The policy data cycle in practice

The policy data cycle has four stages: Issue Identification → Analysis → Recommendation → Monitoring. In theory these stages are sequential. In practice they overlap and loop back on each other. A monitoring result from Stage 4 often opens a new issue in Stage 1.

Stage 1 — Issue Identification

Someone notices a problem: turnout dropped in a cluster of precincts, or a proposed district map looks strange, or a demographic group is suddenly underrepresented in a legislative chamber. The issue might come from a news story, a complaint, a community meeting, or your own data exploration. The output of Stage 1 is a well-formed research question—specific enough to be answerable with available data.

Stage 2 — Analysis

You gather the relevant data, clean it, and perform the analysis needed to answer the question. This might be a simple turnout calculation, a comparison of two districts' demographic compositions, or a full ecological inference analysis of racially polarized voting. The output of Stage 2 is a finding: a factual statement about what the data shows.

Stage 3 — Recommendation

You translate the finding into a policy recommendation. What should a legislator, agency, or court do in response to this finding? A good recommendation is specific, actionable, and grounded in the data—it does not overclaim what the analysis can support.

Stage 4 — Monitoring

After a policy change, you track whether outcomes changed as expected. Monitoring requires a baseline (data from before the change) and a comparison period (data after the change). Without monitoring, you cannot know whether the policy worked.

Georgia redistricting example (pre-Callais)

In 2021, the Georgia General Assembly redrew state legislative districts following the 2020 Census. Advocates used the policy data cycle:

1. Issue ID: Proposed maps appeared to fragment Black-majority precincts in southwest Georgia, potentially reducing the number of majority-Black districts.
2. Analysis: Researchers calculated CVAP (Citizen Voting Age Population) percentages in proposed vs. previous districts and ran racially polarized voting tests.
3. Recommendation: Redraw specific districts to maintain or create districts where Black voters could elect their preferred candidate.
4. Monitoring: After new maps took effect, analysts tracked registration and turnout in affected precincts across the 2022 and 2024 election cycles.

This four-stage cycle is the backbone of the Policy Data Brief you will build in this course.