Data Source: Voting records from Legistar, the municipal government software used by hundreds of U.S. cities. We extracted individual roll-call votes and linked them to person records.
Officials Included: We tracked 1,524 officials who cast at least 3 votes on development-related matters. This threshold ensures a minimum sample while including part-time and recently-elected officials. Officials with fewer than 20 votes should be interpreted with caution—their rates may change significantly with additional data.
Topic Classification: Votes were categorized by topic using matter type classifications from each municipality. We distinguish between administrative votes (meeting minutes, ceremonial items, routine approvals—13,132 votes at 97.1% approval) and substantivevotes (budget, zoning, environmental, residential, commercial, infrastructure—3,073 votes at 93.8% approval). The analysis focuses on substantive votes where policy disagreement is meaningful.
Approval Rate: Calculated as (Yes Votes) / (Yes Votes + No Votes). Abstentions are excluded from the denominator.
Polarization Index: For councils with 3+ tracked officials, we calculated the difference between the highest and lowest individual approval rates.
Selection Bias: Our dataset only includes cities that use Legistar and publish machine-readable roll-call votes. This over-represents larger cities with sophisticated record-keeping and excludes municipalities using other systems. The cities in our analysis self-selected by having accessible data infrastructure.
Limitations: Some cities don’t record roll-call votes for routine matters. Vote counts may include procedural votes alongside substantive ones. The sources cited validate that housing and development debates are active in these cities, but do not independently verify individual vote tallies—those come directly from Legistar records.
Date Range: Records span 2023 through January 2025, with most data from 2024.