A python package for Experiment-as-Market (EXAM), an experiment design algorithm which:
-
produces a Pareto efficient allocation of treatment assignment probabilities;
-
is asymptotically incentive compatible for preference elicitation;
-
unbiasedly estimates any causal effect estimable with standard RCTs.
Please refer to the pseudocode of EXAM for more details of the algorithm.
This package is written and maintained by Yusuke Narita (yusuke.narita@yale.edu).
The latest release of the package can be installed with pip:
pip install experimentasmarket
Any published release can also be installed from source:
pip install git+https://github.com/zliu392/exam-project
The following example demonstrates how to apply EXAM in experiment design applications.
import experimentasmarket
Yusuke Narita. Toward an Ethical Experiment, available at [SSRN].