Contamination bias in the estimation of child maltreatment causal effects on adolescent internalizing and externalizing behavior problems.

BACKGROUND: When unaddressed, contamination in child maltreatment research, in which some proportion of children recruited for a nonmaltreated comparison group are exposed to maltreatment, downwardly biases the significance and magnitude of effect size estimates. This study extends previous contamination research by investigating how a dual-measurement strategy of detecting and controlling contamination impacts causal effect size estimates of child behavior problems. METHODS: This study included 634 children from the LONGSCAN study with 63 cases of confirmed child maltreatment after age 8 and 571 cases without confirmed child maltreatment. Confirmed child maltreatment and internalizing and externalizing behaviors were recorded every 2 years between ages 4 and 16. Contamination in the nonmaltreated comparison group was identified and controlled by either a prospective self-report assessment at ages 12, 14, and 16 or by a one-time retrospective self-report assessment at age 18. Synthetic control methods were used to establish causal effects and quantify the impact of contamination when it was not controlled, when it was controlled for by prospective self-reports, and when it was controlled for by retrospective self-reports. RESULTS: Rates of contamination ranged from 62% to 67%. Without controlling for contamination, causal effect size estimates for internalizing behaviors were not statistically significant. Causal effects only became statistically significant after controlling contamination identified from either prospective or retrospective reports and effect sizes increased by between 17% and 54%. Controlling contamination had a smaller impact on effect size increases for externalizing behaviors but did produce a statistically significant overall effect, relative to the model ignoring contamination, when prospective methods were used. CONCLUSIONS: The presence of contamination in a nonmaltreated comparison group can underestimate the magnitude and statistical significance of causal effect size estimates, especially when investigating internalizing behavior problems. Addressing contamination can facilitate the replication of results across studies.

Evaluating Discrete Time Methods for Subgrouping Continuous Processes.

Rapid developments over the last several decades have brought increased focus and attention to the role of time scales and heterogeneity in the modeling of human processes. To address these emerging questions, subgrouping methods developed in the discrete-time framework-such as the vector autoregression (VAR)-have undergone widespread development to identify shared nomothetic trends from idiographic modeling results. Given the dependence of VAR-based parameters on the measurement intervals of the data, we sought to clarify the strengths and limitations of these methods in recovering subgroup dynamics under different measurement intervals. Building on the work of Molenaar and collaborators for subgrouping individual time-series by means of the subgrouped chain graphical VAR (scgVAR) and the subgrouping option in the group iterative multiple model estimation (S-GIMME), we present results from a Monte Carlo study aimed at addressing the implications of identifying subgroups using these discrete-time methods when applied to continuous-time data. Results indicate that discrete-time subgrouping methods perform well at recovering true subgroups when the measurement intervals are large enough to capture the full range of a system's dynamics, either via lagged or contemporaneous effects. Further implications and limitations are discussed therein.

A Square-Root Second-Order Extended Kalman Filtering Approach for Estimating Smoothly Time-Varying Parameters.

Researchers collecting intensive longitudinal data (ILD) are increasingly looking to model psychological processes, such as emotional dynamics, that organize and adapt across time in complex and meaningful ways. This is also the case for researchers looking to characterize the impact of an intervention on individual behavior. To be useful, statistical models must be capable of characterizing these processes as complex, time-dependent phenomenon, otherwise only a fraction of the system dynamics will be recovered. In this paper we introduce a Square-Root Second-Order Extended Kalman Filtering approach for estimating smoothly time-varying parameters. This approach is capable of handling dynamic factor models where the relations between variables underlying the processes of interest change in a manner that may be difficult to specify in advance. We examine the performance of our approach in a Monte Carlo simulation and show the proposed algorithm accurately recovers the unobserved states in the case of a bivariate dynamic factor model with time-varying dynamics and treatment effects. Furthermore, we illustrate the utility of our approach in characterizing the time-varying effect of a meditation intervention on day-to-day emotional experiences.

Affect and Personality: Ramifications of Modeling (Non-)Directionality in Dynamic Network Models.

The use of dynamic network models has grown in recent years. These models allow researchers to capture both lagged and contemporaneous effects in longitudinal data typically as variations, reformulations, or extensions of the standard vector autoregressive (VAR) models. To date, many of these dynamic networks have not been explicitly compared to one another. We compare three popular dynamic network approaches-GIMME, uSEM, and LASSO gVAR-in terms of their differences in modeling assumptions, estimation procedures, statistical properties based on a Monte Carlo simulation, and implications for affect and personality researchers. We found that all three approaches dynamic networks provided yielded group-level empirical results in partial support of affect and personality theories. However, individual-level results revealed a great deal of heterogeneity across approaches and participants. Reasons for discrepancies are discussed alongside these approaches' respective strengths and limitations.

A Limited Information Estimator for Dynamic Factor Models.

Structural equation modeling (SEM) is an increasingly popular method for examining multivariate time series data. As in cross-sectional data analysis, structural misspecification of time series models is inevitable, and further complicated by the fact that errors occur in both the time series and measurement components of the model. In this article, we introduce a new limited information estimator and local fit diagnostic for dynamic factor models within the SEM framework. We demonstrate the implementation of this estimator and examine its performance under both correct and incorrect model specifications via a small simulation study. The estimates from this estimator are compared to those from the most common system-wide estimators and are found to be more robust to the structural misspecifications considered.

What ergodicity means for you.

This paper explores the relation between within-person and between-person research designs using the concept of ergodicity from statistical mechanics in physics. We demonstrate the consequences of ergodicity using several real data examples from previously published studies. We then create several simulated examples that illustrate the independence of within-person processes from between-person differences, and pair these examples with analytic results that reinforce our conclusions. Finally, we discuss the plausibility of ergodicity being the general rule rather than the exception for social and behavioral processes, address common arguments against heeding the implications of ergodicity for behavioral research, and offer several possible solutions.

Associations between somatomotor-putamen resting state connectivity and obsessive-compulsive symptoms vary as a function of stress during early adolescence: Data from the ABCD study.

Obsessive-compulsive symptoms (OCS) are relatively common during adolescence although most individuals do not meet diagnostic criteria for obsessive-compulsive disorder (OCD). Nonetheless, OCS during adolescence are associated with comorbid psychopathologies and behavioral problems. Heightened levels of environmental stress and greater functional connectivity between the somatomotor network and putamen have been previously associated with elevated OCS in OCD patients relative to healthy controls. However, the interaction of these factors within the same sample of individuals has been understudied. This study examined somatomotor-putamen resting state connectivity, stress, and their interaction on OCS in adolescents from 9-12 years of age. Participants (n = 6386) were drawn from the ABCD Study 4.0 release. Multilevel modeling was used to account for nesting in the data and to assess changes in OCS in this age range. Stress moderated the association between somatomotor-putamen connectivity and OCS (ß = 0.35, S.E. = 0.13, p = 0.006). Participants who reported more stress than their average and had greater somatomotor-left putamen connectivity reported more OCS, whereas participants who reported less stress than their average and had greater somatomotor-left putamen connectivity reported less OCS. These data suggest that stress differentially affects the direction of association between somatomotor-putamen connectivity and OCS. Individual differences in the experience or perception of stress may contribute to more OCS in adolescents with greater somatomotor-putamen connectivity.

Blind Subgrouping of Task-based fMRI.

Significant heterogeneity in network structures reflecting individuals' dynamic processes can exist within subgroups of people (e.g., diagnostic category, gender). This makes it difficult to make inferences regarding these predefined subgroups. For this reason, researchers sometimes wish to identify subsets of individuals who have similarities in their dynamic processes regardless of any predefined category. This requires unsupervised classification of individuals based on similarities in their dynamic processes, or equivalently, in this case, similarities in their network structures of edges. The present paper tests a recently developed algorithm, S-GIMME, that takes into account heterogeneity across individuals with the aim of providing subgroup membership and precise information about the specific network structures that differentiate subgroups. The algorithm has previously provided robust and accurate classification when evaluated with large-scale simulation studies but has not yet been validated on empirical data. Here, we investigate S-GIMME's ability to differentiate, in a purely data-driven manner, between brain states explicitly induced through different tasks in a new fMRI dataset. The results provide new evidence that the algorithm was able to resolve, in an unsupervised data-driven manner, the differences between different active brain states in empirical fMRI data to segregate individuals and arrive at subgroup-specific network structures of edges. The ability to arrive at subgroups that correspond to empirically designed fMRI task conditions, with no biasing or priors, suggests this data-driven approach can be a powerful addition to existing methods for unsupervised classification of individuals based on their dynamic processes.

Estimating both directed and undirected contemporaneous relations in time series data using hybrid-group iterative multiple model estimation.

Researchers across varied fields increasingly are collecting and analyzing intensive longitudinal data (ILD) to examine processes across time at the individual level. Two types of relations are typically examined: lagged and contemporaneous. Lagged relations capture how variables at a prior time point can be used to explain variance in variables at a later time point. These are always modeled using auto- and cross-regressions by means of vector autoregression (VAR). By contrast, there are two types of relations commonly used to model the contemporaneous relations, which model how variables relate instantaneously. Until now, researchers must opt to either model contemporaneous relations as undirected relations among residuals (e.g., partial or full correlations) or as directed relations among the variables (e.g., paths or regressions). The choice for how to model contemporaneous relations has implications for inferences as well as the potential to introduce bias in the VAR lagged relations if the wrong type of relation is used. This article introduces a novel data-driven method, hybrid-group iterative multiple model estimation (GIMME), that provides a solution to the problem of having to choose one or the other type of contemporaneous relation to model. The modeling framework utilized in hybrid-GIMME allows for both types of contemporaneous relations in addition to the standard VAR relations. Both simulated and empirical data were used to test the performance of hybrid-GIMME. Results suggest this is a robust method for recovering contemporaneous relations in an exploratory manner, particularly with an ample number of time points per person. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

Contamination in Observational Research on Child Maltreatment: A Conceptual and Empirical Review With Implications for Future Research.

Contamination is a methodological phenomenon occurring in child maltreatment research when individuals in an established comparison condition have, in reality, been exposed to maltreatment during childhood. The current paper: (1) provides a conceptual and methodological introduction to contamination in child maltreatment research, (2) reviews the empirical literature demonstrating that the presence of contamination biases causal estimates in both prospective and retrospective cohort studies of child maltreatment effects, (3) outlines a dual measurement strategy for how child maltreatment researchers can address contamination, and (4) describes modern statistical methods for generating causal estimates in child maltreatment research after contamination is controlled. Our goal is to introduce the issue of contamination to researchers examining the effects of child maltreatment in an effort to improve the precision and replication of causal estimates that ultimately inform scientific and clinical decision-making as well as public policy.

Penalized Estimation and Forecasting of Multiple Subject Intensive Longitudinal Data.

Intensive longitudinal data (ILD) is an increasingly common data type in the social and behavioral sciences. Despite the many benefits these data provide, little work has been dedicated to realize the potential such data hold for forecasting dynamic processes at the individual level. To address this gap in the literature, we present the multi-VAR framework, a novel methodological approach allowing for penalized estimation of ILD collected from multiple individuals. Importantly, our approach estimates models for all individuals simultaneously and is capable of adaptively adjusting to the amount of heterogeneity present across individual dynamic processes. To accomplish this, we propose a novel proximal gradient descent algorithm for solving the multi-VAR problem and prove the consistency of the recovered transition matrices. We evaluate the forecasting performance of our method in comparison with a number of benchmark methods and provide an illustrative example involving the day-to-day emotional experiences of 16 individuals over an 11-week period.

An introduction to model implied instrumental variables using two stage least squares (MIIV-2SLS) in structural equation models (SEMs).

Structural equation models (SEMs) are widely used to handle multiequation systems that involve latent variables, multiple indicators, and measurement error. Maximum likelihood (ML) and diagonally weighted least squares (DWLS) dominate the estimation of SEMs with continuous or categorical endogenous variables, respectively. When a model is correctly specified, ML and DWLS function well. But, in the face of incorrect structures or nonconvergence, their performance can seriously deteriorate. Model implied instrumental variable, two stage least squares (MIIV-2SLS) estimates and tests individual equations, is more robust to misspecifications, and is noniterative, thus avoiding nonconvergence. This article is an overview and tutorial on MIIV-2SLS. It reviews the six major steps in using MIIV-2SLS: (a) model specification; (b) model identification; (c) latent to observed (L2O) variable transformation; (d) finding MIIVs; (e) using 2SLS; and (f) tests of overidentified equations. Each step is illustrated using a running empirical example from Reisenzein's (1986) randomized experiment on helping behavior. We also explain and illustrate the analytic conditions under which an equation estimated with MIIV-2SLS is robust to structural misspecifications. We include additional sections on MIIV approaches using a covariance matrix and mean vector as data input, conducting multilevel SEM, analyzing categorical endogenous variables, causal inference, and extensions and applications. Online supplemental material illustrates input code for all examples and simulations using the R package MIIVsem. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

Detecting Task-Dependent Functional Connectivity in Group Iterative Multiple Model Estimation with Person-Specific Hemodynamic Response Functions.

Introduction: Group iterative multiple model estimation (GIMME) has proven to be a reliable data-driven method to arrive at functional connectivity maps that represent associations between brain regions across time in groups and individuals. However, to date, GIMME has not been able to model time-varying task-related effects. This article introduces HRF-GIMME, an extension of GIMME that enables the modeling of the direct and modulatory effects of a task on functional magnetic resonance imaging data collected by using event-related designs. Critically, hemodynamic response function (HRF)-GIMME incorporates person-specific modeling of the HRF to accommodate known variability in onset delay and shape. Methods: After an introduction of the technical aspects of HRF-GIMME, the performance of HRF-GIMME is evaluated via both a simulation study and application to empirical data. The simulation study assesses the sensitivity and specificity of HRF-GIMME by using data simulated from one slow and two rapid event-related designs, and HRF-GIMME is then applied to two empirical data sets from similar designs to evaluate performance in recovering known neural circuitry. Results: HRF-GIMME showed high sensitivity and specificity across all simulated conditions, and it performed well in the recovery of expected relations between convolved task vectors and brain regions in both simulated and empirical data, particularly for the slow event-related design. Conclusion: Results from simulated and empirical data indicate that HRF-GIMME is a powerful new tool for obtaining directed functional connectivity maps of intrinsic and task-related connections that is able to uncover what is common across the sample as well as crucial individual-level path connections and estimates. Impact statement Group iterative multiple model estimation (GIMME) is a reliable method for creating functional connectivity maps of the connections between brain regions across time, and it is able to detect what is common across the sample and what is shared between subsets of participants, as well as individual-level path estimates. However, historically, GIMME does not model task-related effects. The novel HRF-GIMME algorithm enables the modeling of direct and modulatory task effects through individual-level estimation of the hemodynamic response function (HRF), presenting a powerful new tool for assessing task effects on functional connectivity networks in functional magnetic resonance imaging data.

An Instrumental Variable Estimator for Mixed Indicators: Analytic Derivatives and Alternative Parameterizations.

Methodological development of the model-implied instrumental variable (MIIV) estimation framework has proved fruitful over the last three decades. Major milestones include Bollen's (Psychometrika 61(1):109-121, 1996) original development of the MIIV estimator and its robustness properties for continuous endogenous variable SEMs, the extension of the MIIV estimator to ordered categorical endogenous variables (Bollen and Maydeu-Olivares in Psychometrika 72(3):309, 2007), and the introduction of a generalized method of moments estimator (Bollen et al., in Psychometrika 79(1):20-50, 2014). This paper furthers these developments by making several unique contributions not present in the prior literature: (1) we use matrix calculus to derive the analytic derivatives of the PIV estimator, (2) we extend the PIV estimator to apply to any mixture of binary, ordinal, and continuous variables, (3) we generalize the PIV model to include intercepts and means, (4) we devise a method to input known threshold values for ordinal observed variables, and (5) we enable a general parameterization that permits the estimation of means, variances, and covariances of the underlying variables to use as input into a SEM analysis with PIV. An empirical example illustrates a mixture of continuous variables and ordinal variables with fixed thresholds. We also include a simulation study to compare the performance of this novel estimator to WLSMV.

Latent variable GIMME using model implied instrumental variables (MIIVs).

Researchers across many domains of psychology increasingly wish to arrive at personalized and generalizable dynamic models of individuals' processes. This is seen in psychophysiological, behavioral, and emotional research paradigms, across a range of data types. Errors of measurement are inherent in most data. For this reason, researchers typically gather multiple indicators of the same latent construct and use methods, such as factor analysis, to arrive at scores from these indices. In addition to accurately measuring individuals, researchers also need to find the model that best describes the relations among the latent constructs. Most currently available data-driven searches do not include latent variables. We present an approach that builds from the strong foundations of group iterative multiple model estimation (GIMME), the idiographic filter, and model implied instrumental variables with two-stage least squares estimation (MIIV-2SLS) to provide researchers with the option to include latent variables in their data-driven model searches. The resulting approach is called latent variable GIMME (LV-GIMME). GIMME is utilized for the data-driven search for relations that exist among latent variables. Unlike other approaches such as the idiographic filter, LV-GIMME does not require that the latent variable model to be constant across individuals. This requirement is loosened by utilizing MIIV-2SLS for estimation. Simulated data studies demonstrate that the method can reliably detect relations among latent constructs, and that latent constructs provide more power to detect effects than using observed variables directly. We use empirical data examples drawn from functional MRI and daily self-report data. (PsycINFO Database Record (c) 2020 APA, all rights reserved).

Assessing the robustness of cluster solutions obtained from sparse count matrices.

Psychological researchers often seek to obtain cluster solutions from sparse count matrices (e.g., social networks; counts of symptoms that are in common for 2 given individuals; structural brain imaging). Increasingly, community detection methods are being used to subset the data in a data-driven manner. While many of these approaches perform well in simulation studies and thus offer some improvement upon traditional clustering approaches, there is no readily available approach for evaluating the robustness of these solutions in empirical data. Researchers have no way of knowing if their results are due to noise. We describe here 2 approaches novel to the field of psychology that enable evaluation of cluster solution robustness. This tutorial also explains the use of an associated R package, perturbR, which provides researchers with the ability to use the methods described herein. In the first approach, the cluster assignment from the original matrix is compared against cluster assignments obtained by randomly perturbing the edges in the matrix. Stable cluster solutions should not demonstrate large changes in the presence of small perturbations. For the second approach, Monte Carlo simulations of random matrices that have the same properties as the original matrix are generated. The distribution of quality scores ("modularity") obtained from the cluster solutions from these matrices are then compared with the score obtained from the original matrix results. From this, one can assess if the results are better than what would be expected by chance. perturbR automates these 2 methods, providing an easy-to-use resource for psychological researchers. We demonstrate the utility of this package using benchmark simulated data generated from a previous study and then apply the methods to publicly available empirical data obtained from social networks and structural neuroimaging. (PsycINFO Database Record (c) 2019 APA, all rights reserved).