Model Misspecification Research Articles

Environmental Degradation and its Effects on Health in Bangladesh

This study aims to analyze the impact of environmental degradation on health using the Autoregressive Distributed Lag (ARDL) method with data from 1976 to 2020. The primary focus is on the influence of CO2 emissions on life expectancy in Bangladesh, evaluating the short-term and long-term relationships among key variables such as CO2 emissions, urban population growth (UPG), money supply (M2), and renewable energy consumption (REC). The study employs unit root tests, including the Augmented Dickey-Fuller (ADF) and Phillips-Perron (PP) tests, to determine the stationarity of the time-series data. Results indicate mixed integration orders, necessitating the use of the ARDL bounds testing approach to confirm cointegration. The findings reveal that CO2 emissions have a significant negative effect on life expectancy, while M2 and REC exhibit varying impacts. The ARDL model confirms a long-term relationship between the variables, with a significant speed of adjustment towards equilibrium. Diagnostic tests validate the model’s robustness, ensuring no issues with heteroskedasticity, autocorrelation, or model misspecification. The study underscores the importance of effective governance in mitigating the adverse effects of environmental degradation on health, advocating for stricter environmental regulations and improved public health policies. These findings have significant implications for policymakers aiming to enhance public health outcomes amidst growing environmental challenges..

Doubly Robust Augmented Model Accuracy Transfer Inference with High Dimensional Features

Transfer learning is crucial for training models that generalize to unlabeled target populations using labeled source data, especially in real-world studies where label scarcity and covariate shift are common. While most research focuses on model estimation, there is limited literature on transfer inference for model accuracy despite its importance. We introduce a novel Doubly Robust Augmented Model Accuracy Transfer Inferen Ce (DRAMATIC) method for point and interval estimation of commonly used classification performance measures in an unlabeled target population with labeled source data. DRAMATIC derives and evaluates a potentially misspecified risk model for a binary response, leveraging high-dimensional adjustment features from both source and target data. It builds on an imputation model for the response mean and a density ratio model to characterize distributional shifts. The method constructs doubly robust estimators that are valid when either model is correctly specified and certain sparsity assumptions hold. Simulations show negligible bias in point estimation and satisfactory empirical coverage levels in confidence intervals. The utility of DRAMATIC is illustrated by transferring a genetic risk prediction model and its accuracy evaluation for type II diabetes across two patient cohorts in Mass General Brigham (MGB). Supplementary materials for this article are available online, including a standardized description of the materials available for reproducing the work.

An efficient workflow for modelling high-dimensional spatial extremes

We develop a comprehensive methodological workflow for Bayesian modelling of high-dimensional spatial extremes that lets us describe both weakening extremal dependence at increasing levels and changes in the type of extremal dependence class as a function of the distance between locations. This is achieved with a latent Gaussian version of the spatial conditional extremes model that allows for computationally efficient inference with R-INLA. Inference is made more robust using a post hoc adjustment method that accounts for possible model misspecification. This added robustness makes it possible to extract more information from the available data during inference using a composite likelihood. The developed methodology is applied to the modelling of extreme hourly precipitation from high-resolution radar data in Norway. Inference is performed quickly, and the resulting model fit successfully captures the main trends in the extremal dependence structure of the data. The post hoc adjustment is found to further improve model performance.

Testing Model Specification in Approximate Bayesian Computation Using Asymptotic Properties

We present a novel procedure to diagnose model misspecification in situations where inference is performed using approximate Bayesian computation (ABC). Unlike previous procedures, our proposal is based on the asymptotic properties of ABC. We demonstrate theoretically, and empirically that our procedure can consistently detect the presence of model misspecification. The examples demonstrate that our proposal shows good finite-sample properties, outperforming existing approaches. An empirical application to modeling exchange rate log returns using a g-and-k distribution completes the article. Supplementary materials for this article are available online.

Joint monitoring of location and scale for modern univariate processes

Autocorrelated sequences of individual observations arise in many modern-day statistical process monitoring (SPM) applications. Often times, interest involves jointly monitoring both process location and scale. To jointly monitor autocorrelated individuals data, it is common to first fit a time series model to the in-control process and subsequently use this model to de-correlate the observations so that traditional individuals and moving-range (I-MR) charts can be applied. If the time series model is correctly specified such that the resulting residuals are normal and independently distributed, then applying I-MR control charts to the residual process should work well. However, if the residual process deviates from normality and/or, due to time series model misspecification, contains levels of autocorrelation, the false alarm rate of such a strategy can dramatically rise. In this paper we propose improvements to a recently published distribution-free joint monitoring strategy that permits tighter control of the in-control average run length when model misspecification is a concern. We evaluate its average run length performance and conclude that the improved joint monitoring strategy proposed in this work is a very useful tool for today’s modern SPM practitioner. The proposed joint monitoring scheme is then applied to a real additive manufacturing process to illustrate its implementation in modern practice.

Simulating realistic patient profiles from pharmacokinetic models by a machine learning postprocessing correction of residual variability.

We address the problem of model misspecification in population pharmacokinetics (PopPK), by modeling residual unexplained variability (RUV) by machine learning (ML) methods in a postprocessing step after conventional model building. The practical purpose of the method is the generation of realistic virtual patient profiles and the quantification of the extent of model misspecification, by introducing an appropriate metric, to be used as an additional diagnostic of model quality. The proposed methodology consists of the following steps: After developing a PopPK model, the individual residual errors IRES = DV-IPRED, are computed, where DV are the observations and IPRED the individual predictions and are modeled by ML to obtain IRESML. Correction of the IPREDs can then be carried out as DVML = IPRED + IRESML. The methodology was tested in a PK study of ropinirole, for which a PopPK model was developed while a second deliberately misspecified model was also considered. Various supervised ML algorithms were tested, among which Random Forest gave the best results. The ML model was able to correct individual predictions as inspected in diagnostic plots and most importantly it simulated realistic profiles that resembled the real data and canceled out the artifacts introduced by the elevated RUV, even in the case of the heavily misspecified model. Furthermore, a metric to quantify the extent of model misspecification was introduced based on the R2 between IRES and IRESML, following the rationale that the greater the extent of variability explained by the ML model, the higher the degree of model misspecification present in the original model.

Individual-level models of disease transmission incorporating piecewise spatial risk functions

Modelling epidemics is crucial for understanding the emergence, transmission, impact and control of diseases. Spatial individual-level models (ILMs) that account for population heterogeneity are a useful tool, accounting for factors such as location, vaccination status and genetic information.Parametric forms for spatial risk functions, or kernels, are often used, but rely on strong assumptions about underlying transmission mechanisms. Here, we propose a class of non-parametric spatial disease transmission model, fitted within a Bayesian Markov chain Monte Carlo (MCMC) framework, allowing for more flexible assumptions when estimating the effect on spatial distance and infection risk.We focus upon two specific forms of non-parametric spatial infection kernel: piecewise constant and piecewise linear. Although these are relatively simple forms, we find them to produce results in line with, or superior to, parametric spatial ILMs. The performance of these models is examined using simulated data, including under circumstances of model misspecification, and then applied to data from the UK 2001 foot-and-mouth disease.

Strategyproof Decision-Making in Panel Data Settings and Beyond

We consider the problem of decision-making using panel data, in which a decision-maker gets noisy, repeated measurements of multiple units (or agents). We consider the setup used in synthetic control methods, where there is a pre-intervention period when the principal observes the outcomes of each unit, after which the principal uses these observations to assign a treatment to each unit. Unlike this classical setting, we permit the units generating the panel data to be strategic, i.e. units may modify their pre-intervention outcomes in order to receive a more desirable intervention. The principal's goal is to design a strategyproof intervention policy, i.e. a policy that assigns units their utility-maximizing interventions despite their potential strategizing. We first identify a necessary and sufficient condition under which a strategyproof intervention policy exists, and provide a strategyproof mechanism with a simple closed form when one does exist. Along the way, we prove impossibility results for strategic multiclass classification, a natural extension of the well-studied (binary) strategic classification problem to the multiclass setting, which may be of independent interest. When there are two interventions, we establish that there always exists a strategyproof mechanism, and provide an algorithm for learning such a mechanism. For three or more interventions, we provide an algorithm for learning a strategyproof mechanism if there exists a sufficiently large gap in the principal's rewards between different interventions. Finally, we empirically evaluate our model using real-world panel data collected from product sales over 18 months. We find that our methods compare favorably to baselines which do not take strategic interactions into consideration, even in the presence of model misspecification.

Loss-Based Variational Bayes Prediction

We propose a new approach to Bayesian prediction that caters for models with a large number of parameters and is robust to model misspecification. Given a class of high-dimensional (but parametric) predictive models, this new approach constructs a posterior predictive using a variational approximation to a generalized posterior that is directly focused on predictive accuracy. The theoretical behavior of the new prediction approach is analyzed and a form of optimality demonstrated. Applications to both simulated and empirical data using high-dimensional Bayesian neural network and autoregressive mixture models demonstrate that the approach provides more accurate results than various alternatives, including misspecified likelihood-based predictions. Supplementary materials for this article are available online.

Model-based bioequivalence approach for sparse pharmacokinetic bioequivalence studies: Model selection or model averaging?

Conventional pharmacokinetic (PK) bioequivalence (BE) studies aim to compare the rate and extent of drug absorption from a test (T) and reference (R) product using non-compartmental analysis (NCA) and the two one-sided test (TOST). Recently published regulatory guidance recommends alternative model-based (MB) approaches for BE assessment when NCA is challenging, as for long-acting injectables and products which require sparse PK sampling. However, our previous research on MB-TOST approaches showed that model misspecification can lead to inflated type I error. The objective of this research was to compare the performance of model selection (MS) on R product arm data and model averaging (MA) from a pool of candidate structural PK models in MBBE studies with sparse sampling. Our simulation study was inspired by a real case BE study using a two-way crossover design. PK data were simulated using three structural models under the null hypothesis and one model under the alternative hypothesis. MB-TOST was applied either using each of the five candidate models or following MS and MA with or without the simulated model in the pool. Assuming T and R have the same PK model, our simulation shows that following MS and MA, MB-TOST controls type I error rates at or below 0.05 and attains similar or even higher power than when using the simulated model. Thus, we propose to use MS prior to MB-TOST for BE studies with sparse PK sampling and to consider MA when candidate models have similar Akaike information criterion.

Curvilinearity in the Reference Composite and Practical Implications for Measurement

AbstractItem difficulty and dimensionality often correlate, implying that unidimensional IRT approximations to multidimensional data (i.e., reference composites) can take a curvilinear form in the multidimensional space. Although this issue has been previously discussed in the context of vertical scaling applications, we illustrate how such a phenomenon can also easily occur within individual tests. Measures of reading proficiency, for example, often use different task types within a single assessment, a feature that may not only lead to multidimensionality, but also an association between item difficulty and dimensionality. Using a latent regression strategy, we demonstrate through simulations and empirical analysis how associations between dimensionality and difficulty yield a nonlinear reference composite where the weights of the underlying dimensions change across the scale continuum according to the difficulties of the items associated with the dimensions. We further show how this form of curvilinearity produces systematic forms of misspecification in traditional unidimensional IRT models (e.g., 2PL) and can be better accommodated by models such as monotone‐polynomial or asymmetric IRT models. Simulations and a real‐data example from the Early Childhood Longitudinal Study—Kindergarten are provided for demonstration. Some implications for measurement modeling and for understanding the effects of 2PL misspecification on measurement metrics are discussed.

The Fundamental Role of Character Coding in Bayesian Morphological Phylogenetics.

Phylogenetic trees establish a historical context for the study of organismal form and function. Most phylogenetic trees are estimated using a model of evolution. For molecular data, modeling evolution is often based on biochemical observations about changes between character states. For example, there are 4 nucleotides, and we can make assumptions about the probability of transitions between them. By contrast, for morphological characters, we may not know a priori how many characters states there are per character, as both extant sampling and the fossil record may be highly incomplete, which leads to an observer bias. For a given character, the state space may be larger than what has been observed in the sample of taxa collected by the researcher. In this case, how many evolutionary rates are needed to even describe transitions between morphological character states may not be clear, potentially leading to model misspecification. To explore the impact of this model misspecification, we simulated character data with varying numbers of character states per character. We then used the data to estimate phylogenetic trees using models of evolution with the correct number of character states and an incorrect number of character states. The results of this study indicate that this observer bias may lead to phylogenetic error, particularly in the branch lengths of trees. If the state space is wrongly assumed to be too large, then we underestimate the branch lengths, and the opposite occurs when the state space is wrongly assumed to be too small.

Robust analysis of stepped wedge trials using composite likelihood models.

Stepped wedge trials (SWTs) are a type of cluster randomized trial that involve repeated measures on clusters and design-induced confounding between time and treatment. Although mixed models are commonly used to analyze SWTs, they are susceptible to misspecification particularly for cluster-longitudinal designs such as SWTs. Mixed model estimation leverages both "horizontal" or within-cluster information and "vertical" or between-cluster information. To use horizontal information in a mixed model, both the mean model and correlation structure must be correctly specified or accounted for, since time is confounded with treatment and measurements are likely correlated within clusters. Alternative non-parametric methods have been proposed that use only vertical information; these are more robust because between-cluster comparisons in a SWT preserve randomization, but these non-parametric methods are not very efficient. We propose a composite likelihood method that focuses on vertical information, but has the flexibility to recover efficiency by using additional horizontal information. We compare the properties and performance of various methods, using simulations based on COVID-19 data and a demonstration of application to the LIRE trial. We found that a vertical composite likelihood model that leverages baseline data is more robust than traditional methods, and more efficient than methods that use only vertical information. We hope that these results demonstrate the potential value of model-based vertical methods for SWTs with a large number of clusters, and that these new tools are useful to researchers who are concerned about misspecification of traditional models.

Covariate adjustment in randomized controlled trials: General concepts and practical considerations.

There has been a growing interest in covariate adjustment in the analysis of randomized controlled trials in past years. For instance, the US Food and Drug Administration recently issued guidance that emphasizes the importance of distinguishing between conditional and marginal treatment effects. Although these effects may sometimes coincide in the context of linear models, this is not typically the case in other settings, and this distinction is often overlooked in clinical trial practice. Considering these developments, this article provides a review of when and how to use covariate adjustment to enhance precision in randomized controlled trials. We describe the differences between conditional and marginal estimands and stress the necessity of aligning statistical analysis methods with the chosen estimand. In addition, we highlight the potential misalignment of commonly used methods in estimating marginal treatment effects. We hereby advocate for the use of the standardization approach, as it can improve efficiency by leveraging the information contained in baseline covariates while remaining robust to model misspecification. Finally, we present practical considerations that have arisen in our respective consultations to further clarify the advantages and limitations of covariate adjustment.

Nonparametric Machine Learning for Stochastic Frontier Analysis: A Bayesian Additive Regression Tree Approach

The stochastic frontier model is widely used in economics, finance, and management to estimate the production function and efficiency of a firm or industry. In the current literature of stochastic frontier analysis, parametric forms of production functions, such as Cobb-Douglas and Translog, are often assumed a priori without validation, which may suffer from model misspecification and lead to biased estimates of efficiency. To address this issue, a new set of stochastic frontier models is proposed, namely, 1) monBART-SFM that constructs a monotone-constrained nonparametric production function via an extension of the monotone Bayesian Additive Regression Tree (monBART) framework; 2) softBART-SFM that is built upon softBART, a smooth version of BART, where both models offer greater flexibility in modeling complex and nonlinear production functions that may have high dimensional inputs. The performance of the proposed method is illustrated through simulation studies and a real data application.

Bias versus variance when fitting multi-species molecular lines with a non-LTE radiative transfer model

Context. Robust radiative transfer techniques are requisite for efficiently extracting the physical and chemical information from molecular rotational lines. Aims. We study several hypotheses that enable robust estimations of the column densities and physical conditions when fitting one or two transitions per molecular species. We study the extent to which simplifying assumptions aimed at reducing the complexity of the problem introduce estimation biases and how to detect them. Methods. We focus on the CO and HCO+ isotopologues and analyze maps of a 50 square arcminutes field. We used the RADEX escape probability model to solve the statistical equilibrium equations and compute the emerging line profiles, assuming that all species coexist. Depending on the considered set of species, we also fixed the abundance ratio between some species and explored different values. We proposed a maximum likelihood estimator to infer the physical conditions and considered the effect of both the thermal noise and calibration uncertainty. We analyzed any potential biases induced by model misspecifications by comparing the results on the actual data for several sets of species and confirmed with Monte Carlo simulations. The variance of the estimations and the efficiency of the estimator were studied based on the Cramér-Rao lower bound. Results. Column densities can be estimated with 30% accuracy, while the best estimations of the volume density are found to be within a factor of two. Under the chosen model framework, the peak 12CO (1 – 0) is useful for constraining the kinetic temperature. The thermal pressure is better and more robustly estimated than the volume density and kinetic temperature separately. Analyzing CO and HCO+ isotopologues and fitting the full line profile are recommended practices with respect to detecting possible biases. Conclusions. Combining a non-local thermodynamic equilibrium model with a rigorous analysis of the accuracy allows us to obtain an efficient estimator and identify where the model is misspecified. We note that other combinations of molecular lines could be studied in the future.

Quantile control via random forest

This paper studies robust inference procedure for treatment effects in panel data with flexible relationship across units via the random forest method. The key contribution of this paper is twofold. First, we propose a direct construction of prediction intervals for the treatment effect by exploiting the information of the joint distribution of the cross-sectional units using random forest. In particular, we propose a Quantile Control Method (QCM) using the Quantile Random Forest (QRF) to accommodate flexible cross-sectional structure as well as high dimensionality. Second, we establish the asymptotic consistency of QRF under the panel/time series setup with high dimensionality, which is of theoretical interest on its own right. In addition, Monte Carlo simulations are conducted and show that prediction intervals via the QCM have excellent coverage probability for the treatment effects comparing to existing methods in the literature, and are robust to heteroskedasticity, autocorrelation, and various types of model misspecifications. Finally, an empirical application to study the effect of the economic integration between Hong Kong and mainland China on Hong Kong’s economy is conducted to highlight the potential of the proposed method.

A structural after measurement approach to structural equation modeling.

In structural equation modeling (SEM), the measurement and structural parts of the model are usually estimated simultaneously. In this article, we revisit the long-standing idea that we should first estimate the measurement part, and then estimate the structural part. We call this the "structural-after-measurement" (SAM) approach to SEM. We describe a formal framework for the SAM approach under settings where the latent variables and their indicators are continuous. We review earlier SAM methods and establish how they are specific instances of the SAM framework. Decoupled estimation for the measurement and structural parts using SAM possesses three key advantages over simultaneous estimation in standard SEM. First, estimates are more robust against local model misspecifications. Second, estimation routines are less vulnerable to convergence issues in small samples. Third, estimates exhibit smaller finite sample biases under correctly specified models. We propose two variants of the SAM approach. "Local" SAM expresses the mean vector and variance-covariance matrix of the latent variables as a function of the observed summary statistics and the parameters of the measurement model. "Global" SAM holds the parameters of the measurement part fixed while estimating the parameters of the structural part. Our framework includes two-step corrected standard errors, and permits computing both local and global fit measures. Nonetheless, the SAM approach is an estimation strategy, and should not be regarded as a model-building tool. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

The Growth Effect of Foreign Direct Investment: Evidence from Time-Frequency Analysis

The impact of FDI inflow on the economy of the host country had not been clearly defined, especially in the empirical literature. Most of the empirical results from past studies are distorted by model misspecification and endogeneity problems. To find a cure for these deficiencies, this study uses an unconventional methodology called wavelet coherence to examine the nexus between FDI inflows and per capita income growth in three African countries (The Gambia, Ghana, and Senegal. Wavelet coherence is a localized correlation in time-frequency space that examines the dynamic nexus between two time series. Most traditional econometric methods that were used to study the nexus between FDI inflows and income assumed that the series are stationary, however, most economic time series are not stationary, hence, this makes most traditional methods ineffective in finding the nexus between FDI and income. The use of wavelet coherence can overcome this challenge because it does not require the assumption of stationarity of the data.The empirical results of the study showed that the impact of FDI inflows depends on the degree to which the FDI is a complement or a substitute for domestic investment. Senegal where FDI complements domestic investment in enhancing growth derived more benefit from FDI inflows than The Gambia where FDI inflows substitute domestic investment and Ghana where FDI inflows had no impact on domestic investment. The causality between per capita income growth and per stock of FDI inflows runs from per stock of FDI inflows or the two series move together.

Small-Sample Bias Correction of Inequality Estimators in Complex Surveys

Income inequality estimators are biased in small samples, leading generally to an underestimation. This aspect deserves particular attention when estimating inequality in small domains and performing small area estimation at the area level. We propose a bias correction framework for a large class of inequality measures comprising the Gini Index, the Generalized Entropy, and the Atkinson index families by accounting for complex survey designs. The proposed methodology does not require any parametric assumption on income distribution, being very flexible. Design-based performance evaluation of our proposal has been carried out using EU-SILC data, their results show a noticeable bias reduction for all the measures. Lastly, an illustrative example of application in small area estimation confirms that ignoring ex-ante bias correction determines model misspecification.