How we used data to select our Priority states for the 2018 elections

Bringing startup scale to progressive and centrist politics

Last week, we published Tech for Campaigns’ 2018 Priorities list for State Legislatures, detailing the fourteen states we are focused on for state legislative races. This week, we wanted to zoom out — and in — on how TFC established and is executing on its 2018 strategy. (Plus, I’ll be hosting a Facebook Live on April 4th at 5PM to chat more about it — check it out!)

In 2018, TFC is aiming to allocate its 70% of its volunteer efforts to state legislature races and 30% to federal and statewide races (U.S. House, U.S. Senate, Governor, Lt. Governor, etc.) These categories include thousands of campaigns, so sorting through them is no trivial task!

For federal and statewide races, between outlets like FiveThirtyEight and The Cook Political Report, there is no shortage of predictions and assessments. For state legislatures, though, there is no canonical public accounting of different races nationwide. Given our focus on helping hundreds of Democrats across multiple states in 2018, we needed to establish our own point of view and organizational framework — and data helped us do it.

How we did it

Being techies, we turned to data and predictive modeling to help us out. Our Build The List volunteer project gathered data from a variety of formats across government and commercial/nonprofit sources, normalized it, and used features like demographics, past election results, and incumbency to predict Democratic and Republican results in each district for 2018. The results provide Tech for Campaigns with a single baseline to evaluate and prioritize races across states and begin to layer in qualitative assessments.

Two main constraints on the list were finding routes to influence the 2020 redistricting process on way to flipping the U.S. House and TFC’s volunteer model. These drove us to de-prioritize state legislatures with only a single U.S. Congressional seat, state legislatures with multi-member election methods, state legislatures that have minimal influence over redistricting, and other factors.

The results

From this process, we emerged with fourteen 2018 Priority states for state legislative races, covering four key objectives:

• Priority Flip: states where our modeling indicated that Democrats have a significant shot of retaking at least one state chamber in 2018;
• Potential Flip: states where our modeling indicated that favorable trends are in place to help make significant gains in the 2018 elections on the way towards a flip in 2020;
• Defends: states where a chamber has a thin Democratic majority; and
• Break Supermajorities: states where the Republican party controls a supermajority in State Senates, giving them veto-proof control over budgets and legislation without the natural consensus achieved by needing some Democratic votes.

The predictive modeling process was key to informing the likelihoods of success of each of these four objectives, driving the selection of eight out of ten of our non-exploratory 2018 states.

Prioritizing individual races

With our initial 2018 State Prioritization complete, our focus will turn to our work with state parties and caucuses and prioritizing individual races. The modeling process is again key; it provides an excellent baseline to layer on qualitative assessments to determine the 10–15 races in each state where TFC volunteers would be best positioned to help.

For non-prioritized races, we will be offering tools and materials any progressive campaign can use to upgrade their digital presence.

Going forward

Tech for Campaigns was founded just over one year ago, in February 2017. Since then, we’ve completed 30+ projects and helped eight seats in the Virginia House of Delegates.

The modeling, inputs, and framework established in this process provide Tech for Campaigns with a repeatable organizational framework and baseline understanding of progressive political chances nationwide. Our prioritization process will be the center of our efforts for years to come as we build the digital arm for progressive politics in the United States.