Data Analysis for PolicyMaking in Georgia
Module 6 · Writing the Policy Data Brief › 6.4 Policy recommendation
Subsection 6.4

Policy recommendation

~3 min

Reading

The policy recommendation is the final section of the brief, but it is not an afterthought. It is the reason the brief exists. Everything you have built—the research question, the data, the finding—points toward a specific thing that someone with power can do.

What makes a recommendation actionable

A recommendation is actionable when it:

  • Names the actor: The Georgia General Assembly should... / The Georgia Secretary of State should... / The U.S. Department of Justice should... A recommendation addressed to "Georgia" or "policy makers" is too vague.
  • Names the specific action: Amend SB 202 to require automatic notice-and-cure for mail ballot signature mismatch. Not: "improve the mail voting system."
  • Is scoped to what the data actually supports: If your finding is about mail ballot rejection rates in one election cycle, your recommendation should be about mail ballot procedures—not about redistricting, school funding, or anything else your data doesn't address.
  • Is realistically achievable: A recommendation that requires amending the U.S. Constitution is technically actionable but practically unhelpful for a one-page brief aimed at a state legislator.

What policymakers can actually do with your finding

In Georgia's context, the following actions are realistic given the current policy landscape:

  • State legislative action: amend Georgia election code (Title 21)
  • Secretary of State guidance: issue administrative rules on signature matching standards
  • U.S. DOJ: open an investigation or file a complaint under other VRA provisions or the Help America Vote Act (HAVA)
  • Litigation: file a Section 2 complaint (now with intent evidence post-Callais), an Equal Protection challenge, or a HAVA complaint
  • County election board: adopt internal policies (within state law)

Georgia redistricting post-Callais — recommendation example

Research question: Did turnout decline in Georgia precincts that were moved from majority-Black legislative districts to majority-white districts after 2021 redistricting?

Recommendation example: "The Georgia General Assembly should commission an independent analysis—using precinct-level turnout data and post-Callais intent-evidence standards—of the 2021 redistricting's effect on voter participation in majority-Black communities. If the analysis confirms a turnout decline attributable to district boundary changes, the Assembly should convene a remedial redistricting process with explicit evaluation of boundary options that maintain or restore minority voting opportunity, documented with the kind of historical and legislative record evidence required under Callais v. Landry."

Why this recommendation works

It names the actor (Georgia General Assembly). It describes a specific first action (commission an analysis) before a larger action (remedial redistricting). It is scoped to the data. It explicitly acknowledges the post-Callais legal standard, showing the policymaker you understand the current legal landscape.

Policy Recommendation — Weak vs. Strong Weak (do not write this) "Georgia should do something about mail ballot rejection rates. Voters deserve fair treatment." Problems: ✗ No named actor ✗ No specific action ✗ "Do something" is not a policy ✗ Not grounded in the finding ✗ Aspirational sentiment, not recommendation Strong (aim for this) "The Georgia General Assembly should amend Title 21 to require that county election offices automatically notify voters within 24 hours of a mail ballot signature mismatch and provide a 7-day cure window." Why it works: ✓ Named actor (GA General Assembly) ✓ Specific action (amend Title 21) ✓ Scoped to the signature mismatch finding ✓ Realistically achievable by a state legislature
Diagram 6.4 · Weak vs. strong policy recommendation. A strong recommendation names the actor, specifies the action, and is grounded in the finding.