🔍 What was tested

Group discussion in legislative caucuses and its effect on policy coalitions—specifically whether talking about bills in a bipartisan caucus shifts legislators' support and changes legislative outcomes.

🧪 How the experiments were run

• Two field experiments conducted with state legislators.
• In caucus meetings, legislators spoke about randomly selected bills (treatment) while other bills were not discussed (control).
• Conversations took place in bipartisan settings to observe within- and cross-party influence.

📈 Key findings

• Chosen bills received roughly twice as much support from caucus members as unselected bills.
• Estimated effects were similar in magnitude for discussions within the same party and across party lines.
• Discussion did not appear to change bill-level outcomes such as bill content or passage, although limited statistical power makes detecting effects on these policy-level outcomes infeasible.

🧩 What explains the influence

• Qualitative data from the experiments point to reputation and reciprocity as central mechanisms that help convert discussion into increased support.
• However, partisan voters and media pressure were identified as obstacles to sustained bipartisanship in the contemporary American legislature.

💡 Why it matters

These results demonstrate that caucus talk can substantially shape legislators' expressed support and coalition-building, even when measurable changes in legislation or passage are not detected. The findings highlight how interpersonal norms (reputation, reciprocity) operate inside legislatures and how external partisan forces can limit the translation of caucus consensus into policy change.