you want to know, then which respondents and variables, and click Run Analysis. Expand for details.
1. Pick what you want to know
How did people answer? — Frequencies. Which variables best separate the outcome groups on their own? — Predictor Strength. Which questions moved together? — Correlations. How did one group compare to another? — Crosstabs (pick exactly two questions).
2. Pick respondents and variables
Optionally restrict to a subgroup, then check the survey questions you want to look at. Cards start empty — for Approval, click Load default model to load the published model’s predictors as a starting point, or use the search box and category buttons to find variables.
3. Run and read
Click Run Analysis. Results appear below — scroll down. Results use the age-by-sex analysis weight described in the paper. The archived materials do not permit reconstruction of the survey's complete original sample design, so the weighted results should not be interpreted as incorporating all original selection, clustering, stratification, or nonresponse adjustments.
Advanced features
Pin runs to compare
Click Pin this run in the results header to save a snapshot. Pinned runs collect at the bottom — expand to compare previous results side by side.
Filter and search
Search by keyword or question code (e.g. "q31" or "Khrushchev"). Category buttons narrow the list by topic. Click × on any active filter chip to clear it.
Go deeper
Variables you find here can go straight into the Model Builder to test what predicts any of the outcomes — Approval, Vote Intention (head-to-head or the three-way leaned race), or Tax Cut Support. The Explorer is a good first step before modeling.