Sunlight Best Friends

Project Overview

Very brief version

How does influence work in politics?

We think that money changes how members of congress do this stuff over time.

That could mean they do stuff in the presence of money that they wouldn't have done otherwise, or it could mean they don't do things in the presence of money that they would have otherwise.

There are essentially four ways members of congress work towards legislation where we can easily collect data:

• They introduce bills
• They co-sponsor bills
• They speechify on the floor
• They vote for bills.

What we want to observe are the ways in which money changes how they do these things.

Slightly less brief version

Our basic theory is that members' decisions on which legislation to introduce and co-sponsor, how (and how much) to speak about legislation on the floor, and how to vote are the result of many factors, but one of these factors is how much money they receive from different interests. While it is pretty much impossible to assess money's influence on members on any one piece of legislation, our belief is that over time, we can see money have a gravitational pull on how members behave. And that often, what members don't do is as important to understand as what they do do.

The big challenge is pulling all the relevant data into properly structured forms, and then using these data to develop models.

Social scientists who have tried to measure influence have struggled because, to paraphrase Weber, influence, like politics is “a strong and slow boring of hard boards.” Influence relationships take time to build, and members of Congress are often cross-pressured by multiple interests. Influence must be modeled as a never-ending multi-dimensional tug of war, not a vending machine of simple transactions. We ought to think of influence as a physics problem, with time being a key dimension.

This is admittedly an ambitious undertaking. But it also has the potential to be truly groundbreaking. It is also an undertaking that is only now becoming possible because only now are the data available and in properly structured forms.

Resources

Code & Data

• Github Repo for project-related code
• The Overview Project, a text-mining application created by the AP to explore large bodies of text
• Sunlight Labs Hub for the Sunlight Foundation's data apis
• Sunlightlabs legislators dump
• OpenSecrets OpenData Portal for OpenSecrets.org's OpenData Initiative
• Govtrack.us bulk data
• Information about salaries, travel, financial disclosures, and foreign gifts for top Congressional staff

Reference Info

• How Laws are Made THOMAS' comprehensive guide to the legislative process
• Sunlight's training resources

People

• Jake Porway (jake@datakind.org), DataKind Sensei
• Cathy O'Neil (cathy.oneil@gmail.com), Data Ambassador
• Lee Drutman (ldrutman@sunlightfoundation.com), Sunlight Liason
• Peter Darche (pdarche@gmail.com), contributor
• Kevin Wilson, contributor

Tasks

• Legislator Timeline
• Follow the Money
• Follow the Policy Proposal
• Follow the Legislation
• Follow the Lobbyists
• Follow the Speeches
• Follow the Votes

• Model the Industries
• Model the Influence

Additional Project Ideas

Have a great idea that isn't part of one of the above projects? Post it here. Dig into a dataset and come across some really interesting patterns or questions? Post them here.