How we calculate Bridge Grades

The detailed methodology we use to calculate letter grades that sort Congressional bridgers from dividers in both parties

As July comes to a wrap, so does Congress whose members flee DC and return to their home districts for their summer recess. Meanwhile, we’re busy compiling the final data from the first seven months of their term, and will soon¹ reveal our first look at the Bridge Grades for this 119th Congress.

In advance of the reveal of these mid-term report cards, now’s the perfect time to get into exactly how this whole Bridge Grades² thing works.

What Bridge Grades measures

In collaboration with IU Innovate (part of Indiana University’s Luddy School of Informatics, Computing, and Engineering), we’re building an open source system described below.

We compile public data from credible and trusted 3rd party sources that measure what a person says (rhetoric: public statements, speeches, social media), and what a person does (legislative record: authoring and sponsoring bipartisan bills).

We process publicly available data along four core collaboration dimensions

1. Consensus solutions (legislative record: bipartisan legislation)

2. Coalition building (legislative record: working with others, builds cross party alliances)

3. Civil discourse (rhetoric: unifying vs dividing rhetoric)

4. Courage (bridging even when it would be easier not to)

---

Detailed Methodology

1. Aggregate datasets from multiple public sources (listed below) for each member of Congress on their legislative records and rhetoric:

   1. Count of the number of bills they’ve written that earns co-sponsorship from a member of the opposite party

   2. Count of the number of bills they’ve co-sponsored that were written by a member from the opposite party

   3. Count of number of public statements they’ve made about bipartisanship topics

   4. Share of public statements made that were about bipartisanship topics

   5. Count of number of public statements made that were characterized as personal attacks

   6. Share of public statements made that were characterized as personal attacks

2. Normalize each dataset to a 0-100 distribution

3. Apply weights to each normalized distribution (principle: weight of legislative record exceeds weight of rhetoric)

4. Sum the weighted scores into a subtotal

5. Add courage bonus (degree of difficulty multiplier) to reward bridging behavior from members who represent strong partisan leaning districts

6. Add centrist bias adjustment (courage bonus) to reward bridging by ideologues

7. Add coalition building bonus (House Reps only) for members of the bipartisan Problem Solvers Caucus

8. Sum final totals and normalize into a 0-100 scale (the Bridge Score)

9. Apply a forced grading curve and assign Bridge Grades for each member. The top 50% (bridgers) earn As and Bs, while the bottom half (dividers) earn Cs and Fs. One standard deviation away from the mean are As and Fs.

---

Our data sources

Bipartisan legislative record powered by Open Plural and sourced through govtrack.us.

• Earn points for authoring legislation that gets sponsored by the other party

• Earn points for sponsoring legislation that was authored by the other party

---

Rhetoric analysis powered by data from America’s Political Pulse (from Polarization Research Lab)

• Earn points for public comments about bipartisanship

• Lose points for making public comments that are personal attacks

---

Coalition points powered by problemsolverscaucus.house.gov

• Bonus points for members of bipartisan Problem Solvers Caucus (House only)

---

Courage bonus powered by Cook Political Partisan Voting Index (Cook PVI)

• Degree-of-difficulty multiplier applied for bridging in partisan districts (House) or states (Senate).

---

Courage bonus powered by voteview (hosted by UCLA's Department of Political Science and Social Science Computing)

• Degree-of-difficulty multiplier for bridging when far from the ideological center to mitigate centrist bias.

---

Notes:

• The grading rubric (data sources used and relative weights applied to each source) will evolve and improve over time as additional objective and reliable data sources become known and available.

• Bridge Grades are assessed specifically and uniquely for each term served. All scores were reset to zero on January 3, 2025 at the beginning of the 119th Congress.

• Legislative record and rhetoric scores are based only on observable behaviors that have occurred specifically during their term of service (from January 3, 2025 to date). Past data and historical records are ignored.

• Remember that each set of grades and scores are relative to one’s peers, and are in no way comparable against past Congresses.

---

Thank you to our collaborators

Hat-tip and kudos to the many contributors whose fingerprints are on this most current version of our Bridge Grades methodology. This system is a collaboration with IU Innovates (Indiana University’s Data Science as a Service) thanks to program sponsor Dr. David Wild at Luddy, program design mentor Kyle Stirling, Maria Aroca, and more than a dozen Masters candidates across two cohorts (Liz, Andy, Tyler, Siddharth, Utkarsh, Sai, Paul, Jason, Jaimie, Katrina, Bryant, Sanjana, Joy, Abubakar). Also thanks to Dr. Royce Carroll for poly-sci sanity checks and Rich Hansen for our very first alpha model. Special thank you to Universidad del Norte, Data Science undergraduate phenoms Emanuel Carbonell Naranjo and Zharick Molina for stitching it all together under Maria Aroca’s exemplary leadership and oversight.

---

Bridge Grades is like Rotten Tomatoes for Congress. Our non-partisan polarization report card scores members of Congress on how collaboratively or divisively they govern. We aggregate objective 3rd party data and compile it into a scoring system that sorts legislative “bridgers” from “dividers.”

Because political ideology and collaboration are independent variables.

1. Target: 3rd week in August.

2. Bridge Grades is like Rotten Tomatoes for Congress. Our non-partisan polarization report card scores members of Congress on how collaboratively or divisively they govern. We aggregate objective 3rd party data and compile it into a scoring system that sorts legislative “bridgers” from “dividers.”

• Bridgers build win-win consensus solutions through cross-partisan coalitions and collaboration for the benefit of common interests

• Dividers pursue zero-sum game governance, engage in personal attacks, and demonstrate predictably partisan legislative records