UP Assembly 2022 · Normal CAPI surveys
Survey weight accuracy
Compare seat calls and alliance support across weight versions. Each survey keeps its own full eligible seat coverage.
—eligible seats in this survey
No results match your filters.
How to read these results
- Survey winners combine member-party support into alliances first. Actual winners are resolved as individual candidates, then mapped to alliances.
- Support error is survey minus actual, in percentage points. Positive means overestimation. MAE is the average absolute error.
- Summary MAE gives equal weight to BSP, INDI, NDA and SP+. Other eligible named parties still participate in winner competition.
- R1 does not have MLA choice. Generic base female/GD versions remain excluded until their question lineage is established.
- Filters select existing results. They do not recompute winners or change a saved metric’s denominator.
- No sign-in is enabled. Anyone with access to this website can download its aggregate data.