Faceted comparison
Repeat one chart across groups so differences can be read on common scales.
% saying 'too little' is being spent. Dots show the estimate; whiskers show the supplied 95% confidence interval.
Democrat
Independent
Republican
Chart data · 6 rows
| Group | Wording | Estimate (%) | Lower 95% | Upper 95% | N |
|---|---|---|---|---|---|
| Democrat | Assistance to the poor | 88.52 | 84.39 | 92.64 | 429 |
| Democrat | Welfare | 51.59 | 44.91 | 58.27 | 468 |
| Independent | Assistance to the poor | 72.13 | 67.34 | 76.92 | 689 |
| Independent | Welfare | 32.69 | 27.92 | 37.46 | 707 |
| Republican | Assistance to the poor | 51.41 | 44.88 | 57.93 | 419 |
| Republican | Welfare | 16.56 | 11.62 | 21.51 | 399 |
Page 1 of 1
The item labels are “Welfare” and “Assistance to the poor.” These are separate survey wording conditions, not before/after answers from the same people. Party panels omit missing party identification. No interval for the difference is inferred from the two marginal intervals.
General Social Survey · 1984–2024 · captured wording-condition estimates
Preview
How to read it
Compare the same position across panels. Shared domains are essential: equal distances must mean equal changes.
When to use it
Wording conditions by party, or relationships that may differ across groups.
Watch out for
Avoid independent panel autoscaling, tiny unreadable panels, and comparisons of incompatible quantities. Different point estimates alone do not prove a significant difference.
Data shape & units
For estimate panels: { label, estimate, lo?, hi?, n? }[]. Group rows before rendering one ForestPlot per panel, with a common domain. ScatterRegression also supports facet and shared scales.
Missing values
Omit unavailable panels with an explanation; never manufacture an interval. Null group labels must not silently become a named group.
Controls & accessibility
Year and breakdown controls update all panels together. Tables retain exact values and bounds. Panels stack at narrow widths.
Use the component
Install from the CTZN registry, then pass your prepared rows. Examples below illustrate the component contract.
npx shadcn add @ctzn-viz/forest-plot{groups.map(group => <ForestPlot key={group}
data={rowsByGroup[group]} domain={[0, 100]}
formatValue={{ kind: "percent", digits: 1 }}
ariaLabel={group + " spending estimates"} />)}Examples using this form
General Social Survey
“Welfare” or “assistance to the poor”?
How different are spending priorities when the survey uses a different label?
CDC PLACES
Obesity and diabetes, county by county
Does the association look similar in counties with different 2020 vote directions?
World Values Survey
Values and economic development
How do country-level attitudes vary with national income?
