causal inference

Estimating causal effects from observational data requires selecting appropriate control and instrumental variables that satisfy causal identification conditions—a challenging task often reliant on strong domain knowledge or ad hoc assumptions. Method: This paper proposes the first end-to-end joint learning framework that automatically identifies valid combinations of control and instrumental variables. Grounded in conditional independence testing, the method integrates nonparametric dependence measures with structural search optimization, ensuring statistical consistency in variable selection under mild regularity conditions. Contribution/Results: Unlike conventional approaches requiring prespecified variable sets or strong prior assumptions, our framework is fully data-driven. In simulations, it achieves significantly higher variable identification accuracy. Empirically, applied to the Job Corps study, its estimated treatment effect closely aligns with results from the randomized controlled trial—demonstrating both validity and robustness in real-world causal inference.

Existing visualization design models overemphasize technical solution-finding while neglecting the dynamic construction of problem framing and reflective practice. Method: Through a mixed-methods approach—including design challenges, reflective diaries, and semi-structured interviews—we conducted reflexive thematic analysis with 11 visualization experts to examine how problem and solution spaces co-evolve in expert practice. Contribution/Results: We reconceptualize “framing” as an ongoing, situated activity—not merely an initial step—and identify key problem-reframing strategies: metaphorical transfer, heuristic probing, and sketch-based iteration. The study extends visualization design theory by integrating explanatory judgment, ethical narrative, and systematic reflective practice into the design framework. This shifts the paradigm from “problem-solving–oriented” to “meaning-making–oriented” design, emphasizing iterative sensemaking, epistemic responsibility, and context-sensitive interpretation throughout the design process.

Estimating the causal effect of beliefs on decision-making is often confounded by omitted variables. Conventional information-provision experiments, while enabling randomized belief manipulation, frequently violate the exclusion restriction assumption due to source effects—introducing new confounders—and raise ethical concerns related to deception. To address these challenges, this paper introduces the Anchoring-Based Causal Design (ABCD), which randomly induces belief variation using non-informative numerical anchors and estimates causal effects via instrumental variable methods. By avoiding substantive informational interventions, ABCD satisfies the exclusion restriction, eliminates source bias, and circumvents ethical issues associated with deception. Eight pre-registered experiments consistently demonstrate that ABCD robustly identifies the causal impact of beliefs, substantially enhancing both internal validity and methodological ethics in belief–behavior causal inference.

Existing causal models struggle to distinguish between the immediate and persistent effects of interventions in time-dynamic systems, particularly when such interventions alter the system’s equilibrium behavior. This work proposes a novel paradigm grounded in system and state representations, integrating causal directed acyclic graphs, the potential outcomes framework, and dynamic systems theory. By introducing an equilibrium-state assumption and employing state-space modeling, the study reformulates the causal inference framework to better capture temporal dynamics. It innovatively defines an equilibrium-oriented “zero effect” concept and combines it with a strategic selection of time points to enable valid identification of time-varying causal parameters. The approach establishes clear criteria for categorizing causal effects under dynamic interventions, substantially enhancing the interpretability and practical utility of causal inference in equilibrium analysis.

Prior visualization research has predominantly examined framing’s rhetorical effects on audiences, overlooking its central role in the design process itself. Method: This study conducts the first systematic qualitative analysis of reflective narratives from over 80 professional visualization designers drawn from podcasts and book chapters, employing thematic coding across heterogeneous sources to identify recurring patterns and iterative dynamics of framing. Contribution/Results: We reveal framing as a pervasive, recursive strategic activity across problem scoping, data interpretation, and narrative structuring. The study identifies critical triggers for frame revision—including stakeholder conflicts and data unexpectedness—as well as corresponding designer strategies. Framing is thus established as an intrinsic strategic dimension of visualization design, shifting the field’s emphasis from technical implementation toward design judgment and strategic reasoning. These findings provide empirically grounded theoretical foundations for pedagogy and the development of human-AI co-design tools.

This study addresses the limitations of traditional control-based causal inference methods—such as matching and difference-in-differences—in settings characterized by pervasive or structurally ambiguous spillover effects, where reliance on uncontaminated control units impedes accurate identification of both average direct and spillover effects. Within the potential outcomes framework, this work provides the first systematic comparison between control-based and prediction-based counterfactual approaches—including interrupted time series and machine learning control—in terms of their identification capabilities. Through simulation and empirical analyses, the authors demonstrate that in environments with widespread interference, prediction-based methods can more reliably estimate certain causal parameters over short horizons, circumventing the stringent assumption of unperturbed units and thereby offering a promising alternative for causal inference under complex interference.

When confronted with complex problems, individuals often favor direct interventions that merely alleviate symptoms over indirect interventions targeting root causes. This study employs cognitive psychology experiments in a hoof-trimming context to systematically investigate whether the level of detail in causal mechanism information and its presentation format—as causal diagrams within multifactorial scenarios—enhances acceptance of indirect interventions. Results indicate that although providing mechanistic information slightly reduced skepticism toward indirect approaches, it did not significantly shift the prevailing preference for symptomatic treatment. These findings challenge the common assumption that merely supplying causal mechanisms promotes rational decision-making, revealing for the first time that such information alone is insufficient to foster choices favoring sustainable interventions.

This study addresses a critical yet previously unrecognized issue in observational causal inference: measurement-induced confounding, wherein latent variables—such as motivation or self-efficacy—are imperfectly measured, leading to biased estimates of adjusted causal effects. The authors formally identify and name this problem, moving beyond conventional two-stage adjustment approaches. They propose a novel Bayesian joint estimation framework that simultaneously models the latent variable’s measurement structure, the treatment assignment mechanism, and the potential outcomes model. This integrated approach effectively corrects bias in average treatment effect estimation and restores the nominal coverage of uncertainty intervals, thereby substantially enhancing the reliability of causal inferences drawn from observational data with error-prone proxies for unobserved confounders.

Traditional human factors analysis relies on correlation-based testing, which only addresses descriptive “what is” questions and fails to resolve causal “if–then” queries. It remains vulnerable to confounding and collider variables, leading to biased decision-making. This paper proposes a paradigm shift from descriptive analysis to causal inference. Leveraging post-occupancy evaluation data from built environments, we integrate causal discovery algorithms, structural equation modeling, and counterfactual intervention analysis to construct directed causal networks among variables—explicitly distinguishing descriptive from interventional queries. Our key contribution is the first systematic application of a rigorous causal inference framework to the human factors domain, enabling robust identification of intervention priorities and hierarchical causal pathways. Empirical validation demonstrates that this approach significantly enhances the accuracy and scientific rigor of human factors–driven system optimization decisions, establishing a novel, interpretable, and actionable causal analysis paradigm for human factors engineering.

Identifying temporal causal effects in Summary Causal Graphs (SCGs) remains challenging: conventional backdoor criteria fail for cyclic structures and temporally abstracted nodes, while existing identifiability conditions are overly complex and yield only limited valid adjustment sets. Method: We propose an equivalent simplification of the identifiability criterion and establish a generalized adjustment criterion that systematically characterizes all valid covariate adjustment sets. We formally define and construct variance-optimal nearly-optimal adjustment sets. Leveraging directed graph theory, path analysis, and asymptotic statistical inference—under the no-hidden-confounding assumption—we rigorously characterize adjustment-set properties constrained by SCGs. Contribution/Results: Our method substantially reduces computational complexity, enables flexible selection of low-variance adjustment sets, and enhances both efficiency and practicality of causal estimation in dynamic systems. It provides the first unified framework for valid and optimal adjustment in SCGs, bridging structural causality with statistical estimation theory.