The Correspondence Between Causal and Traditional Mediation Analysis: the Link Is the Mediator by Treatment Interaction

Mediation analysis is a methodology used to understand how and why behavioral phenomena occur. New mediation methods based on the potential outcomes framework are a seminal advancement for mediation analysis because they focus on the causal basis of mediation. Despite the importance of the potential outcomes framework in other fields, the methods are not well known in prevention and other disciplines. The interaction of a treatment (X) and a mediator (M) on an outcome variable (Y) is central to the potential outcomes framework for causal mediation analysis and provides a way to link traditional and modern causal mediation methods. As described in the paper, for a continuous mediator and outcome, if the XM interaction is zero, then potential outcomes estimators of the mediated effect are equal to the traditional model estimators. If the XM interaction is nonzero, the potential outcomes estimators correspond to simple direct and simple mediated contrasts for the treatment and the control groups in traditional mediation analysis. Links between traditional and causal mediation estimators clarify the meaning of potential outcomes framework mediation quantities. A simulation study demonstrates that testing for a XM interaction that is zero in the population can reduce power to detect mediated effects, and ignoring a nonzero XM interaction in the population can also reduce power to detect mediated effects in some situations. We recommend that prevention scientists incorporate evaluation of the XM interaction in their research.

The use of traditional and causal estimators for mediation models with a binary outcome and exposure-mediator interaction

An important recent development in mediation analysis is the use of causal mediation analysis. Causal mediation analysis decomposes the total exposure effect into causal direct and indirect effects in the presence of exposure-mediator interaction. However, in practice, traditional mediation analysis is still most widely used. The aim of this paper is to demonstrate the similarities and differences between the causal and traditional estimators for mediation models with a continuous mediator, a binary outcome, and exposure-mediator interaction. A real-life data example, analytical comparisons, and a simulation study were used to demonstrate the similarities and differences between the traditional and causal estimators. The causal and traditional estimators provide similar indirect effect estimates, but different direct and total effect estimates. Traditional mediation analysis may only be used when conditional direct effect estimates are of interest. Causal mediation analysis is the generally preferred method as its casual effect estimates help unravel causal mechanisms.

Causal Mediation Analysis in the Presence of Post-treatment Confounding Variables: A Monte Carlo Simulation Study

In many disciplines, mediating processes are usually investigated with randomized experiments and linear regression to determine if the treatment affects the outcome through a mediator. However, randomizing the treatment will not yield accurate causal direct and indirect estimates unless certain assumptions are satisfied since the mediator status is not randomized. This study describes methods to estimate causal direct and indirect effects and reports the results of a large Monte Carlo simulation study on the performance of the ordinary regression and modern causal mediation analysis methods, including a previously untested doubly robust sequential g-estimation method, when there are confounders of the mediator-to-outcome relation. Results show that failing to measure and incorporate potential post-treatment confounders in a mediation model leads to biased estimates, regardless of the analysis method used. Results emphasize the importance of measuring potential confounding variables and conducting sensitivity analysis.

Exposure to ethnic discrimination in social media and symptoms of anxiety and depression among Hispanic emerging adults: Examining the moderating role of gender

METHOD

Two hundred Hispanic emerging adults from Arizona (n = 99) and Florida (n = 101) completed a cross-sectional survey, and data were analyzed using hierarchical multiple regression and moderation analyses.

RESULTS

Higher social media discrimination was associated with higher symptoms of depression and generalized anxiety. Moderation analyses indicated that higher social media discrimination was only associated with symptoms of depression and generalized anxiety among men, but not women.

CONCLUSION

This is likely the first study on social media discrimination and mental health among emerging adults; thus, expanding this emerging field of research to a distinct developmental period.

Causal Mediation Programs in R, Mplus, SAS, SPSS, and Stata

Mediation analysis is a methodology used to understand how and why an independent variable (X) transmits its effect to an outcome (Y) through a mediator (M). New causal mediation methods based on the potential outcomes framework and counterfactual framework are a seminal advancement for mediation analysis, because they focus on the causal basis of mediation analysis. There are several programs available to estimate causal mediation effects, but these programs differ substantially in data set up, estimation, output, and software platform. To compare these programs, an empirical example is presented, and a single mediator model with XM interaction was estimated with a continuous mediator and a continuous outcome in each program. Even though the software packages employ different estimation methods, they do provide similar causal effect estimates for mediation models with a continuous mediator and outcome. A detailed explanation of program similarities, unique features, and recommendations are discussed.

Correlation Database of 60 Cross-Disciplinary Surveys and Cognitive Tasks Assessing Self-Regulation

Self-regulation is studied across various disciplines, including personality, social, cognitive, health, developmental, and clinical psychology; psychiatry; neuroscience; medicine; pharmacology; and economics. Widespread interest in self-regulation has led to confusion regarding both the constructs within the nomological network of self-regulation and the measures used to assess these constructs. To facilitate the integration of cross-disciplinary measures of self-regulation, we estimated product-moment and distance correlations among 60 cross-disciplinary measures of self-regulation (23 self-report surveys, 37 cognitive tasks) and measures of health and substance use based on 522 participants. The correlations showed substantial variability, though the surveys demonstrated greater convergent validity than did the cognitive tasks. Variables derived from the surveys only weakly correlated with variables derived from the cognitive tasks (M = .049, range = .000 to .271 for the absolute value of the product-moment correlation; M = .085, range = .028 to .241 for the distance correlation), thus challenging the notion that these surveys and cognitive tasks measure the same construct. We conclude by outlining several potential uses for this publicly available database of correlations.

A Viable Alternative When Propensity Scores Fail: Evaluation of Inverse Propensity Weighting and Sequential G-Estimation in a Two-Wave Mediation Model

Two methods from the potential outcomes framework - inverse propensity weighting (IPW) and sequential G-estimation - were evaluated and compared to linear regression for estimating the mediated effect in a two-wave design with a randomized intervention and continuous mediator and outcome. Baseline measures of the mediator and outcome can be considered confounders of the follow-up mediator - outcome relation for which adjustment is necessary to eliminate bias. To adjust for baseline measures of the mediator and outcome, IPW uses stabilized inverse propensity weights whereas sequential G-estimation uses regression adjustment. Theoretical differences between the models are described, and Monte Carlo simulations compared the performance of linear regression; IPW without weight truncation; IPW with weights truncated at the 1st/99th, 5th/95th, and 10th/90th percentiles; and sequential G-estimation. Sequential G-estimation performed similarly to linear regression, but IPW provided a biased estimate of the mediated effect, lower power, lower confidence interval coverage, and higher mean squared error. Simulation results show that IPW failed to fully adjust the follow-up mediator - outcome relation for confounding due to the baseline measures. We then compared the mediated effect estimates using data from a randomized experiment evaluating a steroid prevention program for high school athletes. Implications and future directions are discussed.

A Psychometric Analysis of the Brief Self-Control Scale

The Brief Self-Control Scale (BSCS) is a widely used measure of self-control, a construct associated with beneficial psychological outcomes. Several studies have investigated the psychometric properties of the BSCS but have failed to reach consensus. This has resulted in an unstable and ambiguous understanding of the scale and its psychometric properties. The current study sought resolution by implementing scale evaluation approaches guided by modern psychometric literature. Additionally, our goal was to provide a more comprehensive item analysis via the item response theory (IRT) framework. Results from the current study support both unidimensional and multidimensional factor structures for the 13-item version of the BSCS. The addition of an IRT analysis provided a new perspective on item- and test-level functioning. The goal of a more defensible psychometric grounding for the BSCS is to promote greater consistency, stability, and trust in future results.

A Psychometric Evaluation of the Short Grit Scale: A Closer Look at its Factor Structure and Scale Functioning

Grit, the passion and perseverance for long-term goals, has received attention from personality psychologists because it predicts success and academic achievement. Grit has also been criticized as simply another measure of self-control or conscientiousness. A precise psychometric representation of grit is needed to understand how the construct is unique and how it overlaps with other constructs. Previous research suggests that the Short Grit Scale (Grit-S) has several psychometric limitations, such as uncertain factor structure within and across populations, uncertainty about reporting total or subscale scores, and different assessment precision at low and high levels on the construct. We conducted modern psychometric techniques including parallel analysis, measurement invariance, extrinsic convergent validity, and Item Response Theory models on two American samples. Our results suggest that the Grit-S is essentially unidimensional and that there is construct overlap with the self-control construct. Subscale factors were the result of an item doublet, where two items had high correlated uniquenesses, showed similar item information, and were more likely to exhibit measurement bias. Findings replicated across samples. Finally, we discuss recommendations for the use of the Grit-S based on the theoretical interpretation of the unidimensional factor and our empirical findings.

Mediation analysis with binary outcomes: Direct and indirect effects of pro-alcohol influences on alcohol use disorders

A risk factor or intervention (an independent variable) may influence a substance abuse outcome (the dependent variable) indirectly, by affecting an intervening variable (a mediator) that in turn affects that outcome. Mediation analysis is a statistical method commonly used to examine the interrelations among independent, mediating, and dependent variables to obtain the direct and indirect effects of an independent variable on a continuous dependent variable. However, mediation analysis may also be used with binary outcomes, such as a diagnosis of an alcohol use disorder (AUD). Study 1 demonstrated methods of mediation analysis with binary outcomes by examining the direct and indirect effects of pro-alcohol social influences on an AUD, as a function of: (a) the distribution of the independent variable (binary vs. continuous), (b) the frequency of the outcome (non-rare vs. rare), and (c) the effect metric (probability vs. odds ratio). Study 2 was a Monte Carlo (simulation) study of bias in the indirect effects based on estimates from the first study. These methods have wide applicability in addictions research because many key outcomes are binary, and mediation analysis is frequently used to study the causal mechanisms by which interventions and risk factors affect substance abuse.

Uncovering the structure of self-regulation through data-driven ontology discovery

Psychological sciences have identified a wealth of cognitive processes and behavioral phenomena, yet struggle to produce cumulative knowledge. Progress is hamstrung by siloed scientific traditions and a focus on explanation over prediction, two issues that are particularly damaging for the study of multifaceted constructs like self-regulation. Here, we derive a psychological ontology from a study of individual differences across a broad range of behavioral tasks, self-report surveys, and self-reported real-world outcomes associated with self-regulation. Though both tasks and surveys putatively measure self-regulation, they show little empirical relationship. Within tasks and surveys, however, the ontology identifies reliable individual traits and reveals opportunities for theoretic synthesis. We then evaluate predictive power of the psychological measurements and find that while surveys modestly and heterogeneously predict real-world outcomes, tasks largely do not. We conclude that self-regulation lacks coherence as a construct, and that data-driven ontologies lay the groundwork for a cumulative psychological science.

Scientific progress relies on integrating and building on existing knowledge. Here, the authors propose improving cumulative science by developing data-driven ontologies, and they apply this approach to understanding the construct of self-regulation.

The (Lack of) Replication of Self-Reported Mindfulness as a Mechanism of Change in Mindfulness-Based Relapse Prevention for Substance Use Disorders

The development and evaluation of mindfulness-based interventions for a variety of psychological and medical disorders has grown exponentially over the past 20 years. Yet, calls for increasing the rigor of mindfulness research and recognition of the difficulties of conducting research on the topic of mindfulness have also increased. One of the major difficulties is the measurement of mindfulness, with varying definitions across studies and ambiguity with respect to the meaning of mindfulness. There is also concern about the reproducibility of findings given few attempts at replication. The current secondary analysis addressed the issue of reproducibility and robustness of the construct of self-reported mindfulness across two separate randomized clinical trials of mindfulness-based relapse prevention (MBRP), as an aftercare treatment for substance use disorder. Specifically, we tested the robustness of our previously published findings, which identified a latent construct of mindfulness as a significant mediator of the effect of MBRP on reducing craving following treatment. First, we attempted to replicate the findings in a separate randomized clinical trial of MBRP. Second, we conducted sensitivity analyses to test the assumption of the no-omitted confounder bias in a mediation model. The effect of MBRP on self-reported mindfulness and overall mediation effect failed to replicate in a new sample. The effect of self-reported mindfulness in predicting craving following treatment did replicate and was robust to the no-omitted confounder bias. The results of this work shine a light on the difficulties in the measurement of mindfulness and the importance of examining the robustness of findings.

Large-scale analysis of test-retest reliabilities of self-regulation measures

The ability to regulate behavior in service of long-term goals is a widely studied psychological construct known as self-regulation. This wide interest is in part due to the putative relations between self-regulation and a range of real-world behaviors. Self-regulation is generally viewed as a trait, and individual differences are quantified using a diverse set of measures, including self-report surveys and behavioral tasks. Accurate characterization of individual differences requires measurement reliability, a property frequently characterized in self-report surveys, but rarely assessed in behavioral tasks. We remedy this gap by (i) providing a comprehensive literature review on an extensive set of self-regulation measures and (ii) empirically evaluating test-retest reliability of this battery in a new sample. We find that dependent variables (DVs) from self-report surveys of self-regulation have high test-retest reliability, while DVs derived from behavioral tasks do not. This holds both in the literature and in our sample, although the test-retest reliability estimates in the literature are highly variable. We confirm that this is due to differences in between-subject variability. We also compare different types of task DVs (e.g., model parameters vs. raw response times) in their suitability as individual difference DVs, finding that certain model parameters are as stable as raw DVs. Our results provide greater psychometric footing for the study of self-regulation and provide guidance for future studies of individual differences in this domain.

Benchmark validation of statistical models: Application to mediation analysis of imagery and memory

This article describes benchmark validation, an approach to validating a statistical model. According to benchmark validation, a valid model generates estimates and research conclusions consistent with a known substantive effect. Three types of benchmark validation-(a) benchmark value, (b) benchmark estimate, and (c) benchmark effect-are described and illustrated with examples. Benchmark validation methods are especially useful for statistical models with assumptions that are untestable or very difficult to test. Benchmark effect validation methods were applied to evaluate statistical mediation analysis in eight studies using the established effect that increasing mental imagery improves recall of words. Statistical mediation analysis led to conclusions about mediation that were consistent with established theory that increased imagery leads to increased word recall. Benchmark validation based on established substantive theory is discussed as a general way to investigate characteristics of statistical models and a complement to mathematical proof and statistical simulation. (PsycINFO Database Record (c) 2018 APA, all rights reserved).

Sensitivity Analysis of the No-Omitted Confounder Assumption in Latent Growth Curve Mediation Models

Latent growth curve mediation models are increasingly used to assess mechanisms of behavior change. For latent growth mediation model, like any another mediation model, even with random treatment assignment, a critical but untestable assumption for valid and unbiased estimates of the indirect effects is that there should be no omitted variable that confounds indirect effects. One way to address this untestable assumption is to conduct sensitivity analysis to assess whether the inference about an indirect effect would change under varying degrees of confounding bias. We developed a sensitivity analysis technique for a latent growth curve mediation model. We compute the biasing effect of confounding on point and confidence interval estimates of the indirect effects in a structural equation modeling framework. We illustrate sensitivity plots to visualize the effects of confounding on each indirect effect and present an empirical example to illustrate the application of the sensitivity analysis.

Bias, Type I Error Rates, and Statistical Power of a Latent Mediation Model in the Presence of Violations of Invariance

When testing a statistical mediation model, it is assumed that factorial measurement invariance holds for the mediating construct across levels of the independent variable X. The consequences of failing to address the violations of measurement invariance in mediation models are largely unknown. The purpose of the present study was to systematically examine the impact of mediator noninvariance on the Type I error rates, statistical power, and relative bias in parameter estimates of the mediated effect in the single mediator model. The results of a large simulation study indicated that, in general, the mediated effect was robust to violations of invariance in loadings. In contrast, most conditions with violations of intercept invariance exhibited severely positively biased mediated effects, Type I error rates above acceptable levels, and statistical power larger than in the invariant conditions. The implications of these results are discussed and recommendations are offered.

Reasons for Testing Mediation in the Absence of an Intervention Effect: A Research Imperative in Prevention and Intervention Research

OBJECTIVE

Mediation models are used in prevention and intervention research to assess the mechanisms by which interventions influence outcomes. However, researchers may not investigate mediators in the absence of intervention effects on the primary outcome variable. There is emerging evidence that in some situations, tests of mediated effects can be statistically significant when the total intervention effect is not statistically significant. In addition, there are important conceptual and practical reasons for investigating mediation when the intervention effect is nonsignificant.

METHOD

This article discusses the conditions under which mediation may be present when an intervention effect does not have a statistically significant effect and why mediation should always be considered important.

RESULTS

Mediation may be present in the following conditions: when the total and mediated effects are equal in value, when the mediated and direct effects have opposing signs, when mediated effects are equal across single and multiple-mediator models, and when specific mediated effects have opposing signs. Mediation should be conducted in every study because it provides the opportunity to test known and replicable mediators, to use mediators as an intervention manipulation check, and to address action and conceptual theory in intervention models.

CONCLUSIONS

Mediators are central to intervention programs, and mediators should be investigated for the valuable information they provide about the success or failure of interventions.

Applying novel technologies and methods to inform the ontology of self-regulation

Self-regulation is a broad construct representing the general ability to recruit cognitive, motivational and emotional resources to achieve long-term goals. This construct has been implicated in a host of health-risk behaviors, and is a promising target for fostering beneficial behavior change. Despite its clear importance, the behavioral, psychological and neural components of self-regulation remain poorly understood, which contributes to theoretical inconsistencies and hinders maximally effective intervention development. We outline a research program that seeks to define a neuropsychological ontology of self-regulation, articulating the cognitive components that compose self-regulation, their relationships, and their associated measurements. The ontology will be informed by two large-scale approaches to assessing individual differences: first purely behaviorally using data collected via Amazon's Mechanical Turk, then coupled with neuroimaging data collected from a separate population. To validate the ontology and demonstrate its utility, we will then use it to contextualize health risk behaviors in two exemplar behavioral groups: overweight/obese adults who binge eat and smokers. After identifying ontological targets that precipitate maladaptive behavior, we will craft interventions that engage these targets. If successful, this work will provide a structured, holistic account of self-regulation in the form of an explicit ontology, which will better clarify the pattern of deficits related to maladaptive health behavior, and provide direction for more effective behavior change interventions.

A Tutorial in Bayesian Potential Outcomes Mediation Analysis

Statistical mediation analysis is used to investigate intermediate variables in the relation between independent and dependent variables. Causal interpretation of mediation analyses is challenging because randomization of subjects to levels of the independent variable does not rule out the possibility of unmeasured confounders of the mediator to outcome relation. Furthermore, commonly used frequentist methods for mediation analysis compute the probability of the data given the null hypothesis, which is not the probability of a hypothesis given the data as in Bayesian analysis. Under certain assumptions, applying the potential outcomes framework to mediation analysis allows for the computation of causal effects, and statistical mediation in the Bayesian framework gives indirect effects probabilistic interpretations. This tutorial combines causal inference and Bayesian methods for mediation analysis so the indirect and direct effects have both causal and probabilistic interpretations. Steps in Bayesian causal mediation analysis are shown in the application to an empirical example.

SAS® Macros for Computing Causal Mediated Effects in Two- and Three-Wave Longitudinal Models

Mediation analysis is a statistical technique for investigating the extent to which a mediating variable transmits the effect of an independent variable to a dependent variable. Because it is used in many fields, there have been rapid developments in statistical mediation. The most cutting-edge statistical mediation analysis focuses on the causal interpretation of mediated effects. Causal inference is particularly challenging in mediation analysis because of the difficulty of randomizing subjects to levels of the mediator. The focus of this paper is on updating three existing SAS® macros (%TWOWAVEMED, %TWOWAVEMONTECARLO, and %TWOWAVEPOSTPOWER, presented at SAS® Global Forum 2017) in two important ways. First, the macros are updated to incorporate new cutting-edge methods for estimating longitudinal mediated effects from the Potential Outcomes Framework for causal inference. The two new methods are inverse-propensity weighting, an application of propensity scores, and sequential G-estimation. The causal inference methods are revolutionary because they frame the estimation of mediated effects in terms of differences in potential outcomes, which align more naturally with how researchers think about causal inference. Second, the macros are updated to estimate mediated effects across three waves of data. The combination of these new causal inference methods and three waves of data enable researchers to test how causal mediated effects develop and maintain over time.

Tutorial: The practical application of longitudinal structural equation mediation models in clinical trials

The study of mediation of treatment effects, or how treatments work, is important to understanding and improving psychological and behavioral treatments, but applications often focus on mediators and outcomes measured at a single time point. Such cross-sectional analyses do not respect the implied temporal ordering that mediation suggests. Clinical trials of treatments often provide repeated measures of outcomes and, increasingly, of mediators as well. Repeated measurements allow the application of various types of longitudinal structural equation mediation models. These provide flexibility in modeling, including the ability to incorporate some types of measurement error and unmeasured confounding that can strengthen the robustness of findings. The usual approach is to identify the most theoretically plausible model and apply that model. In the absence of clear theory, we put forward the option of fitting a few theoretically plausible models, providing a type of sensitivity analysis for the mediation hypothesis. In this tutorial, we outline how to fit several longitudinal mediation models, including simplex, latent growth and latent change models. This will allow readers to learn about one type of model that is of interest, or about several alternative models, so that they can take this sensitivity approach. We use the Pacing, Graded Activity, and Cognitive Behavioral Therapy: A Randomized Evaluation (PACE) trial of rehabilitative treatments for chronic fatigue syndrome (ISRCTN 54285094) as a motivating example and describe how to fit and interpret various longitudinal mediation models using simulated data similar to those in the PACE trial. The simulated data set and Mplus code and output are provided. (PsycINFO Database Record

Family Support and Family Negativity as Mediators of the Relation between Acculturation and Postpartum Weight in Low-Income Mexican-Origin Women

BACKGROUND

Obesity presents a significant health concern among low-income, ethnic minority women of childbearing age.

PURPOSE

The study investigated the influence of maternal acculturation, family negativity, and family support on postpartum weight loss among low-income Mexican-origin women.

METHODS

Low-income Mexican-origin women (N=322; 14% born in the U.S.) were recruited from a prenatal clinic in an urban area of the Southwest U.S. Acculturation was assessed during a prenatal home visit (26-38 weeks gestation), and post-birth family support and general family negativity were assessed at 6 weeks postpartum. Objective maternal weight measures were obtained at five time points across the first postpartum year.

RESULTS

Higher acculturation predicted higher family support and family negativity. Higher family support predicted decreasing weight across the first postpartum year, and higher family negativity predicted higher weight at 6 weeks postpartum and increasing weight across the first postpartum year. In combination, family negativity and support mediated the impact of acculturation on postpartum weight gain.

CONCLUSIONS

Cultural and family-related factors play a significant role in postpartum weight gain and loss for low-income Mexican-origin women.

Confounding in statistical mediation analysis: What it is and how to address it

Psychology researchers are often interested in mechanisms underlying how randomized interventions affect outcomes such as substance use and mental health. Mediation analysis is a common statistical method for investigating psychological mechanisms that has benefited from exciting new methodological improvements over the last 2 decades. One of the most important new developments is methodology for estimating causal mediated effects using the potential outcomes framework for causal inference. Potential outcomes-based methods developed in epidemiology and statistics have important implications for understanding psychological mechanisms. We aim to provide a concise introduction to and illustration of these new methods and emphasize the importance of confounder adjustment. First, we review the traditional regression approach for estimating mediated effects. Second, we describe the potential outcomes framework. Third, we define what a confounder is and how the presence of a confounder can provide misleading evidence regarding mechanisms of interventions. Fourth, we describe experimental designs that can help rule out confounder bias. Fifth, we describe new statistical approaches to adjust for measured confounders of the mediator-outcome relation and sensitivity analyses to probe effects of unmeasured confounders on the mediated effect. All approaches are illustrated with application to a real counseling intervention dataset. Counseling psychologists interested in understanding the causal mechanisms of their interventions can benefit from incorporating the most up-to-date techniques into their mediation analyses. (PsycINFO Database Record

Empiric validation of a process for behavior change

Most behavior change trials focus on outcomes rather than deconstructing how those outcomes related to programmatic theoretical underpinnings and intervention components. In this report, the process of change is compared for three evidence-based programs' that shared theories, intervention elements and potential mediating variables. Each investigation was a randomized trial that assessed pre- and post- intervention variables using survey constructs with established reliability. Each also used mediation analyses to define relationships. The findings were combined using a pattern matching approach. Surprisingly, knowledge was a significant mediator in each program (a and b path effects [p<0.01]). Norms, perceived control abilities, and self-monitoring were confirmed in at least two studies (p<0.01 for each). Replication of findings across studies with a common design but varied populations provides a robust validation of the theory and processes of an effective intervention. Combined findings also demonstrate a means to substantiate process aspects and theoretical models to advance understanding of behavior change.