EUVIMEDEUVIMED entdecken
Mixed-Methods-StudieOpen Access

Expanding the Data Science Toolkit: Mixed Methods for Improving Causal Inference

Noor Qaragholi, Elizabeth A Stuart

Harvard Data Science Review · 31.10.2025 · 7 (4)

DOI: 10.1162/99608f92.2eb6655b ↗

Kurz gefasst

Mixed-Methods-Studie in der Zeitschrift Harvard Data Science Review, online erschienen am 31.10.2025. Verfasst von Noor Qaragholi und 1 weiteren Autor:innen. Beteiligte Einrichtung: Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland. Die Arbeit wurde bisher 1-mal zitiert (Crossref). Der Volltext ist frei in PubMed Central verfügbar. Gefördert von: NIMH NIH HHS.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Causality experts from different disciplines gathered at the Museo Galileo to discuss the field's past, present, and future, as well as hot topics in causality and causal inference in the new era of data science.Hosting the symposium at the Museo Galileo was especially fitting, as Galileo's pioneering work in experimentation, observation, and the search for underlying causal laws laid the very foundations of modern scientific reasoning.In January 2024, HDSR launched a new column called Catalytic Causal Conversations.The column's description, in the words of column co-editors Iavor Bojinov and Francesca Dominici, reads: Causal inference is a pivotal element in disciplines focused on unraveling the cause-and-effect relationships that underpin outcomes.The 'Catalytic Causal Conversations' column provides approachable insights and varied perspectives on the evolving landscape of causal inference.Its primary objective is to clarify and explore the complexities of causal inference, thereby making it relatable and compelling for a broad audience, and to showcase cutting-edge methods and practical applications in this ever-evolving field.In this issue of the journal, we are announcing a special series of Catalytic Causal Conversations column articles that aims to dive a bit deeper into some of the matters discussed at the Symposium on Causality.This article, authored by Noor Qaragholi and Elizabeth A. Stuart, is the first of the special series.The authors argue for integrating qualitative and quantitative research-mixed methods-into causal inference to enhance the validity and interpretability of findings in complex policy, health, and social research.It highlights that causal inference often relies on untestable assumptions and that qualitative insights can help identify relevant causal questions, clarify mechanisms, and assess confounding.Through examples like Rosenbaum and Silber's (2001) hospital mortality study and McGinty et al.'s (2018) opioid policy evaluation, Qaragholi and Stuart demonstrate how mixed methods can refine definitions, improve data interpretation, and strengthen causal conclusions.The article also explores emerging directions, including using generative AI for qualitative data analysis, balancing prespecification and flexibility in research design, and expanding training for future data scientists to integrate mixed-methods thinking.Ultimately, it calls for bridging disciplinary divides to build more credible, contextually grounded causal science.More column articles are coming!They will discuss key topics of causal inference, including foundational, historical, and policy considerations.Happy reading!

Schlagwörter

Stichwörter der Autor:innen

causal assumptionsevaluationmixed methodsqualitative research

Autor:innen und Institutionen

  • Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland.
  • The Policy & Research Group, New Orleans, Louisiana.
  • Department of Biostatistics, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland.
  • Department of Mental Health, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland.
  • Department of Health Policy and Management, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland.

Förderung

  • NIMH NIH HHS

Zitieren

APA

Qaragholi, N., & Stuart, E. A. (2025). Expanding the Data Science Toolkit: Mixed Methods for Improving Causal Inference. Harvard Data Science Review, 7(4). https://doi.org/10.1162/99608f92.2eb6655b

Vancouver

Qaragholi N, Stuart EA. Expanding the Data Science Toolkit: Mixed Methods for Improving Causal Inference. Harvard Data Science Review. 2025;7(4). doi:10.1162/99608f92.2eb6655b

BibTeX
@article{Qaragholi2025,
  author = {Qaragholi, Noor and Stuart, Elizabeth A},
  title = {Expanding the Data Science Toolkit: Mixed Methods for Improving Causal Inference},
  journal = {Harvard Data Science Review},
  year = {2025},
  volume = {7},
  number = {4},
  doi = {10.1162/99608f92.2eb6655b}
}
Kostenlos starten →