Data Analysis for PolicyMaking in Georgia
Module 1 · The Data-Driven Policy Cycle › 1.4 Data types in policy work
Subsection 1.4

Data types in policy work

~3 min

Reading

Policy analysis draws on several types of data, each with distinct strengths, limitations, and access pathways. Knowing which type of data you need—and what you cannot learn from it—is a foundational skill.

Quantitative data

Numbers you can count, aggregate, and compare: voter registration counts, turnout rates, demographic percentages, district area measurements. Quantitative data is strong for identifying patterns at scale and detecting disparities. Georgia's voter file is a large quantitative dataset—millions of rows, each representing a registered voter with fields like county, registration date, and voting history.

Qualitative data

Text, interviews, observations, and narratives: community testimony about a polling place closure, a precinct worker's account of equipment failures, written comments submitted to a redistricting commission. Qualitative data provides context that numbers alone cannot supply. In redistricting litigation, community testimony about "communities of interest" is a form of qualitative evidence.

Administrative data

Records generated by government operations: voter files, election results certified by the Secretary of State, census block data used by redistricting commissions, mail ballot tracking files. Administrative data is usually large, systematic, and relatively standardized—but it reflects what agencies chose to record, which means gaps and inconsistencies are built in.

Survey data

Responses collected by sampling a population: exit polls, voter attitude surveys, polling on redistricting preferences. Survey data can capture things administrative data cannot (like why someone didn't vote), but it requires careful sampling design and carries uncertainty that administrative data does not.

Geospatial data

Data with a geographic component: census block boundary shapefiles, precinct boundary maps, polling place locations, congressional and state legislative district lines. Geospatial data is essential for redistricting analysis and for proximity analyses like "how far is the nearest polling place from this census block?"

Georgia examples for each type

  • Quantitative: Statewide voter registration file with ~8 million rows (sos.ga.gov)
  • Qualitative: Public comments submitted to the Georgia Senate Reapportionment Committee, 2021
  • Administrative: Certified 2024 general election results by county and precinct
  • Survey: Georgia exit poll data, Edison Research, 2024 general election
  • Geospatial: Georgia House and Senate district boundary shapefiles (Redistricting Data Hub)
Five Data Types in Policy Work — Georgia Examples Type What it captures Georgia example Quantitative Counts, rates, percentages — detects patterns at scale GA voter file, ~8M rows Qualitative Narratives & testimony — supplies context numbers cannot Reapportionment committee comments Administrative Govt. records — large, systematic, but reflects agency choices Certified 2024 election results Survey Sampled responses — captures attitudes & self-reported behavior GA exit poll, Edison Research 2024 Geospatial Place-coded data — district lines, polling places, proximity GA district shapefiles, Redistricting Data Hub Each type has distinct uses. Most policy analyses combine two or more.
Diagram 1.4 · Five data types in policy work. Each type has distinct uses. Most policy analyses combine two or more.