Randomized experiments are one of the most reliable ways to investigate causality. However, their use is limited because a lack of foundational econometric and statistical theory. Current theory requires that studied causal factors are simple and that potential interactions between participants are known and well-structured. Many pressing questions in the social sciences do not conform to these requirements, forcing researchers to use less reliable empirical methods.

This project aims to develop methods and theory that will facilitate the design of randomized experiments in unconventional and complex settings. The primary goal is to develop experimental designs that allow for identification and unbiased estimation of causal effect in the wide range of settings researchers face in the social sciences. A secondary goal is to develop designs that provide maximum precision by reducing the unsystematic estimation error.

There are three parts of the project:

1. Experiments with complex causal factors, including continuous treatment values and settings with intricate interactions between participants.

2. Experiments with severe external restrictions due to ethical considerations or concerns of stakeholders.

3. Adaptive experimental designs in settings with continuous recruitment of participants when detailed background information is available.

This is an interdisciplinary project, combining and advancing insights from the social sciences, statistics, mathematics and computer science.