Ratings that are trusted by voters across the political spectrum
Candidates receive an overall rating based on their scores, making it easily understandable by the average voter.
Our rating process is 100% non-partisan. We are committed to complete transparency around how the scores are derived.
Categories of divisive content
There are five types of divisive language:
“Us vs. them” framing - Accusing others of nefarious intentions and framing them as the enemy
Tribalism - Judging a policy based on who is associated with it, rather than on its merits
Labeling - Applying an ideological label to a policy to induce a judgment about it without assessing its benefits or drawbacks
Trigger words - Using specific words designed to trigger an emotional response and stoke outrage
Civil ≠ Centrist
A candidate’s rating is based solely on the language they use and is not impacted by their ideology, no matter how extreme. They can be conservative or progressive, extremist or moderate.
Image: Daryl Davis, African-American R&B and blues musician, worked to improve race relations by personally engaging with KKK leader Roger Kelly and other Klan members.
Focus on the primaries
We do not expect voters to cross party lines. However, when given a choice between different candidates with similar policy positions in their own party’s primary, we believe a significant number of voters will rally behind the one that is trying to get things done, not tear the country apart.
Join the team
Are you interested in volunteering to help? We are looking for motivated people to help us flag divisive posts and public statements, and manually train our machine-learning system.