The data is here! The data is here!
Finally, the detailed numbers needed for redistricting were made available Thursday afternoon, with the U.S. Census Bureau hosting a “data reveal party” (aka a news conference). The data—normally required to be delivered by March 30 but delayed because of the coronavirus—will tell us how many people live in every nook and cranny of the nation, along with some basic characteristics: voting age, race and ethnicity (Hispanic or not Hispanic). And we’ll have access to the current population of every city, town and hamlet.
For redistricters, the data drop is just the beginning of the redistricting cycle. Data keepers will need to take a few steps first. So, even though we’ve been in a “hurry up” phase, we’ll have to wait a bit longer before the policymakers can start drawing maps.
Even before the data arrived, anyone could look at the new “ideal population” numbers for their state’s electoral districts—congressional, state house and state senate. You can see these at Election Data Services (Table 5 on page 19), or you can do it yourself: Divide the “total resident population” by the number of members in the target chamber. The result is the ideal population for each district for this coming decade. (Note that the denominator is NOT total state population, the headline number released by the bureau in April and used to divvy up congressional seats among the states. That number includes overseas and military citizens. Redistricting is done with the slightly smaller total resident population.)
With the numbers now in hand, here’s what may be happening in your state in the coming days and weeks:
Step one: The data will be retrieved from the bureau’s portal and dumped into the state’s database.
Step two: The database will be tested to ensure that everything is aggregating and disaggregating correctly. Better to find glitches now rather than once mapmaking has begun.
Step three: Census tracts from 2010 will be compared with current ones. This comparison gives a detailed look at what parts of each state are growing most rapidly and where the population may be shrinking. (Census tracts change very little over the decades, whereas census blocks do change.)
Step four: The same will be done for incorporated areas and counties. Did they grow as expected? If any are way off from previous data or from expectations, they’ll get a closer look. Maybe that city increased its boundaries to take in some substantial suburbs—or maybe there’s an error.
Step five: Finally, it’s time to look at census blocks—and to stay calm. Some will look odd, especially those with populations under 250. Ron Jarmin, the Census Bureau’s acting director, wrote on July 28 that some blocks may show occupied housing units but a population of zero; others may show residents under 18 without any adults; and still others may show dozens of people but less than a handful of housing units. These are known oddities and result from the bureau’s decision to use differential privacy to protect respondents’ information. “Though unusual, situations like these in the data help confirm that confidentiality is being protected,” Jarmin wrote.
On that last point, aggregating the blocks into districts will address these oddities: “Instead of looking for precision in an individual block, we strongly encourage data users to aggregate, or group, blocks together,” he wrote. “As blocks are grouped together, the fuzziness disappears. And when you step back with more blocks in view, the details add together and make a sharp picture.” Redistricters will be testing this approach for sure.
Only after all this (and quite a bit more data preparation) is done can the real work of redistricting begin. For legislators, it may seem tough to wait for staff to run the data through these paces—but it’s better than finding problems later in the process.
Wendy Underhill is director of NCSL’s Elections and Redistricting Program.