Emil Niclas Meyer-Hansen
23/5/2025
The Synthetic Causal Framework (SCF) is a novel methodological framework for deriving reliable and valid causal inference for non-manipulable phenomena through use of synthetic units. These units are generated by synthesizing simulated data (i.e., units with simulated characteristics proportional to the population of interest) with real-world data (i.e., human characteristics expressed in text) in the form of life-histories generated by Large Language Models (LLMs).
The research paper produced from this project, The Synthetic Causal Framework: An LLM-based Solution to the Fundamental Problem of Causal Inference (newest version published 18th March, 2025), reviews and develops the theory underlying this procedure and validates its reliability and validity against robust causal results derived from human populations. This paper subscribes to the open science standard and is made freely available.
Materials are also made available on its Open Science Framework (OSF) project page (DOI: 10.17605/OSF.IO/5P3M2).
Except where otherwise indicated, all contents of this document and associated files are licensed under the Creative Commons Attribution 4.0 International License (CC BY 4.0). All software, source code, executable code, code snippets, code chunks, algorithms, and/or scripts within this document and associated files are expressly excluded from the foregoing license, and unless otherwise indicated, are instead licensed under the GNU General Public License, version 3 (GPL-3.0). By engaging with this document and/or any associated files, which include, but are not necessarily limited to, downloading, using, and/or distributing any of them, in parts of whole, you agree to comply with the applicable license terms for the respective content types.