Instrumental variable (IV) strategies are widely used in political science to establish causal relationships. However, the identifying assumptions required by an IV design are demanding, and it remains challenging for researchers to assess their validity. In this paper, we replicate 67 papers published in three top journals in political science during 2010-2022 and identify several troubling patterns. First, researchers often overestimate the strength of their IVs due to non-i.i.d. errors, such as a clustering structure. Second, the most commonly used t-test for the two-stage-least-squares (2SLS) estimates often severely underestimates uncertainty. Using more robust inferential methods, we find that around 19-30% of the 2SLS estimates in our sample are underpowered. Third, in the majority of the replicated studies, the 2SLS estimates are much larger than the ordinary-least-squares estimates, and their ratio is negatively correlated with the strength of the IVs in studies where the IVs are not experimentally generated, suggesting potential violations of unconfoundedness or the exclusion restriction. To help researchers avoid these pitfalls, we provide a checklist for better practice.
相關內容
Real-world data, such as administrative claims and electronic health records, are increasingly used for safety monitoring and to help guide regulatory decision-making. In these settings, it is important to document analytic decisions transparently and objectively to ensure that analyses meet their intended goals. The Causal Roadmap is an established framework that can guide and document analytic decisions through each step of the analytic pipeline, which will help investigators generate high-quality real-world evidence. In this paper, we illustrate the utility of the Causal Roadmap using two case studies previously led by workgroups sponsored by the Sentinel Initiative -- a program for actively monitoring the safety of regulated medical products. Each case example focuses on different aspects of the analytic pipeline for drug safety monitoring. The first case study shows how the Causal Roadmap encourages transparency, reproducibility, and objective decision-making for causal analyses. The second case study highlights how this framework can guide analytic decisions beyond inference on causal parameters, improving outcome ascertainment in clinical phenotyping. These examples provide a structured framework for implementing the Causal Roadmap in safety surveillance and guide transparent, reproducible, and objective analysis.
Observational studies are the primary source of data for causal inference, but it is challenging when existing unmeasured confounding. Missing data problems are also common in observational studies. How to obtain the causal effects from the nonignorable missing data with unmeasured confounding is a challenge. In this paper, we consider that how to obtain complier average causal effect with unmeasured confounding from the nonignorable missing outcomes. We propose an auxiliary variable which plays two roles simultaneously, the one is the shadow variable for identification and the other is the instrumental variable for inference. We also illustrate some difference between some missing outcomes mechanisms in the previous work and the shadow variable assumption. We give a causal diagram to illustrate this description. Under such a setting, we present a general condition for nonparametric identification of the full data law from the nonignorable missing outcomes with this auxiliary variable. For inference, firstly, we recover the mean value of the outcome based on the generalized method of moments. Secondly, we propose an estimator to adjust for the unmeasured confounding to obtain complier average causal effect. We also establish the asymptotic results of the estimated parameters. We evaluate its performance via simulations and apply it to a real-life dataset about a political analysis.
Monocular 3D human pose and shape estimation is an ill-posed problem since multiple 3D solutions can explain a 2D image of a subject. Recent approaches predict a probability distribution over plausible 3D pose and shape parameters conditioned on the image. We show that these approaches exhibit a trade-off between three key properties: (i) accuracy - the likelihood of the ground-truth 3D solution under the predicted distribution, (ii) sample-input consistency - the extent to which 3D samples from the predicted distribution match the visible 2D image evidence, and (iii) sample diversity - the range of plausible 3D solutions modelled by the predicted distribution. Our method, HuManiFlow, predicts simultaneously accurate, consistent and diverse distributions. We use the human kinematic tree to factorise full body pose into ancestor-conditioned per-body-part pose distributions in an autoregressive manner. Per-body-part distributions are implemented using normalising flows that respect the manifold structure of SO(3), the Lie group of per-body-part poses. We show that ill-posed, but ubiquitous, 3D point estimate losses reduce sample diversity, and employ only probabilistic training losses. Code is available at: //github.com/akashsengupta1997/HuManiFlow.
A contiguous area cartogram is a geographic map in which the area of each region is proportional to numerical data (e.g., population size) while keeping neighboring regions connected. In this study, we investigated whether value-to-area legends (square symbols next to the values represented by the squares' areas) and grid lines aid map readers in making better area judgments. We conducted an experiment to determine the accuracy, speed, and confidence with which readers infer numerical data values for the mapped regions. We found that, when only informed about the total numerical value represented by the whole cartogram without any legend, the distribution of estimates for individual regions was centered near the true value with substantial spread. Legends with grid lines significantly reduced the spread but led to a tendency to underestimate the values. Comparing differences between regions or between cartograms revealed that legends and grid lines slowed the estimation without improving accuracy. However, participants were more likely to complete the tasks when legends and grid lines were present, particularly when the area units represented by these features could be interactively selected. We recommend considering the cartogram's use case and purpose before deciding whether to include grid lines or an interactive legend.
Gun violence is a major problem in contemporary American society. However, relatively little is known about the effects of firearm injuries on survivors and their family members and how these effects vary across subpopulations. To study these questions and, more generally, to address a gap in the causal inference literature, we present a framework for the study of effect modification or heterogeneous treatment effects in difference-in-differences designs. We implement a new matching technique, which combines profile matching and risk set matching, to (i) preserve the time alignment of covariates, exposure, and outcomes, avoiding pitfalls of other common approaches for difference-in-differences, and (ii) explicitly control biases due to imbalances in observed covariates in subgroups discovered from the data. Our case study shows significant and persistent effects of nonfatal firearm injuries on several health outcomes for those injured and on the mental health of their family members. Sensitivity analyses reveal that these results are moderately robust to unmeasured confounding bias. Finally, while the effects for those injured are modified largely by the severity of the injury and its documented intent, for families, effects are strongest for those whose relative's injury is documented as resulting from an assault, self-harm, or law enforcement intervention.
Under two-phase designs, the outcome and several covariates and confounders are measured in the first phase, and a new predictor of interest, which may be costly to collect, can be measured on a subsample in the second phase, without incurring the costs of recruiting subjects. By using the information gathered in the first phase, the second-phase subsample can be selected to enhance the efficiency of testing and estimating the effect of the new predictor on the outcome. Past studies have focused on optimal two-phase sampling schemes for statistical inference on local ($\beta = o(1)$) effects of the predictor of interest. In this study, we propose an extension of the two-phase designs that employs an optimal sampling scheme for estimating predictor effects with pseudo conditional likelihood estimators in case-control studies. This approach is applicable to both local and non-local effects. We demonstrate the effectiveness of the proposed sampling scheme through simulation studies and analysis of data from 170 patients hospitalized for treatment of COVID-19. The results show a significant improvement in the estimation of the parameter of interest.
The intersection of ageism and sexism can create a hostile environment for veteran software developers belonging to marginalized genders. In this study, we conducted 14 interviews to examine the experiences of people at this intersection, primarily women, in order to discover the strategies they employed in order to successfully remain in the field. We identified 283 codes, which fell into three main categories: Strategies, Experiences, and Perception. Several strategies we identified, such as (Deliberately) Not Trying to Look Younger, were not previously described in the software engineering literature. We found that, in some companies, older women developers are recognized as having particular value, further strengthening the known benefits of diversity in the workforce. Based on the experiences and strategies, we suggest organizations employing software developers to consider the benefits of hiring veteran women software developers. For example, companies can draw upon the life experiences of older women developers in order to better understand the needs of customers from a similar demographic. While we recognize that many of the strategies employed by our study participants are a response to systemic issues, we still consider that, in the short-term, there is benefit in describing these strategies for developers who are experiencing such issues today.
When an exposure of interest is confounded by unmeasured factors, an instrumental variable (IV) can be used to identify and estimate certain causal contrasts. Identification of the marginal average treatment effect (ATE) from IVs relies on strong untestable structural assumptions. When one is unwilling to assert such structure, IVs can nonetheless be used to construct bounds on the ATE. Famously, Balke and Pearl (1997) proved tight bounds on the ATE for a binary outcome, in a randomized trial with noncompliance and no covariate information. We demonstrate how these bounds remain useful in observational settings with baseline confounders of the IV, as well as randomized trials with measured baseline covariates. The resulting bounds on the ATE are non-smooth functionals, and thus standard nonparametric efficiency theory is not immediately applicable. To remedy this, we propose (1) under a novel margin condition, influence function-based estimators of the bounds that can attain parametric convergence rates when the nuisance functions are modeled flexibly, and (2) estimators of smooth approximations of these bounds. We propose extensions to continuous outcomes, explore finite sample properties in simulations, and illustrate the proposed estimators in a randomized field experiment studying the effects of canvassing on resulting voter turnout.
The rapid recent progress in machine learning (ML) has raised a number of scientific questions that challenge the longstanding dogma of the field. One of the most important riddles is the good empirical generalization of overparameterized models. Overparameterized models are excessively complex with respect to the size of the training dataset, which results in them perfectly fitting (i.e., interpolating) the training data, which is usually noisy. Such interpolation of noisy data is traditionally associated with detrimental overfitting, and yet a wide range of interpolating models -- from simple linear models to deep neural networks -- have recently been observed to generalize extremely well on fresh test data. Indeed, the recently discovered double descent phenomenon has revealed that highly overparameterized models often improve over the best underparameterized model in test performance. Understanding learning in this overparameterized regime requires new theory and foundational empirical studies, even for the simplest case of the linear model. The underpinnings of this understanding have been laid in very recent analyses of overparameterized linear regression and related statistical learning tasks, which resulted in precise analytic characterizations of double descent. This paper provides a succinct overview of this emerging theory of overparameterized ML (henceforth abbreviated as TOPML) that explains these recent findings through a statistical signal processing perspective. We emphasize the unique aspects that define the TOPML research area as a subfield of modern ML theory and outline interesting open questions that remain.
A fundamental goal of scientific research is to learn about causal relationships. However, despite its critical role in the life and social sciences, causality has not had the same importance in Natural Language Processing (NLP), which has traditionally placed more emphasis on predictive tasks. This distinction is beginning to fade, with an emerging area of interdisciplinary research at the convergence of causal inference and language processing. Still, research on causality in NLP remains scattered across domains without unified definitions, benchmark datasets and clear articulations of the remaining challenges. In this survey, we consolidate research across academic areas and situate it in the broader NLP landscape. We introduce the statistical challenge of estimating causal effects, encompassing settings where text is used as an outcome, treatment, or as a means to address confounding. In addition, we explore potential uses of causal inference to improve the performance, robustness, fairness, and interpretability of NLP models. We thus provide a unified overview of causal inference for the computational linguistics community.
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191