Field experiments in economics

The relevance of field experiments in economics

In this paper John A. List discusses the increasing significance of field experiments in economics and their potential for future growth. He argues that field experiments, once thought impractical in economics, have gained traction as a vital tool for empirical research and as such they are not merely a passing trend but represent a fundamental shift in how economists approach causal inference and knowledge generation. List emphasizes that field experiments allow researchers to measure causal impacts of treatments and assess their generalizability across different contexts. This capability addresses two critical problems in experimental research:

• Experimental Problem 1: Measuring the causal impact of treatments and identifying relevant mediators and moderators.
• Experimental Problem 2: Predicting the transferability of causal impacts across various environments.

Methodological Framework

• Potential Outcomes Model: The author introduces the potential outcomes framework to establish causality in field experiments, contrasting it with traditional empirical approaches. This model helps clarify how treatment assignments influence outcomes.
• Types of Field Experiments: List categorizes field experiments into three types:
  • Artefactual Field Experiments (AFE): Conventional lab experiments with non-standard subjects.
  • Framed Field Experiments (FFE): Incorporate field context into AFE designs.
  • Natural Field Experiments (NFE): Conducted in natural settings where subjects are unaware they are part of an experiment.

Growth and Future Directions

• Historical Context: The paper traces the evolution of field experimentation in economics, highlighting three waves:
  • The initial wave focused on agricultural productivity without human subjects.
  • The second wave involved large-scale social experiments conducted by government agencies.
  • The current wave integrates laboratory controls into field settings, allowing for rigorous testing of economic theories.
• Recommendations for Future Research: List proposes four methodological paths for leveraging field experiments:
• • Understanding heterogeneity and causal moderation.
  • Exploring mediation paths (mechanisms).
  • Assessing generalizability and scaling of results.
  • Conducting ethically responsible designs.

Conclusion

The author concludes that field experiments hold great promise for advancing economic research. By embracing these methods, economists can generate valuable insights that inform policy and deepen our understanding of economic behaviors.

Field Experiments: Here Today Gone Tomorrow?
Author: John A. List
From: University of Chicago

An opinion formation experiment

Daniel Kashner and Mateusz Stalinski use field experiments to investigate how the order of information presentation influences opinion formation regarding political issues, specifically net neutrality. The study focuses on whether prior engagement with non-partisan information can prevent the formation of partisan divides in policy support.

Methodology

• Participants: The study involved 851 Republicans recruited through the Prolific platform for an online experiment about net neutrality, a topic that was largely unfamiliar to the participants.
• Experimental Design: The researchers conducted a pre-registered experiment where participants were randomly assigned to two groups based on the order of information they received:
• • Partisan-1st Group: Participants first received partisan information highlighting Republican opposition and Democratic support for net neutrality.
  • Partisan-2nd Group: Participants first viewed non-partisan information that presented factual arguments about net neutrality without any party affiliation.
• Outcome Measurement: The primary outcome measured was the amount donated to the Electronic Frontier Foundation (EFF), a charity advocating for net neutrality. Participants decided how to split a small bonus payment between themselves and the EFF after reviewing the respective information blocks.
• Follow-Up Study: To assess the persistence of the treatment effect, a follow-up study was conducted several weeks later, which included additional checks against placebo issues to confirm results.

The findings highlight that presenting non-partisan information before partisan cues can effectively mitigate political polarization on issues like net neutrality. This approach allows individuals to form their opinions based on factual content rather than party affiliation, potentially offering a pathway to reduce polarization in public discourse. However, the lack of similar effects among Democrats raises questions about the generalizability of these results and suggests that further research is needed to explore these dynamics across different political affiliations and contexts.

Preempting Polarization: An Experiment on Opinion Formation
Authors: Daniel Kashner, Mateusz Stalinski
From: University of Chicago, University of Warwick