Three ballots, one city
Three ballots, one city
The geography is familiar across the three votes, but the balance changes. PP was strongest in all three; the national contest was the closest and also drew the highest turnout.
Where each party was strongest
Select a party to see its share of valid votes.
Participation, winners and income
Turnout and household income reveal another persistent divide through the city. The income layer is the same 2023 measure for all three ballots.
Vote share and turnout
The relationship is not uniform across parties.
In the higher-turnout sections, support tended to lean more towards PP and less towards PSOE. The other parties follow more uneven paths.
Vote share and household income
Each dot is one census section. These are spatial associations, not evidence that income causes a particular vote.
PP’s share rises sharply with section income in every election. PSOE moves in the opposite direction; Más Madrid and Sumar are strongest across a broader middle and lower-income range.
Turnout and household income
This is the clearest relationship, and it changes little between the three ballots.
Turnout was consistently higher in sections with higher mean household income. The national election lifted participation across most of the city without erasing that gap.
Sources and method
The analysis is built from official polling-table results, aggregated to the ten-character census-section code.
Definitions
Turnout is valid plus null votes divided by the electoral census. Party share divides candidature votes by valid votes, including blank ballots. Section winners are calculated from every candidacy; exact ties are resolved in favour of the candidacy with the larger citywide total.
The 5% rule is an editorial citywide inclusion rule, not the statutory threshold for seats. It includes Podemos–IU in the regional view because it received 5.16% within Madrid city, although its result across the region was below 5%.
Curves are LOESS fits with span = 0.5. Income and voting are presented as spatial associations; no causal claim is made.
Official sources
Retrieved 29 August 2026. Source URLs, checksums and validation fixtures are stored with the reproducible source.