Announcing Our New Turnout Projection Tool

You’re helping a city council candidate plan their final two weeks. The district leans slightly against you in voter registration, but early signs point to lower-than-usual turnout — maybe around 55 to 60%.

The question: If turnout really is that low, what will the electorate actually look like?

That’s where the Turnout Projection Tool comes in.

Built on the unique voting history of each district, the tool models how the partisan makeup of the electorate is likely to shift as turnout rises or falls. You just enter a projected turnout rate — say, 58% — and it estimates what share of actual voters will come from each party in that specific district.

It’s not a prediction or a guarantee. But it gives you a grounded, data-driven way to test your assumptions — whether you’re running a campaign, weighing in on a local race, or trying to understand how turnout could shape the outcome.

This new feature is now live for all subscribers. For non-subscribers, we invite you to explore it on any of our free sample jurisdiction pages. Just navigate to a jurisdiction and click the "Registration & Turnout" tab to find the interactive widget by clicking on "Projection Tool."

How It Works: A Look Under the Hood

Our tool doesn’t use a generic, one-size-fits-all formula. For every single jurisdiction, we build a custom multiple regression model from scratch, training it on the unique political behavior of that district.

Specifically, we teach the model to predict the turnout rate for each political party (Democrats, Republicans, and Others). That’s possible because The Ballot Book includes detailed, jurisdiction-level registration and turnout data for every California election since 2014 — giving the model a deep and reliable foundation to learn from.

To make these predictions, the model analyzes how party turnout rates respond to three key factors observed in every primary and general election since 2014:

• The Overall Turnout Rate: This is the main driver. How does a party's turnout rate change when overall participation is 30% versus 70%?
• The Election Year: This allows the model to capture long-term trends. Is a party's base in a district becoming steadily more engaged over the past decade, independent of overall turnout? The year helps account for that.
• Presidential vs. Midterm Cycles: Voter behavior is fundamentally different in a presidential election versus a midterm. The model learns these separate patterns to make its projections more context-aware.

Once the model projects the turnout rate for each party, it calculates the Projected Turnout Share — the metric that shows you who is likely to have the advantage among the voters who actually cast a ballot.

How to Use the Tool

Using the Turnout Projection Tool is simple. First, select the type of election you’re analyzing — either a general election or a primary/special election. These election types behave differently, so choosing the right one is essential for an accurate projection.

Next, enter your estimated overall turnout percentage. For example, if you believe 58% of registered voters will participate, just type in “58.”

Then click “Project.” The tool will instantly generate estimates showing what share of voters are likely to come from each political party in that specific jurisdiction — based on historical turnout patterns and the model’s district-specific calculations.

Why Is This Valuable? Moving from Hunches to Strategy

This tool allows you to game out scenarios and stress-test common political theories:

For Campaigns: Quantify your GOTV (Get-Out-The-Vote) goals. Instead of just saying "we need high turnout," you can see that, for example, "unless turnout exceeds 45%, our candidate's path to victory is narrow."

For Journalists & Pundits: Add a layer of data-driven context to your reporting. Instead of relying solely on registration numbers, you can explore how different turnout scenarios might shape the electorate, providing a richer, more nuanced view of a race's dynamics.

For Political Junkies: Finally get a real, data-informed answer to those "what if" questions that dominate political discussions. See for yourself how much turnout matters in the races you care about most.

Knowing the Limitations: A Baseline, Not a Crystal Ball

This tool provides a powerful baseline for analysis by processing historical data, but it is not a crystal ball. Professionals will want to layer their own judgment on top of these projections. The model is a starting point, not the final word.

Here's what it doesn't do:

• It Does Not Predict Specific Events: The model learns from past patterns. It cannot anticipate how turnout might be influenced by future variables—like a uniquely motivating (or demotivating) candidate at the top of the ticket, a high-profile ballot measure, or a dramatic shift in the political landscape.
• It's About Turnout, Not Persuasion: The model projects which registered party members will show up; it does not predict if they will cross party lines to vote for a different candidate.
• Certainty Decreases at the Extremes: The model is most confident within the historical turnout range for a district. If you test a very low (e.g., 5%) or very high (e.g., 95%) turnout, the tool is extrapolating into uncharted territory. It will still provide a plausible estimate, but as the on-screen message warns, the results are less certain.

Give It a Spin

We believe this tool provides an invaluable new perspective on election dynamics. Head over to any sample jurisdiction page to start exploring.

We're excited to hear what you discover.