Regression Discontinuity Approaches to Evaluate Threshold-Guided Interventions in Observational Studies

arXiv:2508.03878 · stat.ME, econ.EM, stat.AP · Submitted 2025-08-05 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Statistics. The day in statistics: methods, theory and the inference under them.

Tom: We're looking at the paper: "Regression Discontinuity Approaches to Evaluate Threshold-Guided Interventions in Observational Studies".

Jane: Here's the gist of the paper. The Regression Discontinuity (RD) design is a research strategy to estimate the causal effect of a treatment on an outcome.

Tom: This design is crucial when random assignment of interventions is not possible, as it removes the effect of unobserved confounders by focusing only on units whose scores are close to the cutoff.

Jane: The Core Mechanism of RD The RD design functions by assigning treatment based on a threshold rule where units receive treatment if their score is equal to or above a known cutoff, but not if their score is below it.

Tom: This strategy overcomes the challenge of observational studies where comparability between groups is not guaranteed because it focuses only on units whose scores are near the cutoff, assuming they are ex-ante comparable regardless of treatment assignment.

Title and authors: Tom: So we’re looking at this paper today, titled "Regression Discontinuity Approaches to Evaluate Threshold-Guided Interventions in Observational Studies," and it’s written by Matias D. Cattaneo and Rocío Titiunik. Jane, what do you see in the title that makes you think this is a big deal right now?

Jane: Well, it sounds really focused on using this regression discontinuity design to study interventions where there's a clear threshold, like if your score crosses a certain number. It suggests they are applying these methods to observational studies where we don't have random assignment, and they are trying to figure out the causal effect of that treatment.

Lu: What I find interesting is how the paper sets up the distinction between sharp and fuzzy regression discontinuity designs, which basically tells you how much you can actually claim about what you’re measuring. It’s not just one simple way to do it; they are looking at different levels of identification.

Meng: So, if we’re talking about a real-world application, what does this mean for us in terms of actually deploying these methods? Are we talking about something that works reliably outside of a controlled lab setting?

Lalam: Based on the title and the authors Cattaneo and Titiunik, it seems they are aiming to provide a practical roadmap for using regression discontinuity designs when you're trying to estimate treatment effects in real-world scenarios. It’s about moving beyond just theoretical concepts into actual application.

The paper's summary: Tom: Okay, so the paper basically lays out how regression discontinuity design can be used to estimate the causal effect of a treatment when you have a cutoff score, and it’s really trying to tackle that issue of unobserved confounders that usually messes up observational studies. Jane, can you give us the plain-language summary?

Jane: Sure. The paper explains that the core idea is assigning treatment based on whether your score is at or above a certain threshold and nothing below it, and they argue this helps isolate causal effects by focusing only on people right around that cutoff point. They’re looking at how this works in both sharp and fuzzy versions of the design to see what kind of effect you can actually estimate.

Lu: They are really digging into the identification strategy here, which is crucial because it determines what you are even allowed to claim about the effect they find. It separates the parameter of interest for sharp designs from the three parameters they look at in fuzzy designs, like take-up and intention-to-treat effects.

Meng: So, when I think about deploying this in a clinical setting, what’s the practical implication of them separating those parameters? Does it mean we get more precise information about how the treatment actually gets used versus what happens if everyone was supposed to get it?

Lalam: It seems they are providing a way to precisely define the different causal effects that matter, which is important because in a real system, the uptake of a treatment can be very different from what you expect based on policy alone. They are trying to capture that difference.

The paper's improvements: Tom: Now, they don’t just stop at describing the design; they suggest ways to improve it. What are these suggested improvements in the paper, and how do those changes actually make the estimation more robust?

Jane: They focus on addressing potential problems with the fundamental assumption that units near a cutoff are comparable regardless of treatment assignment. One suggestion is using frameworks like continuity-based or local-randomization to formally justify why those units should be comparable in the first place.

Lu: And they propose testing this assumption directly by checking for covariate balance near the cutoff, which means seeing if things like age or other baseline characteristics actually stop changing when you zoom in on that narrow region. They also mention using specific tools from the analysis of experiments to leverage the local-randomization framework.

Meng: If we have to check for covariate balance, that sounds like a lot of work at scale. What’s the practical cost of doing those kinds of checks? Does it require massive data cleaning before you can even start estimating anything meaningful?

Lalam: The improvement is about adding rigorous checks to make sure the identification strategy holds up, specifically by testing whether a null treatment effect on baseline covariates exists right near that cutoff point. That’s a concrete way to test if the comparison groups are actually comparable at that critical decision point.

Conclusion: Tom: So, wrapping things up for this paper on "Regression Discontinuity Approaches to Evaluate Threshold-Guided Interventions in Observational Studies," what’s the main message they want us to walk away with? What’s the final implication for researchers?

Jane: The main message is that while regression discontinuity designs are a powerful tool for estimating local causal effects, you have to be very careful about how you interpret those results because they are inherently local. They aren't meant to generalize far away from the cutoff point.

Lu: I agree with Jane. The conclusion emphasizes that the design provides high internal validity for these specific local effects, but external validity can be low if you try to apply it broadly without acknowledging those limitations regarding unobserved confounders or sorting around the threshold.

Meng: From a deployment standpoint, it means we use this method when we know exactly where our intervention is triggered and only expect the results to hold very close to that trigger point. It’s not a general policy tool for everything.

Lalam: To summarize "Regression Discontinuity Approaches to Evaluate Threshold-Guided Interventions in Observational Studies," the paper confirms that by focusing analysis on units with scores near the cutoff, researchers can estimate causal effects under specific assumptions about local randomization or continuity, but they have to be explicit about those boundaries.

Tom: And that’s where we’ll leave it for today. Next up, we’ll look at how other methods are tackling similar causal inference problems in machine learning and experimental settings.

Department of Operations Research and Financial Engineering, Princeton University · Department of Politics, Princeton University

stat.ME, econ.EM, stat.AP

Submitted: 2025-08-05

Updated: 2026-11-02

License: http://creativecommons.org/licenses/by/4.0/

Importance score: 80/100

The gist: The Regression Discontinuity (RD) design is a research strategy to estimate the causal effect of a treatment on an outcome.

Key concepts

Regression Discontinuity (RD) design
A research strategy used when random assignment isn't possible. It assigns treatment based on whether a unit's score is above or below a specific cutoff. This helps isolate the causal effect by focusing only on units immediately surrounding that critical threshold.
Sharp vs. Fuzzy RD
These terms describe the identification strategy for RD designs. Sharp RD estimates the average causal effect near the cutoff, while fuzzy RD allows estimation of three parameters: first-stage (take-up), intention-to-treat, and complier average treatment effects.
Local Randomization Framework
This assumption posits that units with scores close to the cutoff behave as if they were randomly assigned to either the control or treatment group. This framework is used to infer causal effects by assuming comparability in baseline characteristics near the threshold.

Terminology

Summary

The Regression Discontinuity (RD) design is a research strategy to estimate the causal effect of a treatment on an outcome. p. 0

Where the paper says it

2Department of Politics, Princeton University. August 8, 2025 The Regression Discontinuity (RD) design is a research strategy to estimate the causal effect of a treatment on an outcome.

This design is crucial when random assignment of interventions is not possible, as it removes the effect of unobserved confounders by focusing only on units whose scores are close to the cutoff. p. 1
Where the paper says it

The RD design removes the effect of unobserved confounders by focusing only on units whose scores are close to the cutoff. Intuitively, this “localization” is based on the idea that units with very similar scores will be ex-ante comparable regardless of their treatment assignment.

p. 1
Where the paper says it

The RD design removes the effect of unobserved confounders by focusing only on units whose scores are close to the cutoff. Intuitively, this “localization” is based on the idea that units with very similar scores will be ex-ante comparable regardless of their treatment assignment.

The Core Mechanism of RD

The RD design functions by assigning treatment based on a threshold rule where units receive treatment if their score is equal to or above a known cutoff, but not if their score is below it. This strategy overcomes the challenge of observational studies where comparability between groups is not guaranteed because it focuses only on units whose scores are near the cutoff, assuming they are ex-ante comparable regardless of treatment assignment. p. 0

Where the paper says it

When random assignment of interventions is not possible, the study of causal effects must rely on observational studies where the comparability of units assigned to control and units assigned to treatment is not guaranteed (Rosenbaum, 2010). If unobserved confounders systematically differ between these two groups, it is difficult to interpret their outcome differences as a causal effect of the treatment.

Identification Strategies and Parameters

The identification strategy depends on whether the design is sharp or fuzzy, as this dictates which causal effects can be estimated. In the sharp RD design, the parameter of interest is the average causal effect of the treatment on the outcome for units whose scores are close to the cutoff. p. 1

Where the paper says it

What causal effects (estimands) is the strategy used for? In the sharp RD design, the parameter of interest is the average causal effect of the treatment on the outcome for units whose scores are close to the cutoff (Cattaneo et al.

For fuzzy designs, there are three parameters: 1) the first-stage (or take-up) average effect, 2) the intention-to-treat average effect, and 3) the complier average treatment effect. p. 1
Where the paper says it

In the fuzzy RD design, there are three parameters of interest, all defined for units whose scores are near the cutoff. The first is the effect of the treatment assignment on the treatment received, known as the first-stage (or take-up) average effect.

Assumptions and Falsification

The fundamental RD identifying assumption is that units with score near the cutoff have on average the same observable and unobservable baseline characteristics regardless of assignment. This assumption can be justified through two frameworks: 1) the continuity-based framework, which postulates that potential outcomes are continuous near the cutoff, and 2) the local-randomization framework, which posits that units behave as if randomly assigned to control or treatment within a small region. p. 2

Where the paper says it

The local-randomization framework postulates that units with score within a small region near the cutoff behave as-if randomly assigned to control and treatment. This framework relies on the stronger assumption of having an experimental design among units with score near the cutoff, and employs tools from the analysis of experiments (Cattaneo et al.

To assess plausibility of this assumption, researchers can check for a null treatment effect on baseline covariates near the cutoff. p. 2
Where the paper says it

While the fundamental RD assumption of comparability near the cutoff is untestable, there are several falsification tests that researchers can employ to empirically assess its plausibility. The main strategy, akin to studying covariate balance in randomized controlled experiments, is to check for a null treatment effect on baseline covariates.

Limitations and Interpretation

A key limitation of the strategy is that valid RD designs estimate treatment effects that are local in nature, meaning they may not easily generalize to units far away from the cutoff. The interpretation of findings depends on the specific effect estimated; for instance, in one study, patients with D-dimer scores near 500 ng/mL have similar average nephrotoxicity under the assumption they do not differ in unobserved confounders. p. 3

Where the paper says it

(2021) find that patients with D-dimer score barely below and barely above the 500 ng/mL cutoff have similar average nephrotoxicity as measured by eGFR. Under the assumption that these patients do not differ in unobserved confounders, we conclude that intravenous contrast administered for CTPA has no harmful causal effect on kidney function.

The results pertain specifically to patients with D-dimer score near 500, and may differ for patients with lower or higher scores. p. 3
Where the paper says it

Under the assumption that these patients do not differ in unobserved confounders, we conclude that intravenous contrast administered for CTPA has no harmful causal effect on kidney function. This causal finding pertains to patients with D-dimer score near 500, and may be different from the causal effects for patients with lower or higher D-dimer scores.

Estimation Methods

Causal effects are estimated by first keeping only observations with a score in a neighborhood around the cutoff, and then conducting estimation based on the chosen framework. The continuity-based framework employs local linear regression methods tuned to minimize mean square error and robust bias-correction for uncertainty quantification. p. 2

Where the paper says it

The average treatment effects are estimated by first keeping only observations with score in a neighborhood around the cutoff, and then conducting estimation and inference as determined by the framework used. The continuity-based framework employs local linear regression methods, tuned to minimize mean square error, and robust bias-correction for uncertainty quantification (Cattaneo et al.

Conversely, the local-randomization framework utilizes methods from the analysis of experiments. p. 2
Where the paper says it

, 2024, Section 2). The local-randomization framework employs methods for the analysis of experiments (Cattaneo et al.

Evaluation of Identification

Several falsification tests can be employed to empirically assess the plausibility of the fundamental RD assumption. These include checking for covariate balance by observing that differences in baseline characteristics, such as age, disappear when restricting analysis to units with scores near the cutoff. p. 2

Where the paper says it

The main strategy, akin to studying covariate balance in randomized controlled experiments, is to check for a null treatment effect on baseline covariates. In the CTPA study, the average age among all individuals assigned to receive CTPA was roughly 60 years old, while the average age among those who were not assigned to CTPA was 50 years old, a difference that can confound measures of kidney function.

Other strategies involve assessing whether the number of units near the cutoff is roughly equal between groups and varying the neighborhood around the cutoff. p. 2
Where the paper says it

This balancing of baseline observed characteristics for units with score close to the cutoff is expected in a valid RD design. Other falsification strategies include assessing whether the number of units with score close to the cutoff is roughly the same between groups assigned to control and treatment, varying the neighborhood around the cutoff, considering artificial cutoffs, and removing observations closest to the cutoff.

Conclusion on Causal Effects

The RD design provides high internal validity for estimating local treatment effects, but this comes at the expense of potentially low external validity. The conclusion drawn from one application is that intravenous contrast administered for CTPA has no harmful causal effect on kidney function for patients with D-dimer score near 500. p. 3

Where the paper says it

Under the assumption that these patients do not differ in unobserved confounders, we conclude that intravenous contrast administered for CTPA has no harmful causal effect on kidney function. This causal finding pertains to patients with D-dimer score near 500, and may be different from the causal effects for patients with lower or higher D-dimer scores.

This causal finding is specific to patients with D-dimer score near 500 and may not apply broadly. p. 3
Where the paper says it

Under the assumption that these patients do not differ in unobserved confounders, we conclude that intravenous contrast administered for CTPA has no harmful causal effect on kidney function. This causal finding pertains to patients with D-dimer score near 500, and may be different from the causal effects for patients with lower or higher D-dimer scores.

The design allows researchers to conclude that patients with D-dimer score barely below and barely above the 500 ng/mL cutoff have similar average nephrotoxicity. p. 3
Where the paper says it

(2021) find that patients with D-dimer score barely below and barely above the 500 ng/mL cutoff have similar average nephrotoxicity as measured by eGFR. Under the assumption that these patients do not differ in unobserved confounders, we conclude that intravenous contrast administered for CTPA has no harmful causal effect on kidney function.

Acknowledgments

We thank Issa Dahabreh for comments and suggestions. We are indebted to our collaborators Sebastian Calonico, Rajita Chandak, Max Farrell, Yingjie Feng, Brigham Frandsen, Nicolas Idrobo, Michael Jansson, Luke Keele, Xinwei Ma, Ricardo Masini, Kenichi Nagasawa, Filippo Palomba, Jasjeet Sekhon and Gonzalo Vazquez-Bare for their intellectual contributions to our research program on RD designs. p. 3

Where the paper says it

Acknowledgments We thank Issa Dahabreh for comments and suggestions. We are indebted to our collaborators Sebastian Calonico, Rajita Chandak, Max Farrell, Yingjie Feng, Brigham Frandsen, Nicolas Idrobo, Michael Jansson, Luke Keele, Xinwei Ma, Ricardo Masini, Kenichi Nagasawa, Filippo Palomba, Jasjeet Sekhon, Gonzalo Vazquez-Bare, and Ruiqi (Rae) Yu for their intellectual contributions to our research program on RD designs.

Cattaneo and Titiunik gratefully acknowledge financial support from the National Science Foundation (SES-2019432 and SES-2241575). p. 3
Where the paper says it

We are indebted to our collaborators Sebastian Calonico, Rajita Chandak, Max Farrell, Yingjie Feng, Brigham Frandsen, Nicolas Idrobo, Michael Jansson, Luke Keele, Xinwei Ma, Ricardo Masini, Kenichi Nagasawa, Filippo Palomba, Jasjeet Sekhon, Gonzalo Vazquez-Bare, and Ruiqi (Rae) Yu for their intellectual contributions to our research program on RD designs. Cattaneo and Titiunik gratefully acknowledge financial support from the National Science Foundation (SES-2019432 and SES-2241575).

p. 3
Where the paper says it

We are indebted to our collaborators Sebastian Calonico, Rajita Chandak, Max Farrell, Yingjie Feng, Brigham Frandsen, Nicolas Idrobo, Michael Jansson, Luke Keele, Xinwei Ma, Ricardo Masini, Kenichi Nagasawa, Filippo Palomba, Jasjeet Sekhon, Gonzalo Vazquez-Bare, and Ruiqi (Rae) Yu for their intellectual contributions to our research program on RD designs. Cattaneo and Titiunik gratefully acknowledge financial support from the National Science Foundation (SES-2019432 and SES-2241575).

References

Cattaneo, M. p. 0

Where the paper says it

Generalizations include multi-cutoff and multi-score (or geographic) RD designs. See Cattaneo and Titiunik (2022) for a literature review.

p. 0
Where the paper says it

Generalizations include multi-cutoff and multi-score (or geographic) RD designs. See Cattaneo and Titiunik (2022) for a literature review.

D., Idrobo, N. p. 3
Where the paper says it

Acknowledgments We thank Issa Dahabreh for comments and suggestions. We are indebted to our collaborators Sebastian Calonico, Rajita Chandak, Max Farrell, Yingjie Feng, Brigham Frandsen, Nicolas Idrobo, Michael Jansson, Luke Keele, Xinwei Ma, Ricardo Masini, Kenichi Nagasawa, Filippo Palomba, Jasjeet Sekhon, Gonzalo Vazquez-Bare, and Ruiqi (Rae) Yu for their intellectual contributions to our research program on RD designs.

, and Titiunik, R. p. 0
Where the paper says it

Generalizations include multi-cutoff and multi-score (or geographic) RD designs. See Cattaneo and Titiunik (2022) for a literature review.

p. 0
Where the paper says it

Generalizations include multi-cutoff and multi-score (or geographic) RD designs. See Cattaneo and Titiunik (2022) for a literature review.

p. 0
Where the paper says it

Generalizations include multi-cutoff and multi-score (or geographic) RD designs. See Cattaneo and Titiunik (2022) for a literature review.

(2020), A Practical Introduction to Regression Discontinuity Designs: Foundations, New York, NY: Cambridge Elements: Quantitative and Computational Methods for Social Science, Cambridge University Press. p. 3
Where the paper says it

, and Titiunik, R. (2020), A Practical Introduction to Regression Discontinuity Designs: Foundations, New York, NY: Cambridge Elements: Quantitative and Computational Methods for Social Science, Cambridge University Press.

(2024), A Practical Introduction to Regression Discontinuity Designs: Extensions, New York, NY: Cambridge Elements: Quantitative and Computational Methods for Social Science, Cambridge University Press. p. 3

Where the paper says it

(2020), A Practical Introduction to Regression Discontinuity Designs: Foundations, New York, NY: Cambridge Elements: Quantitative and Computational Methods for Social Science, Cambridge University Press. (2024), A Practical Introduction to Regression Discontinuity Designs: Extensions, New York, NY: Cambridge Elements: Quantitative and Computational Methods for Social Science, Cambridge University Press.

Cattaneo, M. D., Keele, L. p. 3

Where the paper says it

Acknowledgments We thank Issa Dahabreh for comments and suggestions. We are indebted to our collaborators Sebastian Calonico, Rajita Chandak, Max Farrell, Yingjie Feng, Brigham Frandsen, Nicolas Idrobo, Michael Jansson, Luke Keele, Xinwei Ma, Ricardo Masini, Kenichi Nagasawa, Filippo Palomba, Jasjeet Sekhon, Gonzalo Vazquez-Bare, and Ruiqi (Rae) Yu for their intellectual contributions to our research program on RD designs.

, and Titiunik, R. (2023), “A Guide to Regression Discontinuity Designs in Medical Applications,” Statistics in Medicine, 42, 4484–4513. p. 3
Where the paper says it

, and Titiunik, R. (2023), “A Guide to Regression Discontinuity Designs in Medical Applications,” Statistics in Medicine, 42, 4484–4513.

Cattaneo, M. D., and Titiunik, R. (2022), “Regression Discontinuity Designs,” Annual Review of Economics, 14, 821–851. p. 3

Where the paper says it

, and Titiunik, R. (2022), “Regression Discontinuity Designs,” Annual Review of Economics, 14, 821–851.

Goulden, R. p. 0

Where the paper says it

See Cattaneo and Titiunik (2022) for a literature review. Goulden et al.

, Rowe, B. H., Abrahamowicz, M. p. 4
Where the paper says it

Goulden, R. , Abrahamowicz, M.

, Strumpf, E. p. 4
Where the paper says it

, Abrahamowicz, M. , Strumpf, E.

, and Tamblyn, R. p. 4
Where the paper says it

, Strumpf, E. , and Tamblyn, R.

(2021), “Association of intravenous radiocontrast with kidney function: a regression discontinuity analysis,” JAMA internal medicine, 181, 767–774. p. 4
Where the paper says it

, and Tamblyn, R. (2021), “Association of intravenous radiocontrast with kidney function: a regression discontinuity analysis,” JAMA internal medicine, 181, 767–774.

Lee, D. S., and Lemieux, T. p. 1

Where the paper says it

Intuitively, this “localization” is based on the idea that units with very similar scores will be ex-ante comparable regardless of their treatment assignment. RD designs are often viewed as “a much closer cousin of randomized experiments than other competing methods” (Lee and Lemieux, 2010, p.

(2010), “Regression Discontinuity Designs in Economics,” Journal of Economic Literature, 48, 281–355. p. 4
Where the paper says it

, and Lemieux, T. (2010), “Regression Discontinuity Designs in Economics,” Journal of Economic Literature, 48, 281–355.

Rosenbaum, P. p. 0

Where the paper says it

1 Why is the Identification Strategy Used? When random assignment of interventions is not possible, the study of causal effects must rely on observational studies where the comparability of units assigned to control and units assigned to treatment is not guaranteed (Rosenbaum, 2010).

R. (2010), Design of Observational Studies, Vol. p. 4
Where the paper says it

Rosenbaum, P. (2010), Design of Observational Studies, Vol.

10, Springer. p. 4
Where the paper says it

(2010), Design of Observational Studies, Vol. 10, Springer.

--- Page 6 ---

(No content provided for Page 6 in the input text)

Summary

The Regression Discontinuity (RD) design is a research strategy to estimate the causal effect of a treatment on an outcome. This design is crucial when random assignment of interventions is not possible, as it removes the effect of unobserved confounders by focusing only on units whose scores are close to the cutoff.

We thank Issa Dahabreh for comments and suggestions. We are indebted to our collaborators Sebastian Calonico, Rajita Chandak, Max Farrell, Yingjie Feng, Brigham Frandsen, Nicolas Idrobo, Michael Jansson, Luke Keele, Xinwei Ma, Ricardo Masini, Kenichi Nagasawa and Filippo Palomba for their intellectual contributions to our research program on RD designs. p. 3

Where the paper says it

Acknowledgments We thank Issa Dahabreh for comments and suggestions. We are indebted to our collaborators Sebastian Calonico, Rajita Chandak, Max Farrell, Yingjie Feng, Brigham Frandsen, Nicolas Idrobo, Michael Jansson, Luke Keele, Xinwei Ma, Ricardo Masini, Kenichi Nagasawa, Filippo Palomba, Jasjeet Sekhon, Gonzalo Vazquez-Bare, and Ruiqi (Rae) Yu for their intellectual contributions to our research program on RD designs.

Cattaneo and Titiunik gratefully acknowledge financial support from the National Science Foundation (SES-2019432 and SES-2241575).

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Improvement of treatment assignment strategy in personalized medical AI systems: Implement a fuzzy Regression Discontinuity design when assigning treatments based on continuous scores, such as D-dimer levels or risk scores near a threshold. This allows the system to capture the first-stage (or take-up) average effect and intention-to-treat average effect, providing granular insights into how treatment assignment influences both uptake and outcome near the cutoff.

  2. Improvement of causal inference in clinical prediction models: Use the local randomization framework when analyzing patient subgroups near a specific clinical threshold to infer causal effects, assuming units with score within a small region near the cutoff behave as-if randomly assigned to control and treatment. This helps isolate the effect of an intervention on kidney function for patients whose scores are close to the cutoff, mitigating confounding from unobserved characteristics.

  3. Improvement of model validation and robustness: Employ falsification tests by checking for a null treatment effect on baseline covariates specifically within the neighborhood around the cutoff. For instance, ensuring that the difference in age disappeared when restricting analysis to patients with D-dimer score near 500, thereby validating that the comparison groups are comparable at the critical decision point.

Abstract

This article provides an introduction to the Regression Discontinuity (RD) design, and its application to empirical research in the medical sciences. While the main focus of this article is on causal interpretation, key concepts of estimation and inference are also briefly mentioned. A running medical empirical example is provided.

Related papers