Column Editor’s Note: Causal inference is a unifying framework across disciplines because most fields care about understanding cause-effect relationships that drive outcomes. Critical decisions in medicine, life sciences, public policy, business, or other areas depend on quantifying the causal impact of interventions and exposures. In this article, we announce the launch of a new column on causal inference. The column, titled “Catalytic Causal Conversations,” will have a consistent format to provide readers with a comprehensive yet accessible and enlightening overview of emerging topics in causal inference.

Keywords: causal inference, experimentation, interdisciplinary

We are excited to announce the launch of a new column on causal inference in Harvard Data Science Review. This column, titled “Catalytic Causal Conversations,” aims to broaden the reach of causal inference topics by providing accessible overviews, discussions, and thought pieces in this rapidly evolving field. Readers should tune in to learn about the latest methodological developments in causal inference and their implications across diverse domains such as healthcare, education, and environmental policy. By synthesizing current advancements with foundational concepts, we hope to keep the data science community engaged and spark further research by identifying open questions.

The new column is supported by the Harvard University Causal Inference Working Group, consisting of Iavor Bojinov,1 Francesca Dominici,2 Kosuke Imai,3 Luke Miratrix,4 and José Zubizarreta,5 which was established in 2022 to unite the causal inference research communities across Harvard University and encourage interdisciplinary collaborations and discussions of emerging work in causal inference. Initially, the Catalytic Causal Conversations column will build upon the success of a seminar series by the Causal Inference Working Group that was supported by a generous grant from the Alfred P. Sloan Foundation by creating accessible educational materials connected to the research presented. We hope to expand the content of the seminar and invite researchers to contribute to the column, sharing their work and diverse perspectives.

The column will have a consistent format to provide readers with an accessible and enlightening overview of emerging topics in causal inference. Each column article will briefly introduce the research area and discuss applications across different domains, highlighting why this topic matters. The bulk of the column will focus on synthesizing the critical contributions of recent work in this area and providing concrete examples to elucidate the methods and findings. Notably, the column will also discuss open questions that still need to be addressed and future directions for research on the topic of causal inference. The goal is to give readers a big-picture understanding of the research landscape and specific insights into state-of-the-art techniques and remaining challenges. The column will engage experts and non-experts by clearly explaining the latest advancements and unknowns.

Sometimes, the column article will led by a postdoctoral researcher or a PhD student under the guidance of at least one faculty co-author. This approach provides an essential professional development opportunity for early career researchers to hone scientific communication and outreach skills. Synthesizing topics at the frontier of causal inference into an accessible narrative requires a deep understanding of the field. Postdocs and students will gain valuable experience conveying complex technical material to broad audiences by taking the lead in writing the column articles. Cultivating these opportunities for future leaders in causal inference is a central goal of the Catalytic Causal Conversations column. The insights gained will serve these individuals well in future careers, and the entire field will benefit from their ability to clearly explain emerging developments.

What Is Causal Inference?

Causal inference is the science of determining cause-effect relationships from data. When we observe patterns and correlations, a key question is whether one factor causes another. There are several excellent introductory books on causal inference from biostatistics (Hernán & Robins, 2024), computer science (Pearl 2000), economics (Angrist & Pischke, 2008), and statistics (Imbens & Rubin, 2015; Rosenbaum, 2017) perspectives, as well as many review articles; see for example Bojinov et al. (2020), Dominici et al. (2021), Hernán et al. (2019), Holland (1986), Li et al. (2023), Pearl (2009), Rubin (2005), and Stuart (2010).

In the context of causal inference for determining cause-effect relationships from data, a couple of questions of interest are: 1) Does attending preschool improve a child's academic outcomes later in life? 2) Does eating more vegetables reduce someone's risk of certain diseases? Causal inference provides principled techniques to answer these questions and quantify the causal effects of a given intervention on an outcome. By leveraging observational data in combination with experimental data and making appropriate assumptions, causal inference methods allow researchers to disentangle causation from mere association. The field has seen significant growth recently, driven by innovations at the intersection of statistics, computer science, and various application domains.

Why Is Causal Inference Cross-Disciplinary?

Causal inference is a unifying framework across disciplines because most fields care about understanding cause-effect relationships that drive outcomes. Critical decisions in medicine, life sciences, public policy, business, or other areas depend on quantifying the causal impact of interventions and exposures. For example, marketers want to know what campaigns increase sales, not just if they are correlated. Regulators need to determine if a new policy indeed improves social welfare. Artificial intelligence researchers need to and should ascertain whether patterns in data reflect accurate causal mechanisms that can be exploited in predictive systems. While the specific applications differ, the fundamental quest to infer causation from data ties together problems across diverse domains. Causal inference provides a shared set of tools and language for tackling these universal challenges.

By reading these Catalytic Causal Conversations column articles, you will learn and appreciate how methods and applications of causal inference ideas can greatly impact many disciplines of scientific inquiry.

Acknowledgments

The creation of the Causal Inference Working Group was supported in part by a generous grant from the Alfred P. Sloan Foundation.

Disclosure Statement

Iavor Bojinov and Francesca Dominici have no financial or non-financial disclosures to share.

References

Angrist, J. D., & Pischke, J.-S. (2008). Mostly harmless econometrics: An empiricist’s companion. Princeton University Press. https://doi.org/10.2307/j.ctvcm4j72

Bojinov, I., Chen, A., & Liu, M. (2020). The importance of being causal. Harvard Data Science Review, 2(3). https://doi.org/10.1162/99608f92.3b87b6b0

Dominici, F., Stoffi, F. J. B., & Mealli, F. (2021). From controlled to undisciplined data: Estimating causal effects in the era of data science using a potential outcome framework. Harvard Data Science Review, 3(3). https://doi.org/10.1162/99608f92.8102afed

Hernán, M. A., Hsu, J., & Healy, B. (2019). A second chance to get causal inference right: A classification of data science tasks. CHANCE, 32(1), 42–49. https://doi.org/10.1080/09332480.2019.1579578

Hernán, M. A., & Robins, J. M. (2024). Causal inference: What if. CRC Press. https://www.hsph.harvard.edu/miguel-hernan/causal-inference-book/

Holland, P. W. (1986). Statistics and causal inference. Journal of the American Statistical Association, 81(396), 945–960. https://doi.org/10.2307/2289064

Imbens, G. W., & Rubin, D. B. (2015). Causal inference for statistics, social, and biomedical sciences: An introduction. Cambridge University Press. https://doi.org/10.1017/CBO9781139025751

Li, F., Ding, P., & Mealli, F. (2023). Bayesian causal inference: A critical review. Philosophical Transactions of the Royal Society A, 381(2247), Article 20220153. https://doi.org/10.1098/rsta.2022.0153

Pearl, J. (2009). Causal inference in statistics: An overview. Statistics Surveys, 3, 96–146. https://doi.org/10.1214/09-SS057

Pearl, J. (2000). Causality: Models, reasoning and inference (1st ed.). Cambridge University Press. https://doi.org/10.1017/S0266466603004109

Rosenbaum, P. (2017). Observation and experiment. Harvard University Press. https://doi.org/10.4159/9780674982697

Rubin, D. B. (2005). Causal inference using potential outcomes: Design, modeling, decisions. Journal of the American Statistical Association, 100(469), 322–331. https://doi.org/10.1198/016214504000001880

Stuart, E. A. (2010). Matching methods for causal inference: A review and a look forward. Statistical Science, 25(1), 1–21. https://doi.org/10.1214/09-sts313

©2024 Iavor Bojinov and Francesca Dominici. This article is licensed under a Creative Commons Attribution (CC BY 4.0) International license, except where otherwise indicated with respect to particular material included in the article.