Effect Estimation Research Articles

Adversarially Balanced Representation for Continuous Treatment Effect Estimation

Individual treatment effect (ITE) estimation requires adjusting for the covariate shift between populations with different treatments, and deep representation learning has shown great promise in learning a balanced representation of covariates. However the existing methods mostly consider the scenario of binary treatments. In this paper, we consider the more practical and challenging scenario in which the treatment is a continuous variable (e.g. dosage of a medication), and we address the two main challenges of this setup. We propose the adversarial counterfactual regression network (ACFR) that adversarially minimizes the representation imbalance in terms of KL divergence, and also maintains the impact of the treatment value on the outcome prediction by leveraging an attention mechanism. Theoretically we demonstrate that ACFR objective function is grounded in an upper bound on counterfactual outcome prediction error. Our experimental evaluation on semi-synthetic datasets demonstrates the empirical superiority of ACFR over a range of state-of-the-art methods.

Continuous Treatment Effect Estimation Using Gradient Interpolation and Kernel Smoothing

We address the Individualized continuous treatment effect (ICTE) estimation problem where we predict the effect of any continuous valued treatment on an individual using ob- servational data. The main challenge in this estimation task is the potential confounding of treatment assignment with in- dividual’s covariates in the training data, whereas during in- ference ICTE requires prediction on independently sampled treatments. In contrast to prior work that relied on regularizers or unstable GAN training, we advocate the direct approach of augmenting training individuals with independently sam- pled treatments and inferred counterfactual outcomes. We in- fer counterfactual outcomes using a two-pronged strategy: a Gradient Interpolation for close-to-observed treatments, and a Gaussian Process based Kernel Smoothing which allows us to down weigh high variance inferences. We evaluate our method on five benchmarks and show that our method out- performs six state-of-the-art methods on the counterfactual estimation error. We analyze the superior performance of our method by showing that (1) our inferred counterfactual re- sponses are more accurate, and (2) adding them to the train- ing data reduces the distributional distance between the con- founded training distribution and test distribution where treat- ment is independent of covariates. Our proposed method is model-agnostic and we show that it improves ICTE accuracy of several existing models.

Exploiting Geometry for Treatment Effect Estimation via Optimal Transport

Estimating treatment effects from observational data suffers from the issue of confounding bias, which is induced by the imbalanced confounder distributions between the treated and control groups. As an effective approach, re-weighting learns a group of sample weights to balance the confounder distributions. Existing methods of re-weighting highly rely on a propensity score model or moment alignment. However, for complex real-world data, it is difficult to obtain an accurate propensity score prediction. Although moment alignment is free of learning a propensity score model, accurate estimation for high-order moments is computationally difficult and still remains an open challenge, and first and second-order moments are insufficient to align the distributions and easy to be misled by outliers. In this paper, we exploit geometry to capture the intrinsic structure involved in data for balancing the confounder distributions, so that confounding bias can be reduced even with outliers. To achieve this, we construct a connection between treatment effect estimation and optimal transport, a powerful tool to capture geometric information. After that, we propose an optimal transport model to learn sample weights by extracting geometry from confounders, in which geometric information between groups and within groups is leveraged for better confounder balancing. A projected mirror descent algorithm is employed to solve the derived optimization problem. Experimental studies on both synthetic and real-world datasets demonstrate the effectiveness of our proposed method.

NESTER: An Adaptive Neurosymbolic Method for Causal Effect Estimation

Causal effect estimation from observational data is a central problem in causal inference. Methods based on potential outcomes framework solve this problem by exploiting inductive biases and heuristics from causal inference. Each of these methods addresses a specific aspect of causal effect estimation, such as controlling propensity score, enforcing randomization, etc., by designing neural network (NN) architectures and regularizers. In this paper, we propose an adaptive method called Neurosymbolic Causal Effect Estimator (NESTER), a generalized method for causal effect estimation. NESTER integrates the ideas used in existing methods based on multi-head NNs for causal effect estimation into one framework. We design a Domain Specific Language (DSL) tailored for causal effect estimation based on causal inductive biases used in literature. We conduct a theoretical analysis to investigate NESTER's efficacy in estimating causal effects. Our comprehensive empirical results show that NESTER performs better than state-of-the-art methods on benchmark datasets.

Instrumental Variable Estimation for Causal Inference in Longitudinal Data with Time-Dependent Latent Confounders

Causal inference from longitudinal observational data is a challenging problem due to the difficulty in correctly identifying the time-dependent confounders, especially in the presence of latent time-dependent confounders. Instrumental variable (IV) is a powerful tool for addressing the latent confounders issue, but the traditional IV technique cannot deal with latent time-dependent confounders in longitudinal studies. In this work, we propose a novel Time-dependent Instrumental Factor Model (TIFM) for time-varying causal effect estimation from data with latent time-dependent confounders. At each time-step, the proposed TIFM method employs the Recurrent Neural Network (RNN) architecture to infer latent IV, and then uses the inferred latent IV factor for addressing the confounding bias caused by the latent time-dependent confounders. We provide a theoretical analysis for the proposed TIFM method regarding causal effect estimation in longitudinal data. Extensive evaluation with synthetic datasets demonstrates the effectiveness of TIFM in addressing causal effect estimation over time. We further apply TIFM to a climate dataset to showcase the potential of the proposed method in tackling real-world problems.

KG-TREAT: Pre-training for Treatment Effect Estimation by Synergizing Patient Data with Knowledge Graphs.

Treatment effect estimation (TEE) is the task of determining the impact of various treatments on patient outcomes. Current TEE methods fall short due to reliance on limited labeled data and challenges posed by sparse and high-dimensional observational patient data. To address the challenges, we introduce a novel pre-training and fine-tuning framework, KG-TREAT, which synergizes large-scale observational patient data with biomedical knowledge graphs (KGs) to enhance TEE. Unlike previous approaches, KG-TREAT constructs dual-focus KGs and integrates a deep bi-level attention synergy method for in-depth information fusion, enabling distinct encoding of treatment-covariate and outcome-covariate relationships. KG-TREAT also incorporates two pre-training tasks to ensure a thorough grounding and contextualization of patient data and KGs. Evaluation on four downstream TEE tasks shows KG-TREAT's superiority over existing methods, with an average improvement of 7% in Area under the ROC Curve (AUC) and 9% in Influence Function-based Precision of Estimating Heterogeneous Effects (IF-PEHE). The effectiveness of our estimated treatment effects is further affirmed by alignment with established randomized clinical trial findings.

Linear-Time Algorithms for Front-Door Adjustment in Causal Graphs

Causal effect estimation from observational data is a fundamental task in empirical sciences. It becomes particularly challenging when unobserved confounders are involved in a system. This paper focuses on front-door adjustment – a classic technique which, using observed mediators allows to identify causal effects even in the presence of unobserved confounding. While the statistical properties of the front-door estimation are quite well understood, its algorithmic aspects remained unexplored for a long time. In 2022, Jeong, Tian, and Bareinboim presented the first polynomial-time algorithm for finding sets satisfying the front-door criterion in a given directed acyclic graph (DAG), with an O(n³(n+m)) run time, where n denotes the number of variables and m the number of edges of the causal graph. In our work, we give the first linear-time, i.e., O(n+m), algorithm for this task, which thus reaches the asymptotically optimal time complexity. This result implies an O(n(n+m)) delay enumeration algorithm of all front-door adjustment sets, again improving previous work by a factor of n³. Moreover, we provide the first linear-time algorithm for finding a minimal front-door adjustment set. We offer implementations of our algorithms in multiple programming languages to facilitate practical usage and empirically validate their feasibility, even for large graphs.

Uncertainty Quantification in Heterogeneous Treatment Effect Estimation with Gaussian-Process-Based Partially Linear Model

Estimating heterogeneous treatment effects across individuals has attracted growing attention as a statistical tool for performing critical decision-making. We propose a Bayesian inference framework that quantifies the uncertainty in treatment effect estimation to support decision-making in a relatively small sample size setting. Our proposed model places Gaussian process priors on the nonparametric components of a semiparametric model called a partially linear model. This model formulation has three advantages. First, we can analytically compute the posterior distribution of a treatment effect without relying on the computationally demanding posterior approximation. Second, we can guarantee that the posterior distribution concentrates around the true one as the sample size goes to infinity. Third, we can incorporate prior knowledge about a treatment effect into the prior distribution, improving the estimation efficiency. Our experimental results show that even in the small sample size setting, our method can accurately estimate the heterogeneous treatment effects and effectively quantify its estimation uncertainty.

Intergenerational Transmission of Welfare Benefit Receipt: Evidence from Germany

We study the intergenerational transmission of welfare benefit receipt in Germany. We first describe the correlation between welfare receipt experienced in the parental household and subsequent own welfare receipt of young adults. In a second step, we investigate whether the observed correlations reflect causal effects of past welfare experience. We use family fixed effects estimations and Gottschalk's (1996) approach and take advantage of the long‐running German Socio‐Economic Panel Survey to contribute to a sparse literature. We find strong positive correlations between parental and own welfare receipt. These patterns do, however, not persist after controlling for unobserved heterogeneities. Therefore, our results suggest that the strong intergenerational correlation of welfare benefit receipt is determined by family background rather than by the experience of parental welfare benefit receipt.

Uncertainty and cash holdings: the moderating role of political connections

PurposeThis paper aims to investigate the relationship between uncertainty and corporate cash holdings with the moderating role of political connections.Design/methodology/approachWe employ fixed effects estimation on a panel dataset of 669 Vietnamese listed firms over the 2010–2020 period, with one- and two-way standard error clustering. We conduct various robustness tests, including two-stage least squares/instrumental variable and generalized method of moments regressions, alternative cash holding measure, and additional controls for macroeconomic conditions and ownership types.FindingsThe effect of uncertainty on cash holdings is weakened for firms with political connections relative to those without the connections. Although general firms depend on cash flows to adjust their cash holding behavior when uncertainty increases, our findings suggest that politically connected firms do not rely on internal cash flows to accumulate cash when confronted high uncertainty.Practical implicationsOur findings on the role of political connections in moderating the relationship between cash holding and economic policy uncertainty have practical implications for policymaking. Since political connections serve as a buffer for a firm’s liquidity, firms may want to seek those connections, which can, in turn, lead to increasing informal costs and unfair business environment.Originality/valueThis is the first study investigating the role of political connections to the nexus of cash, cash flow and uncertainty, providing novel evidence regarding the less dependence on internal cash flows to save cash by politically connected firms. Second, the paper enriches the literature on the motives of cash holdings by proposing a modified agency view in the context of weak investor protection. Therefore, our findings strengthen the explanation for the positive effect of uncertainty on firms’ cash holdings in emerging markets.

Performance evaluation of parameters as estimators of seismic site effects in northern South America

In this study, we investigate the performance of several parameters as estimators of seismic site effects for sites in northern South America. Investigated parameters include the conventional time-averaged shear wave velocity in the upper 30 m (VS30), the predominant period of the site, and horizontal-to-vertical spectral ratios. To assess the performance of these parameters, we estimate the amplification of seismic waves within shallow soil layers (i.e., site effects) at recording stations across northern South America. The site amplification is estimated by analyzing the residuals calculated from ground-motion records with respect to model predictions for hard-rock site conditions. Several functional forms were introduced to express the linear site amplification in terms of the explored parameters. We evaluate the efficacy of the parameters, as well as their advantages and limitations, by considering their impact in the site-to-site standard deviation of the residuals between observations and predictions using the introduced functional forms.

Optimal allocation strategies in platform trials with continuous endpoints.

Platform trials are randomized clinical trials that allow simultaneous comparison of multiple interventions, usually against a common control. Arms to test experimental interventions may enter and leave the platform over time. This implies that the number of experimental intervention arms in the trial may change as the trial progresses. Determining optimal allocation rates to allocate patients to the treatment and control arms in platform trials is challenging because the optimal allocation depends on the number of arms in the platform and the latter typically varies over time. In addition, the optimal allocation depends on the analysis strategy used and the optimality criteria considered. In this article, we derive optimal treatment allocation rates for platform trials with shared controls, assuming that a stratified estimation and a testing procedure based on a regression model are used to adjust for time trends. We consider both, analysis using concurrent controls only as well as analysis methods using concurrent and non-concurrent controls and assume that the total sample size is fixed. The objective function to be minimized is the maximum of the variances of the effect estimators. We show that the optimal solution depends on the entry time of the arms in the trial and, in general, does not correspond to the square root of allocation rule used in classical multi-arm trials. We illustrate the optimal allocation and evaluate the power and type 1 error rate compared to trials using one-to-one and square root of allocations by means of a case study.

Simultaneous treatment effect estimation and variable selection for observational data

Due to the difficulties inherent in conducting controlled experiments, recent causal inference studies have focused on developing treatment effect estimation using observational data. One major difficulty in causal inference from observational data is that the underlying causal structure is unknown. This may result in the misidentification of potential sources of causal estimation bias, such as confounders, which must be controlled for accurate estimation. To consider all possible confounders and other relevant information, conventional methods predominantly use all observed covariates indiscriminately. However, previous studies have shown that including all covariates without considering their causal relationships may deteriorate the estimation performance. Although several data-driven variable selection methods have been proposed for treatment effect estimation, they cannot distinguish the confounders from other outcome-related covariates and are limited to simple regression forms. In this study, we propose a method called the Variable Selection Causal Estimation Network (VSCEN) that performs treatment effect estimation and causal variable selection simultaneously. Through end-to-end differentiable training, the VSCEN selects only the covariates beneficial for effect estimation and uses only those selected for effect estimation. Experimental evaluations on fully synthetic, semi-synthetic, and real datasets demonstrate the VSCEN’s superior performance in conditional average treatment effect estimation and competitive performance in average treatment effect estimation along with its accurate variable selection capabilities.

Contrast-specific propensity scores for causal inference with multiple interventions.

Existing methods that use propensity scores for heterogeneous treatment effect estimation on non-experimental data do not readily extend to the case of more than two treatment options. In this work, we develop a new propensity score-based method for heterogeneous treatment effect estimation when there are three or more treatment options, and prove that it generates unbiased estimates. We demonstrate our method on a real patient registry of patients in Singapore with diabetic dyslipidemia. On this dataset, our method generates heterogeneous treatment recommendations for patients among three options: Statins, fibrates, and non-pharmacological treatment to control patients' lipid ratios (total cholesterol divided by high-density lipoprotein level). In our numerical study, our proposed method generated more stable estimates compared to a benchmark method based on a multi-dimensional propensity score.

An operational approach for accurate 177Lu and 177mLu activity quantifications to comply with the environmental release criteria: the role of GEANT4 for efficiency curve and True Coincidence Summing effect estimation

The increasing use of 177Lu-labelled agents for targeted radionuclide cancer therapy highlights the radiation protection challenges in the management of radioactive waste due to 177mLu impurities. To ensure safe disposal, compliance with clearance criteria must be verified using calibrated systems, such as Hyper Pure Germanium (HPGe) spectrometers. This work aims to validate a customized GEANT4 model of our HPGe system in Marinelli beaker geometry to propose an operational approach properly quantifying 177Lu and 177mLu in waste samples. The system's efficiency curve was calculated by modelling gamma sources in the energy range of interest and validated by spectra measurements of 177Lu and 99mTc sources. Correction factors accounting for true coincidence summing (TCS) effect were simulated for 177Lu, 177mLu and 99mTc and they were applied to the spectrum measurement of a waste sample model with known activities of 177Lu and 177mLu. Thus, an operational approach for activities quantification was tested comparing the results with the nominal values. No significant differences were observed between simulated and measured efficiency values. TCS correction factors are significant only for 177mLu (1.6 at 112.95 keV and 204.11 keV). Eventually, the proposed framework to quantify 177Lu in a waste sample allowed to estimate the 177Lu and 177mLu component activities within a maximum 16% uncertainty. Results show that the HPGe model could be a powerful resource for a wide range of applications in daily clinical routine and it could be used to build a simple quality assurance program to monitor the detector response constancy in time.

Does household’s food and nutrient acquisition capacity predict linear growth in children? Analysis of longitudinal data from rural and small towns in Ethiopia

Nutritious foods are unaffordable to the majority of households in lower-income countries. Hence, starchy staples tend to be the main source of diets for men, women, and children, implying poor diet quality and deficiency in key nutrients that are crucial for child’s physical growth and cognitive development. Much of existing studies on child linear growth or stunting examined the impact of current dietary consumption on physical growth even though the effect of diets and nutrition would only be observed with time lag. This study examined the associations between sustained consumption of animal sourced foods (ASFs) and adequate nutrient intakes (measured by the mean nutrient adequacy (MPA)) and linear growth differences among children in rural and small towns in Ethiopia. Using nationally representative longitudinal data, results of fixed effects estimations suggested that a standard deviation increase in the household MPA was associated with a 10–11% increase in the height-for-age z-score; and a 10% increase in the household MPA was associated with a 3.2 – 5.6% reduction in stunting rate in children 6–59 months of age. Results also indicated some associations between past ASFs consumption and current linear growth outcomes in younger children. The gap in the household MPA was over 41%, indicating the need for nutrient-dense ASFs in the household diets despite the sustainability concerns around ASFs. Key implication is that policies and interventions aimed at improving linear growth outcomes of children in rural and small towns in Ethiopia should improve household’s access to ASFs and nutrient acquisition capacity.

The anti-inflammatory effect of acupuncture on peripheral inflammatory cytokines in patients with major depressive disorder: A systematic review and meta-analysis

There is robust evidence that inflammation is closely associated with major depressive disorder (MDD). The antidepressant effect of acupuncture has been associated with favorable changes in peripheral levels of inflammatory cytokines. However, the findings of these studies were not consistent. This systematic review and meta-analysis were performed to explore the peripheral inflammatory cytokines of acupuncture in treating MDD. The study protocol was registered at PROSPERO, CRD42021289207. We conducted a systematic search for eligible randomized controlled trials (RCTs) that reported acupuncture as an intervention for patients with MDD. The outcome of interest was changes in inflammatory cytokine levels measured before and after the intervention. We searched electronic databases including PubMed, Embase, Cochrane, SinoMed, Wanfang, China National Knowledge Infrastructure (CNKI), and Chongqing VIP (CQVIP). Nine studies including 848 patients were eligible and included. The findings from two studies suggest that acupuncture may lead to changes in inflammatory cytokine levels, irrespective of comparisons with sham acupuncture or Selective Serotonin Reuptake Inhibitors (SSRIs) and Selective Serotonin-Norepinephrine Reuptake Inhibitors (SNRIs). While the small sample size, no firm conclusions can be drawn yet. Compared with SSRIs/SNRIs alone, the pooled effect estimation indicates acupuncture combined with SSRIs/SNRIs decreased levels of pro-inflammatory cytokines IL-6, TNF-α and IFN-γ, and elevated levels of anti-inflammatory cytokines IL-4 and IL-10. This study demonstrates that acupuncture appears to have an efficacy in producing anti-inflammatory effects in individuals with MDD.

Spatio-temporal modeling of high-throughput multispectral aerial images improves agronomic trait genomic prediction in hybrid maize.

Design randomizations and spatial corrections have increased understanding of genotypic, spatial, and residual effects in field experiments, but precisely measuring spatial heterogeneity in the field remains a challenge. To this end, our study evaluated approaches to improve spatial modeling using high-throughput phenotypes (HTP) via unoccupied aerial vehicle (UAV) imagery. The normalized difference vegetation index was measured by a multispectral MicaSense camera and processed using ImageBreed. Contrasting to baseline agronomic trait spatial correction and a baseline multitrait model, a two-stage approach was proposed. Using longitudinal normalized difference vegetation index data, plot level permanent environment effects estimated spatial patterns in the field throughout the growing season. Normalized difference vegetation index permanent environment were separated from additive genetic effects using 2D spline, separable autoregressive models, or random regression models. The Permanent environment were leveraged within agronomic trait genomic best linear unbiased prediction either modeling an empirical covariance for random effects, or by modeling fixed effects as an average of permanent environment across time or split among three growth phases. Modeling approaches were tested using simulation data and Genomes-to-Fields hybrid maize (Zea mays L.) field experiments in 2015, 2017, 2019, and 2020 for grain yield, grain moisture, and ear height. The two-stage approach improved heritability, model fit, and genotypic effect estimation compared to baseline models. Electrical conductance and elevation from a 2019 soil survey significantly improved model fit, while 2D spline permanent environment were most strongly correlated with the soil parameters. Simulation of field effects demonstrated improved specificity for random regression models. In summary, the use of longitudinal normalized difference vegetation index measurements increased experimental accuracy and understanding of field spatio-temporal heterogeneity.

No star is good news: A unified look at rerandomization based on [formula omitted]-values from covariate balance tests

Randomized experiments balance all covariates on average and are considered the gold standard for estimating treatment effects. Chance imbalances are nonetheless common in realized treatment allocations. To inform readers of the comparability of treatment groups at baseline, contemporary scientific publications often report covariate balance tables with not only covariate means by treatment group but also the associated p-values from significance tests of their differences. The practical need to avoid small p-values as indicators of poor balance motivates balance check and rerandomization based on these p-values from covariate balance tests (ReP) as an attractive tool for improving covariate balance in designing randomized experiments. Despite the intuitiveness of such strategy and its possibly already widespread use in practice, the literature lacks results about its implications on subsequent inference, subjecting many effectively rerandomized experiments to possibly inefficient analyses. To fill this gap, we examine a variety of potentially useful schemes for ReP and quantify their impact on subsequent inference. Specifically, we focus on three estimators of the average treatment effect from the unadjusted, additive, and interacted linear regressions of the outcome on treatment, respectively, and derive their asymptotic sampling properties under ReP. The main findings are threefold. First, the estimator from the interacted regression is asymptotically the most efficient under all ReP schemes examined, and permits convenient regression-assisted inference identical to that under complete randomization. Second, ReP, in contrast to complete randomization, improves the asymptotic efficiency of the estimators from the unadjusted and additive regressions. Standard regression analyses are accordingly still valid but in general overconservative. Third, ReP reduces the asymptotic conditional biases of the three estimators and improves their coherence in terms of mean squared difference. These results establish ReP as a convenient tool for improving covariate balance in designing randomized experiments, and we recommend using the interacted regression for analyzing data from ReP designs.

Power Analyses for Estimation of Complier Average Causal Effects Under Random Encouragement Designs in Education Research: Theory and Guidance

Random encouragement designs evaluate treatments that aim to increase participation in a program or activity. These randomized controlled trials (RCTs) can also assess the mediated effects of participation itself on longer term outcomes using a complier average causal effect (CACE) estimation framework. This article considers power analysis methods for such CACE analyses for a range of RCT designs, including nonclustered, clustered, and random block designs. The focus is on behavioral encouragements to promote action, such as text messaging, that are increasingly being tested in education trials. We derive asymptotic distributions of the CACE estimators using generalized estimating equations theory, which underlie the power formulas. We incorporate noncompliance from both the actual receipt of the encouragement and participation itself. An illustrative power analysis provides sample size guidance using an available Shiny R dashboard.