Assessing moderated mediation in linear models requires fewer confounding assumptions than assessing mediation

It is well known from the mediation analysis literature that the identification of direct and indirect effects relies on strong no unmeasured confounding assumptions of no unmeasured confounding. Even in randomized studies the mediator may still be correlated with unobserved prognostic variables that affect the outcome, in which case the mediator's role in the causal process may not be inferred without bias. In the behavioural and social science literature very little attention has been given so far to the causal assumptions required for moderated mediation analysis. In this paper we focus on the index for moderated mediation, which measures by how much the mediated effect is larger or smaller for varying levels of the moderator. We show that in linear models this index can be estimated without bias in the presence of unmeasured common causes of the moderator, mediator and outcome under certain conditions. Importantly, one can thus use the test for moderated mediation to support evidence for mediation under less stringent confounding conditions. We illustrate our findings with data from a randomized experiment assessing the impact of being primed with social deception upon observer responses to others’ pain, and from an observational study of individuals who ended a romantic relationship assessing the effect of attachment anxiety during the relationship on mental distress 2 years after the break‐up.     

A general approach to causal mediation analysis

Traditionally in the social sciences, causal mediation analysis has been formulated, understood, and implemented within the framework of linear structural equation models. We argue and demonstrate that this is problematic for 3 reasons: the lack of a general definition of causal mediation effects independent of a particular statistical model, the inability to specify the key identification assumption, and the difficulty of extending the framework to nonlinear models. In this article, we propose an alternative approach that overcomes these limitations. Our approach is general because it offers the definition, identification, estimation, and sensitivity analysis of causal mediation effects without reference to any specific statistical model. Further, our approach explicitly links these 4 elements closely together within a single framework. As a result, the proposed framework can accommodate linear and nonlinear relationships, parametric and nonparametric models, continuous and discrete mediators, and various types of outcome variables. The general definition and identification result also allow us to develop sensitivity analysis in the context of commonly used models, which enables applied researchers to formally assess the robustness of their empirical conclusions to violations of the key assumption. We illustrate our approach by applying it to the Job Search Intervention Study. We also offer easy-to-use software that implements all our proposed methods.     

纵向数据的中介效应分析

目前中介效应检验主要是基于截面数据,但许多时候截面数据的中介分析不适合进行因果推断,因而需要收集历时性的纵向数据,进行纵向数据的中介分析。评介了基于交叉滞后面板模型、多层线性模型和潜变量增长模型的纵向数据的中介分析方法及其四个发展。第一,中介效应随时间变化,如连续时间模型、多层时变系数模型。第二,中介效应随个体变化,如随机效应的交叉滞后面板模型和多层自回归模型。第三,中介模型的整合,如交叉滞后面板模型与多层线性模型整合为多层自回归模型。第四,中介检验方法的发展,建议使用Monte Carlo、Bootstrap和贝叶斯法进行纵向数据的中介分析。总结出一个纵向数据的中介分析流程并给出相应的Mplus程序。随后展望了纵向数据的中介分析的拓展方向。     

Nonlinear Predictive Models for Multiple Mediation Analysis: With an Application to Explore Ethnic Disparities in Anxiety and Depression Among Cancer Survivors

Mediation analysis allows the examination of effects of a third variable (mediator/confounder) in the causal pathway between an exposure and an outcome. The general multiple mediation analysis method (MMA), proposed by Yu et al., improves traditional methods (e.g., estimation of natural and controlled direct effects) to enable consideration of multiple mediators/confounders simultaneously and the use of linear and nonlinear predictive models for estimating mediation/confounding effects. Previous studies find that compared with non-Hispanic cancer survivors, Hispanic survivors are more likely to endure anxiety and depression after cancer diagnoses. In this paper, we applied MMA on MY-Health study to identify mediators/confounders and quantify the indirect effect of each identified mediator/confounder in explaining ethnic disparities in anxiety and depression among cancer survivors who enrolled in the study. We considered a number of socio-demographic variables, tumor characteristics, and treatment factors as potential mediators/confounders and found that most of the ethnic differences in anxiety or depression between Hispanic and non-Hispanic white cancer survivors were explained by younger diagnosis age, lower education level, lower proportions of employment, less likely of being born in the USA, less insurance, and less social support among Hispanic patients.     

Using session-by-session measurement to compare mechanisms of action for acceptance and commitment therapy and cognitive therapy

Debate continues about the extent to which postulated mechanisms of action of cognitive behavior therapies (CBT), including standard CBT (i.e., Beckian cognitive therapy [CT]) and acceptance and commitment therapy (ACT) are supported by mediational analyses. Moreover, the distinctiveness of CT and ACT has been called into question. One contributor to ongoing uncertainty in this arena is the lack of time-varying process data. In this study, 174 patients presenting to a university clinic with anxiety or depression who had been randomly assigned to receive either ACT or CT completed an assessment of theorized mediators and outcomes before each session. Hierarchical linear modeling of session-by-session data revealed that increased utilization of cognitive and affective change strategies relative to utilization of psychological acceptance strategies mediated outcome for CT, whereas for ACT the mediation effect was in the opposite direction. Decreases in self-reported dysfunctional thinking, cognitive "defusion" (the ability to see one's thoughts as mental events rather than necessarily as representations of reality), and willingness to engage in behavioral activity despite unpleasant thoughts or emotions were equivalent mediators across treatments. These results have potential implications for the theoretical arguments behind, and distinctiveness of, CT and ACT.     

Job security and work performance in Chinese employees: The mediating role of organisational identification

This research focuses on investigating whether organisational identification mediates the effects of job security on in‐role behaviour and extra‐role behaviour and how these mediation mechanisms differ according to gender. Through analysing 212 supervisor‐subordinate dyads from a Chinese air transportation group, the research indicated that organisational identification partially mediated the effect of job security on in‐role behaviour and fully mediated the effect of job security on extra‐role behaviour. A multi‐group analysis also showed that there were significant differences between male and female employees in these relationships. In addition, moderated mediation analyses showed that gender moderated the indirect effects of job security on in‐role behaviour and extra‐role behaviour through organisational identification. Limitations and implications of these findings are discussed.     

Bias in cross-sectional analyses of longitudinal mediation

Most empirical tests of mediation utilize cross-sectional data despite the fact that mediation consists of causal processes that unfold over time. The authors considered the possibility that longitudinal mediation might occur under either of two different models of change: (a) an autoregressive model or (b) a random effects model. For both models, the authors demonstrated that cross-sectional approaches to mediation typically generate substantially biased estimates of longitudinal parameters even under the ideal conditions when mediation is complete. In longitudinal models where variable M completely mediates the effect of X on Y, cross-sectional estimates of the direct effect of X on Y, the indirect effect of X on Y through M, and the proportion of the total effect mediated by M are often highly misleading.     

Time-Varying Treatments in Observational Studies: Marginal Structural Models of the Effects of Early Grade Retention on Math Achievement

One of the main objectives of many empirical studies in the social and behavioral sciences is to assess the causal effect of a treatment or intervention on the occurrence of a certain event. The randomized controlled trial is generally considered the gold standard to evaluate such causal effects. However, for ethical or practical reasons, social scientists are often bound to the use of nonexperimental, observational designs. When the treatment and control group are different with regard to variables that are related to the outcome, this may induce the problem of confounding. A variety of statistical techniques, such as regression, matching, and subclassification, is now available and routinely used to adjust for confounding due to measured variables. However, these techniques are not appropriate for dealing with time-varying confounding, which arises in situations where the treatment or intervention can be received at multiple timepoints. In this article, we explain the use of marginal structural models and inverse probability weighting to control for time-varying confounding in observational studies. We illustrate the approach with an empirical example of grade retention effects on mathematics development throughout primary school.     

Estimation of Controlled Direct Effects in Longitudinal Mediation Analyses with Latent Variables in Randomized Studies

Abstract

In a randomized study with longitudinal data on a mediator and outcome, estimating the direct effect of treatment on the outcome at a particular time requires adjusting for confounding of the association between the outcome and all preceding instances of the mediator. When the confounders are themselves affected by treatment, standard regression adjustment is prone to severe bias. In contrast, G-estimation requires less stringent assumptions than path analysis using SEM to unbiasedly estimate the direct effect even in linear settings. In this article, we propose a G-estimation method to estimate the controlled direct effect of treatment on the outcome, by adapting existing G-estimation methods for time-varying treatments without mediators. The proposed method can accommodate continuous and noncontinuous mediators, and requires no models for the confounders. Unbiased estimation only requires correctly specifying a mean model for either the mediator or the outcome. The method is further extended to settings where the mediator or outcome, or both, are latent, and generalizes existing methods for single measurement occasions of the mediator and outcome to longitudinal data on the mediator and outcome. The methods are utilized to assess the effects of an intervention on physical activity that is possibly mediated by motivation to exercise in a randomized study.     

Bayesian Causal Mediation Analysis for Group Randomized Designs with Homogeneous and Heterogeneous Effects: Simulation and Case Study

A fully Bayesian approach to causal mediation analysis for group-randomized designs is presented. A unique contribution of this article is the combination of Bayesian inferential methods with G-computation to address the problem of heterogeneous treatment or mediator effects. A detailed simulation study shows that this approach has excellent frequentist properties, particularly in the case of small sample sizes with accurate informative priors. The simulation study also demonstrates that the proposed approach can take into account heterogeneous treatment or mediator effects without bias. A case study using data from a school-based randomized intervention designed to increase parent social capital leading to improved behavioral and academic outcomes in children is offered to illustrate the Bayesian approach to causal mediation in group-randomized designs.     

A Regression Framework for Causal Mediation Analysis with Applications to Behavioral Science

We introduce and extend the classical regression framework for conducting mediation analysis from the fit of only one model. Using the essential mediation components (EMCs) allows us to estimate causal mediation effects and their analytical variance. This single-equation approach reduces computation time and permits the use of a rich suite of regression tools that are not easily implemented on a system of three equations. Additionally, we extend this framework to non-nested mediation systems, provide a joint measure of mediation for complex mediation hypotheses, propose new visualizations for mediation effects, and explain why estimates of the total effect may differ depending on the approach used. Using data from social science studies, we also provide extensive illustrations of the usefulness of this framework and its advantages over traditional approaches to mediation analysis. The example data are freely available for download online and we include the R code necessary to reproduce our results.     

Forgive them for I have sinned: The relationship between guilt and forgiveness of others' transgressions

We propose that guilt leads to forgiveness of others' transgressions. In Study 1, people prone to experience guilt (but not shame) were also prone to forgive others for past misdeeds. In Study 2, we manipulated harm‐ and inequity‐based guilt; both increased forgiveness of others' transgressions. Further, the effect of guilt on forgiveness was mediated by identification with the transgressor. In Study 3, we replicated the guilt–forgiveness relationship and examined three other plausible mediators: capability for similar wrongdoing, empathic understanding, and general identification; only identification with the transgressor satisfied the criteria for mediation. In Study 4, we induced guilt by asking participants to harm a friend or stranger. Guilt induced by harming a friend led to greater forgiveness of third‐party transgressors, and again, identification with the transgressor mediated the effect. We discuss the implications of these results for understanding how the prosocial effects of guilt extend beyond the boundaries of a single interpersonal relationship.     

An Alternative Framework for Defining Mediation

The present article provides an alternative framework for evaluating mediated relationships. From this perspective. a mediated process is a chain reaction, beginning with an independent variable that affects a mediator that in turn affects an outcome. The definition of mediation offered here, presented for stage sequences, states three conditions for establishing mediation: (a) the independent variable affects the probability of the sequence no mediator to mediator to outcome; (b) the independent variable affects the probability of a transition into the mediator stage; (c) the mediator affects the probability of a transition into the outcome stage at every level of the independent variable. This definition of mediation is compared and contrasted with the well-known definition of mediation for continuous variables discussed in Baron and Kenny (1986), Judd and Kenny (1981), and Kenny, Kashy, and Bolger (1997). The definition presented in this article emphasizes the intraindividual, time-ordered nature of mediation.     

Schemas Mediate the Link Between Procrastination and Depression: Results from the United States and Pakistan

The current study extended the Procrastination-Health model by examining a multiple mediation model, with two cognitive schemas (defectiveness; insufficient self-control) serving as mediators. The models were as follows: procrastination → defectiveness → depression; procrastination → insufficient self-control → depression. Participants included 412 (271 women, 141 men) United States (US) and 240 (107 women, 133 men) Pakistani college students, who responded via self-report questionnaires. In the US sample, results revealed a non-significant direct effect between procrastination and depression after consideration for the two cognitive schemas, suggesting the schemas completely mediated the model. Both defectiveness and insufficient self-control schemas were significant individual mediators. In the Pakistani model, results revealed a significant direct effect and indirect effect through the two cognitive schemas, indicating partial mediation. Only the indirect path through defectiveness schemas was significant in the Pakistani model. Given slight differences in the two models, a moderated-mediation model was analyzed to determine if the strength of the direct and indirect effects varied by nationality. The strength of the direct and indirect effects was not moderated by nationality. Overall, this is the first study to identify cognitive mediators in the procrastination-depression relationship. Such findings represent a significant extension of the Procrastination-Health model and offer some unique cognitive insights into culturally sensitive conceptualizations and treatments for depression.     

Attitude similarity and attraction: Validation,positive affect,and trust as sequential mediators

Effectance motivation—an urge for certainty and a feeling of being able to know, predict, and control one's environment—was initially proposed as the mechanism underlying attitude similarity effects on attraction. However, this motivation was discarded as an explanation when positive affect was identified. The presence of alternative mechanisms did not deny a role for the validation of attitudes in attraction. Therefore, we investigated the validation of one's views by those of peers as an additional mediator and its relation with two previously known mediators of positive affect and trust. As hypothesized, validation mediated attitude similarity effects when measured alone (Experiment 1) and within sequential mediation patterns involving positive affect (Experiment 2A) and trust (Experiments 2B and 2C).     

基于结构方程模型的多重中介效应分析

多重中介模型是指存在多个中介变量的模型。多重中介模型可以分析特定中介效应、总的中介效应和对比中介效应。指出了目前多重中介模型分析普遍存在的问题,包括分析不完整、使用Sobel检验带来的局限。建议通过增加辅助变量的方法进行完整的多重中介效应分析,使用偏差校正的Bootstrap方法进行中介检验。总结出一个多重中介SEM分析流程,并有示例和相应的MPLUS程序。随后展望了辅助变量和中介效应检验方法的发展方向。     

Latent Class Dynamic Mediation Model with Application to Smoking Cessation Data

Traditional mediation analysis assumes that a study population is homogeneous and the mediation effect is constant over time, which may not hold in some applications. Motivated by smoking cessation data, we propose a latent class dynamic mediation model that explicitly accounts for the fact that the study population may consist of different subgroups and the mediation effect may vary over time. We use a proportional odds model to accommodate the subject heterogeneities and identify latent subgroups. Conditional on the subgroups, we employ a Bayesian hierarchical nonparametric time-varying coefficient model to capture the time-varying mediation process, while allowing each subgroup to have its individual dynamic mediation process. A simulation study shows that the proposed method has good performance in estimating the mediation effect. We illustrate the proposed methodology by applying it to analyze smoking cessation data.

     

The Mediated Moderation Model of Interactivity

This paper argues for enhanced consideration of third variables in interactivity research and proposes a “mediated moderation” model to bring increased sophistication to bear on the study of information technology effects. Interactivity, a central phenomenon in new media research, is an elusive concept that has enduringly intrigued and confused scholars. Extant conceptualizations have produced incomplete causal models and have generally ignored the effect of third variables. We conceptualize interactivity as technological attributes of mediated environments that enable reciprocal communication or information exchange, which afford interaction between communication technology and users, or between users through technology. Specifying roles for mediator and moderator variables, this paper proposes a model that incorporates interactive attributes, user perceptions (mediators such as perceived interactivity), individual differences (moderators such as Internet self-efficacy), and media effects measures to systematically examine the definition, process, and consequences of interactivity on users. Lastly, statistical procedures for testing mediated moderation are described.     

BANOVA: Bayesian Analysis of Experiments in Consumer Psychology

This article introduces a Bayesian extension of ANOVA for the analysis of experimental data in consumer psychology. The approach, called BANOVA (Bayesian ANOVA), addresses some common challenges that consumer psychologists encounter in their experimental work, and is specifically suited for the analysis of repeated measures designs. There appears to be a recent surge in interest in those designs based on the recognition that they are sensitive to individual differences in response to experimental treatments and that they offer advantages for assessing causal mediating mechanisms, even at the individual level. BANOVA enables the analysis of repeated measures data derived from mixed within–between‐subjects experiments with Normal and nonNormal‐dependent variables and accommodates unobserved individual differences. It allows for the calculation of effect sizes, planned comparisons, simple effects, spotlight and floodlight analyses, and includes a wide range of mediation, moderation, and moderated mediation analyses. An R software package implements these analyses, and aims to provide a one‐stop shop for the analysis of experiments in consumer psychology. The package is illustrated through applications to a number of data sets from previously published studies.     

How do psychological treatments work? Investigating mediators of change

Little is known about how psychological treatments work. Research on treatment-induced mediators of change may be of help in identifying potential causal mechanisms through which they operate. Outcome-focused randomised controlled trials provide an excellent opportunity for such work. However, certain conceptual and practical difficulties arise when studying psychological treatments, most especially deciding how best to conceptualise the treatment concerned and how to accommodate the fact that most psychological treatments are implemented flexibly. In this paper, these difficulties are discussed, and strategies and procedures for overcoming them are described.