Twenty policy questions, asked the same way for fifteen years
The CES put identical policy items to hundreds of thousands of respondents from 2006 to 2021 — a controlled experiment in opinion change that most survey series can't offer. Pick any issue: the level moves slowly, the party gap fast.
Share who support universal background checks, by party identification. The gap between the party lines — not the level — is what changed over these years.
- Democrats
- Republicans
- Independents
CES cumulative file via issue-almanac/v1. Policy items asked 2006–2021 (CES restructured its policy module after 2021).
Democratic minus Republican support, in percentage points, when the item was first asked (open dot) and last asked (filled). Nearly every arrow points away from zero — the gaps widened almost everywhere.
- Democrats higher
- Republicans higher
CES cumulative file via issue-almanac/v1. Policy items asked 2006–2021 (CES restructured its policy module after 2021).
Within-state coefficient of education on supporting legal abortion in all cases, one model per state. Education pushes the same direction almost everywhere — but two to three times harder in some states than others.
- Northeast
- Midwest
- South
- West
issue-almanac/v1 regressions. Linear probability model per state; coefficient scaled to percentage points per education step.